Policy Analysis and Economics

Recent Economic Thought Series

Editor:

Warren J. Samuels
Michigan State University
East Lansing, Michigan, U.S.A.

Previously published books in the series:

Feiwel, G.: Samuelson and Neoclassical
Economics
Wade, L.: Political Economy: Modern Views
Zimbalist, A.: Comparative Economic Systems:
Recent Views
Darity, W.: Labor Economics: Modern Views
Jarsulic, M.: Money and Macro Policy
Samuelson, L.: Microeconomic Theory
Bromley, D.: Natural Resource Economics:
Policy Problems and Contemporary Analysis
Mirowski, P.: The Reconstruction of Economic
Theory
Field, A.: The Future of Economic History
LOWry, S.: Pre-Classical Economic Thought
Officer, L.: International Economics
Asimakopulos, A.: Theories of Income
Distribution
Earl, P.: Psychological Economics:
Development, Tensions, Prospects
Thweatt, W.: Classical Political Economy
Peterson, W.: Market Power and the Economy
DeGregori, T.: Development Economics
Nowotny, K.: Public Utility Regulation
Horowitz, I.: Decision Theory
Mercuro, N.: Law and Economics
Hennings, K. and Samuels, W.: Neoclassical
Economic Theory, 1870 to 1930
Samuels, W.: Economics as Discourse
Lutz, M.: Social Economics

Policy Analysis and Economics

Developments, Tensions, Prospects

Edited by
David L. Weimer

"
~.

Springer Science+Business Media, LLC

Ubrary of Congress Cataloging-in-Publication Data

Policy analysis and economics: developments, tensions, prospects/
[edited by] David L. Weimer.
p.
cm.-(Recent economic thought series)
Includes bibliographical reterences and index.
ISBN 978-94-010-5720-2
ISBN 978-94-011-3866-6 (eBook)
DOI 10.1007/978-94-011-3866-6
1, Policy sciences. 2. Economics. 1. Weimer, David Leo.
II. Series.

H97.P637 1991
338.9-dc20

91-3197
CIP

Copyright 1991 Springer Science+Business Media New York

Originally published by Kluwer Academic Publishers in 1991
Softcover reprint of the hardcover 1st edition 1991
AII rights reserved. No pari of this publication may be reproduced, stored
in a retrieval system or transmitted in any torm or by any means,
mechanical, photo-copying, recording, or otherwise, without the prior
written permission of the publisher, Springer Science+Business Media, LLC .

Prinled on acid-free paper.

Contents

Contributing Authors

vii

Preface

ix

Economists as Policy Analysts: Historical Overview

Robert H. Nelson

2
Continuing Controversies in Policy Analysis
Hank Jenkins-Smith

3
Beyond Efficiency: Economics and Distributional Analysis
Alphonse G. Holtmann

4
Beyond Self-Interest
Amitai Etzioni

Economic Theories of Decision Making Under Uncertainty:

Implications for Policy Analysis
W. Kip Viscusi

The New Institutional Economics: Implications for Policy Analysis

Howard Frant

7
Macroeconomics and Macroeconomists as Instruments of Policy
George Horwich

23

45

65

85

111

127

vi

POLICY ANALYSIS AND ECONOMICS

8
The New Trade Theory: Implications for Policy Analysis
John Pomery

159

Policy Research: A Withering Branch of Economics?

187

Index

205

w. Lee Hansen

Contributing Authors

Amitai Etzioni
University Professor
Gelman Library, Room 714
George Washington University
Washington, D.C. 20052
Howard Frant
College of Business and Public Administration
453 B Harvill Building
University of Arizona
Tucson, Arizona 85721
W. Lee Hansen
Department of Economics
1180 Observatory Drive
University of Wisconsin
Madison, Wisconsin 53706
Alphonse G. Holtmann
Department of Economics
P.O. Box 248126
University of Miami
Coral Gables, Florida 33124
George Horwich
Department of Economics
Krannert Building
vii

viii
Purdue University
West Lafayette, Indiana 47907
Hank Jenkins-Smith
Department of Political Science
University of New Mexico
Albuquerque, New Mexico 87106
Robert H. Nelson
Office of Policy Analysis
U.S. Department of the Interior
18th and C Streets N.W.
Washington, D. C. 20240
John Pomery
Department of Economics
Krannert Building
Purdue University
West Lafayette, Indiana 47907
W. Kip Viscusi
Department of Economics
Duke University
Durham, North Carolina 27706
David L. Weimer
Public Policy Analysis Program
University of Rochester
Rochester, New York 14627

POLICY ANALYSIS AND ECONOMICS

Preface

Long before policy analysis emerged as a separate profession with its own
graduate schools, economists offered advice about government policies.
Positive economics provides the tools for predicting the impacts of proposed policies; normative economics, especially welfare theory, offers a
framework for valuing the impacts of policies in terms of efficiency and
simple notions of equity. With the expansion of economic theory into ever
wider fields of human behavior, it is no wonder that economists have
prominence as teachers and practitioners of policy analysis. Indeed, many
economists see policy analysis as essentially applied economics. Though
other social scientists might object to this somewhat parochial view,
economics and policy analysis share much in commom in terms of development and prospects. The purpose of this volume is to trace these interrelationships and explore the tensions that they create.
Tensions arise for several reasons. Changes in the discipline of economics affect the findings, methods, and personnel offered to policy analysis. For example, on the one hand, the "new institutional economics"
appears to be extending the influence of economists to questions involving
nonmarket oranizations, while on the other hand, the apparently growing
emphasis within the economics profession on creating rather than empirically testing theory suggests that fewer of the best young scholars will be
drawn to policy-relevant research. Within the schools of policy analysis,
the drift toward public management may reduce the demand for traditional
economic training. These changes may complicate the long-standing debate among policy analysts over the adequacy of the self-interest paradigm
for predicting behavior and the appropriateness of welfare economics as a
normative foundation for public policy.
ix

POLICY ANALYSIS AND ECONOMICS

Economists as Policy Analysts

The Progressive Movement, which emphasized the separation of technical
expertise from politics, opened the door for the participation of economists
in the formulation and evaluation of federal policy. In chapter 1, Robert
H. Nelson traces the changing role of economists in policy-making from
this early involvement to the present. He notes the spread of economists
from agencies such as the Treasury Department, the Council of Economic
Advisers, and the Federal Reserve, which deal directly with economic
issues, to all the major units of the federal government. Either in terms of
job classification or self-identification, the economists in federal executive
agencies now number in the thousands. Add to these the economists
serving the legislative branch in the Congressional Budget Office, the
Congressional Research Service, and committee staffs, as well as the
growing number of economists in "think tanks" and consulting firms who
contribute research and analysis to the policy process, and one might
conclude that economic analysis now makes an important contribution to
public policy.
Though Nelson points to these numbers as a rough measure of the
influence of economic analysis on federal policy, he notes several factors
that have limited, and are likely to continue to limit, the effectiveness of
economists as policy analysts.
First, as economists move beyond the traditional concerns of economic
policy, they often must compete more directly with other sorts of expertise.
At the Environmental Protection Agency, for instance, physical scientists
have a claim to expertise about the substance of policy and lawyers have a
claim to expertise about the policy-making process. Nelson notes that the
limited influence of economists in this area is not only indicated by the
rejection of effluent taxes, markets in pollution rights, and other policy
instruments commonly advanced by economists but also by congressional
prohibitions in some circumstances against the use of benefit-cost analysis,
the primary technique of applied economic analysis. One might speculate
that the prospects for the increased influence of economists in such
substantive policy areas may be as much a function of the penetration of
economics into the curriculum of the competing professions as it is a
function of the direct participation of economists.
Second, the increasingly theoretical orientation of graduate training in
economics does not sufficiently encourage the development of the practical
skills most useful in policy analysis. Graduates with fields in public finance
may never have been asked to do a benefit-cost analysis; those with fields
in econometrics may know much about asymptotic theory but little about

PREFACE

xi

the nuts-and-bolts of collecting and using data; few graduates in any fields
have experience explaining economic concepts in nontechnical terms.
Nelson worries that, if trends continue, graduate programs in economics
may no longer be a primary source of policy analysts. (In chapter 9, W. Lee
Hansen considers the match between graduate education in economics and
policy research of various kinds.)
Third, economists often have little perparation for dealing with the
ambiguity of goals found in most areas of public policy. On the one hand,
taking the concept of economic efficiency from welfare economics as the
only goal of public policy puts the economist at odds with the distributional
goals that loom so large in politics. On the other l)and, it is naive to think
that the political process can routinely provide suffiCiently clear goals and
their relative weights so that policy analysis can be viewed as an optimization problem. Nelson argues that economists, and other experts, are likely
to enjoy even less professional authority in the future so that to be
influential they will have to address the full range of values held by
participants in the policy process.
Welfare Economics as a Normative Foundation for
Policy Analysis
Efficiency informs normative economic analysis. In chapter 2, Hank
Jenkins-Smith discusses the evolution of this value from its utilitarian
foundations to its development in welfare economic theory as Pareto
efficiency and in application as potential Pareto efficiency. He notes the
recurring criticisms of these concepts as the normative basis for policy
analysis: Accepting preferences as given ignores the possible role of
education and public discourse in shaping values; defining social value
only in terms of the preferences of individuals slights socially desirable
processes such as democratic choice; Pareto efficiency as a theoretical
concept takes the initial endowment as a given; use of the potential Pareto
criterion as a practical rule does not guarantee that adopted policies will
actually leave everyone better off; and intangible costs and benefits tend to
be excluded in application of the potential Pareto criterion.
In addition to these criticisms, which have been discussed extensively in
the policy literature, Jenkins-Smith identifies two others that are likely to
be the focus of future debate. One criticism is the issue of standing in
benefit-cost analysis. Whose preferences count? Economists usually measure costs and benefits in terms of national polities. But should willingness
to pay be aggregated for citizens, residents, illegal aliens, or criminals?

xii

POLICY ANALYSIS AND ECONOMICS

How should the preferences of future generations be counted? Economists

have given little attention to resolving these questions of practical importance to policy analysis.
The second criticism concerns the inadequacy of welfare economics as a
framework for evaluating alternative institutional arrangements. Changes
in technology and tastes may create opportunities for increasing aggregate
welfare through the alteration of the system of entitlements, rights, and
obligations. Jenkins-Smith reviews the arguments made by Daniel W.
Bromley that proper evaluation of such "institutional transactions" demands a broader concept of social welfare than Pareto efficiency. Bromley
argues that some reference must be made to a social welfare function for
the determination of social efficiency. Jenkins-Smith worries that Bromley's approach might leave economic analysis so indeterminate as to make
it ineffective.
In chapter 3, Alphonse G. Holtmann considers the development of the
concept of the social welfare function in modern welfare economics. He
focuses on the question: What can economists contribute to the discussion
of distributional issues in public policy? Although it is unrealistic to expect
the political system to provide economists with social welfare functions for
evaluating policies with both efficiency and distributional effects, economists may be able to make a contribution by determining what form the
social welfare function would have to take to justify the selection of one
policy rather than another. They can also contribute by moving discussion
beyond consideration of only the static consequences of policies to such
dynamic topics as econonic growth and intergenerational equity.
Holtmann is optimistic that positive economics can play an increasing
role in informing public policy. The greater availability of data and
increasing computing power make distributional analysis more feasible.
Further, as economic theory advances it should permit better predictions
of the effects of alternative policies.

Economics as a Positive Basis for Polley Analysis

The belief in the adequacy of positive economics is not universally shared,
however. In chapter 4, Amitai Etzioni challenges the basic assumption of
neoclassical economics that individuals act in their own self-interest. He
does not wish to replace the self-interest assumption but rather to expand
behavioral motivations to include moral and social considerations. So, for
example, he rejects explanations of altruism in terms of self-interest as
reductionist and uninformative. He argues that concerns about what one

PREFACE

xiii

ought to do not only explain altruistic acts but also the much more common
obedience to laws, observance of conventions, and general respect for
others that make society civil.
Accepting the positive proposition that moral and social considerations
do affect behavior in important ways has important normative implications. Public policies should not just be evaluated in terms of the material
incentives they provide, but also in terms of their contributions to personal
and social norms that promote socially desirable behavior. In other words,
preferences should not be taken as given, and the nature of social relationships should not be ignored.
The major problem with the socioeconomic approach to policy analysis
advocated by Etzioni is our relative inability to predict behavior in terms
of noneconomic variables that can be manipulated by government. If
socioeconomics is to replace neoclassical economics as the dominant
paradigm for policy analysis, it must provide a theory that offers better
predictions of the consequences of policy alternatives. Until a track record
develops, it is likely to serve largely as an important reminder that selfinterest does not explain all policy relevant behavior.
Even this role is hindered by a lack of theory. Consider, for example,
Etzioni's critique of educational choice and educational vouchers. His
criticisms fall largely within the standard economic framework: Parents
face an informational asymmetry in evaluating schools, and wealthier
parents enjoy a larger choice set. Yet one might be inclined to make
socio-economic arguments in favor of parental choice and vouchers: They
invigorate the desirable social norm of parental involvement in education,
and they increase the opportunity for parents to express a demand for
moral education consistent with their beliefs.
Less fundamental, but nevertheless policy-relevant, challenges can be
made to positive economics. In chapter 5, W. Kip Viscusi considers
decision making under uncertainty, a topic in positive economics especially
important to public policy in the areas of health, safety, and environment.
Despite the considerable success of the expected utility model in predicting
behavior in situations of uncertainty, accumulating evidence from laboratory experiments and observation of insurance markets brings its adequacy
into question. It is not just that people sometimes deviate from the
predictions of the expected utility model, but that they often seem to do
so in systematic ways such as overestimating very small risks and underestimating relatively larger ones. Obviously, a theory that can encompass
these behaviors without forfeiting the successes of expected utility theory is
very desirable for predicting the consequences of public policies.
Viscusi rejects prospect theory, which allows the utility of lotteries to

xiv

POLICY ANALYSIS AND ECONOMICS

depend on functions of the probabilities and utilities of prizes, because it

can accommodate almost any patterns of behavior and thus offers few
predictions. Instead, he proposes what he calls "prospective reference
theory," which is based on Baysian updating of information about probabilities. Individuals hold beliefs about the probabilities of future events. As
they discover more information, they update their beliefs, moving closer,
but not necessarily all the way, to beliefs consistent with the new information. Viscusi argues that prospective reference theory explains most of the
documented anomalies in risk behavior while preserving the expected
utility framework as a basis of ex ante prediction.
Critics of the neoclassical paradigm have also argued that it largely
ignores the problem of institutional design. The view that market competition weeds out less efficient forms of organization has made the problem of
institutional design seem unimportant to many economists until recently.
Of course, the limited competition faced by most government organizations should have attracted attention to the problem of nonmarket supply,
but the dominant tradition in public finance, the field of most direct
relevance, was to focus on market failures as justifications for government
rather than the more general question of appropriate institutional design.
In recent years, however, several intellectual streams have coalesced into
what has become known as the new institutional eonomics. In chapter
6, Howard Frant outlines these intellectual streams and considers the
relevance of the new institutional economics for policy analysis.
One stream takes the transaction as the basic unit of analysis. Beginning
with Ronald Coase in the 1930s, a small number of economists have
attempted to understand the choice between production and contracting
in terms of the minimization of transaction costs. The work of Oliver
Williamson, for example, points to opportunism, asset specificity, and
bounded rationality as important attributes of the contracting process.
The second stream considers the problem of agency: How can contracts
between principals and agents be structured so as to minimize the sum of
discretionary and monitoring costs? Frant notes that it shares with the
transaction cost framework a central concern with opportunism. Nevertheless, because it posits self-interested behavior and explicitly models asymmetric information, it has been much more readily accepted by economists
than the transaction cost framework.
The new institutional economics that is taking shape seems to be
encompassing these streams within a game theoretic framework. It is likely
to interest more economists in a number of questions cutting across
substantive areas of public policy. Frant points to a better understanding of
privatization and the functioning of the nonprofit sector as areas where

PREFACE

xv

economists are just starting to make policy-relevant contributions. More

generally, he sees the new institutional economics as expanding the range
of policy instruments commonly considered by economists to include
changes in institutional structures as well as changes within existing
structures.
Applications: Macroeconomics and Trade Theory
The relationship between economics and public policy exhibits different
trends in different areas of application. Economists have played, and will
continue to play, a dominant role in the area of macroeconomics. In the
area of trade polky, however, the weakening of the apparent professional
consensus in support of free trade may make economists relatively less
influential in the future than they have been in the past.
In chapter 7, George Horwich traces the development of macroeconomic theory and its influence on public policy from the Great Depression
to the current debate over the significance of the federal budget deficit.
He argues that, just as the comparative statics of supply and demand
provide a powerful framework for micropolicy, comparative statics with
appropriately defined aggregate supply and aggregate demand can playa
similar role in macropolicy.
It is only relatively recently, however, that aggregate supply has received serious treatment from macroeconomists involved in policy debates.
The Keynesian system, which focuses attention on the problems created
by inadequate aggregate demand, served as the basis for the large-scale
macroeconomic models that became major analytical tools in the federal
agencies during the 1970s. It is not surprising that these models did not deal
well with the oil price shocks of that decade-economic shocks having their
primary effects through shifts in aggregate supply. The Council of Economic Advisers, though less bound to any particular model, also viewed the
oil price shocks as having adverse effects largely through reductions in
aggregate demand. Indeed, CEA at various times supported petroleum
price controls, strongly rejected by most microeconomists, because they
might limit the transfer of funds to oil companies and a subsequent
reduction in aggregate demand.
A major macroeconomic controversy continues to surround the implications of the apparently large federal deficits of the 1980s. Horwich sees this
controversy continuing. He argues that resolution demands fundamental
changes in the way the deficit is measured.
Overall, the story that Horwich tells is one of relatively slow shifts

xvi

POLICY ANALYSIS AND ECONOMICS

in macroeconomic policy. Nevertheless, despite considerable controversy

over the relevant theory, economists usually play a central role in the
policy process. Aside from the fact that the substance of the macropolicy
falls squarely within the purview of economics, one can speculate that the
diffuse costs and benefits of macropolicy leave economists less constrained
by the competition of organized interests within the political process.
In the case of trade policy, however, costs and benefits tend to be
concentrated for industries so that they have an incentive to seek various
forms of protection from the government. Until recently, a near consensus
among economists in favor of free trade has provided a strong conceptual
bulwark to blunt somewhat the concentrated interests seeking protectionism. The new trade theory, which starts from the perspective of imperfect
competition, has begun to break the consensus among economists for free
trade as a first "rule of thumb."
In chapter 8, John Pomery reviews the growing controversy among
trade theorists over the free trade prescription. Even economists who start
from the assumption of competitive world markets have long admitted the
possibility that in some circumstances, such as when an importer has
monopsony power or domestic markets are distorted, protective measures
can be welfare enhancing. But the new trade theory admits many more
possibilities. Critics charge, however, that its conclusions are too sensitive
to small changes in assumptions for it to be a reliable guide for public
policy. They also fear that the very breakdown of near consensus among
economists will clear the way for special interests to secure protections.
After clarifying the issues in this ongoing debate, Pomery turns in his
last sections to several fundamental questions in the philosophy of science
that are relevant not only to trade theory but more generally to other areas
of economic theory with policy implications. Indeed, the questions he
raises provide the broadest conceptual framework for the topic of this
volume and therefore deserve consideration even from those readers not
specifically interested in trade theory.
Pomery suggests that a conflict in theoretical perspectives underlies the
debate. Just as Newtonian and quantum mechanics each have validity in
certain circumstances, so too might the competitive perspective of the old
trade theory and game theoretic perspective of the new trade theory. If
each has validity in some circumstances, more and better empirical evidence will be unlikely to resolve the debate. A clearer specification of both
what is inside and outside of each perspective is needed for deciding which
is appropriate for informing any particular policy issue.
Making perspectives more comprehensive may not necessarily be the
best approach for policy purposes. Pomery notes, for example, the concep-

PREFACE

xvii

tual puzzle that arises in interpreting models that endogenously predict

trade policy for each country based on strategic interaction among governments and domestic interest groups. If such models describe and predict
well, by what means would one expect an exogenous policy recommendation to have any effect? Pomery points out that endogenizing political
behavior and information acquisition push potential leverage points on
policy to areas such as ideology, ideas, and institutions.

Economics Training and Polley Analysis

In the final chapter of the volume, W. Lee Hansen considers the prospects
for policy research remaining a major concern of economists. He notes
factors that suggest the declining importance of policy research for the
academic economists who train and influence the coming generations of
economists. Recent years have witnessed less availability of funding for
empirical policy research. The very proliferation of economists within
government has reduced the demand by agencies for economics faculty as
consultants and visitors, thus weakening the attraction of policy research
and reducing the opportunity for policy experience. Perhaps of greatest
significance, however, is the apparent mismatch between the knowledge
and skills emphasized in the most highly ranked doctoral programs in
economics and the skills most useful to practicing policy researchers and
analysts.
Hansen draws on recent findings of the American Economic Association's Commission on Graduate Education in Economics to characterize
doctoral training and its applicability for policy practitioners. A majority of
current faculty believe that knowledge of theory is, and should be, emphasized. A majority of surveyed Ph.D.s from the 1977-1978 cohort report
that, even though theory was emphasized in their own doctoral training,
the emphasis in their current jobs, both academic and nonacademic, is on
application, The apparent mismatch is even more striking with respect
to skills. Whereas the educational experience emphasizes analytics and
mathematics, practitioners point to the importance of communication skills
and critical judgment in their current jobs.
How important is the mismatch? On the one hand, the continued
demand for economists as policy analysts suggests that the knowledge and
skills currently emphasized in doctoral training have intrinsic value, provide a good foundation for acquiring the more immediately useful skills, or
at least signal the ability to be analytical. Indeed, formal training in most
professions emphasizes theory over practice so as to provide as solid

xviii

POLICY ANALYSIS AND ECONOMICS

foundation for self-learning. Further, the graduates of public policy schools

serve as good substitutes for economists in situations where the more
practical skills take precedence. On the other hand, if the trend away from
policy research in graduate programs continues, even fewer students with
an interest in policy will be recruited, possibly leading to a poor correspondence between the interests of new Ph.D.s and the jobs that many will
inevitably take outside of academia. Hansen notes that this trend, along
with the reduced opportunity for academic economists to gain policy
experience, suggests that public policy schools will find it more difficult to
recruit appropriate economists for their faculties in the future, thus
perhaps threatening the quality of their products.
Economics and economists now make essential contributions to policy
analysis. It is inconceivable that economics will not continue to be one of
the foundations of policy analysis. The declining involvement of academic
economists in policy research may be but a temporary phenomenon. Yet it
might signal a more significant trend toward a sharp division of labor
between the faculty of economics departments who speak mainly to other
academic economists about puzzles within the body of economic theory
and economists in professional schools, government, and think-tanks who
speak mainly to issues of public policy. Such a division of labor need not
concern either economists or policy analysts as long as the division is not so
complete as to preclude intellectual exchange.
David L. Weimer
University of Rochester

Policy Analysis and Economics

ECONOMISTS AS POLICY
ANALYSTS: HISTORICAL OVERVIEW
Robert H. Nelson

1. Introduction
Through the mid-nineteenth century a major task of American colleges
was to train students for the ministry; they read Adam Smith and other
economists in their study of moral philosophy. In the late nineteenth
century, however, higher education acquired a new mission (Bledstein,
1976; Buck, 1965). In the future the university would be the training
ground for a scientific elite to manage American society. The social
sciences were required along with many other specialized disciplines to
provide the technical knowledge to serve this purpose. The role of the
economist as policy analyst took shape in this era.
The underlying belief motivating these changes was that social and
individual behavior obeyed scientific laws that were discoverable and
usable, much as in the realm of physical science. If a correct scientific
understanding of the laws that controlled social and individual behavior
could be achieved, it would be possible to build a new and better societyone reflecting the improvements that physical science was generating in the
physical world. It was the heyday of faith in human progress based on
scientific knowledge-hence, the characterization of the "progressive" era
(Waldo, 1984).
1

POLICY ANALYSIS AND ECONOMICS

The creation of the American Economic Association (AEA) in 1885

reflected these beliefs (Coats, 1960; Farmer, 1975; Rader, 1966). The first
AEA statement of principles indicated the existence of "a vast number of
social problems, whose solution requires the united efforts, each in its own
sphere, of the church, of the state, and of science" (Ely, 1977: 140).
Richard Ely, the founding father of the AEA, was also a leading member
of the social gospel movement who sought to improve the conditions of the
poor, to create a more equal distribution of income, and to serve other
progressive causes, all based on a scientific analysis of American society.
Improvements in the condition of American labor, for example, depended
on assembling facts and figures relating to the actual condition of labor and
then applying this scientific knowledge to develop a program of social
reform. Moreover, Ely was no academic recluse; he wrote widely, including popular books and magazine articles and took his case for progressive
reforms directly to the American public. His efforts were so successful that
a later historian would find that Ely had exerted "a pervasive influence on
[American] social thought and the course of social reform both during the
80's and 90's and during subsequent periods of reform" (Fine, 1964: 238).
Ely's activism and explicit grounding of his efforts in an ethical set of
convictions, however, generated unease among many of his fellow economists. Frequently, they did not share Ely's particular convictions and in
any case saw a conflict between the objectivity that economic science was
seeking and the enlisting of economics in a particular reform movement.
By 1892 pressures of fellow economists drove Ely out as secretary (chief
administrative officer) of the AEA. There was increasing acceptance
among American economists that a sharper dichotomy should be maintained between scientific and political activities.
2. The Progressive Model

The scientific-political dichotomy would be central to the governing

theories of the American progressive movement of the early twentieth
century. According to progressive writers, government should be divided
into two distinct domains, one of politics and a second of technical
expertise (Goodnow, 1967). Politics would be characterized by subjective
considerations involving social values, ethical beliefs, and other matters
unsuitable for scientific determination. Taking such factors into account,
political leadership would set the overall social goals. Then, in the second
domain of expertise, decisions would be made to implement these goals,
based on objective knowledge derived from scientific research. Much of

ECONOMISTS AS POLICY ANALYSTS: HISTORICAL OVERVIEW

the routine administration of government was believed by progressives to

be the proper responsibility of professional experts. In such matters special
interests and other political influences should be strictly excluded.
Economists had a critical role in the progressive vision because they
were the keepers of the particular field of expert knowledge concerned
with material growth and production. Indeed, the diminishing of material
scarcity was critical to progressive thinking because it would reduce conflict
over possession of material goods-and ultimately, many hoped, such
conflict would become unnecessary and irrelevant. The striving for greater
possessions and riches in the past had been the source of much of the wars,
hatreds, and other human misbehavior in history. If modern progress now
seemed to offer the possibility of material abundance, the future might
hold a new heavenly condition on earth, a kind of secular salvation of
humankind. The economics profession would become a leading priesthood
guiding the way.
Such a vision appeared in the writings of the professional economist who
achieved the greatest visibility in the progressive era. A leading historian of
progressivism describes Thorstein Veblen as having had "the most brilliant
mind of his time" (Wiebe, 1967: 153). If outside the professional mainstream and unique in many respects, Veblen was nevertheless associated
closely with the institutional school of economics, which also included John
R. Commons, John Wesley Mitchell, and a number of other American
economists. This school contended with neoclassical economists for
supremacy in American economics during the first two decades of this
century but then receded sharply in numbers and influence (Dorfman,
1959).
Veblen portrayed in graphic form many of the themes that were
characteristic of progressive thought. The management of an increasingly
complex society required "a core of highly trained and specially gifted
experts, of diverse and various kinds," who would constitute the "indispensable General Staff of the industrial system. " They would be possessed
of "technological knowledge" of a character that was "exactingly specialized, endlessly detailed, reaching out into all domains of empirical fact."
Even in business, Veblen believed, technically trained engineers should be
given final control of enterprises because they were not "in any degree
benefited by any supervision or interference on the side of the owners" and
thus should operate "unhampered by commercial considerations and reservations" (Veblen, 1965: 68-70).
Although progressive thinking developed the foundations for the creation of the American welfare state in the twentieth century, and for the
role that economists would play in administering it, only the first steps were

POLICY ANALYSIS AND ECONOMICS

taken in the progressive era itself. Agencies such as the Civil Service
Commission (1883), Bureau of Labor Statistics (1884), Interstate Commerce Commission (1887), Forest Service (1905), Food and Drug
Administration (1906), Federal Reserve Bank (1913), Federal Trade
Commission (1914), and Bureau of the Budget (1922) paved the way for
what would later in the century become an even greater proliferation of
federal agencies and responsibilities. Initially, the role of economists in the
new progressive agencies was limited. Professional foresters, for example,
dominated the new Forest Service, which had few ecomomists and did little
economic analysis (Nelson, 1985). And professionals in the field of public
administration held sway at the Bureau of the Budget. At the Federal
Reserve System, however, economics professionals played a major role
from the beginning.
Within cabinet agencies, the Bureau of Agricultural Economics, created
in 1922 to advise the secretary of Agriculture, became a leading center for
economic research and policy study. This office may well have been the
first of the many policy "shops" that are found now in virtually every
federal agency (Barber, 1981, 1985). In the 19205 and 19305 economists in
government exerted probably more influence over agricultural policy than
over any other area of federal policy. The efforts of government economists laid the groundwork for new American farm policies of the 19305.
Some, such as E. J. Working, Mordecai Ezekiel, and Frederick Waugh,
were also prominent leaders in American agricultural economics and
contributed significantly in broader professional fields such as the development of statistical methods and econometrics in the United States
(Fox, 1986).

3. The Spread of Economists In Government

The depression of the 19305, and then to even a greater degree the
demands of World War II, brought economists in growing numbers to the
federal government (Stein, 1986). Many economists participated individually in developing legislation for and administering New Deal programs. A Federal Reserve economist, Lauchlin Currie, did much in the
19305 to introduce Keynesian ideas into Washington policy debates. During the later New Deal, economists also began asserting a greater institutional and formal presence in government. Two of the most important
current economic staffs were created in 1938, the Office of Tax Analysis of
the Treasury Department and the economic staff of the Antitrust Division
of the Justice Department. The Office of Tax Analysis in the 19405

ECONOMISTS AS POLICY ANALYSTS: HISTORICAL OVERVIEW

consisted of around 25 economists and at one time or another included a

number of leading future members of the American economics profession
-among them, E. Cary Brown, Richard Goode, Milton Friedman, C.
Lowell Harriss, Walter Heller, Joseph Pechman, Nancy Ruggles, and Carl
Shoup. Paul Samuelson served for a time as a consultant.
Another important development during the New Deal years was the
government's assuming of responsibility for assembling and publishing
national income statistics (Duncan and Shelton, 1978). Initially, this effort
was privately undertaken by the National Bureau of Economic Research
(NBER) under the leadership of John Wesley Mitchell. Founded in 1919
by Mitchell, the NBER worked during the 1920s to solve theoretical and
practical problems of national income accounting. In the early 1930s,
Simon Kuznets, an NBER employee on loan to the Commerce Department, spearheaded the publication of the first comprehensive national
income estimates of the federal government. The Bureau of Economic
Analysis in the Commerce Department, which is today responsible for
GNP, national income, and other basic economic statistics, traces its
origins to these developments in the early 1930s.
World War II gave a further large boost to the employment of government economists, as federal planners assumed control over many economic
tasks. Economists were required to assist in economic planning and
coordination to meet military production requirements, to impose price
controls, to collect new taxes, to supply allies, and to study enemy
economic capabilities. During the war years the number of economists in
the State Department, for example, increased from about 100 to about 400
and the Bureau of Economic Analysis was created, in 1944.
The visibility of government economists rose still further with the
enactment of the Employment Act of 1946, creating the Council of
Economic Advisers (CEA) (Flash, 1965). The council consists of a chairman, who reports directly to the President, and two other council members. They are assisted by a small staff of 15 to 20 economists, typically
recruited from leading universities for temporary stays in government. The
CEA was given the role of promoting high employment with stable prices
and otherwise overseeing the macroeconomic management of the U.S.
economy. Influenced by newly spreading Keynesian ideas, and desiring to
avoid any repetition of the depression years, Congress accepted that the
government could in principle manage scientifically the future overall
directions of the American economy. This new charge to American
economists involved an abandonment of earlier laissez-faire views of a
self-regulating economic system; instead, the members of the American
economics profession and the expert knowledge they possessed were now
to be responsible for assuring a happy national macroeconomic result.

POLICY ANALYSIS AND ECONOMICS

4. Economists and Politics: An Uneasy Relationship

The agencies in which economists served and the expectations of economists themselves were generally based on progressive political theories and
writings. Yet, as more economists entered government service, they often
found their experiences at odds with progressive formulations. Other social
science and professional experts often had similar experiences. Historian
Barry Karl (1963: 222) writes of the 19308 that "some of those who had
considered their mission to Washington a new step forward in the relations
between government and the academic community left with the feeling
that they had, in part at least, been subjected to a political trick." It turned
out that democratic politicians were much less respectful of the boundaries
of professional expertise than the experts coming into government had
been led to expect. Indeed, many concluded that "political elites only use
social science for ends that still fit into the jungle politics [that] science was
supposed to be eliminating" (Seidelman, 1985: 148).
As the focal point for the activities of economists in government, the
Council of Economic Advisers would offer important lessons for economists concerning the relationship of economic professionals with the political
process. The first CEA chairman, Edwin Nourse, sought to adhere closely
to the progressive model (Nourse, 1953). The CEA should supply the
President and other top political leaders of the executive branch with data
on economic trends, possible danger signs in the economy, alternative
economic policy measures, and other important economic information.
Reviewing the menu of options developed by his economic advisers, the
President would then make a choice that in part reflected social values
(e.g., the pain caused by unemployment versus the harm of inflation), his
own political aims, and other subjective factors. Nourse went so far in
adhering to this view as to decline even to participate in some key policy
discussions that he regarded as having an excessively political content.
Nourse was rewarded for his loyalty to progressive prescriptions with a
reputation for ineffectiveness and weakness. His successors would not
make the same mistake. Leon Keyserling, and all subsequent CEA chairmen, have found the maintenance of a strict dichotomy of politics and
economic expertise to be unworkable. For one thing, economic policymaking raises political and value concerns that frequently seem to be
inextricably interwoven with the economic elements. To advocate an
economic policy effectively, therefore, also requires advocating the political and value consequences. Economists in government could not credibly
suggest to political leaders that noneconomic matters were outside their
domain but were the responsibility of someone else. Successful advocacy

ECONOMISTS AS POLICY ANALYSTS: HISTORICAL OVERVIEW

of sound economic policies depended on the ability to package economic

policies with political, administrative, legal, and other attractive features.
In choosing an economic policy measure, a government economist might
well be compelled to pick second-best economic answers because they had
superior political prospects.
The CEA achieved perhaps its greatest influence under the chairmanship of Walter Heller in the 1960s. Heller successfully promoted the
"Kennedy tax cut," which helped to spur rapid economic growth in the
mid-1960s and raised the public reputation of government economists to
new heights. Heller was a skillful player in the internal politics of the
Kennedy administration and also was adept at public persuasion, especially
using the medium of television. (He would eventually appear on "Meet the
Press" 12 times, a frequency exceeded by only three other public figures.)
Heller later explained that "the detached, Olympian, take-it-or-Ieave-it
approach to presidential economic advice-the dream of the logical
positivist-simply does not accord with the demands of relevance and
realism." Instead, the economic adviser in government "presses the case
for some measures and against others" (Heller, 1967: 15). This approach
may involve pointing out not only economic consequences but also "political hazards." As Gardner Ackley, Heller's successor as CEA chairman,
put it, "If his economic adviser refrains from advice on the gut questions of
policy, the President should and will get another" (Ackley, 1966: 176).
Not only the three CEA members but the professional staff as well
found that in practice a direct involvement with politics was common place
and perhaps inevitable. Reflecting on his experiences on the CEA staff in
1963-1964, Burton Weisbrod observed that almost every staff member
became in part an "amateur political strategist." They would "think about
how we could get this thing through that group, who would accept what,
and what kind of friends and allies you needed to do what" (Allen, 1977:
67-68). The choice of topics, the factors taken as given, and the policy
solutions proposed could all be significantly influenced by political, administrative, and other noneconomic considerations.

5. The Demise of the Progressive Model

The experiences of economists mirrored changing views among political
scientists and other students of American government in the years following World War II. Dwight Waldo, for example, wrote in 1948 that "either
as a description of the facts or a scheme of reform, any simple division of
government into politics-and-administration is inadequate." Instead, "the

POLICY ANALYSIS AND ECONOMICS

governing process is a 'seamless web of discretion and action'" (Waldo,

1984: 128). Another leading political scientist, David Truman (1951),
portrayed American politics in terms of interest-group pressures exerted
throughout the affairs of government; special interests could not be
confined to a domain of politics. Indeed, by the late 19608 Theodore Lowi
would conclude that many students of government saw no alternative to,
and were now defending the legitimacy of, a governing system of "interestgroup liberalism" (Lowi, 1969: 68). The proper goal of American politics
was to incorporate all the affected interests into the decision-making
process; the social acceptability and legitimacy of government decisions
depended on achieving a satisfactory compromise among all the parties.
In one of the most famous social science articles of the post-war era,
Charles Lindblom (1959) wrote of government as "the science of 'muddling through.''' Contrary to progressive expectations for comprehensive
planning and rational decision making, government seldom was able to set
clear goals and frequently failed to evaluate systematically the alternative
courses of action. Democratic political preferences and social values could
seldom be established in advance to guide administrative and other expect
decisions. Instead, the values themselves were an outcome of the muddling
process, typically revealed only after the fact. Lindblom found that "pluralism" best described the practices of American government, a pluralism
of interest-group pressures, conflicting ideologies, bureaucratic incentives,
and still other inputs.
In such a process economists became yet another voice pressing their
case, although in this instance the objective was not to establish an
interest-group agenda but to advance particular economic policies and
more generally an economic way of thinking. Thus, commenting on his
experiences as director of the Bureau of the Budget in the mid-1960s,
Charles Schultze observed that "political values permeate every aspect of
the decision-making process in the majority of federal domestic programs.
There is no simple division of labor in which the politicians 'achieve'
consensus on an agreed-on set of objectives while the 'analysts' design and
evaluate-from efficiency and effectiveness criteria-alternative means
of achieving these objectives" (Schultze, 1968: 2-3). Schultze saw the
necessity for economists to enter directly and energetically into the pluralistic processes by which decisions were made in American government.
They would function not as disinterested technicians but as active advocates for an economic point of view, the official spokespersons in internal
government debate for more efficient programs and policies. As Schultze
now concluded, the proper role of economists in government would be to
serve as "partisan efficiency advocates" (Schultze, 1968: 101).

ECONOMISTS AS POLICY ANALYSTS: HISTORICAL OVERVIEW

In the mid-1960s Schultze was taking action to put his views into practice
not only at the Bureau of the Budget (BOB) but more broadly throughout
government by means of the Planning, Programming, and Budgeting
System (PPBS). When BOB was founded in 1922, its efforts were
grounded in the precepts of scientific administration, taught at leading
American universities in professional schools of public administration
(Berman, 1979). The emphasis was on achieving greater efficiency through
a better understanding of the proper span of control, personnel policies,
and other administrative mechanics. In the 1960s, however, BOB shifted
its emphasis to an economic understanding of efficiency, reflected in the
preparation of benefit-cost studies, the undertaking of cost-effectiveness
analyses, and generally the use of economic methods and tools. The first
director of BOB in the Kennedy administration was an economist, Kermit
Gordon. Including Gordon, four of the past nine directors have been
economists.
This ascendancy of economic thinking and professional economists at
the budget office reflected a governmentwide trend in the 19605 and 1970s.
Economists began to analyze and recommend policy in fields where they
had previously had little input. Partly for this reason, the number of
economists in government, which had already grown rapidly after World
War II, continued its rate of increase in this period.
Statistics on the numbers of economists in government are derived in
several ways. For one set of calculations, the Office of Personnel Management (OPM) (1987) compiles data on the numbers of federal employees
who occupy positions formally classified "economist." By this standard,
the number of federal economists totaled 2,221 in 1947 and rose by 57
percent to reach 3,480 in 1961. Over the next 16 years it rose another
52 percent, yielding 5,298 economists in 1977. However, growth slowed
considerably in the next decade, with OPM reporting 5,748 federal economists in 1987. These economists represented 0.4 percent of all federal
civilian employment in 1987. The State Department had the largest number (1,167), followed by the Labor Department (1,069), Agriculture
Department (837), Commerce Department (516), Department of the
Treasury (306), and Department of the Army (301). Women represented
22 percent of federal economists in 1987; 65 percent of federal economists
worked in the Washington, D.C. area; and 18 percent had a Ph.D.
Another source of data on the number of federal economists is the
National Science Foundation (NSF, 1988). The NSF survey is not based on
formal job classifications but reflects self-identification of job status. It thus
includes many who work on policy analysis, program planning, policy
evaluation, and other such tasks yet who consider themselves economists.

10

POLICY ANALYSIS AND ECONOMICS

For 1986 NSF estimated that there were 12,800 federal economists, more
than twice the number identified by OPM. According to NSF, federal
economists represented about 8 percent of the economics profession
nationwide. About 35 percent were engaged in "reporting [of economic
data], statistical work, and computing"; 28 percent in "management!
administration"; 27 percent in "research and development"; and the
remaining 10 percent in other duties.
The post-World War II rise of the American welfare state thus stimulated a demand for the skills of many economists in government, performing a wide range of tasks. Many were engaged in the generation of
economic statistics, production of economic reports, and other activities
that did not routinely bring them into direct contact with top policymakers. Other economists served as close advisers to high-level officials
and in some cases assumed top political positions themselves. Initially,
economists were concentrated in agencies with specific economic responsibilities. However, an important development since the 19608 has been the
use of economic analysis and the spread of policy analysts into many fields
where they previously had had little involvement.
6. The Old and the New Polley Economics
Economists were traditionally found in important policy-making positions
in areas where the economics profession had a long research interest and
where the objective of the government agency tended to be narrowly
economic. Thus in addition to the Council of Economic Advisers, the
agencies in which economists had the largest institutional presence over the
years were the Treasury, Department of Commerce, and Federal Reserve.
The Department of Agriculture and the Department of Labor also traditionally included large numbers of economists in statistical and economic
research departments, although their impact on policy was less assured and
was often considerably diluted by strong interest-group pressures, competing professional groups, and the diverse viewpoints found in these
agencies. In the 19608, however, economists began to move outside of
traditional domains to take on whole new fields. It was the beginning of a
period that in the social sciences has since been characterized as exhibiting
a pervasive "economic imperialism," challenging the basic methods and
outlooks of other fields of social science inquiry (Radnitzky and Bernholz,
1987). There were related developments within government as well, as
economists in new policy fields challenged other professional groups for
their traditional dominance.

ECONOMISTS AS POLICY ANALYSTS: HISTORICAL OVERVIEW

11

The formal introduction of economic and systems analysis in the early

1960s in the Defense Department opened one major new policy area
in which economists had previously made a limited contribution. The
development of the Planning, Programming, and Budgeting System
(PPBS) in the mid-1960s was modeled after Defense Department efforts.
Although PPBS was dismantled early in the Nixon administration, its goals
and purposes survived in a less formal and more incremental way. In the
1970s offices of policy analysis, program planning, and similar designations
were created throughout the federal government (Meltsner, 1976).
Although there were predecessor offices in many cases, the new offices
tended to give a greater emphasis to, and to ground their thinking in,
economics (Nelson, 1989; Wildavsky, 1979: 407-419). By means of such
offices, economic approaches to policy issues obtained an institutional base
in diverse agencies such as the Department of Housing and Urban
Development; the Department of Health, Education and Welfare; the
Environmental Protection Agency; the Department of Transportation;
and ultimately almost every major federal agency. The creation of the
Congressional Budget Office in 1974 similarly gave the Congress a much
increased capability to obtain its own independent sources of economic
analysis.
The new agencies in which economics was now being applied often
differed in a number of respects from the traditional economic agencies.
Frequently the new agencies had to confront value questions more explicitly and in ways that did not arise, for instance, in designing a price index
to measure inflation. Economists might propose vouchers as a means of
using the market mechanism to improve American education, but such a
proposal would inevitably be challenged as undermining the sense of
national unity promoted by a common system of public education. The
very legitimacy of a market approach and method thus was an issue in the
education field. Similarly, health policy operated in a setting in which a
market allocation of access to health care was increasingly rejected; within
a bureaucratic system of health care, rationing access according to economic calculations of benefits and costs was also widely unacceptable. In the
environmental field Congress in some cases expressly prohibited federal
agencies from even considering benefits and costs in deciding allowable
pollution levels and in making other environmental policy decisions.
In these new arenas for the application of economic ideas, economists
also frequently encountered other professional groups with their own
long-standing policy preferences, administrative approaches, and world
views. Doctors, educators, foresters, military officers, environmental scientists, and other professionals-to say nothing of the lawyers found

12

POLICY ANALYSIS AND ECONOMICS

everywhere-often clashed with economists. Noneconomists, for example,

often proclaimed a standard of maximum physical output at the highest
quality as the proper social goal. In the Forest Service, professional
foresters stated that maximizing the physical volume of wood was the
appropriate output objective for timber harvest policy in the national
forests. Foresters, like noneconomics professionals in most fields, often
strongly opposed introducing discount rates, applying benefit-cost tests,
and using other economic methods. In some cases, remarkable as it was to
economists first encountering these fields, noneconomics professionals
were even reluctant to enter explicitly the costs of production into their
calculations with respect to socially appropriate output levels.
The role of interest groups was often greater in policy-making outside
traditional economic agencies. When the Federal Reserve adopts a monetary policy, or the Council of Economic Advisers recommends a fiscal
policy, the impacts tend to fall over broad areas of the American economy.
Because of the breadth of the impacts, the President and other high-level
officials often become involved. Although tax policy has traditionally been
filled with special-interest provisions, the enactment of the Tax Reform
Act of 1986 demonstrated that in this area as well strong presidential and
other high-level leadership could overcome the many demands for special
tax treatment. The federal agencies responsible for collection of economic
statistics have been successful over the years in defending their activities
from interest-group pressures. By contrast, when the Corps of Engineers
decides whether to build a dam in a particular county, or when the Bureau
of Land Management decides whether to increase fees for grazing cattle on
public land, special-interest pressures are frequently decisive. In such areas
intensely interested parties have well-developed skills and a long record of
success in obtaining their particular needs from government.
In the newer agencies all these factors tended to impose on economists
a less secure status and greater obstacles to exerting a policy impact.
Confronted with direct challenges to their influence and sometimes to their
very presence, economists have had to try to develop skills over a wider
range. Economists in the newer agencies were more likely to see themselves in terms of policy "entrepreneurs." They have had to "sell" more
effective policies in their agencies, exhibiting an entrepreneurial resourcefulness and frame of mind (Leman and Nelson, 1981).
Economists in the newer agencies have also been pressed in recent years
to defend explicitly the way of thinking that is implicit in the application of
economic analysis (Rhoads, 1985). Such a defense takes economists well
beyond the realms of economic or even political analysis, into the domains
of philosophy and cultural analysis. To argue that an economic approach is

ECONOMISTS AS POLICY ANALYSTS: HISTORICAL OVERVIEW

13

an appropriate way of thinking is to make a claim with respect to core

assumptions and values on which society is based. In short, one of the
newer roles of a government economist might be said to be that of
"ideological combatant," recognizing that by ideology is meant a perspective or way of thinking about the world such as economics represents.
In the new agencies the educational backgrounds of those who applied
economic tools often differed from those in the traditional economic
agencies. In the latter a graduate degree in economics, often a Ph.D., has
frequently been expected. In the new policy and program evaluation
offices, however, there is much greater variety, reflecting the wider range
of skills sought. In many cases staff members have degrees from public
policy schools whose intent has been to provide a basis not only for
effective economic analysis but also for the adoption and implementation
of this analysis in the policy arena. Other practitioners of the new policy
economics may not have formal training in either economics or public
policy. In the newer agencies there is less confidence that anyone form of
education or type of knowledge is uniquely suited or even necessary to
being an effective policy economist.

7. The Impact of Policy Economists

The introduction of policy economists in most Washington agencies has
by now generated a considerable literature concerning their impact and
contribution to policy formulation (Nelson, 1987). On the whole, the
reports are mixed. Policy economists have failed to realize the high
expectations of the 1960s, when at least some proponents expected that the
spreading use of economic analysis would sharply increase the overall
rationality and efficiency of government. Yet, there have been some major
successes and many minor triumphs.
The deregulation of several key industries and tax reform, two of the
most important policy developments of the past 15 years, were significantly
attributable to the efforts of government economists. The foundations for
deregulation were laid in a series of economic studies--many sponsored by
the Brookings Institution and later the American Enterprise Institutefrom 1960 onward. By the mid-1970s the political climate had shifted to the
point where deregulatory measures were becoming feasible. In the late
19705 two professional economists, Alfred Kahn and Darius Gaskins,
became chairmen of the Civil Aeronautics Board and the Interstate
Commerce Commission, respectively. In these positions they led successful efforts to enact legislation significantly deregulating the American

14

POLICY ANALYSIS AND ECONOMICS

transportation industry. They were assisted by many other government

economists in less visible staff positions. Economists also played important
roles in the deregulation of the communications and energy industries.
Reviewing these developments, Martha Derthick and Paul Quirk (1985:
246) conclude that "if economists had not made the case for pro-competitive
deregulation, it would not have occurred-at least on the scale the nation
has witnessed."
As director of the economic studies program at Brookings, Joseph
Pechman was a key figure in supporting the groundwork not only for
deregulation but also for tax reform. Economists had long criticized the tax
system for setting high marginal rates and then undermining these rates
through numerous loopholes. The consequence was both to offend a
widely held American sense of equity and to seriously distort the efficiency
of national investment. In 1977 David Bradford, a Princeton economist
then serving as director of the Office of Tax Analysis in the Treasury
Department, organized the writing of the Blueprints for Basic Tax Reform,
which stimulated considerable discussion of basic changes in tax law. In
1984 another economist in the Office of Tax Analysis, Eugene Steuerle,
organized the preparation of a comprehensive proposal for tax reform.
Many political compromises later, much of this effort survived in the
historic Tax Reform Act of 1986 (Minarek, 1989).
There have also been areas in which government economists have been
less successful. Since at least the early 1970s, many government and other
economists have argued that U. S. environmental policies should rely less
on command-and-control methods and more on market incentives. Most
economists initially favored taxes on pollution, but in recent years a
number have seen equal or greater merit (and political feasibility) in a
system of marketable permits to pollute. Although 1990 legislation to
amend the Clean Air Act includes a market permit system, the use of the
market mechanism for environmental purposes has on the whole made
little headway. Economists have generally had a peripheral role in the
development of environmental policy over the past quarter century. Their
proposals to apply benefit-cost tests to the development of regulatory
rules, for instance, have not met much greater success than their proposals
to use the market mechanism (Anderson and Ostro, 1983).
The problems of the savings and loan industry illustrate a failure of
another type among policy economists. In this case few economists recognized the ingredients brewing for a national economic loss of huge magnitude or forcefully took the argument for remedial actions to the public and
political leadership. Although a handful of economists were in the field and
had an understanding of the problems, these economists tended to see

ECONOMISTS AS POLICY ANALYSTS: HISTORICAL OVERVIEW

15

their obligations in a limited professional context. It was enough to write

articles or books making the information, and a professional diagnosis
of it, available to policy-makers. It was not necessary~r perhaps even
compatible with professional norms-to became a public crusader for
savings and loan reform in the way that an Alfred Kahn had done with
respect to the deregulation of key industries.
In the defense area, where there were such high expectations for
economic analysis in the 1960s, the results have fallen well short. A
number of weapons systems have, however, been altered or canceled,
based at least partly on economic arguments. Reviewing the three decades
over which the use of economic and systems analysis has become institutionalized within the Defense Department, Peter deLeon (1987: 122, 124)
found that "in many instances, one can plausibly argue that analysis has
played a substantial, beneficial role in defense policy-making." Yet, there
have also been many failures, partly because "the key to successful policy
analysis and formulation is the definition of objectives. Yet, ... there is
nothing resembling agreement on this central issue, either in terms of ends
or means." As have economists working in areas of domestic policy,
deLeon (124-125) concludes that in the future "defense analysts must
explicitly incorporate the social and political aspects of policy into the
defense decision-making calculus" and that "defense analysts are being
forced to incorporate contexts, assumptions, and approaches well beyond
the system analysis techniques once thought to be the cardkeys to policymaking chambers."
In the 1960s a number of economists began to argue forcefully that
controlled experiments were required to resolve key issues that underlay
policy developments in areas such as welfare, health, education, and
housing. With strong support from economists serving in government
policy offices, a series of major social experiments were undertaken,
eventually receiving more than $1 billion (1983 dollars) in federal funds
(Greenberg and Robins, 1986: 357-360). These experiments yielded estimates superior to those that had been available concerning matters such as
the impact of a negative income tax on work effort. Yet, different analysts
often achieved different statistical results, and a "litany of problems"
clouded the overall reliability (Haveman, 1987: 193).
Haveman (175-199) suggests that on balance the social experiments
exerted a constructive influence on policy-making. Greenberg and Robins
(1986: 351) are somewhat less enthusiastic, concluding that "the lessons
provided so far by previous social experiments are not especially encouraging." One problem was that "the researchers have sometimes appeared to
be more concerned with methodological issues than with policy issues,"

16

POLICY ANALYSIS AND ECONOMICS

an orientation promoted by the disinclination of professional journals

to publish articles that "merely report results using standard evaluation
techniques." Partly for this reason, the level of researcher communication
with the ostensible beneficiaries of their efforts, the "government agencies,
Congress, the administration and the general public" has generally been
poor and "has sometimes been nonexistent" (346-347). Contrary to many
initial expectations, the overall effect of social experiments tended to
dampen enthusiasm of political leaders for social policy reforms; by raising
complex questions, and then providing answers that were inconclusive or
cast doubt on proposed solutions, policy-makers were more likely to shy
away from bold departures.
A systematic examination of the experiences of a range of economists
serving at high policy-making positions in government was undertaken
by William Allen (1977). From 1972 to 1976 Allen taped interviews with
about 60 economists who had had such experience. Typical of many
responses, W. Lee Hansen, who had worked on the staff of the Council of
Economic Advisers, saw the role of the policy economist as "just goodcommon-sense economics.... The kind of basic analytical framework that
we all sort of got in Econ 101. ... Simple supply and demand and benefitcost. ... If you can just keep these things in your mind, plus if you are
open to seeing how they might have to be modified in the light of
institutional constraints and considerations, then, I think that's the game
really" (Allen, 1977: 70-71).
Summing up the overall sense of the many economists interviewed,
Allen offered a pessimistic assessment of the influence of high-level
economists in the normal course of events:
In speaking with economists who are or have been in government, one obtains a
picture and gains an impression which is sobering. The government economist
typically is not a highly independent researcher and analyst, free first to pick
many of his subjects and entirely free to broadcast generally the results of his
labors. He is a member of an organization, commonly devoting the bulk of his
time to topics specified from on high-the specification often being enunciated
only a few days (or, indeed, hours) before the deadline; conscious of a
prevailing orientation and purpose on the part of these administrative superiors
who constitute his main audience; conscious, also, that the decisionmakers he is
more or less directly advising are themselves subject to constraints of worldly
realism and political feasibility-along with innocence in the area of economic
analysis; bringing to his task an accumulated intellectual capital which, even if
impressive at the outset of his government work, may not thereafter be greatly
enlarged or even well maintained; having more or less available a corpus of
theory and an arsenal of techniques which, for all their elegance, refinement, and

ECONOMISTS AS POLICY ANALYSTS: HISTORICAL OVERVIEW

17

academic glamor, are often too time-consuming for purposes of shooting from
the hip and too esoteric for the data, the colleagues, and the audience; and
having little reason to suppose that his work has significant impact in the making
of policy, being largely confined to support of programs and procedures determined earlier and by others and for which he may have only modest sympathy.
[pp. 86-87]

8. The Future of Policy Economics

In the current period there are trends both favorable and unfavorable to
the future role and influence of policy economists in government decision
making (Rivlin, 1987; Pechman, 1989). For the past 15 years energy policy,
deregulation, tax reform, and deficit reduction-all in significant degree
economic issues-have been at the center of national policy debates.
Indeed, economists have had the opportunity to play visible and important
roles in these national debates. Economics has become the leading currency of Washington policy discussion. Even when political factors may be
the driving force, proponents of policy initiatives are today commonly
expected to show how these initiatives will contribute to higher incomes
and the growth of U.S. economic productivity. Thus, current proponents
of major reforms in the U.S. education system frequently argue that these
reforms are needed to improve international economic competitiveness
and to make the best possible investment in the future of the nation.
Although conflicts in the Middle East still seem to follow older patterns,
the waning of the cold war suggests that the international politics of
military alliances will receive declining attention in the future, while
international economic relations (and environmental relations) may become more important. Recent events in communist nations, including
the virtual renunciation of communist economic organization in Eastern
Europe, have focused the attention of the world on the problems of
establisliing a market economy and generally building a satisfactory economic system. The leading message of the economics profession for many
years, the efficiency of using the market mechanism, has now become the
message of political leaders all over the world. Numerous economists are
being consulted and their services sought for tasks that may be critical to
the future peace and well-being of the world.
Other trends, however, suggest a less promising future for policy
economists. The reputation of economists with the American public has
eroded sharply since the high point of the mid-1960s. Predictive and other
failures have diminished public confidence in the scientific qualities of

18

POLICY ANALYSIS AND ECONOMICS

economic knowledge. The tendency of economists to divide among themselves and seemingly to quarrel endlessly over so many policy matters is
perhaps even more damaging to the scientific prestige of economists. If
such erosion in its scientific status continues for long, difficult questions
concerning the legitimacy and proper role of economics in government
may be raised.
The forces buffeting the economics profession are symptomatic of
developments affecting all American professions in recent years. The past
20 years has seen a sharp decline in public confidence in experts and in
almost any type of professional who makes claims to special authority
based on expertise. Experts are seen as commonly asserting much more
knowledge than they actually possess. Experts are accused of often failing
to maintain objectivity and neutrality; instead, they offer value-laden
conclusions, even while presenting these conclusions in many cases as
scientific truths. Thus not only economists but lawyers, doctors, psychologists, foresters, and many other professionals have been facing a growing
public distrust.
The future world in which policy economics operates will seemingly be
a world in which economists, and other professionals as well, will have
to persuade top officials, not by asking these officials to trust their
professional authority, but by persuading in the language and in the
manner of argument that are the daily fare of ordinary government debate
and public discussion. The persuasive skills called for will involve the
ability to draw convincing historical analogies; to take diverse events and
integrate them into a big picture; to show that economic policy proposals
make sense from political, moral, legal, philosophical, and other perspectives; to find simple statistics and data that effectively make a point; and to
convert broad policy principles into feasible short-tenn measures that over
time will serve to implement these principles.
This exercise will be much closer to art than to science (Wildavsky,
1979). The need for all these and other skills will impose new demands and
pressures on the educational institutions responsible for training the policy
economists of the future. What will be needed will be graduates who have
at least some command of many disciplines and who have the ability to see
connections and to tie diverse strands together. In a way it will be a return
to an emphasis on the liberal arts, requiring a reversal of the trend to
specialization that has marked university life for many years-indeed,
since the progressive era. The challenge will be, as higher education once
saw its mission, to instill and support the qualities of judgment, resourcefulness, commitment, and the capacity to grasp the larger dimensions of
political and economic issues.

ECONOMISTS AS POLICY ANALYSTS: HISTORICAL OVERVIEW

19

Meeting this challenge, to be sure, will not be easy. Indeed, recent

trends in economics education at the graduate level seem, if anything, to be
moving in the opposite direction, toward specialization and emphasis on
mathematical and statistic skills over breadth of knowledge. In the worst
case, the result could be a group of economics professionals who are suited
only for teaching other professionals in the nuances of a self-contained
system. If the economics profession is itself unable to supply the need for
policy-relevant economic advice, society will undoubtedly look elsewhere
-perhaps to public policy schools, but there are a number of other
possible sources. Both the economics profession and the quality of economic advice might eventually prove to be the worse for this result.
References
Ackley, Gardner, "The Contribution of Economists to Policy Formation," The
Journal of Finance 21:2 (May 1966), 169-177.
Allen, William R., "Economics, Economists, and Economic Policy: Modern
American Experiences," History of Political Economy 9:1 (Spring 1977), 48-88.
Anderson, Robert C, and Ostro, Bart, "Benefits Analysis and Air Quality
Standards," Natural Resources Journal 23:3 (July 1983), 565-575.
Barber, William J., From New Era to New Deal: Herbert Hoover, the Economists
and American Economic Policy, 1921-1933 (New York: Cambridge University
Press, 1985).
- - , "The United States: Economists in a Pluralist Society," in A. W. Coats,
ed., Economists in Government: An International Comparative Study (Durham,
N.C: Duke University Press, 1981).
Berman, Larry, The Office of Management and Budget and the Presidency, 19211979 (Princeton: Princeton University Press, 1979).
Bledstein, Burton J., The Culture of Professionalism: The Middle Class and the
Development of Higher Education in America (New York: Norton, 1976).
Buck, Paul, Social Sciences at Harvard: From Inculcation to the Open Mind
(Cambridge: Harvard University Press, 1965).
Coats, A. W., "The First Two Decades of the American Economic Association,"
American Economic Review 50:4 (September 1960), 555-574.
deLeon, Peter, "The Influence of Analysis on U.S. Defense Policy," Policy
Sciences 20:2 (1987), 105-128.
Derthick, Martha, and Quirk, Paul J., The Politics of Deregulation (Washington,
D.C: The Brookings Institution, 1985).
Dorfman, Joseph, The Economic Mind in American Civilization: Volumes Nand
V, 1918-1933 (New York: Viking, 1959).
Duncan, Joseph W., and Shelton, William C, Revolution in United States Government Statistics, 1926-1976 (Washington, D.C: Government Printing Office,
1978).

20

POLICY ANALYSIS AND ECONOMICS

Ely, Richard T., Ground Under Our Feet: An Autobiography (New York: Arno
Press, 1977; first ed. 1938).
Farmer, Mary 0., Advocacy and Objectivity: A Crisis in the Professionalization of
American Social Science, 1865-1905 (Lexington: University of Kentucky Press,
1975).
Fine, Sidney, Laissez-Faire and the General Welfare State: A Study of Conflict in
American Thought, 1865-1901 (Ann Arbor: University of Michigan Press,
1964).
Flash, Edward S., Jr., Economic Advice and Presidential Leadership: The Council
of Economic Advisors (New York: Columbia University Press, 1965).
Fox, Karl A., "Agricultural Economists as World Leaders in Applied Econometrics, 1917-1933," American Journal of Agricultural Economics 68:2 (May
1986), 381-386.
Goodnow, Frank J., Politics and Administration: A Study in Government (New
York: Russell and Russell, 1967; first ed. 1900).
Greenberg, David H., and Robins, Philip K., "The Changing Role of Social
Experiments in Policy Analysis," Journal of Policy Analysis and Management
5:2 (Winter 1986), 340-362.
Haveman, Robert H., Poverty Policy and Poverty Research: The Great Society and
the Social Sciences (Madison: University of Wisconsin Press, 1987).
Heller, Walter W., New Dimensions of Political Economy (New York: Norton,
1967).
Karl, Barry Dean, Executive Reorganization and Reform in the New Deal: 19321940 (Cambridge, Mass.: Harvard University Press, 1963).
Leman, Christopher K., and Nelson, Robert H., "Ten Commandments for Policy
Economists," Journal of Policy Analysis and Management 1:1 (Fall 1981),
97-117.
Lindblom, Charles E., "The Science of 'Muddling Through,''' Public Administration Review 19:2 (Spring 1959), 79-88.
Lowi, Theodore J., The End of Liberalism: Ideology, Policy and the Crisis of
Public Authority (New York: W. W. Norton, 1969).
Meltsner, Arnold J., Policy Analysts in the Bureaucracy (Berkeley: University of
California Press, 1976).
Minarek, Joseph J., "How Tax Reform Came About," in David Colander and
A. W. Coats, eds., The Spread of Economic Ideas (New York: Cambridge
University Press, 1989).
National Science Foundation, Profiles-Economics: Human Resources and Funding,
special report NSF 88-333 (Washington, D.C.: November 1988).
Nelson, Robert H., "The Office of Policy Analysis in the Department of the
Interior," Journal of Policy Analysis and Management 8:3 (Summer 1989),
395-410.
- - , "The Economics Profession and the Making of Public Policy," Journal of
Economic Literature 25:1 (March 1987), 49-91.
- - , "Mythology Instead of Analysis: The Story of Public Forest Management," in Robert T. Deacon and Bruce M. Johnson, eds., Forestlands: Public

ECONOMISTS AS POLICY ANALYSTS: HISTORICAL OVERVIEW

21

and Private (San Francisco: Pacific Institute for Public Policy Research, 1985).
Nourse, Edwin G., Economics in the Public Service: Administrative Aspects of the
Employment Act (New York: Harcourt Brace, 1953).
Office of Personnel Management, Occupations of Federal White-Collar and BlueCollar Workers, Pamphlet 56-20, September 30, 1987.
Pechman, Joseph A., "The United States," in Joseph A. Pechman, ed., The Role
of the Economist in Government: An International Perspective (New York: New
York University Press, 1989).
Rader, Benjamin G., The Academic Mind and Reform: The Influence of Richard T.
Ely in American Life (Lexington: University of Kentucky Press, 1966).
Radnitzky, Gerard, and Bernholz, Peter, Economic Imperialism: The Economic
Method Applied Outside the Field of Economics (New York: Paragon House,
1987).
Rhoads, Steven E., The Economist's View of the World: Government, Markets, and
Public Policy (New York: Cambridge University Press, 1985).
Rivlin, Alice M., "Economics and the Political Process," The Amercian Economic
Review 77:1 (March 1987), 1-10.
Schultze, Charles L., The Politics and Economics of Public Spending (Washington,
D.C.: Brookings Institution, 1968).
Seidelman, Raymond, Disenchanted Realists: Political Science and the American
Crisis, 1884-1984 (Albany: State University of New York Press, 1985).
Stein, Herbert, "The Washington Economics Industry," American Economic
Review 76:2 (May 1986), 1-9.
Truman, David B., The Governmental Process: Political Interests and Public
Opinion (New York: Alfred A. Knopf, 1951).
Veblen, Thorstein, The Engineers and the Price System (New York: Augustus M.
Kelley, 1965; first ed. 1921).
Waldo, Dwight, The Administrative State: A Study of the Political Theory of
American Public Administration (New York: Holmes and Meier, 1984; first ed.
1948).
Wiebe, Robert H., The Search for Order, 1877-1920 (New York: Hill and Wang,
1967).
Wildavsky, Aaron, Speaking Truth to Power: The Art and Craft of Policy Analysis
(Boston: Little Brown, 1979).

CONTINUING CONTROVERSIES
IN POLICY ANALYSIS
Hank Jenkins-Smith

1. Introduction

Contemporary public policy analysis is built on a foundation of principles

at the heart of welfare economics: Analysts are typically trained to rely on
aggregation of individual preferences to specify social value, generally in
terms of economic surplus. While there are indeed other foundations
of policy analysis and there are numerous challenges to the primacy
of welfare economics (Kelman, 1981; DeLeon, 1989; Brown, 1991), the
dominant paradigm for public policy remains deeply imbedded in the
concepts and principles of microeconomics and the primacy it places on
efficiency (Jenkins-Smith, 1990).
Despite their continuing prominence in the foundations of policy analysis, controversy over welfare economic principles has been both vigorous
and sustained. Furthermore, the critiques of the welfare economic bases of
policy analysis have become more fundamental and-in this writers viewmore compelling as critics have applied closer scrutiny to contemporary
analytical techniques.
This chapter will assess those controversies in the application of welfare

23

24

POLICY ANALYSIS AND ECONOMICS

economic principles in policy analysis that have shown greatest tenacity.

Discussion will begin with a brief review of the development of utility
theory and the concepts of efficiency as the cornerstones of contemporary
policy analysis and then tum to those critiques who most directly challenge
these cornerstones. The more venerable charges of distortion of public
preferences and violations of the process of public discourse will be
reviewed but primary attention will be given to more recent critiques who
challenge both the application and the normative bases of efficiency as the
primary criterion for choice in public policy. The chapter will conclude
with a discussion of the implications of these critiques for the practice of
policy analysis.
2. The Development of Utility Theory and Efficiency

The concept of efficiency has evolved over the past two centuries, beginning as a key element of a reformist political movement and undergoing
transformation toward an ever more technical and "content-free" analytical concept. Central to the concept of efficiency is the theory of utility as
developed by economists from foundations first laid by Jeremy Bentham.
According to Bentham, human experience provides utility when it
produces "benefit, advantage, pleasure, good, or happiness" or when it
prevents "mischief, pain, evil, or unhappiness" (Bentham, [1789] 1970).
Individual behavior can be understood as the pursuit of utility, based on a
hedonistic calculus designed to maximize pleasure and minimize pain. The
value of an act or experience was to be defined in terms of the pleasures
and pains it brought to individuals; individuals were thus the ultimate
arbiters of value. Drawing on these premises, Bentham called for enlightened formulation of public policy based on the principles of utility (Bentham, [1776] 1977).
For Bentham the concept of utility was rich in substantive content. The
chief dimensions of utility were (1) intensity, (2) duration, (3) certainty
or uncertainty, and (4) propinquity or remoteness. Also important were
fecundity, meaning the likelihood that the pleasure or happiness would be
followed by more of the same, and purity, meaning the likelihood that the
pleasure would not be followed by pain (Bentham, [1776] 1977: 38-41).
Bentham lists on less than 26 categories of pleasures and pains making up
the "simple" roots of utility to which all specific pains and pleasures can be
reduced, including (but not limited to) sense, wealth, skill, amity, good
name, power, piety, benevolence, and association (Bentham, [1776] 1977:
chap. 5).

CONTINUING CONTROVERSIES IN POLICY ANALYSIS

25

Much like economists who followed him, Bentham employed money as

a common metric in the measurement of utility, and he argued that the
value of money itself decreases as the amount of money held increases-a
phenomenon later dubbed "the decreasing marginal utility of money." As
Bentham argued, "the quantity of happiness produced by a particle of
wealth (each particle being of the same magnitude) will be less and less at
every particle" (Bentham, 1843: 229).
For Bentham, the development and use of utility theory in the formulation of public policy rested on an unresolved ambiguity. On one hand,
Bentham held that it would be impossible to determine the quantity of
utility that an individual obtained from an experience or to compare the
gains or losses of utility of any two individuals. On the other hand, unless
one could reasonably compare the utilities of individuals, the utility
calculus would provide little aid to the formulation of public policy.
Bentham's solution was simply to assume utilities to be comparable and to
justify the assumption as providing a useful tool for legislators that would
be, at worst, far better than any basis for legislation that failed to take
utility into account.
A central weakness in Bentham's calculus was the incomparability of
individual utilities in prescribing legislation for the "greatest good." Given
that the pursuit of the greatest good for the greatest number would often
involve utility losses for some, could we be sure that the added value to the
gainers would offset the losses to the losers? The incomparability of
individual utilities would seem to preclude such assurance. Utility theorists
following Bentham struggled with this problem and generally rejected
Bentham's contention that utilities could be assumed comparable.
More recent utility theorists have broadened and generalized the theory
of utility, effectively purging it of the substantive form specified by
Bentham. Political economists Stanley Jevons (1911), Leon Walras (1926),
and Carl Menger (1871), among others, altered Bentham's formulations to
focus on the satisfaction to be obtained from the possession of individual
goods (see Stigler, 1975: 220-241, for an overview). Among these early
theorists the concept of diminishing marginal utility (holding that the
increment of utility gained from possession of an added unit of a good
declines as the number of units possessed increases) was deemed a general
law of utility. Furthermore, most early marginal utility theorists held total
utility of the individual to be derived from simple summation of the utilities
gained from the goods possessed by the individual.
This formulation, however, ignored complementary and substituting
goods, which, when consumed, change the utility to be gained from
consumption of other goods. This problem led theorists to specify a more

26

POLICY ANALYSIS AND ECONOMICS

general individual utility function that would allow for utilities to be

affected by such relationships among goods (Edgeworth, 1881). The
general utility function eliminated the need to assume diminishing marginal utility for consumption of individual goods in order to assure equilibrium
in economic trades; indeed, diminishing marginal utility was shown to be
neither a necessary nor sufficient condition for achievement of equilibrium
conditions (Stigler, 1965: 126). Still later development of utility theory,
based on the concept of indifference schedules, took a further leap toward
generality; the conception of individual utility itself that had underlaid the
development of utility theory was abandoned as too "metaphysical" and
restrictive. 1
Utility theory has thus been marked by a progressive development away
from the substantively grounded, limited theory of utility proposed by
Jeremy Bentham; successively, interpersonal comparisons of utility, additive individual utility functions, the requirement of diminishing marginal
utility, and even the "metaphysical" conception of utility itself have been
shed in favor of increasingly general propositions.
Despite the increasing generality of the theory, the prescriptive "legislative" role of utility theory envisioned by Bentham has been preserved.
That prescriptive role is embodied in the concept of economic efficiency.
The normative heart of the concept of efficiency (derived straight from
Bentham) holds that a social system or policy ought to be designed to
maximize the satisfaction of individual wants, subject to certain limitations
on the analyst's ability to specify what constitutes an "improvement" in
overall want satisfaction. Thus efficiency analysis, like Bentham's calculus
of pleasure and pain, is based solidly on an individualistic and utilitarian
view of human activity. It is assumed at the outset that the system is made
up of individuals and that each individual has an ordinally ranked set
of wants or "preferences." "Society" is thus strictly an aggregation of
individuals, and social welfare is no more or less than the aggregation
of the welfare of the individuals that make up the society.
The chief limitation confronting the efficiency analyst is the premise that
each individual holds a possibly unique schedule of preferences; without
omniscient capabilities it is impossible to determine whether satisfaction of
one individual's wants adds more to total well-being (or social welfare)
than would satisfaction of some other individual's wants. Rather than
confront the problem of the incomparability of individual utilities directly,
the criterion of efficiency seeks a way around this problem.
Simply stated, an individual's well-being can occur either (a) without
reduction (and perhaps with improvement) in the well-being of others or
(b) with a concomitant decrease in the well-being of others. In the latter

CONTINUING CONTROVERSIES IN POLICY ANALYSIS

27

case, due to the assumed incomparability of individual utilities, it is

impossible to determine whether society was a whole is made better off by
the change. In the former case, when one person's gain is not offset by the
losses of others, the sum of the individuals' welfare can unambiguously be
said to have increased. This represents a clear-cut increase in the efficiency
with which wants are satisfied.
According to this strict criterion of efficiency, dubbed the Pareto
optimality criterion, if at any time any individual can be made better off
without reducing the well-being of others, the system is not operating
strictly efficiently. There exists no uniquely efficient allocation of resources; beginning from any suboptimal distribution of wealth, movement
toward efficiency can be achieved by distributing the added increment of
well-being in countless ways among the members of society. Thus many
Pareto superior solutions exist, differing according to the distributions of
goods achieved.
Unbending application of the strict Pareto optimality criterion would
impose severe restrictions on government action: No policy, no matter how
beneficial it might be to the community as a whole, could be taken that
diminished the well-being of even one individual. In the extreme, even an
action required to maintain the viability of the community (say, maintenance of law and oder) would violate the criterion if any person, on
the basis of his or her own assessment of the effect of the action on his or
her utility, determined that the action would leave the person worse off
than would no action at all. But clearly much of government action does
just that.
The "potential Pareto principle" allows redistributions that increase net
welfare such that those who gain from the redistribution could compensate
those who lose, restoring the losers to their prior level of well-being, while
the winners retain enough of their gains to be better off than they would
have been without the redistribution (Hicks, 1939). In this case individuals'
utilities need not be compared: as long as a common value of the good
redistributed is available, such as the money value of the gains and losses to
the individuals involved, it is possible to determine just how much of the
improvement to the gainers must be given in compensation to the losers to
assure that no one is worse off than they would have been without the
redistribution. Thus the criterion specifies that a system that results in an
allocation of goods from which subsequent redistribution could result in net
increases in social well-being is not efficient. Furthermore, in the comparison of policies, the relevant criterion is: Which policy option serves to
create the largest net gain in social well-being?
The potential Pareto criterion is the central normative standard in

28

POLICY ANALYSIS AND ECONOMICS

contemporary public policy analysis. Texts in policy analysis, and curricula

in the major schools of public policy in the United States, widely invoke
efficiency as the most prominent, and least ambiguous, of the guides for
public policy analysts (Jenkins-Smith, 1990; Friedman, 1987; Stokes, 1986;
Weimer and Vining, 1989).

3. Normative Critiques of Contemporary Policy Analysis

Drawing on its origins in welfare economics, policy analysis would provide
objective analysis and advice; that is, for a given policy option, the results
of analysis would reflect the maximized sum of objectively measured and
aggregated net benefits for the individuals constituting society. Thus, in
an important sense, analysis based on welfare economic principles is
antipolitical; it is designed to remove social decision making "from the
tumult of politics to the domain of putatively scientific, dispassionate
inquiry" (Nieman, 1984). Policy analysis is to be a vehicle whereby
political conflict over public policy is reduced. This element of contemporary policy analysis reflects, at least in part, the stance of the prophets of the
"end of ideology," contending that the withering of ideological differences
will lead to a commonality of ends, wherein only means will remain dispute
(Bell, 1960; Bell and Kristol, 1965). According to Edith Stokey and
Richard Zeckhauser (1978: 261),
one objective of descriptive analysis is to narrow areas of disagreement in policy
disputes.... Policy disagreements would lessen-and perhaps vanish-if we
could predict with certainty the safety consequences of the breeder reactor, or
the cost of annual upkeep of clay [tennis] courts, or whether a special shuttle bus
for the elderly would be heavily used.

Analysis can thus reduce political conflict by moving the debate from
argument about values-about which "men can only ultimately fight"
(Friedman, 1953: 5}-to discussion of predictions about means, which can
be resolved through analysis.
But critics contend that the antipolitical mode of policy analysis may
serve to distort public policy by focusing on the end result of policy
(maximizing net social welfare) to the exclusion of focus on the procedure
by which policy choice is made. That procedure is held to be of intrinsic
value in democratic societies, independently of the end result of policy.
Policy analysts, the critics argue, typically justify process
[e]ither in purely formal, positivist terms or in terms of a superior tendency to
maximize aggregate satisfaction in the end, rather than in terms intrinsic to the

CONTINUING CONTROVERSIES IN POLICY ANALYSIS

29

process itself in its constitutive function of defining substantive human roles,

rights and relationships and structuring their evolution over time. [Tribe, 1972:
82]

Perhaps most notable, the legitimacy accorded decisions reached

through sanctioned procedures may be largely independent of the substantive content of policy. Cognizant only of end results, analysts may fail to
capture an essential element of value. In this view the value of a policy
choice cannot be ascertained without considering the procedure from
which it is derived.
Equally troubling to these critics, the very application of techniques of
the contemporary policy analysis to values subject to discontinuous preference orders serves to erode those values. Kelman (1981) argues that widely
held social rights or decision procedures are withheld from the run of more
common questions by excluding them from the benefit-cost calculus (also
see Okun, 1975, and Brown, 1990). To attempt to impute a dollar value to
such "specially valued things" eliminates their special status, thus reducing
their value. "Cost-benefit analysis thus may be like the thermometer that,
when placed in a liquid to be measured, itself changes the liquid's temperature" (Kelman, 1981: 38). In this view, analysis not noly distorts human
values but destroys them as well.
Critics are concerned that contemporary policy analysis may erode more
than end values, however. An increased reliance on the essentially static
and passive measures of individual preferences, as typically conducted for
benefit-cost analysis, diminishes or eliminates the role of active debate and
expression in a public forum as a means to develop and channel citizen
preferences (see, e.g., Cummings et aI., 1986; Brookshire and Coursey,
1987). Critics argue that application of the efficiency criterion is predicated
on the existence of a universe of individual preference functions that may
be tapped in order to "discover" optimum policies. The role of the analyst,
using shadow prices to inform willingness-to-pay analysis, is to determine
the value of particular options based on the given distribution of tastes.
This vision of democracy, with its emphasis on accurate reflection of
individual tastes in public policies, radically departs from the conception of
the process of democracy-and participation in particular-as important to
the formation of tastes, the legitimization of public choice, and the
full development of the political individual (Pateman, 1970; Benello and
Roussopoulus, 1971). At bottom, the emphasis on process in democracy
is based on the contention that the deliberative process is more than a
simple summation of tastes. It is a process by which individuals develop
policy preferences through exposure to the preferences of others and
reasoned discourse, or through the competition for votes and necessity

30

POLICY ANALYSIS AND ECONOMICS

for compromise. Preferences are formed, as well as expressed, as they

are identified through political processes.
The policy analytic techniques employed to measure citizen preferences
depart from the process views of democracy in a number of ways. Benefitcost analysis, for example, commonly employs market prices, or estimated
shadow prices, to calculate the value of benefits produced or resources
used in public policy. Taking a more direct approach, many major cities
now incorporate annual or biennial surveys of citizen opinions on tax and
expenditure issues as a routine part of policy development (Hess, 1980;
Stipak, 1980). Thus indirect measure through price or price estimates, or
direct solicitation of preferences through surveys, supplements or replaces
more traditional processes of politics.
How have proponents and practitioners responded to this critique?
Quite broadly, they have argued that the critics misconstrue the role to be
played by the analyst in the policy process. According to Allan Williams
(1977: 519-545), rather than precluding the processes of politics, the use of
analytic techniques can usefully serve as but one "loop" in the iterative
process of arriving at decision. Thus policy goals are specified and provided
to analysts, who then determine what policy options best achieve these
goals. Options in hand, the analysts submit their results to policy-makers
who again assess goals in light of necessary means. Thus analysis plays a
constructive and informative role in the policy process.
Furthermore, empirical research on the applications of policy analysis
indicates that analysis has become well integrated into the policy process,
with analysis employed by many sides in the political debate (JenkinsSmith and Weimer, 1985; Jenkins-Smith, 1990). As policy issues arise,
individuals, interest groups, or public agencies tend to mobilize analytical
resources to define the problem, define the causes, and specify probable
policy solutions. Others, whose interests may be harmed by the proposed
solutions, mobilize additional analytical resources to redefine the problem,
identify other causes, or propose "better" policy solutions. In the process
analysis becomes employed on two (or more) sides to the debate. Thus
analysis is often integrated into the policy process, rather than supplanting
it as the critics charge.
4. Public Versus Private Values

One controversy cen(ers on the use of prices, either actual or inferred, as

indices of citizen preferences for public policies. Market prices result from
a myriad of private, individual choices regarding consumption and produc-

CONTINUING CONTROVERSIES IN POLICY ANALYSIS

31

tion of the good in question. In using those prices, critics contend, the
analyst wrongly presumes that the citizen values goods exchanged in purely
private transactions identically with the value of those things in public use.
Thus, Kelman (1981: 38) contends, policy analysts "insidiously" assume
that "there should be no difference between private behavior and the
behavior we display in public life." Social values that for some reason are
not expressed in private behavior are excluded from the calculus of public
decision, and therefore the valuations reached through benefit-cost analysis are flawed. 2 More importantly, the use of private preference and
behavior reflected in price as a guide for public decision would seem to lock
public decision into the pattern set by private behavior; rather than
exploring what values ought to be served, public policy would mimic
existing private behavior. The formative role of politics as a shaper of
public values is thus eroded.
The use of citizen surveys as a device for policy formation is similarly
susceptible to attack for its exclusion of the formative qualities of process.
Surveys are akin to snapshot photographs; a well-designed survey, asking
the right questions, may measure the preferences of the respondents at the
time of the survey and, if not methodologically flawed, adequately reflect
the likely responses of the broader population. Surveys are passive measures, however, for which a sample of citizens' answers are solicited to
policy questions about which the respondents mayor may not have
devoted significant thought and reflection. The important presumption
underlying the use of surveys in policy formulation is that coherent
preferences on policy issues actually exist, preferences that are susceptible
to measurement and that are reasonably stable. Critics have argued that
these presumptions are in error for a broad array of public issues, including
the most important ones (Bogart, 1976; Achen, 1975; Farlie, 1978).
On most complex policy issues, the nature of public opinion may be
better described as a "natural force," like "currents of the air or ocean,
constantly changing in their contours and directions" (Bogart, 1967: 334).
Uncertainty, lack of information, the compelling novelty of the survey
situation, and question construction and phrasing often make "public
opinion" on policy issues unintelligible if not misleading. The point of this
line of criticism is that the well-formed public opinion presumed to exist
will in many cases be absent; the development of stable and intelligible
public preferences occurs through the workings of the political process
-the public forum for raising, defining and debating public issues.
Surveys used in absence of this process, or surveys used to replace this
process, will fail to find-or what is worse, will fabricate-what does not
exist.

32

POLICY ANALYSIS AND ECONOMICS

5. Inadequacies of Utilitarianism
What may be the most persistent controversy surrounding the application
of welfare economic bases of policy analysis takes aim at the philosophical
roots of policy analysis. Critics of policy analysis argue that contemporary
policy analysis itself is a purveyor of a well-developed ideological
perspective-that of individualistic utilitarianism (Bromley, 1990; Brown,
1991). Further, utilitarianism itself has been under extensive attack within
the field of moral philosophy (Gorovitz, 1977; Smart and Williams, 1973).
One serious reservation concerning utilitarianism derives from its lack of
distinction among values-push-pin and poetry are accorded equal status. Within the general critique of utilitarianism, the difficulty in handling
such concepts as rights and obligations, which have traditionally found
bases outside the language of preference and utility, has received significant attention; can lying, unjustly punishing the innocent, or even repression of a small minority be justified on the basis that doing so increases
aggregate well-being?
These criticisms have led Steven Kelman (1981: 34) to exclaim that "it is
amazing that economists can proceed in unanimous endorsement of costbenefit analysis as if unaware that their conceptual framework is highly
controversial in the discipline from which it arose-moral philosophy."
Peter Brown (1991: 6) adds that "utilitarianism leaves too much up for
grabs.... It is compatible in principle with severe violations of human
rights if total utility will be maximized by the rights violation."
Another cause for concern among critics of the utilitarian basis of policy
analysis is grounded in perceived flaws in methods used for the aggregation
and comparison of values. Ideal policy analysis would express the value of
the costs and benefits of policy options in a common metric, preferably
dollars. These valuations would be accomplished through determination of
individuals' willingness to pay to obtain the benefit or willingness to accept
to bear a loss. In the ideal this approach would permit the comparison of a
wide range of policies over all affected values. The policy option that
maximizes aggregate well-being (dollars) is selected, and those who lose
would (ideally) be compensated through lump-sum transfers. Critics contend, however, that this approach ignores the structure of values; it ignores
the possibility that certain values may have a lexicographical ordering in
which some minimum level of good x must be obtained before any x will be
traded for good y. Put differently, not amount of y can compensate for the
loss of good x below some threshold. According to Laurence Tribe (1972:
87-88), in such circumstances the "very concept of proper distribution (of
x and y) must now be defined not with respect to the single homogeneous
entity called 'wealth' but with respect to the enjoyment of these rights as
such." Where such orderings hold, Tribe argues, the principle of com-

CONTINUING CONTROVERSIES IN POLICY ANALYSIS

33

pensation of the losers by the winners fails; discontinuous preference

orderings for rights cannot be reduced to the same "undifferentiated mass"
of total welfare as continuous preference orderings for marketable goods
and services.
Although recent utilitarian theorists have sought to find a natural basis
in enlightened self-interest for rights and obligations (e.g., derived from
within Rawls' "veil of ignorance"), utilitarianism remains a contentious
basis for prescriptive political theory (Harsanyi, 1985; Rawls, 1971; Barry,
1973; Bluhm, 1983). Richard Zerbe (1988) makes a very cogent argument
that use of a conceptual device like the veil of ignorance can permit
utilitarianism to handle concepts like rights and justice. Though many
rights cannot be derived from within the language of utilitarianism, the
application of efficiency criteria can be made consistent with a specification
of those rights. For example, once property rights are defined, benefit-cost
analysis could impute a value of zero to stolen goods. Critics (e.g.,
Bromley 1990; Brown, 1991), however, maintain that the utilitarian bases
of efficiency analysis are incorrigible and, when employed as the normative core of policy analysis, will seriously distort analysis, mislead policymakers, and ultimately diminish the use of policy analysis.
In a recent and illustrative debate, for instance, a critic of the techniques
of policy analysis charged that the underlying philosophy of benefit-cost
analysis-individualistic utilitarianism-would permit what are considered
to be morally reprehensible acts in the interest of maximizing utility
(Kelman, 1981). A defender of those techniques responded
[the critic] ... hints that "economists" are so morally numb as to believe that a
routine cost-benefit analysis could justify killing widows and orphans, or abridging freedom of speech, or outlawing simple evidences of piety or friendship. But
there is nothing in the theory or practice of cost-benefit analysis to justify that
judgment. Treatises on the subject make clear that certain ethical or political
principles may irreversibly dominate the advantages and disadvantages capturable by cost-benefit analysis. [Solow, 1981: 41]

In other words, the economists and policy analysts, like other people, take
recourse to "other ethical or political principles" that may overwhelm the
results of benefit-cost analysis.

6. Hard Versus Soft Values

One of the most common criticisms of contemporary policy analysis is
that, in the zeal to quantify, analysts tend to underemphasize those costs
and benefits that are intangible or that are external to market valuation
(Tribe, 1971, 1972; Kelman, 1981: 33-40; Paris and Reynolds, 1983: 118123). Though it may eventually prove technically feasible, valuation of

34

POLICY ANALYSIS AND ECONOMICS

nonmarketed goods, such as the value of a pleasant view or of the right to

life (Zeckhauser, 1975), has proved difficult, and consensus regarding
appropriate techniques elusive. Critics of analysis contend that these
"soft" values tend to be ignored, whereas easily measured values are
emphasized; when attempts at valuation are made, that valuation may
be more akin to "subjective guesses" than objective science (Paris and
Reynolds, 1983: 120).
7. The Problem of "Standing" in Policy Analysis
Taking a somewhat different tack, one particularly compelling critique of
benefit-cost analysis points out that, for analysis to be applied to a specific
policy, the analyst must decide which preferences count, that is, which
preferences have "standing" in benefit-cost analysis (MacRae and Whittington, 1986 and 1988; Brown, 1986). In the abstract, benefit-cost analysis
should count all benefits and costs associated with the policy under
scrutiny, regardless of who bears those benefits and costs. In point of fact,
however, analysts routinely reject many preferences as illegitimate;
the lost value to criminals of illegal activity is usually not counted as a cost
in analyses of criminal justice policies3 nor are the benefits or costs of
immigration policies for illegal aliens. Debates over how to count benefits
and costs for future generations, fetuses, nonhumans (e.g., endangered
species), inanimate objects (e.g., rivers), and others are indicative of the
breadth of the standing problem (Weimer and Vining, 1989: 78-79).
For contemporary policy analysis, the standing issue presents a particularly thorny problem: How are analysts to decide which preferences to
include or exclude? Given that some preferences are socially (and politically) unacceptable, the analyst must make such a decision, and making
that decision forces the analyst outside of the (relatively) safe domain of
efficiency analysis. Some criteria of justice or equity must be invoked, at
least implicitly. The great danger here is that the analyst's criteria of justice
will masquerade behind a facade of objectivity and thus be opaque to
decision makers, the public, and even the policy analyst.
8. Efficiency and Institutional Transactions
One of the most important recent contributions to the debate over application of welfare economic principles to policy analysis is made by Daniel
Bromley (1989a). Bromley argues that a critical class of economic and
political decisions are made regarding the choice sets of individuals and

CONTINUING CONTROVERSIES IN POLICY ANALYSIS

35

groups and that those choice sets are specified and modified through
"institutional transactions" such as the specification of rights and obligations. Two of Bromley's points are of particular relevance here: that the
specification of efficiency is often predicated on existing institutional
designs, which are themselves based on prior institutional transactions or
accident; and that the traditional welfare economic dichotomy of policy
change resulting in either efficiency gains or redistribution of wealth is too
restrictive. These arguments, according to Bromley, undermine the conventional uses of welfare economic concepts in policy analysis.
For Bromley, public policy is best understood as a matter of defining the
choice sets of individuals (or classes of individuals). To specify a new right
(such as the right not to breath cigarette smoke-filled air on passenger
airplanes) is to redefine the choice sets of individuals. What was once a
nonright for the non-smoker, and a privilege for the smoker, becomes a
right (to clean air) for the former and an obligation (not to smoke in certain
areas) for the latter. More broadly, choices are made within a complex of
institutions (defined as conventions and rules or entitlements) that establish what is a right, obligation, privilege, or non-right. Institutional transactions take place when these institutions are altered.
Applied to the concept of efficiency, Bromley makes a simple point:
What is "efficient" is dependent on the existing set of institutions. First,
Bromley argues (following Kahneman and Tversky, 1983) that individuals
and decision makers tend to treat expected losses and gains differently.
With respect to the domain of gains, risk aversion predominates: Individuals are prone to take the "sure thing" even if the expected value is less
than some riskier choice. Regarding the domain of losses, however, risk
acceptance is prevalent: The chooser tends to gamble on that option that
might result in least loss, even if the expected losses of the option are
greater than those for some other, less risky option. That means that
decision makers will generally accept riskier strategies, and perhaps incur
greater costs, to avoid a probabilistic loss than they will to obtain a
probabilistic benefit.
Bromley points out that, for many policy issues, what counts as a gain
or loss is dependent on existing institutions. For example, reductions in
existing environmental or safety standards are generally considered losses
to those currently "entitled" to them, whereas increases in such standards
would be considered a gain. Given that gains and losses tend to be valued
differently, efficiency analysis of a change in standards may be dependent
on the nature of the existing set of standards, entitlements, rights, and
obligations. Thus the current structure of institutions is critical for specification of costs and benefits. Finally, Bromley points out that existing
institutions themselves result from prior "institutional transactions" or

36

POLICY ANALYSIS AND ECONOMICS

even historical accident. Why should such institutions be considered sacrosanct? And, if they are not sacrosanct, the analyst would be remiss not
to consider the efficiency implications of a policy from the standpoint
of different institutional arrangements. Thus Bromley's charge is that
efficiency analysis, as conventionally employed, is inherently "conservative" in that it imputes a special (and unjustifiable) status to the existing
constellation of institutions.
Bromley (1989b) carries his point further by arguing that the conventional dichotomy between policies that promote efficiency gains (which
enlarge the pie) and transfers of wealth (which redivide the pie) is
underspecified. Given this dichotomy, institutional transactions can be
either efficient or merely redistributive. Bromley argues that this conventional view ignores the possibility that a reallocative policy may be socially
efficient. What might appear to be "merely redistributive," as when mine
employees demand greater expenditures for mine safety, may be stimulated by a shift in social preferences toward greater mine safety. If so, the
reallocation from mine owners to mine workers, though redistributive,
might be socially efficient. More generally, demands for policy change can
result from changes in collective attitudes about the weights and valuations
of elements of a nation's full consumption set. Note that changes in these
weights and valuations may be incremental (e.g., as a society gradually
becomes more concerned about environmental quality) or abrupt (as when
the franchise is enlarged, adding new citizens who carry different weights
and valuations). In either case, the change results in demands for institutional changes that might add to net social utility and should not be
branded as "rent seeking." Bromley dubbs such changes "reallocations of
economic opportunity." Such policies should be kept conceptually distinct
from reallocative policies that are not driven by changes in collective
attitudes or relative prices, policies that Bromley calls "reallocations of
economic advantage."
One import of Bromley's reformulation of the efficiency and redistributive effects of policy change is that it requires the policy analyst to attend to
the shape and content of the social welfare and utility functions. In his
view, the analyst cannot legitimately escape the need to compare welfare
changes across groups, or across types of gains and losses, by reference to
the strict Pareto criterion of efficiency. To do so is to ignore and denigrate a
vast domain of potential social welfare gains that public officials can and do
reference in the formulation of public policy. In addition, to rely on
conventional efficiency analysis to dodge the issue is to accept the implicit
sanctification of the existing institutional order as it defines choice sets and
the distribution of income. If economists are to influence the policy
process, according to Bromley, they must explicitly grapple with the claims

CONTINUING CONTROVERSIES IN POLICY ANALYSIS

37

of various groups on that process and assess the origin of the claim for a
new institutional transaction; Is a claim based on a shift in the social
welfare function, changes in social utility relations, or alterations in the
social weighting of welfare across different groups? Or is it merely the
seeking of "economic advantage" by one group over others? And, in
conducting such analyses, the analyst must not fall into the trap of
accepting as sacrosanct the existing institutional order.
In advocating this shift, Bromley is asking the analyst to step from under
the protective umbrella of the efficiency criterion. Remember that Pareto
optimality and potential Pareto optimality provide the analyst with a
way around the analytically difficult problem of comparisons of individual
utility. Bromley's argument is that the umbrella was a mere fiction anyway
and that analysts must come to grips with the real engines of institutional
change if they are to make a contribution to public policy.
This is a tall order. First, shifting the focus of policy analysis to
specification of the social welfare function, or changes in social utility
relations, will result in a much wider range of plausible analytical claims on
behalf of a wide array of policy positions. Social welfare functions are
notoriously hard to specify and may well be subject to constant change.
Furthermore, to assess the efficiency of policy change from the standpoint
of potential as well as existing institutional arrangements raises a practical
and a philosophical problem. The number of potential institutional
arrangements is, of course, infinite; which ones will the analyst employ in
assessing the social value of change? Given that not all can be used, the
choice of the subset to use will be important because (as Bromley demonstrates) the assumed institutional arrangements may well determine the
analytical outcome. Is that better than the use of the existing institutional
order? That leads to an underlying philosophical question: Should policy
analysts accord the results of past and current political and economic
processes (i.e., the existing package of institutional arrangements) as
legitimate and therefore a sound base for policy comparisons? Or should
the analyst invent, and compare, an array of alternative social arrangements for purposes of assessing the merits of policy change? This issue, it
seems to me, is akin to the debate between the Burkean conservative, who
accords great legitimacy to the existing order, and the welfare state liberal,
who sees the need for extensive public intervention to achieve social ends.
9. Summary: The Dynamic of Controversies In
Polley Analysis

Although far from exhaustive of the controversies focused on applications

of contemporary policy analysis, the list provided here includes those that

38

POLICY ANALYSIS AND ECONOMICS

have been or (in this author's view) will be, the most vigorous. To
summarize the major themes: The standard uses of efficiency criteria in
policy analysis are criticized because they exclude what critics contend is
the essential role in citizen preference formation performed by the policy
process. Exclusion of that step confuses public and private choice, excludes
the possibility of reasoned debate over what values ought to be inculcated
(i.e., What kind of people should we become?), and in many cases inhibits
the very formulation of "public opinion." In addition, analysts are forced
to make decisions on the standing of preferences, requiring choice among
criteria of justice-a task for which the analyst is not prepared through
training in contemporary policy analysis. Furthermore, it is charged, the
techniques of analysis distort existing preferences, reducing complex relations within and among human values to a structureless mass. "Soft" or
intangible values are ignored or underemphasized when traded off against
more tangible values. Finally, analysts routinely employ an inappropriate
base of comparison in assessing the benefits and costs of policy change
(existing institutional arrangements) and have underspecified the engines
of policy change (acknowledging only efficiency and redistributive
reasons). The result is that many policy initiatives that may result in net
social improvement are classified by analysts as mere "rent seeking."
For some of these critics-most notably Peter Brown and Daniel
Bromley-the flaws of contemporary policy analysis have resulted in the
mininal influence of policy analysis in the formulation of public policy.
According to Brown, analysis has not penetrated because it has been
imbued with a different philosophical underpinning than the ones typically
employed in personal and public choice. For Bromley, attempts to provide
only dispassionate advice in the interest of maximizing net utility, ignoring
the interests and influence of the various actors in the political process as
they translate into social utility, relegates policy analysts to an insignificant
role in the political process. Other critics-such as Kelman and Tribeargue that we are in danger of using too much analysis and that the
tel:hniques of analysis must be transformed to overcome "the ideological
structure of particular errors that ... have flowed from the basic axioms of
policy analysis and related techniques" (Tribe, 1972: 106).4 For these
critics, analysis distorts public preferences in formulation and expression,
thereby threatening the effective operation of the democratic process.
How have defenders of contemporary policy analysis responded? The
broadest rejoinder to the critics has been that, though there may be
limitations or flaws in the techniques of analysis, tough decisions regarding
the use of scarce resources must be made, and such techniques as benefitcost analysis are useful aids to such decisions. As three federal analysts
argued in defense of benefit-cost analysis in a recent debate:

CONTINUING CONTROVERSIES IN POLICY ANALYSIS

39

[W]e do not dispute that cost-benefit analysis is highly imperfect. We would

welcome a better guide to public policy, a guide that would be efficient, morally
attractive, and certain to ensure that government follow the dictates of the
governed. [Butler, Calter, and Ippolito, 1981: 41-42]

But, they add, no such better guide is evident. Says another defender of
benefit-cost analysis:
[Benefit-cost analysis] is not the way to perfect truth, but the world is not a
perfect place, and I regard it as the height of folly to react to the greater (though
still incomplete) rigour which [benefit-cost analysis] requires of us by shrieking
"1984" and putting our heads hopefully back into the sand (or the clouds).
[Williams, 1977: 543]

The critics of analysis are thus viewed as hypercritical, rejecting the

promising techniques of contemporary policy analysis for fear of overblown imperfections, and as having nothing to offer in their stead.
In a nutshell, proponents, critics, and defenders seem to evoke different
visions of the role analysis is to play. Following Bentham, the proponents
of policy analysis have tended to be reformists. The foundations of policy
analysis militate toward changes in political institutions to better map the
social welfare function into public policies. At the very least, the techniques of policy analysis are seen as a significant corrective for the practice
of politics as usual. For their part, the critics of contemporary policy
analysis take the proponents (or perhaps, the more extreme proponents) at
face value; what would the world look like, they ask, should contemporary
policy analysis be fully implemented without restraint in public policymaking? Their conclusion is that such a world would be dreadful. Defenders, finally, scoff at the critics' concern, knowing full well that analysis and
analytical techniques are far from singularly influential in public decisions.
They point out that policy analysis competes with other institutions,
theories, and methods for policy-making that-though inferior-predominate now and will probably in the future as well. The defenders need not
take the critics seriously because analysis is seen by the defenders as but
one all-too-insignificant piece among many in the Rube Goldberg policy
machine; in the defenders' view, the critics' excesses stem from erroneously believing that economists are "morally numb" and from "having their
heads in the sand (or clouds)."

Notes
1. As Vilfredo Pareto, cited in Stigler (1965: 126), wrote early in this century:

The entire theory [now) rests on a fact of experience, that is to say, on the determination
of the quantities of goods which constitute combinations that are equivalent for the

40

POLICY ANALYSIS AND ECONOMICS

individual. The theory of economic science thus acquires the rigor of rational mechanics; it
deduces its results from experience, without the intervention of any metaphysical entity.

2. Strictly speaking, this is a mistaken view of benefit-cost analysis. A perfect analysis

would capture these excluded social values and would treat them analogously to externalities
of market price. While conceptually straightforward, this would prove difficult to accomplish
in practice, and therefore the criticism may have more weight in practice than it does in
theory.
3. A noteworthy exception is Long, Mallar, and Thornton (1981).
4. Tribe suggests several reforms: (1) elimination of the attempt to provide objective
analysis in favor of a more impassioned and self-consciously value-laden approach and (2) the
adoption of a "subtler, more holistic and more complex style of problem solving, ... relying at
each stage on the careful articulation of a wide range of interrelated values and constraints
through the development of several distinct 'perspectives' on a given problem, each couched
in an idiom true to its internal structure rather than translated into some 'common denominator'" (1972: 107). Thus Tribe has in mind a radical transformation of policy analysis.

References
Achen, Christopher, "Mass Political attitudes and the Survey Response," American Political Science Review, 69(3) (Fall 1975), 1218-1231.
Barry, Brian, The Liberal Theory of Justice (Oxford: Clarendon Press, 1973).
Bell, Daniel, The End of Ideology (New York: Free Press, 1960).
Bell, Daniel, and Irving Kristol, The Public Interest 1(1), (1965), 3-5.
Benello, C. George, and Dimitrios Roussopoulus, eds., The Case for Participatory
Democracy (New York: Grossman, 1971).
Bentham, Jemery, Fragment on Government, in The Collected Works of Jeremy
Bentham, J. H. Bums and F. Rosen, (eds.) (London: Althone Press, 1977).
- - - , Introduction to the Principles of Morals and Legislation, J. H. Bums
and H. L. A. Hart, (eds.) (London: Athlone Press, 1970).
Bentham, Jeremy, Works of Jeremy Bentham, vol. 3 (Edinburgh: Tait, 1843).
Bluhm, William T., "Liberalism as the Aggregation of Individual Preferences:
Problems of Coherence and Rationality in Social Choice" (presented at a
Conference on the Crisis of Liberal Democracy, SUNY-Geneseo, 1983)..
Bluhm, William T., Theories of the Political System (Englewood Cliffs, N.J.:
Prentice-Hall, 1973).
Bogart, Leo, "No Opinion, Don't Know, Maybe No Answer," Public Opinion
Quarterly 31 (Fall 1967), 331-345.
Bromley, Daniel W., "The Ideology of Efficiency: Searching for a Theory of Policy
Analysis," Journal of Environmental Economics and Management 19:1 (July
1990), 86-107.
- - - , Economic Interests and Institutions: The Conceptual Foundations of Public
Policy (Oxford: Blackwell, 1989a).
Bromley, Daniel W., "Institutional Change and Economic Efficiency," Journal of
Economic Issues 23:3 (1989b), 735-769.

CONTINUING CONTROVERSIES IN POLICY ANALYSIS

41

Brookshire David and Don Coursey, "Measuring the Value of a Public Good: An
Empirical Comparison of Elicitation Procedures," American Economic Review
7:4 (September 1987), 554-566.
Brown, Peter, "Ethics and Education for the Public Service in a Liberal State,"
Journal of Policy Analysis and Management 6:1 (Fall 1986), 56-68.
Brown, Peter, "The Failure of Market Failures," (Forthcoming, Journal of SocioEconomics).
Butler, Gerard, John Calter and Pauline Ippolito, "Defending Cost-Benefit Analysis," Regulation 5(2) (March/April 1981), 41-42.
Cummings, Ronald, David Brookshire, and Don Coursey, Valuing Environmental
Goods: An Assessment of the Contingent Valuation Method (Totowa: Rowman
and Allenheld, 1986).
DeLeon, Peter, Advice and Consent (New York: Russell Sage, 1989).
Dumont, Etienne, Bentham's Theory of Legislation, trans. by Charles Atkins
(London: Oxford University Press, 1914).
Edgeworth, Francis, Mathematical Psychics (London: Kegan Paul, 1881).
Farlie, Henry, "Galloping Toward Dead Center," The New Republic 178(14),
Issue 3300, April 8, 1978), 18-21.
Friedman, Lee, "Public Policy Economics: A Survey of Current Pedagogical
Practice," Journal of Policy Analysis and Management 6:3 (Spring 1987),
503-520.
Friedman, Milton, Essays in Positive Economics (Chicago: Chicago University
Press, 1953).
Gorovitz, Samuel (ed.), Utilitarianism with Critical Essays (Indianapolis: BobbsMerrill, 1977).
Harsanyi, John c., "Rule Utilitaranism, Equality, and Justice," Social Philosophy
and Policy 2(2) (Spring 1985), 115-127.
Hayek, F., The Constitution of Liberty (Chicago: University of Chicago Press,
1960).
Hess, F. William, "Listening to the City: Citizen Surveys," Urban Affairs Papers 2
(Summer 1980), 1-9.
Hicks, J. R., "The Foundations of Welfare Economics," The Economic Journal
49:196 (December 1939), 696-712.
Jenkins-Smith, Hank, Democractic Politics and Policy Analysis (Pacific Grove,
Calif.: Brooks/Cole, 1990).
Jenkins-Smith, Hank, and David Weimer, "Analysis and Retrograde Action: The
Case of Strategic Petroleum Reserves," Public Administration Review 45:4
(July-August 1985), 485-494.
Jevons, Stanley, Theory of Political Economy, 4th ed. (London: Macmillan, 1911).
Kahneman, Daniel, and Amos Tversky, "Choices, Values and Frames," American
Psychologist 39:4 (1983), 341-350.
Kelman, Steven, "Cost-Benefit Analysis: An Ethical Critique," Regulation 5:1
(January-February 1981), 34-42.
Long, David, Charles Mallar, and Craig Thornton, "Evaluating the Benefits and
Costs of the Job Corps," Journal of Policy Analysis and Management 1: 1 (Fall
1981), 55-76.

42

POLICY ANALYSIS AND ECONOMICS

MacRae, Jr., Duncan, and Dale Whittington, "Assessing Preferences in CostBenefit Analysis: Reflections on Rural Water Supply Evaluation in Haiti,"
Journal of Policy Analysis and Management 7:2 (Winter 1988), 246-263.
MacRae, Jr., Duncan and Dale Whittington, "The Issue of Standing in CostBenefit Analysis," Journal of Policy Analysis and Management 5:4 (Summer
1986), 665-682.
Menger, Carl, Grundsatze der Volkswirtschaftslhre (Vienna: Braumuller, 1871).
Mill, John S., "Utilitarianism," in Utilitarianism, Liberty, and Representative
Government (New York: E. P. Dutton, 1910).
Nieman, Max, "The Ambiguous Role of Policy Analysis and an Illustrative Look at
Housing Policy" (paper presented at the annual meeting of the Western Political
Science Association, Sacramento, California, 1984).
Okun, Arthur, Equality and Efficiency: The Big Tradeoff (Washington, D.C.:
Brookings, 1975).
Oser, Jacob and William C. Blanchfield, The Evolution of Economic Thought, 3rd
ed. (New York: Harcourt, Brace, Jovanovich, 1975).
Paris, David and James Reynolds, The Logic of Policy Inquiry (New York:
Longman, 1983).
Pateman, Carole, Participation and Democratic Theory (Cambridge: Cambridge
University Press, 1970).
Rawls, John, A Theory of Justice (Cambridge, Mass.: Harvard University Press,
1971).
Riker, William and Peter Ordeshook, An Introduction to Positive Political Theory
(Englewood Cliffs, N.J.: Prentice-Hall, 1973).
Smart, J. J. c., and Bernard Williams (eds.), Utilitarianism: For and Against
(Cambridge, Mass.: MA: Cambridge University Press, 1973).
Solow, Robert M., "Defending Cost-Benefit Analysis," Regulation 5(2)(March/
April 1981), 40-41.
Stigler, George, The Development of Utility Theory in Essays in the History of
Economics. (Chicago: University of Chicago Press, 1965).
Stipak, Brian, "Local Government's Use of Citizen Surveys," Public Administration Review 40 (September/October 1980), 521-525.
Stokes, Donald, "Political and Organizational Analysis of the Policy Curriculum,"
Journal of Policy Analysis and Management 6:1 (Fall 1986), 45-55.
Stokey, Edith, and Richard Zeckhauser, A Primer for Policy Analysis (New York:
W. W. Norton, 1978).
Tribe, Laurence, "Policy Science: Analysis or Ideology?" Philosophy and Public
Affairs 2 (Fall 1972), 67-110.
Tribe, Laurence, "Trial by Mathematics: Precision and Ritual in the Legal Process," Harvard Law Review 84 (1971), 1361-1365.
Walras, Leon, Elements d'economie politique pure (Paris: Pichon & DurandAuzais, 1926).
Weimer, David L. and Aidan R. Vining, Policy Analysis: Concepts and Practice
(Englewood Cliffs, N.J.: Prentice-Hall, 1989).
Williams, Allan, "Cost-Benefit Analysis: Bastard Science? And/Or Insidious

CONTINUING CONTROVERSIES IN POLICY ANALYSIS

43

Poison in the Body Politick?" in Robert Haveman and Julius Margolis, eds.,
Public Expenditure and Policy Analysis, 2nd ed. (Chicago: Rand McNally,
1977), 519-545.
Zeckhauser, Richard, "Procedures for Valuing Lives," Public Policy 23(4) (Fall
1975), 419-464.
Zerbe, Jr., Richard O. "The Ethical Foundations of Benefit Cost Analysis,"
unpublished manuscript, University of Washington, 1988.

BEYOND EFFICIENCY:
ECONOMICS AND
DISTRIBUTIONAL ANALYSIS
Alphonse G. Holtmann

1. Introduction
Distributional analysis, as discussed in this chapter, refers to the distribution of well-being among the members of society. The discussion first
illustrates that efficiency, as defind by most contemporary economists, is
directly related to social well-being and that questions of equity and
efficiency are often inextricably combined when dealing with public policy
issues, requiring that the policy-maker and the analyst deal with both
dimensions of these issues. This is true whether the policy under consideration involves pure transfers within the economy or resource allocation to a
productive project. Suggesting possible means for policy analysts to evaluate programs that influence social welfare, however, requires a better
perspective on the issue of social welfare, defining terms and putting
philosophical, political, and theoretical developments in historical context.
To understand the role of the public policy analyst in dealing with the
distribution of well-being within a society, the chapter will briefly explore
both the nature of modem welfare economics and the recent history of
the interaction between government policy in market economies and the
theoretical work in applied welfare economics or policy analysis. Welfare
45

46

POLICY ANALYSIS AND ECONOMICS

economics is the normative branch of the discipline that deals with

reaching judgments on what ought to be, whereas positive economics is
that branch of the discipline that deals with explaining how the economy
functions. Deciding whether more funds should be spent on a Head Start
program is within the domain of welfare economics. Explaining why oil
prices went up during the last several decades is an exercise in positive
economics.
Individual well-being is the foundation of modern welfare theory, which
is not an unreasonable foundation for a theory of welfare in a society that
relies on markets to distribute most goods and services. The principal
notion of welfare theory is that individuals can order choices available to
them in a fashion that will allow them to gain the greatest satisfaction from
the choices available, usually given some constraints on the choices. If a
fully informed individual, given a choice between remaining at home or
working at the prevailing wage, decides to remain at home, one would
conclude that the individual's welfare or utility would not be increased by
working. The chapter will discuss not only the role of individual satisfaction or choice as the foundation of welfare economics but some of the
difficulties of measuring welfare. National income in this example will be
smaller because of the individual's choice, but well-being will not have
decreased according to individualistic welfare theory. Generally, however,
one would expect a positive relationship between national income and
social welfare.
Individual well-being is associated with social or group well-being
through the Pareto principle: For a given initial distribution of endowments, society's allocation of resources is Pareto optimal if no one can be
made better off without making someone else worse off. Conversely, the
allocation of resources in both production and consumption is not Pareto
optimal if someone's utility can be increased with no decrease in anyone
else's utility. One allocation is said to be Pareto superior to another
allocation if someone's satisfaction is increased by the new allocation
without a decrease in anyone else's satisfaction. The ethical and technical
assumptions underlying these principles are the subject matter of other
chapters in this volume, but it is important to note at least two aspects of
the principles that are important for public policy purposes. First, even for
those who find the Pareto ethic acceptable, it is likely to be too limiting
because many would find it satisfactory to help some in society at the
expense of others. For instance, this author is perfectly content to tax
the extremely wealthy to give to the extremely poor. Second, there are
different Pareto optimal allocations for different initial distributions of
endowments. Economists, unable to make interpersonal comparisons of
satisfaction among individuals, are unable to compare Pareto-optimal

BEYOND EFFICIENCY: ECONOMICS AND DISTRIBUTIONAL EFFICIENCY

47

allocations (sometimes these allocations are called efficient allocations or

Pareto-efficient allocations), leading to the dilemma that economists cannot scientifically choose, at least by means of Pareto principles, among
redistributions of resources that help some at the expense of others.
Apparently, theoretical welfare economics and the Pareto principles are
not helpful in allowing policy analysts to determine the ideal distribution of
well-being in society. But, agreeing that policy analysts cannot determine
the ideal distribution of resources is not to say that they cannot be helpful
to policy-makers whose decisions inevitably involve such matters. To
understand the important role of Pareto efficiency in policy analysis, it
helps to chronicle the development of modern welfare theory.

2. Modern Welfare Economics

As Paul Samuelson (1965) has indicated, economists since the time of
Adam Smith and the Physiocrats have believed that a free competitive
market leads to desirable consequences for the population. The simple
logic behind this conclusion, closely related to our earlier principle that
fully informed individual choice reveals welfare orderings, is that individuals trade when they can gain from the trade. Pareto (1929) and Barone
(see Hayek, 1935), and later Lerner (1934), Hotelling (1938), and Bergson
(1938), were able to show, under the appropriate assumptions, that a
competitive market equilibrium would result in a Pareto efficient allocation
of resources: Given the initial endowments, it is not possible to rearrange
goods and services or production so that someone is helped without hurting
someone else. The proof of this theorem is elegantly derived in Samuelson's Foundations of Economic Analysis (1965).
This close connection between the Pareto efficiency principle and the
conpetitive market no doubt partly accounts for the domination of applied
welfare economics by this criterion of welfare. But there is more to the
story of theoretical developments and our current state of art in policy
analysis. In the 1930s and 1940s a large literature developed-some of
which has been alluded to--that not only shows the relationship between
a competitive equilibrium and Pareto efficiency but rigorously proves all
the theorems of modern welfare economics. The analysis provides the basis
for Samuelson's chapter on welfare economics in 1947 (1965) and J. de V.
Graaff's classic volume Theoretical Welfare Economics ten years later
(1967). One theorem of this theory is that any ethical individualistic
optimum can be attained through lump-sum taxes and subsidies, allowing
the individuals to trade at fixed prices.
Thus after World War II when the major market economies were just

48

POLICY ANALYSIS AND ECONOMICS

reviving on a peacetime basis, welfare theory had seemingly provided a

clear guide for the policy-maker. Whatever the ethical basis for favoring
different individuals in society, lump-sum transfers could in principle be
used to reallocate endowments and a competitive market could then
achieve a Pareto-efficient allocation that would be optimal in the context of
that endowment. It is difficult, however, to suggest practical tax and
subsidy schemes that are lump-sum in nature, that is, taxes and subsidies
that do not affect the market prices facing the consumer.
In addition, as indicated earlier, economists cannot provide the
appropriate ethical standard, sometimes called a welfare function, to index
transfers among individuals. In fact, as Kenneth Arrow (1963) has shown,
it is impossible for any fair collective choice process to provide consistently
an appropriate ordering to resolve any interpersonal difference while at
the same time satisfying certain reasonable axioms. For example, majority
voting is not a satisfactory vehicle for making all social choices because
situations arise where there is no equilibrium. That is, in some cases where
more than two options arise (helping the poor, the middle class, or the
rich, for example), any selected policy can be overturned by another policy
selected by the majority. Such cyclical majority problems may be fairly
common, implying that one needs to know how the vote was reached
before being confident that the outcome reflects the will of the majority.
That economists cannot scientifically specify a welfare function that
selects among the Pareto-efficient allocations is so discouraging to many
welfare analysts that they recommend that economists would be best
advised to spend their efforts on positive economics (Graaff, 1967: 170).
Perhaps many share that belief, but others like the notion that efficiency
and equity can in principle be separated when dealing with public project
evaluations. In any event, many economists became deeply involved in
policy analysis.
With the state of general welfare economics basically settled by the late
1940s, economists in the early 1950s turned their attention to the assumptions that provide the basis for Pareto optimality. That is, even if one
accepts the initial endowments in a static environment and assumes that
the economy is perfectly competitive, it is only possible to reach a
Pareto-efficient allocation of resources when all the assumptions underlying our individual welfare index are met. One crucial assumption is that
the utility function (the mechanism for ordering choices) of each individual
is independent of that of other individuals in the society.
An extreme case of the interdependence of individual utility functions
was explored by Paul Samuelson (1954) in his work on public goods. In
that classic work, Samuelson shows that certain goods provide satisfaction

BEYOND EFFICIENCY: ECONOMICS AND DISTRIBUTIONAL EFFICIENCY

49

to all individuals in society at the same time (national defense being an

example) and that these types of goods will not be efficiently provided in a
competitive market. Of course, the pure public good that enters everyone's
utility function when it enters anyone's utility function is an extreme case
of the problem of external benefits or costs. While your consumption
of medical services, educational services, or gardening services no doubt
increases your satisfaction, they may also increase my satisfaction. I may
be less apt to become ill when you take care of your health; I may find it
less costly to transact my business because you are better educated; and I
may simply enjoy observing your garden.
In all these cases of externalities and in the case of pure public goods,
a competitive market will not assure a Pareto-efficient allocation of
resources. Hence, government might be able to improve social welfare, in
the Pareto sense of helping someone without hurting others, with appropriate taxation, subsidies, or provision of services. In addition, there are other
characteristics of markets that might prevent the economy from achieving a
Pareto-efficient allocation of resources. Economies of scale in production
might lead to monopoly, which is generally inconsistent with Pareto
optimality. Uncertainty, lack of information, and the lack of appropriate
insurance markets might lead to a failure of the market to achieve a
Pareto-efficient result. Though the list of factors that cause market failure
(Bator, 1958), as the failure of the market to achieve a Pareto solution has
come to be known, could be greatly expanded, it is noteworthy that
theoretical analysts turned to these issues of efficiency after 1950, providing some of the incentive for policy analysts to also dwell on efficiency
issues for several decades.

3. Public Expenditures and Human Capital

Before discussing the accomplishments and limitations of the broad spectrum of studies concerning public policy evaluation, three other events
must be noted-events basically concomitant with the theoretical developments on market failure already discussed-that influenced the nature of
public policy studies undertaken in the United States. Two contributions
at the conceptual level influenced policy evaluation around 1960. In 1959
Richard Musgrave published his path-breaking book, The Theory of Public
Finance, which although not unduly long (628 pages) was almost encyclopedic in its coverage of modern public finance, including the theory of
public expenditures. In that volume Musgrave divided the public household, as he called it, into the allocation branch, the distribution branch,

50

POLICY ANALYSIS AND ECONOMICS

and the stabilization branch, stressing the proposition mentioned earlier

that, at least conceptually, one cound separate the issues of equity and
efficiency-the major concerns of the distribution and allocation branches,
respectively. While clearly indicating that there was no simple normative
rule for evaluating public policy, Musgrave suggested that, as an analytical
tool, the allocative operations of government policy could be separated
from the distributional operations. This separation of equity and efficiency
was not new, being implicit in the basic theorem of welfare economics that
lump-sum transfers of income could be used to attain distributional equity,
but Musgrave's book formalized the approach for policy analysts during
this period.
It was natural that the early work in project evaluation was aimed at
services already in the government sector; river development, public
health programs, and education were early candidates for public policy
analysis. Most of these early analyses concentrated on the efficiency
aspects of the evaluation of a project of investment and not the equity
aspects. In a detailed analysis of U.S. river projects, John Krutilla and
Otto Eckstein state, "We take for granted that political, judicial, and other
social processes in democracy tend to adjudicate disputes involving the
distribution of income" (1958: 49). These scholars explicitly adopted the
procedure later suggested by Musgrave.
Of course, some of the early project evaluations considered distributional questions, but these studies clearly attempted to determine the
impact of the distribution of costs or benefits on efficiency rather to analyze
the equity dimensions of the distribution of costs and benefits per se.
For example, Burton Weisbrod (1962), in "Education and Investment
in Human Capital," discusses the possible distribution of benefits from
education among different groups through time and by location, showing
that all beneficiaries are not likely to be able to express appropriately their
demand for education. Although the equity considerations are recognized,
his article emphasizes that external effects can cause efficiency problems,
resulting in over- or underinvestment in education. As this example
demonstrates, there is a distinction between looking at the distribution
of costs and benefits because they influence an efficient allocation of
resources and looking at the costs and benefits because they influence
the equity of the allocation of resources. In any case, it is reasonable to say
that most of the early focus in project evaluation was on efficiency.
Note that the preceding discussion concerns investments in education,
thus relating to the second event that influenced the nature of project
evaluation for the last several decades. In 1961 T. W. Schultz of the
University of Chicago extolled the virtues of particular expenditures on

BEYOND EFFICIENCY: ECONOMICS AND DISTRIBUTIONAL EFFICIENCY

51

people that could be considered as investments: Expenditures made today

result in a stream of future benefits for the individual and perhaps society.
This notion of human capital, though not completely new, captured the
imagination of a large number of economists who quickly set out to
measure the returns to expenditures on health, education, migration,
information, and training in the economy. In 1970 the annotated bibliography Cost-Benefit Analysis and the Economics of Investment in Human
Resources listed 389 papers that had been written on some aspect of policy
analysis in this area (Wood and Campbell, 1970).
This interest in investments in people was of major importance to policy
analysts for two reasons. First, these types of investments in people might
be rather important to the economic growth process. After all, the United
States, the richest country in the world, makes large investments in its
people and thus provides a model for other countries. Second, expenditures on health, education, and other forms of human capital are thought
not only to influence growth in developing economies but also to have
particular consequences for the achievement of Pareto efficiency in any
market economy. Thus the question naturally arises as to whether Pareto
efficiency requires that developed economies like that of the United States
devote more resources to these investments than is generated through
market forces. It is common, even today, to hear individuals speak of
underinvestment in human resources. That is, it might be possible to make
everyone better off by increasing these types of expenditures. Finally,
there is the view that these investments in people could reduce poverty.
Many of these human services are, of course, already provided by
government, and few people question the government role in supporting
them. So the issue arises as to the amount of service provided by government and the level of government-federal versus state and local
government-that should support these services. Should, for example,
Head Start be initiated, and how should it be financed? Today's policy
suggestions concerning child day care and aid to dependent children have
much of the same rationale as these earlier projects. Again, although some
analysts address the issue of the distributional consequences of these
investments, the emphasis of most studies is on the efficiency issue.
4. The Great Society

Many of the Great Society programs brought to fruition during the

Johnson administration (though some had been conceived in the Kennedy
presidency) reflected a fundamental change in the nature of the federal

52

POLICY ANALYSIS AND ECONOMICS

government's social programs. Although these administrations followed

those since the New Deal in expanding and modifying the money transfer
programs, particularly those associated with Social Security, many new
in-kind transfer programs were initiated during this period-a time, as we
indicated earlier, of great activity in project evaluation and analysis. Many
of these programs, of course, continued in the presidencies of Nixon, Ford,
Carter, and Reagan, but the surge in the growth of in-kind transfer
programs took place in the Johnson administration, when in-kind transfers
grew in nominal dollars by 84.6 percent, a growth rate that declined to 4.8
percent by the first term of the Reagan administration (Browning, 1987).
In-kind transfer programs, such as Head Start, Upward Bound, Manpower
Development and Training, Area Redevelopment, and neighborhood
health centers, initiated during the Johnson era were related to investments in the poor, not purely income transfers to the poor, though the
latter were also expanded in the War on Poverty. Many of these programs
had great appeal to policy analysts, particularly economists, because the
programs emphasized investments in people as had been highlighted by
Schultz and given much theoretical and empirical content by such people as
Gary Becker (1964), Jacob Mincer (1962), Burton Weisbrod (1962),
George Stigler (1962), and numerous others and because many of these
investments might be justified on allocative efficiency grounds. Without
stretching the analytical framework, one can imagine that investments in
education, health, and training of the poor might represent a Pareto
improvement in resource allocation, particularly if any adverse changes in
the income distribution might be addressed through other government tax
and subsidy programs. Thus, to a large extent, it was easy to analyze these
programs purely in terms of efficiency: Could some be made better off
without hurting others? Though some analysts mention the possibility of
distributional problems arising and though some commentators stress the
actual redistribution of income that takes place through these programs
because beneficiaries of the programs are not identical to those who pay
the taxes to support them, most studies stress efficiency and concerns with
equity arise somewhat later.
In retrospect, it is perhaps not difficult to see why policy analysts, at
least in the United States, were largely concerned with the efficiency of
in-kind programs in their evaluations. First, the transfer programs under
Social Security emerged when modern welfare economics was rejecting the
idea that interpersonal utility evaluations were possible, making a scientific
evaluation of these transfer programs appear vacuous. In addition, the
Social Security programs seemed to reflect a social consensus that was
intact from the Roosevelt administration until most recently. Second, the

BEYOND EFFICIENCY: ECONOMICS AND DISTRIBUTIONAL EFFICIENCY

53

in-kind programs stressed during the War on Poverty had the charm of
representing the newly rediscovered notion of investing in people, and as
investments, the rates of return to these programs could actually be
calculated and compared to the returns on other investments in the
economy, including other government investments in highways, flood
control, and the like-the redistributional consequences of all government
investments being left to other government devices or, perhaps, considered
as desirable fringe benefits of the programs. In any case, contributions to
national product were most often the primary criterion of early project
evaluations, though certainly nonmeasurable cost and benefits were also
stressed-the possible role of education in promoting democracy, for
example. Finally, expenditures on people, as Schultz had stressed, were
like other government investments, promoting growth in the economy over
time and allowing everyone to be better off in the future.
Although the preceding synopsis has concentrated on the United States
experience, other market economies like those of Great Britain and France
seemed just as concerned with efficiency in their formal project evaluations
and, at the same time, were initiating programs that involved extensive
government determination of resource allocation and income distribution
through government involvement in activities ranging from the provision
of medical care to generating electric power (Nelson, 1964).

5. Linking Equity and Efficiency in Policy Analysis

As has been shown, the particular emphasis given to the Pareto efficiency
concept by the project analyst results from complex interactions between
the apparent needs of public decision makers and the analytical tools
available during a particular period. Nevertheless, it has been recognized
for some time that an appropriate welfare function to aggregate individual
utilities would allow the social planner to select among Pareto-efficient
allocations, thus permitting the evaluation of projects on the basis of their
contribution to social welfare (Samuelson, 1965).
The problem is, and always has been, that scholars cannot agree on
the appropriate welfare function because such a welfare function implies
interpersonal utility comparisons. Though agreement on the appropriate
social welfare function for society is unlikely, this notion of a social welfare
function can be very useful in evaluating public projects because it permits
the combining of the equity dimensions of a project with those of efficiency
in a manner that allows thoughtful consideration of the issues. Also, it is
perhaps worth noting at this point that Arrow's possibility theorem does

54

POLICY ANALYSIS AND ECONOMICS

not rule out the existence of a social welfare function but only indicates
that fair social decision rules cannot consistently produce such a ranking
device.
A social welfare function allows us to translate gains and losses of
satisfaction of different individuals into a measure of social well-being,
allowing us implicitly to make interpersonal utility comparisons. To consider the first of two extremes, we might have a Benthamite welfare
function of the form W = !.U i , which implies that social welfare is
increased when anyone's satisfaction is enhanced and where W is social
welfare, U i is an individual's utility function, and i goes from 1 to n
members of the population, or a Rawlsian welfare function of the form
W = min(U i , ... ,un), which implies that social welfare is improved only
when satisfaction of the person who is worst off is improved. Or, even
more generally, consider a social welfare function of the isoelastic form W
= [1/(1 - p)]!.[(UY-P], which represents the Benthamite form for p = 0
and approaches the Rawlsian form as p ~ 00 (Ballard, 1988: 1028). In the
case of the isoelastic welfare function, aWlaUi = (Ui ) - P, implying that
larger values of p give greater weight to individuals with lower utility levels.
If we were to visualize these welfare functions in terms of social
indifference between the utility of two individuals, the Benthamite case
would generate social indifference curves that are straight lines with a slope
of -1, implying that society is indifferent as to the particular individual
who gains from a program. Alternatively, Rawlsian social indifference
functions form right angles, implying that helping anyone but the least well
off person leaves society on the same indifference function. These examples, of course, form the polar cases of more general isoelastic functions
with less extreme values of p that allow interpersonal comparisons of
individuals' satisfaction and some degree of substitutability of satisfaction
among the individuals.
This approach of integrating efficiency analysis and distributional analysis has been recognized for some time and is becoming increasingly
important in evaluating projects that stem from concerns for either allocative efficiency or distributional justice. Burton Weisbrod (1970) was one of
the earliest proponents of using a social welfare function approach as a
means of making practical policy decisions. In his work Weisbrod suggests
a social welfare function of the form W = aYI + bY2 + ... + mYn, where W
is social welfare, Y is the amount of individual income, and the coefficients
reflect the marginal increase in welfare associated with a change in an
individual's income. Thus in the two-person case the marginal rate of
substitution of income between the two people would be dy 1ldY2 = -bla,
allowing us to evaluate transfers of income among individuals.
The idea behind Weisbrod's analysis, however, is not simply to proclaim

BEYOND EFFICIENCY: ECONOMICS AND DISTRIBUTIONAL EFFICIENCY

55

weights for interpersonal comparisons of income but rather to try to

determine these weights from government rankings of projects that had
been determined through benefit-cost analysis. As he indicates, the virtue
of this procedure is in its confronting the weights used by the decision
makers with those held implicitly by others in the society and in its ability
to forecast the government's choices among programs. Although Weisbrod
illustrated his approach using actual data from federal river projects, the
approach has not been empirically applied, to this writer's knowledge.
Perhaps the limitation of this approach is that it forms a social welfare
function that comes out of the political process, rather than use a social
welfare function to help form the decision maker's views. Of course, to the
extent that there is an iterative process, this might not be a powerful
objection. In any case, this effort was one of the earliest to incorporate
distribution theory into project evaluation in a practical way, though in a
still earlier paper Stephen Marglin (1962) had fully outlined the technical
procedures for introducing distributional criteria into project evaluation.
As Marglin shows, the utilities of those included in the welfare function
can be differentiated by almost any characteristic: race, region, sex,
income group, level of disability, and the like. This type of disaggregation
according to group membership is often of great interest to policy-makers.
As shall be seen, a number of positive distributional studies concern
themselves with the relationship of benefits accruing to whites versus
blacks, young versus old, and so on, suggesting that the distribution of
costs and benefits to certain groups of individuals is crucial for meaningful
public policy. For any particular project, a given division of the population
may appear sensible, although disaggregation by different population
characteristics makes comparisons of different programs more difficult.
Using a social welfare function in a recent analysis of U.S. metropolitan
school fiscal reform, Robert Inman (1978) analyzed six proposed fiscal
reform programs in New York State for three different social welfare
specifications-utilitary (pro-middle class), Rawlsian (pro-poor), and
equal school spending-to evaluate the programs. Among other things, he
finds that there is no dominant reform proposal and concludes that the
legislative process may collapse into a pro-poor versus pro-middle class
debate. Inman suggests that a possible solution to the debate may be
compromise plans that satisfy legislators favoring the poor and legislators
favoring the middle class. This analysis, then, provides policy-makers with
information concerning the implications of their choices for distributional
equity, which allows them to make decisions that can be defended to their
constituencies.
Social welfare functions have also been used in recent studies of the
efficiency costs of redistribution policies. Whereas much of the earlier

56

POLICY ANALYSIS AND ECONOMICS

discussion focused on introducing distributional considerations into analyses that are focused on efficiency, the efficiency costs of pure transfer
programs have become an increasing concern to the social policy analyst.
Programs that attempt to transfer resources from the more affluent to the
poorest often must take more than a dollar from the affluent to transfer a
dollar to the poor. In an early analysis Arthur Okun (1975) refers to this
problem as the leaking bucket, but more recent analysts, refer to it as the
marginal efficiency cost of redistribution (Browning, 1987).
In a 1988 study Charles Ballard uses a computable general equilibrium
model to calculate the marginal efficiency cost of redistributing income
through two programs: pure cash transfer programs and wage subsidy
programs. Assuming that taxes finance the redistribution programs, he
finds that a pure cash transfer program implies that the more affluent are
made worse off by between $1.50 and $2.30 for every dollar transferred to
the poor and that a wage subsidy program tends to allow a dollar for dollar
transfer from the high-income group to the low-income group. Though, as
he indicates, the costs of a cash transfer are high, they are not as high as
earlier estimates of these costs made by Edgar Browning (1987). Nevertheless, how is one to evaluate the cost of the transfer? This is the very
question that Weisbrod raises when he suggests finding a social welfare
function to determine the marginal rate of substitution of income among
individuals.
Indeed, Ballard uses a social welfare function of the isoelasticity form
mentioned earlier to determine critical values of p for which such transfers
would be acceptable, finding that one does not have to be anywhere near
the Rawlsian position to accept the costs of these seemingly expensive cash
transfer programs.
Although a number of other studies that employ a social welfare
function as a means of making interpersonal comparisons among groups in
society can be cited, these few examples are sufficient to make clear that a
social welfare function is a useful tool for applied policy analysis. And
though it is not likely that any particular social welfare function will gain
wide acceptance, alternative specifications of the welfare function provide
the policy-maker with important information in making decisions that will
actually result in gains to some and costs to others.

6. Fairness
Though the preceding social welfare approach deals with questions of
equity in a formal and explicit manner through interpersonal utility comparisons, it is inconsistent with, or at least not in the tradition of, the

BEYOND EFFICIENCY: ECONOMICS AND DISTRIBUTIONAL EFFICIENCY

57

individualistic approach that is the foundation of the Pareto principles and

much of the modern welfare theory discussed earlier. Recent work on
fairness (Baumol, 1986; Varian, 1990), however, raises the possibility that
analysts can combine Pareto notions of efficiency with notions of fairness
to make judgments about efficiency and equity that result in the individual
being better off and judging the result as fair. As expected, there will be
situations in which an allocation results in a Pareto improvement but is not
judged to be fair.
Like the Pareto approach to efficiency, which starts with the "weak"
value judgment that a reallocation that helps some and does no harm to
others is a Pareto-superior allocation, the theory of fairness or equity starts
with the value judgment that an allocation is fair if no agent prefers
another's allocation to his of her own (Varian, 1990). At the outset, it
should be noted that it is not the utility or satisfaction from another's
allocation that one must not envy, but the allocation itself, so there are no
interpersonal utility comparisons involved.
Now to see how this framework functions, consider attempting to
distribute a fixed bundle of goods among n individuals. Theorists in this
area like to use cakes to represent the bundle, which is a matter of taste, of
course. According to the preceding definition, n equal shares of the cake
will be fair because no one envies anyone else's share. But, of course, this
division of the cake need not be Pareto efficient. Will trade result in a
Pareto-efficient allocation?
To see the problem here, consider three individuals who have shared
two cakes, one with nut topping and one with raisin topping, and assume
that the shares are such that each person received one third of each cake.
The allocation is fair because no one would select another's allocation.
Now suppose two of the individuals like nuts and one likes raisins, and
suppose the raisin lover trades his nut-topped piece to one of the nut lovers
for a raisin-topped piece. The allocation will be Pareto optimal, but it will
no longer be fair. The one nut lover will envy the new share of the other
nut lover.
Suppose, however, that the raisin lover cuts his nut-topped piece in half
and trades one-half to each nut lover for each nut-lover's raisin topped
piece. Now note that the resulting allocation is Pareto optimal and fair
according to our definition. In fact, generalizing from this example, it can
be shown that a competitive equilibrium from an equal division must be a
fair allocation. And, of course, we know that a competitive equilibrium is a
Pareto-optimal equilibrium.
This simple cake example (Baumol, 1986) highlights the problem that
economists sometimes encounter when making policy recommendations.
That is, it is clearly possible to make a recommendation that results in

58

POLICY ANALYSIS AND ECONOMICS

a Pareto improvement that is deemed unfair by at least one appealing

measure of fairness. Thus society may be unwilling to make Paretosuperior moves that are deemed unfair by some members of society.
The idea of fairness, as defined in this relatively recent literature and
exemplified by the work of Hal Varian (1990) and William Baumol (1986),
seems to show great promise for being helpful to public policy analysts in
attempting to show the relationships between efficiency and equity in any
particular recommendation. Baumol (1986: 104-107), for example, shows
that fairness (lack of envy) could require overcompensation for air pollution victims, resulting in a welfare loss to society when damages are equally
spread among the population and people reduce their pollution protection
because of the compensation. His purpose is to show that policy analysts
will have to find means of making trade-offs between equity and efficiency
in many public policy decisions because it will be impossible to achieve
maximum efficiency and fairness simultaneously. He also shows that the
concept is useful in analyzing public utility pricing. In any case, it appears
that it might be possible to integrate concepts of fairness (lack of envy)
with Pareto efficiency in a manner that will help policy analysts make
better recommendations to policy-makers.

7. Positive Economics
Policy-makers are likely to be interested in whether a particular policy will
actually do what it is intended to do, who will gain and who will lose from
it, and its implications for growth. Consequently, positive economics can
play an important role in distributional analysis by giving policy-makers the
information they need to make their own distributional assessments.
Thus, even if one eschews the tools of welfare economics, one can still
be helpful to the decision-maker. For example, in recommending criteria
for choosing alternative means for redistributing income, Burton Weisbrod
(1970) mentions six criteria that might be considered when evaluating
a redistributive program: administrative cost, target efficiency, allocative
efficiency, consumer versus taxpayer sovereignty, flexibility over time, and
nondemeaning benefits. (Because stigma may be attached to receiving a
particular government benefit, delivery systems that identify recipients are
sometimes thought to be demeaning; i.e., people might refuse to use food
stamps because redeeming the stamps identifies the user as poor.) Such a
set of criteria might be appropriate for any program undertaken with
government support. Putting aside the question of the particular criteria to
be used, the general idea is that the analyst can point to various aspects of a
program that are of interest to the policy-maker and then allow the

BEYOND EFFICIENCY: ECONOMICS AND DISTRIBUTIONAL EFFICIENCY

59

political process to determine the final outcome. Further, Weisbrod suggests (1978) that information gleaned from questionnaires on the attitudes
of groups affected by a government program would help policy-makers in
reaching decisions.
Much of the recent work in applied welfare economics falls into this
positive genre. For example, James P. Smith and Finis Welch (1989)
recently reported a comprehensive study of the relationship between black
and white males' incomes over the last 40 years, concluding that substantial
progress has been made in narrowing the income gap between these
groups. Obviously, there are dimensions of well-being other than income,
and different individuals will determine progress in closing the income gap
as too fast or too slow, but it is important in evaluating our racial equality
programs to know what is happening. In a similar vein, this author's work
finds that reducing the major diseases in the United States would result in
gains to both white and black males but would not change their relative
economic positions, as measured by the present value of future income
(Holtmann, 1973a). Unlike the Smith-Welch analysis, which provides
positive information on the results of programs aimed at reducing inequality, my results offer information on the impact of inequality from programs
meant to accomplish other ends.
Of course, this type of descriptive information is helpful in evaluating all
types of programs and is becoming increasingly important in evaluating
social programs. As an example of this trend, a recent book by Sheldon
Danziger and Kent Portney (1988) reports 12 essays that deal with what
gets distributed to whom and from whom. These essays cover topics
ranging from the distributional consequence of state tax reform to an
analysis of child support. Thus this volume shows that the long history of
generally positive analysis of government programs can be applied to both
distributional and efficiency questions. Such studies are certain to provide
quantitative measures where only guesses were available before, and they
are likely to provide new methods for measuring difficult-to-measure costs
and benefits from programs. There is therefore a continuing need for this
type of analysis.
8. Organizational Form: Efficiency and Fairness

Until relatively recently, evaluations of projects to implement either

money transfers or in-kind services did not consider the type of organization that would deliver the service or make the transfer. If, for example,
more health care or education is deemed necessary on efficiency grounds,
the type of organization that might deliver the service is completely outside

60

POLICY ANALYSIS AND ECONOMICS

the realm of the analysis. Spurred by. the analysis of Weisbrod (1975) and
Hansmann (1980), the question of organizational efficiency has been raised
in several recent studies, particularly those dealing with not-for-profit
organizations. Thus the project analyst has one more dimension of the
project to deal with: to recommend the appropriate organizational motif to
best accomplish the efficiency or equity goals of a particular program.
The recent literature voices two major concerns about the appropriate
organization for service delivery: Do not-for-profit organizations emerge in
a market economy because of some sort of market failure? And, how does
one choose among government, for-profit, and not-for-profit organizations
in industries where all three exist? Again, efficiency is the main concern in
most studies of not-for-profit firms, with earlier studies suggesting that
these institutions are inefficient and more recent studies suggesting that
they enhance efficiency (Holtmann, 1988). But little has been done to
relate questions of equity, distribution of services, and fairness to the
analysis of not-for-profit organizations.
This area appears to be fruitful for future research. For example, are
different groups of people served by government entities, for-profit firms,
and not-for-profit organizations when they exist in the same industry? The
nursing home and child care industries are examples where all three types
of organizations exist. Are trade-offs between efficiency and equity made
differently in different organizational settings? We know, for example, that
there are more often waiting lists in not-for-profit nursing homes than in
for-profit homes. Also, donations are a major source of revenue for many
not-for-profit organizations, but not for government or for-profit organizations. Do these differences relate to perceived differences in efficiency
among organizations or perceived differences in the way organizations
handle issues of equity?
It seems, then, that many program evaluations should include recommendations concerning the appropricate delivery system to accomplish the
goals of improving efficiency and redistributing income (Nelson, 1977).

9. Beyond Static Efficiency and Equity

To summarize, it is clear that project evaluation and policy analysis have
emerged from a classic beginning based on modern welfare economics and
Pareto efficiency to become a modern discipline of policy analysis (see
Friedman, 1984). This discipline continues to rely on a strong theoretical
foundation associated with more refined concepts of welfare functions and
fairness, allowing the analyst to better interpret empirical estimates of the
consequences of social programs for both efficiency and equity. In addi-

BEYOND EFFICIENCY: ECONOMICS AND DISTRIBUTIONAL EFFICIENCY

61

tion, the large number of data sets that contain information concerning
various social programs and the increasing power of computers allow
better empirical understanding of these social programs. Thus policy
analysis has not stagnated at the level of efficiency analysis but has moved
on from there at an encouraging rate.
Unresolved issues remain, however. Most of our present policy analysis deals with efficiency and equity in a purely static environment. Dynamic
questions are more difficult to handle, but they go to the heart of the
matter for some programs (Stiglitz, 1979). Of course, in the classic purely
competitive economy, prices would exist that allow consumers and producers to make the appropriate intertemporal decisions. But the dynamics of
both public and private decisions mean that those decisions will affect
generations who cannot be directly involved in them, as we have not yet
found a satisfactory means of evaluating, in terms of either efficiency of
equity, these distributions over time.
Nevertheless, many studies, such as those concerning the influence of
pollution on the environment or those conerning the impact of social
security programs on savings (Feldstein, 1976) and retirement decisions
(Burtless and Moffitt, 1984), have already started policy analysts down the
road to further incorporating growth into analyses. For, in many cases, the
policy analyst might have to consider not only trade-offs between equity
and efficiency in a static sense but also the potential impact on growth and
intergenerational equity of any social program.
Intergenerational equity can be considered within the context of a social
welfare function, as interpersonal equity was considered in section 5,
where the satisfaction of future generations is an argument of the function.
Alternatively, studies of completed projects can be used to suggest the
possible impact of current projects on future generations. For example,
Burton Weisbrod (1971) calculated the rate of return from the successful
development of the polio vaccine, and Holtmann (1973a) calculated the
gain from developing the vaccine at a faster rate, providing some guidance
for evaluation of current research and development projects that will
benefit future generations. Hence, as the analysis of distributional questions expands to include future generations, current analytical concepts
and past experience can supplement new methods of analysis.

10. Conclusions
As policy analysts become more involved with issues of equity or fairness,
their role as scientific evaluators versus advocates will become increasingly
important. Though lack of appropriate data and lack of agreement on the

62

POLICY ANALYSIS AND ECONOMICS

appropriate empirical measures of particular costs and benefits (for example, the lack of agreement on the appropriate discount rate to be used in
project evaluations) presently leave ample room for disagreement on the
merits of any particular public program, evaluations based on efficiency
criteria offer the decision makers seemingly objective evidence. After all,
if all the costs and benefits of a program can be measured accurately, the
project in question either offers a potential Pareto-superior allocation for
society or it doesn't. Of course, a potential improvement in welfare is not a
sufficient condition for adopting a program unless all those injured by the
program are in fact compensated for their losses. But, as we indicated, this
need not be an issue for the analyst who logically separates the efficiency
issues from those of equity.
If the incorporation of distributional analysis into project evaluation
leads to highly subjective recommendations, varying greatly from one
analyst to another, many will abandon the field as some early theorists
abandoned welfare economics. In addition, policy-makers will be less
inclined to rely on one evaluation that can be considered objective,
increasing the cost of project evaluation. However, recent developments in
the theory of equity are rooted in traditional welfare theory, suggesting
that analysts will be cautious as they incorporate distributional analysis into
public policy analysis. Hence, though advances in this area may be slower
than we would like, the integrity of the growing field of public policy
analysis will be enhanced.

References
Arrow, Kenneth, Social Choice and Individual Values (New York: Wiley, 1963).
Atkinson, Anthony, and Joseph Stiglitz, Lectures on Public Economics (New
York: McGraw-Hill, 1980).
Ballard, Charles, "The Marginal Efficiency Cost of Redistribution, American
Economic Review 78:5 (1988), 1011-1033.
Bator, Francis, "Anatomy of Market Failure," Quarterly Journal of Economics
72:3 (August 1958), 351-379.
Baumol, William, Superfairness: Applications and Theory (Cambridge, Mass.:
MIT Press, 1986).
Becker, Gary, Human Capital (National Bureau of Economic Research, 1964).
Bergson, Abram, "A Reformulation of Certain Aspects of Welfare Economics,"
Quarterly Journal of Economics 52:2 (1938), 310-334.
Browning, Edgar, "On the Marginal Welfare Cost of Taxation," American Economic Review 77:1 (March 1987), 11-23.
Browning, Edgar, and William Johnson, "The Trade-off between Equality and
Efficiency," Journal of Political Economy 92:2 (1984), 175-203.

BEYOND EFFICIENCY: ECONOMICS AND DISTRIBUTIONAL EFFICIENCY

63

Burtless, Gary, and Robert Moffitt, "The Effect of Social Security Benefits on the
Labor Supply of the Aged," In Henry Aaron and Gary Burtless, eds., Retirement and Economic Behavior (Washington, D.C.: Brookings Institution, 1984),
pp. 135-171.
Danziger, Sheldon, and Kent Portney, The Distributional Impacts of Public Policies (New York: St. Martins Press, 1988).
Feldstein, Martin, "Social Security and Savings: The Extended Life Cycle
Theory," American Economic Review 66:1 (1976), 77-86.
Friedman, Lee, Microeconomic Policy Analysis (New York: McGraw-Hill, 1984).
Graaff, J. de V., Theoretical Welfare Economics (New York: Cambridge University Press, 1967).
Hansmann, Henry, "The Role of Non-Profit Enterprise," Yale Law Journal 89:5
(1980), 835-901.
Hayek, Fredrick, ed., Collectivist Economic Planning (London: Routledge, 1935),
245-290.
Holtmann, Alphonse, "Theories of Non-Profit Institutions," Journal of Economic
Surveys 2:1 (1988), 29-45.
- - , "The Size and Distribution of Benefits from U.S. Medical Research: The
Case of Eliminating Cancer and Heart Disease," Public Finance 28:3-4 (1973a),
353-361.
- - , "On the Optimal Timing of Research Expenditures," Inquiry 10 (March
1973), 47-49.
HotelIing, Harold "The General Welfare in Relation to Problems of Taxation and
of Railway and Utility Rates," Econometrica 6:2 (1938), 242-269.
Inman, Robert, "Optimal Fiscal Reform of Metropolitan Schools," American
Economic Review 68:1 (March 1978), 107-122.
Krutilla, John, and Otto Eckstein, Multiple Purpose River Development (Baltimore: Johns Hopkins Press, 1958).
Lerner, Abba, "Economic Theory and Socialist Economy," Review of Economic
Studies 2:1 (1934), 51-61.
Marglin, Stephen, "Economic Factors Affecting System Design," in Arthur Maass
et aI., ed., Design of Water-Resource Systems (Cambridge, Mass.: Harvard
University Press, 1962), pp. 129-225.
Mincer, Jacob, "On-The-Job Training: Costs, Returns, and Some Implications,"
The Journal of Political Economy 70:5 Part 2 (October 1962), 50-79.
Musgrave, Richard, The Theory of Public Finance (New York: McGraw-Hill,
1959).
Nelson, James, ed., Marginal Cost Pricing in Practice (Englewood Cliffs, N.J.:
Prentice-Hall, 1964).
Nelson, Richard, The Moon and the Ghetto (New York: Norton, 1977).
Okun, Arthur, Equality and Efficiency: The Big Tradeoff (Washington, D.C.:
Brookings Institution, 1975).
Pareto, Vilfredo, Manuel d'Economie Politique (Paris: Marcel Giard, 1929).
Samuelson, Paul, Foundations of Economic Analysis (New York: Atheneum,
1965).

64

POLICY ANALYSIS AND ECONOMICS

Samuelson, Paul, "The Pure Theory of Public Expenditures," Review of Economics and Statistics 36:4 (1954), 387-389.
Schultz, Theodore, "Investment in Human Capital," American Economic Review
51:1 (March 1961), 1-17.
Smith, James, and Finis Welch, "Black Economic Progress after Myrdal, Journal
of Economic Literature 27:2 (June 1989), 519-564.
Stigler, George, "Information in the Labor Market," The Journal of Political
Economy 70:5 Part 2 (1962), 94-105.
Stiglitz, Joseph, "A Neoclassical Analysis of Natural Resources," in Kerry Smith,
ed., Scarcity and Growth Reconsidered (Baltimore: Johns Hopkins Press, 1979),
pp.36-66.
Varian, Hal, Intermediate Microeconomics (New York: Norton, 1990).
Weisbrod, Burton, "Income Redistribution Effects and Benefit-Cost Analysis," in
Samuel Chase, ed., Problems in Public Expenditure Analysis (Washington,
D.C.: Brookings Institution, 1968), pp. 177-208.
Weisbrod, Burton, "Distributional Effects of Collective Goods," Policy Analysis
5:1 (1978), 67-95.
- - , "Toward A Theory of the Voluntary Non-Profit Sector in a Three Sector
Economy," in Edmund Phelps, ed., Altruism, Morality and Economic Theory
(New York: Russell Sage, 1975), 171-195.
- - , "Costs and Benefits of Medical Research: A Case Study of Poliomyelitis,"
Journal of Political Economy 79:3 (May-June 1971), 527-544.
- - , "Collective Action and the Distribution of Income: A Conceptual
Approach, "in Robert Haveman and Julius Margolis, eds., Public Expenditures
and Policy Analysis (Chicago: Markham 1970), pp. 117-141.
- - , "Education and Investment in Human Capital," Journal of Political
Economy 70:5 Part 2 (October 1962), 106-123.
Wood, Donald, and Harry Campbell, Cost-Benefit Analysis and the Economics
of Investment in Human Resources (Kingston, Ontario: Industrial Relations
Centre, Queen's University, 1970).

BEYOND SELF-INTEREST
Amitai Etzioni

1. Introduction

When asked why he contributed to a beggar, and was this not due to
Christ's commandment, Thomas Hobbes responded that he did so "with
the sole intent of relieving his own misery at the sight of the beggar"
(Losco, 1986: 323). This statement captures a belief held for centuries by
many social philosophers and social scientists, by the public-at-Iarge, and
championed by the neoclassical school of thought: that human behavior
can be explained by self-interest. Altruistic acts, to the extent that they are
recognized at all, are depicted merely as another means of pursuing one's
own interest.
Socio-economists strive to correct this way of thinking with a deontological ethic based on a force beyond self-interest-that of moral and social
causes. This ethic presents arguments and evidence that people are not
driven solely by self-interest but are also significantly directed by their
moral considerations. These moral factors represent a source of action and
valuation distinctly different from, and often conflicting with, self-interest.
Acknowledging this distinction, this chapter shall endeavor to show, allows
for more productive conceptualization, is more predictive, and is ethically
65

66

POLICY ANALYSIS AND ECONOMICS

sounder than self-interest-only approaches. The first section is based on

substantive arguments. The following two sections present evidence of
moral behavior in support of these arguments, both in individual and
public choices. The final section illustrates the policy implications of
including these moral factors in one's analyses.
2. As Reflected in Utility Theory

The self-interested, "Hobbesian" view is currently represented in the

neoclassical school of thought. Neoclassicists hold that individuals seek to
"maximize" one utility: their self-interest, also defined as pleasure, happiness, and consumption. This concept of utility, as developed late in the
eighteenth century largely by Jeremy Bentham and widely used by neoclassicists, holds that "[all] actions are directed toward the gain of pleasure
or the avoidance of pain" (Dyke, 1981: 31). Moreover, many utilitarians
grant moral approval to pleasure: pleasure is "good"; pain is "evil."
An analysis of this school, specifically its theory of economic utility,
uncovers sound conceptual reasons for distinguishing moral causes from
those of self-interest. Clearly, self-interest, or pleasure, accounts for a
good deal of human behavior, and to this extent the concept of utility is
logical, proper, and productive. To assume, however, that all (or virtually
all) behavior is explainable in the narrow terms of self-interest is too
simplistic. Many actions are either pleasurable or moral but often not both.
Many times, pursuing one's pleasure is in conflict with carrying out one's
moral duties (an area to be addressed in more detail later). Furthermore,
ascribing a moral goodness to pursuing one's pleasure represents a gross
value judgment which once removed, allows for the recognition of other
sources of value.
When one contemplates the substance of the term "moral behavior,"
the kinds of acts the term encompasses, one finds that living up to one's
moral obligations, discharging one's duties, and doing "what's right"
evokes a feeling distinctly different from that of indulging one of the
senses, satisfying a personal need, or doing something "fun." Indeed,
many moral acts are explicitly based on the denial of pleasure in the name
of the principle evoked. Doing penance, abstention from premarital sex,
and Ramadan fasting are not what most people consider sources of
pleasure. True, acting in line with one's moral precepts produces a kind of
satisfaction, a sense of moral worth, but it is more the kind one gets when
a hard day's work is done than the pleasure of getting off work early with
full pay.

BEYOND SELF-INTEREST

67

Moreover, although pleasure and living up to one's moral commitments

are not always or necessarily in conflict, very often they do conflict in terms
of their call on resources (the time, energy, and assets dedicated to one
behavior are often required by the other as well, a major factor, for
example, in the life of parents who work outside of the home). Quite
frequently pleasure-seeking acts and moral commitments are also incompatible in terms of the behavior expected and rewarded (e.g., when
managers are under great pressure to increase profits but also not to violate
the rules of the game; using kickbacks or bribes to get more business are
cases in point).
In an attempt to counter challenges that the prevailing concept of utility
is amoral, asocial, and counter to basic, observable behavior, neoclassicists
point to a concept of utility that is different but goes under the same name:
one that contains service to others and a commitment to moral values. A
simple conceptual device is used: The consumption or pleasure by others is
made into the actor's pleasure, hence the "interdependent" utility. Essentially, this concept attemtps to explain pleasure, self-interest, and the
opposite kind of behavior--<:aring for others and for the community-in
the same self-interested terms. To convert altruism into self-interest is,
however, to misunderstand altruism. By assuming self-interest to be the
foundation for all behavior, including self-denial and altruism, the interdependent concept fails to account for motives, such as the actor's belief
that the altruistic act was the right thing to do, regardless of whether it
brought him or her pleasure. Also as a result, the concept of interdependent utility ceases to differentiate between people's preferences; moral
preferences are lost among a myriad of others, the preference for honesty
being on par with a taste for peanut butter (McPherson, 1984: 243).
Furthermore, expanding the concept of utility to explain, in this case,
all motives for human activity violates a rule of sound conceptualization.
Once a concept is defined so that it encompasses all the incidents that are
members of a given category, it ceases to enhance one's ability to explain.
According to this concept, there is no behavior that does not seek to
maximize an actor's self-interest. As a result, this "theory" fails to shed
light on motivation. Affixing the labels "utility" or "self-interest" to all
behavior explains nothing. Insight into what motivates behavior is lost, and
the concept becomes unproductive and tautological. In short, the neoclassical approach dilutes the explanatory power of the concept to the point
where it becomes quite meaningless. It thus seems, on grounds of sound
conceptualization, that the quest for self-satisfactions and seeking to serve
others (the public included) out of a sense of moral obligation are best kept
apart.

68

POLICY ANALYSIS AND ECONOMICS

3. What Is Moral?
If a major motivating factor is morality, how is one to distinguish between

it and other factors? After several hundred years, philosophers have yet to
produce a fully satisfactory definition of what is moral. Many, though,
agree that statements are to be viewed as moral if they meet four criteria.
First, individuals who act on the basis of moral meta-preferences sense that
they "must" behave in the prescribed way. Their moral acts are experienced as imperatives, things they must do because they are obligated, duty
bound. Most of us are familiar introspectively with this experience, doing
something because it is right, as distinguishable from doing it because it is
enjoyable.
The notion of an imperative is supported by the observation that people
set aside certain realms as commanding a special, compelling status. Emile
Durkheim (1947) points to the fact that people treat certain acts as
"sacred" (which need not mean religious). Such sacred moral principles
characteristically repudiate the instrumental rationality that includes costbenefit analysis; that is, people feel obligated to save a life or donate blood
without calculating the potential payoffs (Goodin, 1980).
The need for additional criteria to characterize moral acts arises out of
the fact that there can also be nonmoral imperatives, for example, obsessions with a forlorn love or even with an object, such as an illegal drug or a
fetish. The addition of two other criteria helps here to separate moral
imperatives from others (Childress, 1977).
Second, individuals who act morally are able to generalize their
behavior-they are able to justify an act to others and to themselves by
pointing to general rules. Statements such as "because I want it" or "I need
it badly" do not meet this criterion because no generalization is entailed.
"Do unto others as you wish others to do unto you," on the other hand, is a
prime example of a generalized rule.
Third, moral preferences must be symmetrical in that there must be a
willingness to accord other comparable people, under comparable circumstances, the same standing or right. (Otherwise, the moral dictum is
rendered arbitrary. Such an arbitrary rule would state, "This rule applies
to Jane but not to Jim although there is no relevant difference between
them.") Racist ideologies, although they otherwise have the appearance of
a moral system in that they are compelling (to their believers) and possibly
generalizable, fail to qualify as moral by this test.
Finally, moral preferences affirm or express a commitment, rather than
involve the consumption of a good or a service. They are therefore
intrinsically motivated and not subject to means-end analysis (Dyke, 1981:
11). (The fact that there are nonmoral acts that are intrinsically motivated

BEYOND SELF-INTEREST

69

does not invalidate this criterion. It only shows that the universe of such
acts is larger than that of moral acts, and hence, as was already indicated,
this criterion is necessary but not sufficient in itself.) As to the argument
that moral acts themselves are not impulsive acts but reflect deliberations
and judgments (especially evident when one must sort out what course to
follow when one is subject to conflicting moral claims), these deliberations
are not of the same kind as means-end considerations; they require
judgments among moral ends.
4. In Theory of Personality

Moral commitments, especially of the absolute kind, are distinguishable

from pleasurable activities because values are often "internalized." That
is, individuals see values as their own and not as external conditions to
which they merely adapt. Internalization has been defined as part of the
socialization process in which a person learns to "conform to rules in
situations that arouse impulses to transgress and that lack surveillance and
sanctions" (Kohlberg, 1968: 483). Social scientists may trace the genesis of
one's moral commitments to one's parents, culture, or peer or reference
groups; but whatever their source, once internalized moral commitments
become an integral part of the self. Thus those who feel they ought to serve
their country, God, or a cause feel strongly-sometimes despite strong
protestations from spouses, friends, and peers--that such actions are in
line with their values and their duty. Individuals' behavior depends on
whether their consideration of morality is driven by an external constraint
or by their own inner values.
The strength of moral commitments is reflected in studies that show that
the more individuals act under the influence of moral commitments, the
more they are expected to persevere (when circumstances change).
Conversely, the more individuals heed their pleasure or self-interest-for
example, by calculating costs and benefits--the less likely they are to
persevere. As a result, moral commitments are expected to "stretch out"
the learning curve and to increase transaction costs when changes favored
by economic rationality are inconsistent with moral commitments.
More specifically, the learning curve records the time lapse and the costs
involved in improving performance. The level of learning costs, the shape
of the curve, is determined in part by nonmoral factors (for example, the
complexity of the information to be absorbed). However, assuming all
these factors are equal, it is expected that learning will be slower and more
costly the higher the moral objections to what is being taught.
Choices that are relatively heavily "loaded" with moral considerations

70

POLICY ANALYSIS AND ECONOMICS

are expected to be unusually difficult to reverse, to be very "lumpy" (or

highly discontinuous), and to reveal a high "notch-effect" (a resistance
to pass a threshold that makes behavior sticky before it is passed; the
reluctance is greatly diminished or lost once passage is completed).
The concept of "transaction costs" developed to help explain "stickiness," the fact that people do not modify their behavior even when such
modifications are advantageous to them, or when the expected gains of
modification exceed the expected costs. It appears that the stronger the
moral commitment, the higher is the return needed before these individuals will violate their implicit contracts in the face of the changed
economic circumstances that favor such a violation; they will therefore
absorb more of an economic loss to live up to their obligations. For
example, it is considered "improper" for professors to accept an offer from
another university in May to start teaching in September (because it is
difficult to replace faculty so late in the academic year). This sense of
propriety is said to be stronger in some universities (Ivy League?) than in
others (Red Brick?). Hence, to move a professor in May from the first kind
of university would take a larger salary differential than from the second
kind. That is, moral behavior is "stickier" than amoral behavior.
Moral commitments may also reduce what economists call "moral
hazards." Specifically, the stronger the moral underwriting of implicit
contracts, the lower the transaction costs, resulting in less need to buy
hedge protection (in case resources are not delivered, or workers quit, and
so on) and to spend resources on legal action (such as drafting explicit
contracts and litigation to enforce them).
Although this discussion has referred to specific situations, the same
point has been made about whole societies. Because it is not possible to
provide enough police, accountants, and inspectors to verify more than a
small segment of all transactions, economies and societies require that
most transactions be based on voluntary compliance. This compliance, in
turn, is significantly affected by the relative level of morality within a given
society (or subsociety) in a given historical period. For example, when
corruption in the society is high, it acts as a major drag on economic
performance. Stated differently, to the extent that moral commitments
enhance the resources that can be dedicated to economic activity rather
than to supervision and verification, a higher level of morality increases
productivity and the GNP.
Prosocial behavior benefits when violations of moral commitments
cause guilt that leads, among other consequences, to compensatory behavior. For example, in experiments, people who were first induced to tell
a lie were subsequently twice as likely to volunteer to carry out a chore

BEYOND SELF-INTEREST

71

than those who did not lie (Freedman, 1970: 156). The relationship is often
more complex: A single transgression may lead to some guilt, but only the
cumulative effect of several incidents will lead to an act of remorse. The
behavioral consequences include sequences or cycles in which one or more
illicit acts are followed by bouts of morally approved behavior, or an
increased search for, or commitment to, "rehabilitation" (or atonement) via
morally approved behavior.
The concept also holds for joining welfare. Before one joins, a welfare
dollar may be worth only 80 cents, but once one joins, and one is
stigmatized as a "welfare client," the very next dollar may be worth 100
cents. And even if one gets off welfare and no longer receives stigmatized
income, the stigma is likely to linger.
Neoclassical concepts of utility fail to explain the sources of preferences
and the factors that cause them to change. Such explanations are necessary
because people change their preferences as the constraints under which
they "implement" them change. Hence, changes in behavior may be due to
changes in constraints or in preferences, often some combination of the
two. Without a conception and measurement of preference formation and
dynamics, in which moral values playa pivotal role, a satisfactory theory of
behavior is hard to imagine. The assumption that people's preferences are
"set" or "stable," which is central to the neoclassical paradigm because it
allows it to disregard changes in intentions and in those values that affect
preferences, simply ignores the most elementary observations of daily
experience.
Even accepting the gross oversimplification that preferences are set
rather than constantly changing, morality enters into choices because of
the important observation that choices are not simple, one-dimensional,
one-time (or one-stage) events but are in fact multifaceted. Neoclassical
psychology treats individuals as rational bundles of preferences and desires. These psychologists assume that people are unitary, that they are
driven by one overriding preference (usually for "pleasure" or some other
monistic "utile"). Deontological psychology, on the other hand, recognizes that people also have a moral dimension. It is increasingly realized
that humans, unlike animals, are constituted by two layers of preferences,
one (the metapreference) used to evaluate the other (the regular preference), and that when these two conflict, a typical struggle ensues. The
simple statement, "I would Like to go to a movie, but ought to visit my
friend in the hospital," captures the common tension. Moral values are the
most important source of these scrutinizing metapreferences.
While it is possible to study systematically the forces behind the two
levels of preferences, the outcome of the struggle rarely lands on one side

72

POLICY ANALYSIS AND ECONOMICS

or the other. Rather, human behavior is frequently conflicting and not

consolidated; choices are multi-faceted events. If the friend is neglected,
guilt nags; if the movie is skipped today, it may appeal more strongly the
next time a duty calls.

5. Evidence in Private Choices

To this point the argument has been made for recogOlzlOg a moral
motivation beyond self-interest. This section presents empirical evidence,
and evidence from everyday experience, of moral acts. At first one may
wonder, "Why document the obvious?" After all, does anyone really
doubt that a significant part of people's conduct expresses moral commitments? But the fact is that neoclassicists and many others have labored
long and hard to show that practically all behavior is driven by pleasure and
self-interest. Altruistic acts are accounted for as "really" being efforts to
enhance one's reputation, gain social approval, and so on. This section
reveals the existence of genuine (unselfish) altruistic behavior and
confirms, through example, the distinct influence moral concerns have on
private behavior.
A poignant example: By standard exchange theory, husbands and wives
of ailing Alzheimer's patients have no reason to stay with their mates; they
get nothing in return. Victims of Alzheimer's (a disease currently without
as much as a dim hope for cure) slowly lose their memory, first of events
and names of simple objects, until finally they cannot recognize even those
members of their own family who care for them. Eventually, they require
constant care and become incontinent and abusive. This means spouses of
these patients labor and nurse with no promise of reciprocal care, no
thanks, or even worse, with abuse in return. Nevertheless, nearly all
husbands and wives of Alzheimer's patients stay with their spouses
throughout their illnesses. "Exchange" theorists may try to explain this
behavior by arguing that they stay rather than endure the neighbors'
gossip. Yet, weighed against all of the hardships endured in caring for an
Alzheimer's patient, this argument seems quite unconvincing. People seem
to stay mainly because it is the "right" thing to do.
A number of major empirical studies have concluded that Americans
are also giving more to charity. Members of the neoclassical school would
have predicted a drop in donations after the 1986 tax reform, which
included changes that made each charitable dollar more costly to the
donor. Yet in 1985 total charitable donations were $80.31 billion; in 1988
they reach $104.37 billion, an increase of 13.6 percent after adjusting for

BEYOND SELF-INTEREST

73

inflation (Weber, 1989: 11, 14). With the advent of sharp government
cutbacks in social programs in the early 19808, polls also show that the
spirit of voluntarism rose gradually throughout the decade. In 1977, 27
percent of Americans reported that they were involved in a charity or
social service activity; that figure rose to 39 percent in 1987 and to 41
percent in 1989 (Public Opinion, 1987: 40; Gallup, 1989: 20). As Alan
Wolfe concludes, "These latter developments in particular insure that the
present period, characterized as it is by faith in the market, will not be
characterized by greater private niggardliness" (Wolfe, 1989: 88).
A considerable body of experimental data also supports the existence of
significant amounts of altruistic behavior. Several experiments show that
many people mail back "lost" wallets to strangers, cash intact (Hornstein,
Fisch, and Holmes, 1986). In another study, 64 percent of the subjects who
had an opportunity to return a lost contribution to an "Institute for
Research in Medicine" did so (Hornstein, Masor, and Sole, 1971: 110).
The costs are forgoing the found cash, as well as paying for postage and
going to the trouble of mailing the contribution. The reward? Chiefly, the
inner sense of having done what is right. (For more examples of altruistic
behavior, see Janis and Mann, 1977: 27; Schwartz, 19770: 283).
5. 1. A Factor in Economic Behavior

Evidence shows that commitments to moral values also affect economic

activities. Consider saving behavior. Neoclassical economists explain the
level of saving mainly by the size of one's income (the higher one's income,
the more one saves), by the desire to provide for consumption in retirement, and by the level of interest rates. However, studies show that those
who enjoy high incomes, as in the United States, save comparatively less
than do those in less affluent countries (International Monetary Fund,
1989).
The factors cited explain only part of the variance in the amount saved:
the extent to which one believes that it is immoral to be in debt, that one
ought to save for its own sake and to avoid dependence on the government
or one's children, and that one ought to help one's children "start off in
life. "
Much research has shown that firms do not pursue one overarching goal,
profit, but that they have mixed goals; they do not maximize anyone utility
and are internally divided rather than unified actors (Herendeen and
Schechter, 1977: 1514; Bailey and Boyle, 1977: 50; Monsen, Chieu, and
Cooley, 1968: 442). Among the goals that compel executives are those

74

POLICY ANALYSIS AND ECONOMICS

prescribed as morally appropriate by their peers, communities, and society

as a whole. These intentions are rarely limited to maximizing profit
(Donaldson and Lorsch, 1983). Neoclassical economists generally reject
the work of Cyert and March (1963) because it strayed too far from the
optimization paradigm. Wiliamson's work on contracting (transaction cost
economics) has received some recognition because it combines optimization with bounded rationality. Yet not until the development of the
principaVagent model, which allows for optimization in the face of asymmetrical information, did neoclassical economists finally accept the notion
that firms could have goals other than short- or long-term (market share)
profit maximization (see Moe, 1984: 739-777). However, these are but
holding operations that refuse to come to terms with the existence of major
independent sociaVmoral corporate goals.
Edmund Phelps (1975: 5) notes that in the same world in which people
sell unsafe products, gouge, and short-weigh, one also finds "the prevalence of altruistic behavior: A producer might advertise his product
truthfully when he need not, a labor union may refrain from breaking the
law when it could do so for a net gain, a benevolent butcher may abstain
from short-weighing." Kenneth Arrow (1975) goes a step further to argue
that since it is impossible to generate enough reinforcement agents and
incentives for many elements of the implicit and explicit social contracts
that sustain the market (e.g., trust in money), the whole economic system
would break down if it were not for these not-self-enforcing arrangements,
that is, those based on morality and government.
6. Evidence In Public Choices

The utility conception is particularly strained when it is extended to

encompass "public goods" and activities on behalf of the collectivity.
"Public goods" is a term introduced by Paul Samuelson to refer to those
goods that can be "used by many persons at the same time without
reducing the llmount available for any other person" (Alchian and Allen,
1983: 99). Price mechanisms cannot work here to allocate resources
rationally or to ensure "sufficient supply, efficient production, or general
welfare" (Arrow, 1974). These "goods" include much of a society's
culture, heritage, national defense, and environmental protection, as well
as pivotal economic elements such as major segments of scientific knowledge and of the nation's infrastructure. "Activities on behalf of the
collectivity" refers to acts such as voting and doing voluntary work that
serve shared needs.

BEYOND SELF-INTEREST

75

Government coercion and inducements are the major methods used to

ensure attention to, and resources for, these public benefits. Another
major method is to build on moral commitments. Civility, the individual's
moral commitment to shared concerns, is the concept used to refer to the
factor that leads people to contribute to the common good (for additional
discussion of the concept, see Janowitz, 1983). With the concept introduced at home, cultivated in schools, fostered by the news media,
enhanced by voluntary associations, and extolled from presidential and
other civic leaders' "pulpits," citizens of a nation feel obliged to contribute
to the well-being of the community they share.
A large number of experiments, under different conditions, most of
them highly unfavorable to civility, show that people do not take free rides
but pay voluntarily as much as 40 percent to 60 percent of what economists
figure is due to the public till if the person were not to free ride at all. The
main reason: The subjects consider it the "right" or "fair" thing to do. (For
an article about these experiments, as well as a report of their own, see
Marwell and Ames, 1981: 295-310.)
Another series of experiments was specifically designed to show that
egoistic incentives (reputation, reciprocity) are not neccessary to create
cooperation and altruism. In one experiment seven people were seated
together in a room, unable to talk to one another. Each subject, with $5 in
hand, was given a choice to anonymously keep the money or contribute
it toward a group bonus of $10, with the understanding that if three or
more subjects also contributed, everyone (including those who chose
not to contribute) would get the bonus, and if three out of the group
did not contribute, no one would get the bonus, including the one or two
who had contributed. Hence, noncontributors received either $15
or $5, contributors, $10 or $5. The results of the experiment showed
that 51 percent of the groups were "rewarded." The success rate rose
to 100 percent when, in another version of the experiment, members
were allowed a ten-minute discussion (used to elicit a sense of commitment) before making their choice (Dawes, Orbell, and van de Kragt,
1983).
Dawes concludes that special payments (to make people "behave") do
not explain why:
humans often appear to forgo choosing domain strategies in favor of those that
have more benign collective results. People do not rip-off their neighbors
whenever they know they can avoid being caught; airplane squadrons in war fiy
in formation even though the planes at the periphery are most likely to be shot
down, and some soldiers even fall on hand grenades to save their comrades.
[Dawes, 1989: 5]

76

POLICY ANALYSIS AND ECONOMICS

7. Public Polley Implications

The position advanced here is not the opposite of the neoclassical position;
it does not hold that people are impervious to their own self-interest.
A new discipline, socio-economics, suggests that individuals are simultaneously under the influence of two major sets of factors: their pleasure and
their moral duty. It is not suggested that analysis of moral factors ought to
replace considerations of self-interest, or that the latter are inappropriate;
rather, the point is that policy analysis ought to encompass both kinds of
considerations. This socio-economic position is advanced as an alternative
to the reductionism of neoclassicists, to the notion that people act morally
only as long as it makes sense in self-interested terms. Moreover, by
viewing human motivations as codetermined (by a quest for self-interest
and by moral factors), we seek to improve both the descriptive and
normative bases of policy analysis.
7. 1. Descriptive Implications

Voting behavior serves as an example of the socio-economic position.

Neoclassicists argue that normally there is little reason for people to vote
because they get nothing in exchange; they cannot reasonably expect that
their vote will make a difference. They might vote therefore only when
the election is close, when they feel their individual vote will make
a difference. 1 There is, however, only weak evidence in support of this
correlation between voter turnout and closeness of the election. And,
obviously, many millions vote when the elections are not close. The
explanatory and predicting factor, on the contrary, turns out to be a moral
commitment: People vote because they feel it is their "civic duty" (Barry,
1978: 17). Hence, if we seek to increase voter turnout, we need not worry
only or first about length of queues, or providing umbrellas, not to mention
payoffs; we should also consider expanding civic education.
A recent RAND study of the external costs of smoking and drinking
(costs not included in the price, such as making others sick) approaches the
problem from both perspectives (Manning et aI., 1989). The study concludes that from a neoclassical point of view, smokers should not be made
to pay a higher price for their cigarettes, because their costs already more
than pay for their harms. (A pack of cigarettes increases medical costs but
reduces life expectancy by 137 minutes, lowering the amount of pension
payments collected.) Even when the study considered the effects of
smoking upon society (i.e., passive smokers, other than smokers and

BEYOND SELFINTEREST

77

family members of smokers, and victims of fires caused by smokers), they

found the taxes on cigarettes to compensate sufficiently for these deaths
and illnesses. (External costs would more than double, by the way, if the
authors had included those imposed on the smokers' families.)
When the study broadened its purview to encompass "other" criteria (in
effect, morals), however, it openly and explicitly called attention to
considerations that might lead to imposing higher taxes on cigarettes by
drawing on criteria other than economic efficiency. This second approach
argues that even if one accepts uncritically that one ought not interfere in
private choices, there is still some room to consider it morally appropriate
to act to discourage smoking. The authors point out that the young. when
they "choose" to smoke-and become addicted-often are not fully informed about the consequences of their choice; hence to help them stop
smoking is, in effect, to act according to their latent preference (our
terminology, not theirs). Second, the fact that many smokers try to quit
signals their "true" desire. Thus, even if one holds that it is wrong to
interfere with people's choices once their preferences are formed, there is
moral justification to act on their behalf, on needs they express, albeit not
in immediate choices (e.g., not in buying cigarettes but in also "buying"
stop-smoking clinics).
More generally, the neoclassical approach claims to be value-neutral by
assuming that behavior reveals preferences, that the market best serves
people's preferences, or "what the individuals want;" neoclassicists disregard the fact that preferences are to a large extent socially formed and
hence reflect society's values, culture, and power structure. In effect, the
assumption of "consumer sovereignty" reflects a value system and a social,
economic, and political structure-that of mature capitalism. A more
encompassing analysis must include the factors that shape or reshape
preferences.
Deterring crime is another useful example. Neoclassicists analyze crime
in terms of self-interested cost-benefit analysis. They argue that the probability of being arrested and convicted, the size of the penalty, and the size
of the loot-<:>r the "costs" and "benefits"--<:orrelate with the frequency of
a large variety of crimes, including murder and rape (Andreano and
Siegfried, 1980; Rottenberg, 1979). The inferences are subject to considerable methodological controversies but are of no concern here. To the extent
that these data demonstrate that self-interest plays a role in situations
considered the domain of impulsive behavior, neoclassicists provide an
important correction to the over-socialized view of crime, a view that
focuses almost exclusively on the role of education, peer pressure, subculture, and other such factors. Yet to the extent that neoclassicists suggest

78

POLICY ANALYSIS AND ECONOMICS

that self-interest accounts for all or most of the variance, they vastly
overstate their findings (Cook, 1980), and their conclusions tend to mislead policy-makers.
Examples of misleading statements are from Paul Rubin (1980: 13):
"The decision to become a criminal is in principle no different from the
decision to become a brick-layer ... the individual considers the net costs
and benefits of each alternative and makes his decision on this basis," and
"tastes are constant and a change in [criminal] behavior can be explained
by changes in prices [such as penalties]. " They ignore the fact that despite
whatever correlations are found between "prices" and level of criminality, much of the variance (in crime rates) remains unexplained, probably
because moral and other social factors are at work. Such statements also
overlook the fact that the "taste" for crime, like all others, is affected by
normative and other social factors, for example, by the extent to which
the relevant subculture disapproves of the particular kinds of crime
(Grasmick and Green, 1981).
Similarly, statements such as Charles Murray's (1984: 168) that "crime
occurs when the prospective benefits sufficiently outweigh the prospective
costs" are not only formulated in a way that makes falsification impossible (if no crime occurs under an expected set of conditions the benefits
might be said not to "sufficiently" outweigh the costs), but they also tend
to mislead policy-makers into disregarding the role of education, subculture, leadership, and role models. Of special interest in this context is
James Wilson's (1985) discussion of the role of various "impulse control"
movements and organizations in nineteenth-century America. Wilson
points out that as industrialization advanced, youngsters who once left
their own homes only to work in homes under the supervision of other
farmers or artisans then started to reside in boarding houses in cities
without any family bonds or authority. As a result, disorderly conduct
became more widespread. What followed were numerous efforts to advance control of impulse and build up inner control, self-discipline, and
"character." These included Sunday schools, YMCAs, temperance movements, and various religious and secular voluntary associations. Some had
other goals, but impulse control was a successful by-product; others were
aimed directly at, and were effective in, instilling self-discipline.
Another example: Economists recommend various policies to encourage saving and reduce consumption, among them curtailing federal expenditures (a major source of consumption) or taxing consumption. Both
policies have a cost. The first may cause a recession, which exacts huge
human and economic costs; the second is regressive and imposes an unfair
burden on the poor. The fact that policies have costs does not mean that

BEYOND SELF-INTEREST

79

they are necessarily undesirable. It does, however, point to the merit of

considering other perspectives. For example, a more comprehensive policy
would emphasize the accumulation of debt as socially undesirable behavior, behavior that undermines our collective well-being and threatens
our future-just as debt was perceived in the 1950s. To bring about the
change in saving and consumption patterns, the President, community
leaders, and educators would all have a role in changing people's longrange perspectives. Such a policy would cost relatively little, would not be
regressive, and rather than imposing a solution, would draw on people's
values.
The policy point is that operationalizing certain programs requires work
not merely on the self-interest, cost-benefit, deterrence, incentive, and
police side but also on the formation of preferences side, via moral
education, peer culture, community values, and the mobilization of
appropriate public opinion, factors that neoclassicists take for granted.
Consequently, their theories provide no analytic framework to conceptualize the ways in which preferences are formed and might be reformed.

7.2. Social and Moral Factors in the Market

Neoclassical policy analysis typically suggests that we expand market forces

to allow individuals to more freely pursue their own preferences. The idea
of "choice" has recently been advocated by several policy analysts as a way
to "fix" public education: Parents should be free to choose which school
their child will attend. Four states have introduced choice (Arkansas,
Iowa, Minnesota, and Nebraska), and more than half are considering
it. The policy was called the cornerstone of President Bush's 1989 Educational Summit with the governors (Wall Street Journal, 1989: A26).
This idea was previously tried using a voucher system. Gary Becker
(1986) described that system:
A comprehensive voucher system would require public schools to compete for
students and to cover their costs with tuition revenue. Public schools could
compete effectively only if they suppressed violence, instilled discipline,
involved parents, offered a challenging curriculum, and controlled costs. If a
public school could not attract sufficient students with tuition that covered costs,
it would have to close or be sold to a private owner. [Po 19]

(For studies of the voucher system see Doyle and Finn, 1983, and Areen
and Jencks, 1971).
At the same time, experience with vouchers shows the dangers of the

80

POLICY ANALYSIS AND ECONOMICS

simplistic introduction of competition into human service areas in which

consumer knowledge is limited. It is far more difficult for parents to
evaluate education, than, say, a can of beans. Those searching out criteria
on which to base their choice of a school tend to land on measurable figures
that report the results of test scores but neglect much of the other subtler,
deeper, yet highly relevant aspects of education such as curriculum design
and guiding philosophies of schools and teachers.
In this case, giving more information to the "consumers" is a necessary
but insufficient step in implementing choice in schools. Even with all of the
relevant facts in hand, consumers have a hard time evaluating large
amounts of complex information. Thus, giving them scores indicating how
various schools compared on numerous measures is likely to have few
beneficial effects. (Consumers are already overwhelmed by the much
simpler information found on cereal boxes and do not engage in proper
comparative shopping.)
The moral issue still remains: When the market responds to purchasing
power rather than to votes, each of which reflects one person, it never per
se provides for the weaker members. Hence, unless externally affected, the
market simply caters to the rich and neglects the poor. The essence of
public education is, of course, an effort to overcome this market bias. The
more market forces are relied on therefore, the greater the danger that the
poor will receive less than the rich. Hence, choice should be combined with
efforts to increase the resources available to public schools in poorer areas.
7.3. Social and Moral Factors and Institutional Change

Neoclassical analyses tend to focus on transactions among individuals or

small units (such as households and small firm) and their aggregation in
anonymous markets, that is, markets that are assumed to have no collective controls. To the extent that institutions are studied at all within this
paradigm, they are generally perceived as reflecting arrangements made
voluntarily and knowingly by individuals, in line with their interests and
goals. Traditionally, other social sciences tended to view institutions as
reflecting historical (macro-) processes, society-wide values, and power
relations. We seek to encompass the influence of both individuals and that
of society. We attempt to combine aggregative analysis with collective
analysis by assuming that collective factors provide the context and are
"priors" within which individuals act, and which in turn are affected
by them.
The significance of systematically including institutional analyses arises

BEYOND SELF-INTEREST

81

because institutions themselves hinder or assist policy-making, and hence,

even if one does not seek to modify the institutions, their effects on policy
must be taken into account. For example, a multi-year economic policy
formed within the United States (say, a corporate development plan) that
ignored the well-established economic effects of the four-year political
cycle driven by presidential elections is less likely to succeed than an
economic policy that takes the cycle into account. All other things being
equal, the expansive policies of election years provide a much more
hospitable economic environment for a new product, or newly expanded
production capacities, than do those of first-year new administrations.
"Bitter medicine" is usually prescribed during this first year; hence the
period tends to be economically restrictive. The cycle, in turn, reflects the
Constitution and not an aggregation of individual dicisions. Similarly, one
must expect little success for a policy that ignores differences among
institutions-for example, shifting law enforcement functions from the FBI
to local governments-because of the widespread corruption institutionalized in many local police forces. The same must be said about a policy that
shifts responsibilities from the IRS or the Social Security Administration to
local tax collection or welfare agencies.
Beyond accounting for the established features of existing institutions
and the powerful inertia and vested interests they tend to generate, one
must also recognize that institutions can be changed and policy advanced
via such changes. Thus, instead of, or in addition to, using educational
campaigns to encourage many millions of Americans to increase their
saving, one can enhance saving by changing the tax laws, under some
conditions by reducing corporate outlays of dividends (i.e., by increasing retained earnings) or, more effectively, by reducing government
expenditures.
Segregating Social Security from the unified budget, and investing its
surpluses into a portfolio of American and foreign corporate and government bonds, could probably do more for the American savings rate than
would, say, doubling the size of funds individuals can salt away, tax
deferred, in their IRAs. While a constitutional amendment to balance the
budget may well create several new problems, it would modify significantly
the institutional context of the struggle to reduce federal deficits.
Similarly, aside from working on individual incentive schemes, corporations often benefit when they also introduce institutional changes such as
increased cooperation with labor unions (GM-UAW, in recent years),
quality circles, or participatory decision making (Greenberg, 1980). (None
of these are reforms sure to have the desired result; more research is
needed about the conditions of success and failure.) One may argue

82

POLICY ANALYSIS AND ECONOMICS

whether individuals or institutions are more powerful; however, one

conclusion is clear: Policy analysis should consider individual, aggregative,
and institutional factors.

Acknowledgments
The author is indebted to Judith Lurie and Susan McDougall for research
assistance. This chapter draws on the author's book, The Moral Dimension
(1988).

Notes
1. For a more sophisticated argument that tries to reconcile voting with rational choice
theory, see Uhlander (1989). It would take us too far afield here to respond to the numerous
assumptions the article introduces. Even after these are made, the most the author can claim
is that "some data from recent off-year elections are shown to be consistent with the model"
(Uhlander, 1989: 390).

References
Alchian, Armen Albert, and William R. Allen, Exchange and Production: Competition, Coordination, and Control, 2nd & 3rd eds. (Belmont, Calif.: Wadsworth, 1977, 1983).
Andreano, Ralph, and John. J. Siegfried, eds., The Economics of Crime (New
York: John Wiley, 1980).
Areen, Judith, and Christopher Jencks, "Educational Vouchers: A Proposal for
Diversity and Choice," Teachers College Record 72:3 (1971), 327-335.
Arrow, Kenneth, The Limits of Organization (New York: W. W. Norton, 1974).
- - - , "Gifts and Exchanges," in Edmund Phelps, ed., Altruism, Morality, and
Economic Theory (New York: Russell Sage Foundation, 1975), pp. 13-28.
Bailey, Duncan, and Stanley T. Boyle, "Sales Revenue Maximization: An Empirical Investigation," Industrial Organization Review 5:1 (1977), 46-55.
Barry, Brian, Sociologists, Economists, and Democracy (Chicago: University of
Chicago Press, 1978).
Becker, Gary, Business Week (March 24, 1986), 19.
Childress, James F., "The Identification of Ethical Principles," Journal of Religious
Ethics 5:1 (1977), 39-68.
Cook, Philip J., "Punishment and Crime: A Critique of Current Findings Concerning the Preventative Effects of Punishment," in Ralph Andreano and John
Siegfried, eds., The Economics of Crime, (New York: John Wiley, 1980), pp.
127-136.

BEYOND SELF-INTEREST

83

Cyert, Richard M., and James G. March, Behavioral Theory of the Firm (Engelwood Cliffs, N.J.: Prentice Hall, 1963).
Dawes, Robyn M. "Social Dilemmas, Economic Self-Interest, and Evolutionary
Theory," to be published in Research in Psychology: Frontiers of Mathematics
(New York: Springer-Verlag, 1989).
Dawes, Robyn M., John M. Orbell, and Alphons J. C. van de Kragt, "The
Minimal Contributing Set as a Solution to Public Goods Problems," American
Political Science Review 77:1 (1983), 112-122.
Donaldson, Gordon, and Jay W. Lorsch, Decision Making at the Top (New York:
Basic Books, 1983).
Doyle, Dennis, and Chester E. Finn, Jr., "Educational Quality and Choice:
Toward a Statewide Public School Voucher Plan (Washington, D.C.: U.S.
Department of Education, National Institute of Education, November 15,
1983).
Durkheim, Emile, The Division of Labor in Society (Glencoe, Ill.: Free Press,
1947).
Dyke, C., Philosophy of Economics (Engelwood Cliffs, N.J.: Prentice-Hall, 1981).
Etzioni, Amitai, The Moral Dimension: Toward A New Economics (New York:
Free Press, 1988).
Freedman, Johnathan L., "Transgression, Compliance, and Guilt," in J. Macaulay
and L. Berkowitz Altruism and Helping Behavior (New York: Academic Press,
1970).
Gallup Poll (November, 1989).
Goodin, Robert E., "Making Moral Incentives Pay," Policy Sciences 12:2 (1980),
161-178.
Grasmick, Harold G., and Donald E. Green, "Deterrence and the Morally
Committed," Sociological Quarterly 22:1 (1981), 1-14.
Greenberg, Paul D., and Edward M. Glaser, Some Issues in Joint UnionManagement Quality of Worklife Improvement Efforts (Kalamazoo, Mich.:
W. E. Upjohn Institute for Employment Research, 1980).
Herendeen, James B., and Mark C. Schechter, "Alternative Models of the
Corporate Enterprise: Growth Maximization, An Empirical Test," Southern
Economic Journal 43:4 (April 1977), 1505-1514.
Hornstein, Harvey A., E. Fisch, and M. Holmes, "Influence of a Model's Feelings
About His Behavior and His Relevance as a Comparison to Other Observers'
Helping Behavior," Journal of Personality And Social Psychology 10:3 (1968),
222-226.
Hornstein, Harvey A., Hugo N. Masor, and Kenneth Sole, "Effects of Sentiment
and Completion of a Helping Act on Observer Helping: A Case for Socially
Mediated Zeigarnik Effects," Journal of Personality and Social Psychology 17:1
(1971), 107-112.
International Monetary Fund, International Financial Statistics Yearbook, Vol.
XVII, (Washington, D.C.: International Monetary Fund, 1989).
Janis, Irving, and Leon Mann, Decision Making: A Psychological Analysis of
Conflict, Choice, and Commitment (New York: Free Press, 1977).

84

POLICY ANALYSIS AND ECONOMICS

Janowitz, Morris, The Reconstruction of Patriotism: Education for Civic Consciousness (Chicago: University of Chicago Press, 1983).
Kohlberg, Lawrence, "Moral Development," International Encyclopedia of the
Social Sciences, vol. 10, David L. Sills, ed. (New York: MacMillan and Free
Press, 1%8).
Losco, Joseph, "Understanding Altruism: A Critique and Proposal for Integrating
Various Approaches," Political Psychology 7:2 (1986), 323-348.
McPherson, Michael S., "On Shelling, Hirschman and Sen: Revising the Conception of the Self," Partisan Review 51:2 (1984), 236-247.
Manning, Willard G., Emmett B. Keeler, Joseph P. Newhouse, Elizabeth Sloss,
and Jeffrey Wasserman, 'The Taxes of Sin: Do Smokers and Drinkers Pay
Their Way?" Journal of the American Medical Association 261:11 (1989),
1604-1609.
Marwell, Gerald, and Ruth E. Ames, "Economists Free Ride, Does Anyone
Else?" Journal of Public Economics 15:3 (1981), 295-310.
Moe, Terry, "The New Economics of Organization," American Journal of Political
Science 28:4 (1984), 739-777.
Monsen, R. Joseph, John S. Chieu, and David E. Cooley, "The Effect of
Separation of Ownership and Control on Performance of the Large Firm,"
Quarterly Journal of Economics 82:3 (August 1%8), 435-451.
Murray, Charles A., Losing Ground: American Social Policy 1950-1980 (New
York: Basic Books, 1984).
Phelps, Edmund S., "Introduction," Altruism, Morality, and Economic Theory
(New York: Russell Sage Foundation, 1975), pp. 1-9.
Public Opinion (NovemberlDecember, 1987).
Rottenberg, Simon, ed., The Economics of Crime and Punishment (Washington,
D.C.: American Enterprise Institute, 1979).
Rubin, Paul H., "The Economics of Crime," in Ralph Andreano and John J.
Siegfried, eds., The Economics of Crime (New York: John Wiley, 1980), 13-25.
Schwartz, Shalom H., "Elicitation of Moral Obligation and Self-Sacrificing Behavior: An Experimental Study of Volunteering to be a Bone Marrow Donor,"
Journal of Personality and Social Psychology 15:4 (1970), 283-293.
Uhlander, Carole, "Rational Turnout: The Neglected Role of Groups," American
Journal of Political Science 33:2 (1989), 390-422.
Wall Street Journal (October, 20, 1989).
Weber, Nathan, ed., Giving USA (AAFRC Trust for Philanthropy 1989), pp.
11-14.
Wilson, James, Thinking About Crime rev. ed. (New York: Vintage Books, 1985).
Wolfe, Alan, Whose Keeper? Social Science and Moral Obligation (California:
University of California Press, 1989).

ECONOMIC THEORIES
OF DECISION MAKING UNDER
UNCERTAINTY: IMPLICATIONS
FOR POLICY ANALYSIS
W. Kip Viscusi

1. Introduction
If decisions under uncertainty were completely straight-forward, many of

the government programs now in existence would not be necessary.

Regulation of many job risks and product hazards would be superfluous, as
market processes would create adequate incentives for safety. Programs to
provide information about risks to product purchasers and the public at
large also would be unnecessary because individuals would already have
sufficient information to make rational decisions. The expansive scope of
the tort liability system could also be reduced because the presumption that
accident victims engaged in uninformed behavior would lose force. A
world in which markets provided full information about risks, and people
used this information effectively, would demand much less from public
policy. This chapter explores ways in which impediments to rationality
arise and how their shortcomings provide an impetus for several types of
policy intervention.
The actual performance of the economy, unfortunately, often falls far
short of our ideal. Individuals display tendencies to overestimate some
risks and to underestimate others. Market responses to risk range from
85

86

POLICY ANALYSIS AND ECONOMICS

inadequate to wildly alarmist. The same people who fail to buckle their
seatbelts may express great alarm over the threats posed by secondhand
smoke or by two grapes found in Philadelphia with traces of cyanide even
though these latter threats involve less severe risks.
These difficulties have not gone unnoticed by academic researchers. A
flourishing cottage industry now generates an increasing variety of anomalies and inconsistencies in choices under uncertainty. The two main messages of this literature are that people err and that in many cases the
pattern of these errors is quite systematic. This chapter explores some of
the most salient findings in this literature and attempts to indentify the
common themes that have emerged. 1
From the standpoint of public policy, examination of these issues is
relevant for several reasons. First, assessing the pattern of errors in individual choice is important to highlight the context in which government
intervention is warranted. Second, if the character of the market failure
can be identified, particular kinds of policies might be implemented. For
example, if the problem is one of incomplete information, the obvious
remedy is to provide appropriate hazard warnings to the affected group.
Finally, examining the limitations of rational behavior also indicates potential weak links in the implementation of government policies. If individual
responses are not rational, even with the aid of such efforts as welldesigned informational programs, more directive kinds of policy intervention should be considered.
The chapter will first analyze individual decisions to purchase insurance
against losses caused by natural disasters and will then turn to overall
assessment of patterns of risk perception. After considering several of the
main behavioral anomalies indentified in the literature, two models of
choice to address these anomalies will be presented. The chapter will
conclude by indicating the potential role of public provision of information
to remedy inadequacies in the knowledge individuals have about the risks
they face.
2. Responses to Natural Disaster Insurance:
A Second Look

An influential early study of the rationality of choices under uncertainty

was that of Howard Kunreuther (1976), who examined the patterns of
purchases of earthquake and flood insurance. Because these lines of
insurance are heavily subsidized by the U.S. government, Kunreuther
expected widespread purchase by individuals at risk, but his predictions
based on standard insurance economics were not borne out. Moreover, his

ECONOMIC THEORIES OF DECISION MAKING UNDER UNCERTAINTY

Table 5-1.

87

Profile of Knowledge of Insurance Cost Provisions

The Percent of Respondents

Who Do Not Know Terms

Cost of insurance:
Flood
Earthquake
Deductible:
Flood
Earthquake

Insured

Uninsured

17

68
76

11
44

25

82
85

Source: Adapted from Kunreuther (1976: 234).

detailed analysis of the risk perceptions of potential purchasers as well as

their actual purchases provided a fruitful case study for examining how
individuals respond to risk and whether they take advantage of subsidized
insurance.
The principal conclusion reached by Kunreuther (p. 250) was quite
strong: "The expected utility model, as traditionally used by economists,
provides relatively little insight into the individual choice process regarding
the purchase of insurance." This section will examine some of the specific
evidence developed by Kunreuther and, in particular, indicate the extent
to which full rationality appears not to hold.
In assessing these results the historical context in which Kunreuther
worked must be noted. At that time, economists did not question the
assumption of individual rationality. Because the literature dealing with
deficiencies in choice-making under uncertainty was still in its infancy, a
researcher would be expected to place great weight on any flaws found in
the rational choice model. Yet, in the more than a decade since
Kunreuther's article appeared, a substantial literature has attacked the
expected utility model as failing to capture the way in which decisions
actually are made, as opposed to the way in which they should be made.
Thus a retrospective examination of Kunreuther's results would permit a
search for the systematic and rational aspects of the behavior that are
reflected in his results as opposed to dwelling on the shortcomings of the
expected utility framework.
A useful starting point is to examine the individual knowledge of the
insurance terms that were offered. Table 5-1 summarizes the extent to
which the insured and uninsured population groups understand the terms
of insurance.

88

POLICY ANALYSIS AND ECONOMICS

Examination of table 5-1 clearly indicates that the full-information

assumption often implicit in economic analysis does not hold. The large
majority of uninsured do not know the terms of insurance; even among
the insured, many do not know important details such as size of the
deductible. These gaps in knowledge occurred despite the fact that survey
respondents were those in the household most responsible for insurance
purchases.
There are also a number of positive features of the responses. Differential knowledge of the insured and uninsured follow the expected parterns.
First, in every case people who purchased the insurance have a better
understanding of the premiums and the deductibles than those who did
not. Second, the cost of insurance is easier to remember than the deductible. As expected, almost all people who purchased flood or earthquake
insurance remember its premium, even though they may not remember all
of the specific provisions. Third, the fact that people do not know all of the
details regarding the insurance is not necessarily a sign of irrationality.
What is important is that people are being offered subsidized insurance. In
effect, the government has recommended it, and if individuals take this
information as a signal, they may not need to know all of the details
concerning what makes the purchase attractive. The uninsured likewise are
not necessarily irrational for not knowing the terms of the insurance since
they may not have heard of its availability. Such a lack of knowledge may
be more an indication of ineffective marketing of the insurance than of any
failure of individual irrationality.
One could then raise the secondary issue of whether the consumer
search process to learn about such insurance is fully rational. With the
substantial stakes involved in disaster insurance, consumers have some
financial incentive to acquire information about such coverage.
It should also be noted that the fact that not all people choose to
purchase subsidized insurance is not necessarily irrational since individual
risks differ, thus affecting the relative value of insurance. Furthermore, the
prospect of government aid after natural disasters provides for at least
some insurance without any prior purchase, and we would want to know
what these expectations were before pronouncing behavior irrational.
One can obtain a more refined perspective on the expectations of the
different population groups by examining the damage expected from a
severe flood of earthquake (see table 5-2). Perhaps the main anomaly in
this table is that 9 percent of those who purchased flood insurance, and 2
percent of those who purchased earthquake insurance, expected zero
damage. Notwithstanding the careful questionnaire that was administered,
it may be that some individuals did not properly understand the survey,

89

ECONOMIC TIlEORIES OF DECISION MAKING UNDER UNCERTAINTY

Table 5-2.
Earthquake

Individual Expectations of the Damage From Severe Floods or

Earthquake

Flood

%
%
Damage Claims
%
%
Insured/
Insured/
%
%
($)
Insured Uninsured Uninsured Insured Uninsured Uninsured

:s:1O,000
10,001-30,000
>30,000

Don't know

9
22
37
24
8

29
26
24
12
9

.31
.85
1.54
2.00
.89

2
13
32
47
6

12
19
27
34
8

.17
.68
1.19
1.38
.75

Note: In addition to drawing on data in Kunreuther (1976: 234), this table includes
statistics on the "Insured/Uninsured" calculated by the author.

perhaps confusing the probability of severe damage with the absolute level
of the damage. Another possibility that was borne out in the interviews by
Kunreuther and co-workers (1976) is that even when there was no assessed
risk, consumers valued the peace of mind that insurance provided.
More specifically, Kunreuther and his colleagues examined whether
insurance would be attractive from an expected value standpoint by
examining the contingency price ratio, which gives the relative odds that
the adverse event will occur. Many consumers who purchased insurance
even though they understood that it had a low perceived expected payoff
did so because they indicated that having insurance provided reassurance.
This beneficial effect is similar to influences such as anxiety and regret that
have played an important role in the subsequent literature. In particular,
insurance is a multiattribute commodity that provides a diverse set of
consumer benefits not readily captured by the standard expected utility
model.
A principal inference from table 5-2 is that the overall relationship
between insurance purchases and damage percentions is quite plausible. In
terms of distribution of the expected claims for both flood insurance and
earthquake insurance, the groups with the highest expected damage claims
are generally more likely to purchase insurance. The best statistics for
assessing the rationality of these patterns are the rations of the insured to
the uninsured purchasers for any particular damage claim group shown in
the third and sixth columns of table 5-2. The fraction of each damage claim
category purchasing insurance steadily rises as one moves to the higher
claims groups.

90

POLICY ANALYSIS AND ECONOMICS

Kunreuther also explores detailed measures of the rationality of insurance purchases, using information on individual decisions and risk perceptions. Once again, some of the patterns are plausible. For example, the
assessed probability of a severe flood or earthquake is higher for those who
purchase insurance than for those who do not.
Nevertheless, the results do not accord with fully rational insurance
purchases, as Kunreuther demonstrates quite convincingly. The two key
variables driving insurance purchases are the perceived seriousness of the
disaster (often coupled with past experience) and discussions with friends
and neighbors. These variables are consistent with a fully articulated
Bayesian model of insurance decisions and with the literature on the
psychology of risk perceptions. However, the standard expected utility
model in the insurance literature placed little emphasis on these concerns.

3. Aspects of Risk Perceptions

The inadequacies in individual risk perceptions are now well established
in the literature. In particular, individuals may not accurately assess the
probabilities of the adverse events that may affect them. One identified
variation concerns size of the risk. Once a risk is known, people tend to
overestimate low probability events and underestimate larger risks. These
regularities have been documented in detail by Fischhoff and colleagues
(1981). Among the small risks that are overestimated are those of dying
from botulism, tornadoes, floods, pregnancy, and smallpox vaccinations.
The substantially hazardous yet under-estimated risks of dying include
from diabetes, stomach cancer, stroke, heart disease, and homicides.
Combs and Slovic (1974) attribute this effect to an availability bias. The
greater media coverage of some events assists consumers in recalling them.
The media coverage effect is also consistent with a Bayesian learning
model since it represents a potentially valuable source of information?
Unfortunately, this coverage may provide inaccurate information regarding the relative magnitude of risks.
A systematic bias in risk perception assumes, of course, that individuals
are aware that the risk is present at all. Hidden risks, such as the hazards
from unknown carcinogens, may be neglected altogether if individuals
have no awareness that a risk is present. Once they have some awareness
of the risk, however, they tend to overestimate the extent of the hazard for
very small risks. The implications of this result are reflected both in the
actions that individuals take as well as policies that government implements. Low-probability events, such as being killed by a terrorist attack

ECONOMIC THEORIES OF DECISION MAKING UNDER UNCERTAINTY

91

while vacationing in Europe or dying from eating Chilean fruit, often

evoke large responses. In contrast, individuals have been reluctant to wear
seatbelts voluntarily but have been compelled to comply by mandatory
seatbelt requirements in many states, which have not resulted in full
compliance. Similarly, large controllable probabilities such as the risks
posed by our diet often merit little attention, whereas we focus instead on
minuscule hazards such as those that have been targeted by the media.
The pattern of bias in risk perception is not altogether inconsistent with
rational behavior. Consider a simple Bayesian learning model in which we
initially begin with the same risk perception for all classes of risk. As we
acquire additional information about each risk, our perceptions will move
closer toward the truth. Yet, because we have less than complete information, in our move toward the true probabilities we will continue to
overestimate the very small risks and underestimate the very large risks.
Although there are other possible explanations for the size-related bias in
risk perception, this possibility is also quite plausible and is consistent with
rational behavior on the part of imperfectly informed individuals. At this
juncture, we do not know whether the central difficulty is that there are
flaws in individual decision making or whether people simply do not have
full information about the risks they face.
A second aspect of these risk perception patterns is that individuals
appear to overestimate low-probability events and accurately perceive
events of zero risk as truly posing no hazard whatsoever. As a result, there
is a jump in individuals' risk perceptions when they move from no-risk
situations to ones posing very small risks. This pattern has been documented in a variety of studies of risk perception and is reflected in the
emphasis of government policies, such as the Delaney clause prohibiting
the use of carcinogenic food additives even at very small risk levels.
Similarly, most government agencies attempt to eliminate all risks that
are judged to be "significant." As our ability to detect risks has improved,
however, the threshold for significance has steadily declined. We consequently run the risk that the stringency in the regulatory policies will be
dictated by our measurement technologies rather than by the overall merits
of these efforts.
Widespread media attention, often exceeding the relative importance
of risks, contributes to a bias in public policy. 3 The damage inflicted by
tornadoes, floods, and earthquakes is highly touted by the media. Easterners express fear of ever having to move to San Francisco and face
the perils of earthquakes, but they are far less reluctant to live in cities
with higher homicide rates and greater risks of death from air pollution
than posed by earthquake hazards.

92

POLICY ANALYSIS AND ECONOMICS

An excessive emphasis on dramatic and highly publicized risks does not

necessarily imply that individuals are irrational. One can envision a process
whereby one forms risk perceptions by utilizing available information to
update one's probabilistic judgments. The extensive coverage provided to
natural disasters does not indicate the levels of risk involved, only selective
body counts. If the media primarily provide evidence of highly prominent
adverse outcomes, we would expect individuals processing this information
to form correspondingly high risk assessments. In effect, the media are
giving people the numerator of the risk but not the denominator. If
information about the numerator is repeated sufficiently often, particularly
if it is provided to a degree far in excess of its overall relative value, flawed
risk perceptions are likely to result.
A variety of other systematic biases have been observed in the perception of risks by individuals. People tend to underestimate risks that are
within their control. For example, most of us believe we are above-average
drivers, and few of us regard ourselves as being average or below average. 4
Similarly, most parents believe they are above average in the safety
precautions they take for their children. 5
The character of the risk is also of consequence. Risks such as cancer or
the chance of being killed by an explosion evoke substantial fear and
dread. As a consequence, they tend to be relatively overestimated once
they are called to people's attention.
The statistics in table 5-3 illustrate several of these aspects of individual
risk perceptions as they affect economic decisions. Consumers participating in the study were told that the starting risks posed by different pairs of
products were each 15110,000. They were then asked to value incremental
risk reductions of 5/10,000 for each of the two risks. Table 5-3 provides
their valuation of each successive incremental risk reduction. The risks
from household products-toilet bowl cleaner and insecticide-generally
posed temporary injuries such as skin poisoning or child poisoning that
respondents were told would produce nausea and stomach cramps for
several days.
As people purchase successive risk reductions, their willingness to pay
for risk reduction should decline. This pattern is borne out initially. At an
initial risk level of 15/10,000, the willingness to pay for a risk reduction of
5110,000 is greater than it is at 10/10,000. Once we reach a risk level of
5/10,000 and are offered the opportunity to purchase a complete elimination of the risk, however, there is a substantial jump in individual valuations. These results reflect the more general finding that people overassess
low-probability events and place a premium on the elimination of risks
with certainty. There is a substantial discontinuity in risk perceptions once
we move from some small risk to no risk at all.

ECONOMIC THEORIES OF DECISION MAKING UNDER UNCERTAINTY

Table 5-3.

Marginal Valuation of Reducing Risks by 5110,000

Starting Risk
(injuries/l 0, 000
bottles)
15
10
5

93

Incremental Willingness to Pay (dollars/bottle)

InhalationSkin Poisoning

InhalationChild Poisoning

GassingEyeburn

GassingChild Poisoning

1.04

1.84

.34
2.41

.54
5.71

.65
.19
.83

.99
.24
.99

Source: Viscusi, Magat, and Huber (1987: 475).

Table 5-4. Responses to Valuation Questions for Risk Increases of 1/10,000 for
Each Product

Injury Pair
Inhalation-skin poisoning
Inhalation-child poisoning
Eyebums-gasing
Gassing-<:hild poisoning

Percentage for Whom

Product Is Too Risky
to Purchase

Mean Value
($/bottle) of
Positive Responses

77.2
68.1
61.5
74.3

2.86
3.19
5.52
1.28

Note: This question asked subjects what price discount they would require on the new
product to accept an additional risk of 1/10,000 for both injuries, starting with risks of 15
injuries per 10,000 bottles sold for both injuries. See Viscusi, Magat, and Huber (1987: 477).

Moreover, the extent of this overestimation of low-probability events

may not be symmetric. From an economic standpoint, a person's willingness to pay for a sufficiently small risk reduction should have a value
equivalent to his or her willingness to accept a small risk increase. For
larger changes, willingness to pay amounts should be less than willingness
to accept values. Individuals who gave the values for their willingness to
pay for risk decreases were also asked how much of a price discount they
required to be willing to accept a risk increase of 5/10,000. The universal
response was that they would not purchase the product at all. As a result,
the survey was reformulated, inquiring about how much of a price discount
they required to accept a product that posed an extra risk of 1110,000 of
each type.
The results of this exercise are reported in table 5-4. For the overwhelming majority of consumers, the product is too risky to purchase at
any price. The survey even inquired whether individuals were willing to
accept a cash payment and free use of the product. As a result, these

94

POLICY ANALYSIS AND ECONOMICS

responses fully reflect all of the options economists might develop to get
people to display a finite risk-dollar trade-off. Moreover, for those who
were willing to state a price discount they would accept the extent of the
discount given in the final column of table 5-4 greatly exceeds their
willingness to pay for a risk reduction of much greater magnitude described
in table 5-3.
What these findings suggest is that individual inadequacies in risk
perceptions are highly complex and have profound implications for perceptions are highly complex and have profound implications for economic
behavior and risk regulation policies. Most of these results represent a
substantial departure from the full information world of perfect rationality.
Nevertheless, many systematic patterns to these biases highlight situations
in which government intervention can be most profitable. In particular,
these results suggest that political pressures for intervention may be
greatest when there is an upward shift in a small risk. However, it is the
large, stable probabilities of an adverse event that merit the greatest public
concern.

4. Behavioral Anomalies
The literature on the rationality of choice under uncertainty is replete with
examples of behavior that contradicts most models of rational decision
making. The frame of reference for assessing rationality is the expected
utility model. Under the expected utility framework, individuals maximize
a linear weighted average of the utility of different payoffs, where these
weights are the probabilities associated with the outcome. This framework
has strong appeal from a normative standpoint because of the plausibility
of the axioms on which it is based. Some of the implications of the risk
perception patterns noted in the last section are inconsistent with the
maximization of expected utility. Here I consider two additional examples
that illustrate fundamental deviations in behavior from the expected utility
model.
Consider a situation in which you are required to play Russian roulette.
For concreteness, suppose that you are unmarried and have no children
and that three bullets are left in the gun. How much would you pay for the
removal of one bullet? Alternatively, consider a situation in which only
one bullet is in the gun. How much would you be willing to pay for the
removal of this single bullet? In general, people would be willing to pay a
much greater sum for removal of the final bullet than for removal of one
bullet, which would not buy them complete freedom from death.

ECONOMIC THEORIES OF DECISION MAKING UNDER UNCERTAINTY

95

The behavior in this example, which was developed by Richard Zeckhauser, contradicts the predictions of standard expected utility theory. The
purchase of a bullet when many bullets remain in the gun should be more
highly valued because the expected marginal utility of money is much less
when the risk of death is substantial. The opportunity cost of buying back
the bullet is very low if there is a good chance that one may die. The
attractiveness of purchasing the bullet that ensures survival stems from
people's tendency to overestimate low-probability events. As a consequence, the inadequacies in risk perceptions discussed earlier also
emerge as an influential factor that leads to prominent contradictions in
the rationality of economic behavior.
Perphaps the most well known anomaly in expected utility theory is the
Allais Paradox. 6 If people are confronted with two sets of lottery choices,
they often give responses that are not mutually consistent. The particular
example constructed by Allais can be summarized by two equations:
U(l00)

> (.10) U(500) + (.89) U(l00) + (.01) U(O)

(1)

and
(.11) U(l00)

+ (.89)

U(O)

< (.10) U(500) + (.90) U(O).

(2)

In the first case, shown in equation 1, the utility of the certain re~ard of
$100 is preferred to a 0.1 chance of a $500 reward and a 0.89 chance of a
$100 reward, since there is also a 0.01 chance that the person will receive
nothing at all. Even though the expected payoff is much greater under the
dominated option, the chance of losing all one's money leads individuals to
prefer the certain payoff of $100.
If, however, we reduce each of these two prospects by a 0.89 chance of
winning $100 so that the chance of no payoff is 0.89 for the formerly
preferred case and 0.90 for the formerly dominated case, we generally find
the reversal in preferences shown in equation 2. This reversal obtains even
though we have simply subtracted an amount, 0.89 U(l00), from each side
of the equation. Since subtracting any positive amount from an inequality
should leave the direction of the inequality unaffected, there is a clearcut
contradiction.
A wide variety of models have been developed to explain inconsistencies such as the Allais Paradox. 7 Clearly, the expected utility theory is not
adequate and one must either modify our characterization of individual
preferences or our assumptions regarding how probabilistic information
enters individual decisions.
One possibility may be that the utility of receiving a zero payoff is not
equal to zero. Moreover, how we feel about a zero payoff could depend

96

POLICY ANALYSIS AND ECONOMICS

on the other payoffs that we are missing. In a situation in which we are

flipping a coin and have a chance to win a dollar, receiving no reward may
matter little. However, if the reward had been $1 million, the zero
outcome may be devastating. The potential role of regret has been
formalized in a variety of theories. 8
As a practical mater, it is difficult to identify any common choice
situations where the Allais Paradox is encountered. The main thrust of the
various anomalies that have been provided in the literature is primarily to
indicate that choices under uncertainty may be flawed in some manner.
They do not suggest that the particular anomalies that have been identified
are prevalent or of substantial consequence. This result is not entirely
reassuring since the problems that do seem to be prevalent reflect more
fundamental inadequacies in the way people process and act on risk
information.

5. An Example of Alternative Models of Choice:

Prospect Theory
A wide variety of models have been developed to address the problems
arising from choices under uncertainty. One of the most widely discussed
variants of this type is prospect theory, which was developed by psychologists Daniel Kahneman and Amos Tversky (1979). Their analysis consists
both of a summary of the evidence in the psychology of risk literature as
well as development of an alternative model to explain this behavior. Their
analysis can be described as largely descriptive in nature. The emphasis is
on developing central themes in the anomalies that have emerged and then
imposing suitable amendments on the nature of choice and risk perceptions
to reflect these phenomena.
The range of their discussion is quite broad. For example, they review a
variety of anomalies, including the Allais Paradox discussed in the last
section. The generalization that Kahneman and Tversky developed as
characterizing the Allais Paradox as well as similar examples pertains to
the scale of the probabilities involved. In particular, bigger payoffs look
better as the difference in the odds gets smaller. For example, a 0.45
chance of winning $6000 may be viewed as being clearly inferior to a 0.9
chance of winning $3000. Yet, if we shrink these probabilities by dividing
them by 450 so that we have a 0.001 chance of winning $6000 as opposed to
a 0.002 chance of winning $3000, our preference is generally for the chance
of winning $6000. As the probabilities become smaller they tend to look
more similar, and we place greater weight on the payoffs involved.
Kahneman and Tversky incorporate this property within the context of

ECONOMIC THEORIES OF DECISION MAKING UNDER UNCERTAINTY

97

their theory, noting that it is an empirical regularity, not a prediction of

their analysis.
Another phenomenon that they note is what they term the "reflection
effect." As economists have long claimed, people exhibit risk aversion with
respect to potential gains. In contrast, however, when losses are at stake,
individuals display risk seeking in the experimental contexts considered by
psychologists. For example, people may prefer $3000 for sure to a 0.8
chance of winning $4000. Yet, they would rather face a 0.8 chance of losing
$4000 than to incur a sure loss of $3000.
In the hypothetical lotteries presented to student subjects, Kahneman
and Tversky found considerable boldness of this type with respect to
risking losses. Although these results have been readily replicated in other
experimental contexts, they do not seem to accord with actual behavior.
The United States currently has a thriving insurance industry in which
insurance firms offer coverage for a price that greatly exceeds the expected
payoff. If individuals were truly risk seeking with respect to losses, they
would not purchase insurance at all, much less the substantial quantities of
actuarially unfair insurance that they now buy. Individual attitudes toward
losses may be quite different when they are dealing with real stakes as
opposed to paper losses in classroom experiments.
Moreover, even if the reflection effect does hold, it does not undermine
theories of rational decision making. Claims that people are risk averse
with respect to gains and risk loving with respect to losses suggest a
preference pattern that is unusual but not necessarily different from what
economists usually assume. Its only implication is that preferences are
somewhat unconventional.
Another anomaly has more fundamental implications for the structure
of utility functions. Kahneman and Tversky find that experimental subjects
fail to integrate a bonus with the valuation of the lottery with which they
are presented. For example, individuals tend to be unresponsive to having
been told that they will initially be given $1000 before they confront a
hypothetical lottery. That researchers should fail to find a lack of response
to artificial endowments of wealth does not mean that base levels of wealth
are unimportant. Rather, a more reasonable interpretation is that telling
individuals in an experimental context that they are richer is not the same
as actually making them richer. Wealth conveyance must take place before
individuals transform their attitudes toward risk. Economists have produced an abundant literature indicating that there are substantial wealth
effects in terms of attitudes toward risk bearing. 9 This economic evidence
is more plausible than the attempts to simulate changes in wealth in
classroom experiments.
In developing their theory to explain these and other experimental

98

POLICY ANALYSIS AND ECONOMICS

behavioral anomalies, Kahneman and Tversky incorporate a number of

notions developed in the literature on the psychology of risk perception. In
particular, they indicate that people sometimes edit the complex lotteries
with which they are presented before evaluating them. Moreover, they
note that people often have to resort to simplifications and methods of
combining probabilities that may not accord with the pinpoint predictions
of rationality.
Their assessments contain much truth regarding the manner in which
people deal with complex choices under uncertainty. The main remaining
task is to develop systematic frameworks for predicting how people will
interpret complex lottery information. Psychologists know, for example,
that the framework of the information presented is often of consequence.
What they are less able to tell is how particular methods of framing will
alter behavior. Ascertaining results is possible only after the fact. The
major gap is in developing predictive models to assess what will happen
before it has in fact been observed.
The essential elements of the Kahneman and Tversky theory are twofold. First, individuals employ valuations for payoffs that differ from those
in standard expected utility models. Figure 5-1 illustrates the shape of the
valuation function. As indicated in the preceding discussion, individuals
are hypothesized to be risk averse in the domain of gains, and risk loving in
the domain of losses. The peculiar shape of the utility function is not the
only novelty. Kahneman and Tversky also suggest that utility is a function
of the changes in assets offered by the lottery, rather than the net asset
position that will prevail after the lottery. Thus levels of wealth are not of
consequence in their model. As indicated, this structure is not consistent
with existing empirical evidence in actual choice-taking situations. At this
stage, the utility function component of their analysis appears to be
somewhat speculative. Moreover, the essential elements can probably be
reconciled with standard expected utility models by simply letting utility
functions have risk-averse and risk-loving regions.
The second component of the analysis concerns how people process
probabilistic information. Kahneman and Tversky hypothesize that people
attach to the payoffs decision weights that are not equal to probabilities.
Essentially, probabilities are transformed into decision weights by the
function 1T(p). Figure 5-2 sketches the shape of the decision weight
function. Kahneman and Tversky do not indicate precisely how this
function behaves near zero risks, but they do suggest that risk perceptions
increase sharply when one moves away from a zero probability, as a wide
variety of evidence has indicated.
Moreover, their sketching of the shape of the decision weight function

99

ECONOMIC THEORIES OF DECISION MAKING UNDER UNCERTAINTY

RlskAverse
Value V

Gains
Losses

Figure 5-1.

Change in Wealth

Relationship Between Decision Weights and Stated Probabilities

implies an overestimation of low probability events and an underestimation

of larger risks, a finding consistent with other psychological evidence.
There is, however, no formal basis offered for the functional form they
selected other than to note that it is governed by empirical regularities that
have been observed elsewhere.
Although the Kahneman-Tversky prospect theory is not a generally
accepted alternative to expected utility theory, it does serve to synthesize
many of the experimental results. Moreover, consideration of this analysis
suggests how much can potentially be lost by a move from expected utility
theory. The standard expected utility model of economists has the advantage that its assumptions are quite generally appealing. As a model of
behavior, it has strong predictive power. Reconciling actual patterns of
choice with observed behavior tends to yield theories that by their very

100

POLICY ANALYSIS AND ECONOMICS

/
/

/
/

nIp) = p / /
/
/

/
/
/

"

7t(p)

/
/

/
/
/

/
/
/
/

/
/

Decision weight

/
/

nIp)

/
/
/

/
/
/
/

Slated probability p

Figure 5-2.

Shape of Value Function with Respect to Changes in Wealth

nature are less attractive to economists because they deviate from normatively appealing axioms. Additionally, once we begin to alter many of the
essential characteristics of expected utility theory, we lose much of the
power of having a theory. If a theory ultimately consists of a descriptive
analysis that is able to accept any behavioral pattern ex post, after having
observed it, such a theory achieves very little because it has no predictive
power.
Most fundamentally, if the difficulty is that people are simply making
errors but would rationally choose to follow the maxims of expected utility
theory if they were appraised of the impact of their decisions, abandoning
expected utility theory may be too hasty. Expected utility may remain an
excellent reference point for anlyzing whether behavior is rational. We can
then identify departures from expected utility theory as indicating errors in
the ways in which markets function. Rather than trying to justify these
errors with an alternative theory of choice, we might instead choose to
intervene with government policies to correct the inadequacies that are
identified.

ECONOMIC THEORIES OF DECISION MAKING UNDER UNCERTAINTY

101

This author developed framework, termed prospective reference

theory, in an attempt to retain expected utility theory while at the same
time recognizing the impediments to risk perception (Viscusi, 1989). In
particular, this formulation postulates that people confronted with experimentallotteries treat the information provided as partial. In particular,
their perceived probability is not the same as the actual probability but is
instead a weighted average of some prior probability and the information
with which they are provided. For the standard classroom experiment, the
assumption is that a priori people treat all lottery outcomes as being
equally likely and form their posterior probability assessments as linear
weighted averages of both the stated probability and the probability that
would prevail if all outcomes were uniformly distributed. This approach is
consistent with Bayesian learning models (following either normal or beta
distributions) .
This rather simple amendment to expected utility theory recognizes that
people may not treat probabilistic information as being fully informative.
Moreover, it incorporates the salient properties of the evidence with
regard to risk perceptions, as low probabilities are overestimated, high
probabilities are underestimated, and there is a jump in probabilities once
we move from a zero risk situation to one in which there is a small positive
risk.
Consider, for example, how one might apply the prospective reference
theory approach to the Allais Paradox. For concreteness, let us consider
the extreme case in which the study participant places no weight whatsoever on the stated lottery information but instead treats outcomes within a
particular lottery as being equally likely. Other variants of the perceptions
of stated lotteries are also possible, including a weighted average of the
stated probabilities and the equal probability case. For simplicity we will
confine ourselves to the extreme situation in which the stated probabilities
are given zero weight, except for the certainty cases involving probabilities
of 0 and 1.
In the case of equation 1, the probability of receiving a payoff of 100
remains at 1.0. However, the probability of the payoffs on the right side
of the inequality in equation 1 all become 0.33 because there are three
outcomes-all of which are perceived as being equally likely. In the case of
equation 2, all of the lotteries involved are binary lotteries involving two
possible outcomes, so that all of the associated probabilities become 0.5.
Thus we are led to a reformulation of these equations given by the
following expressions:
U(loo) > (.33) U(5OO) + (.33) U(loo) +.33 U(O)

and

(3)

102

POLICY ANALYSIS AND ECONOMICS

(.5) U(lOO)

Setting U(O)
become

+ (.5) U(O) < (.5) U(500) + (.5) U(O)

= 0 without

(4)

loss of generality implies that these conditions

U(lOO) > .33 [U(500) + U(lOO)]

(5)

U(lOO) < U(500)

(6)

and

Since
U(500) > .33 [U(500)

+ U(lOO)]

equalities 5 and 6 not only can occure with prospective reference theory,
but they necessarily must occur for behavior to be consistent. The main
manipulation of the problem is simply to treat the stated probabilities as
being less than fully informative.
The advantage of this theory over models of irrational behavior is that it
predicts almost all of the phenomena that have been identified as anomalies
in the literature. Results such as the Allais Paradox and Kahneman and
Tversky's general principle underlying such violations of the substitution
axiom in expected utility theory are all generated as predictions of this
model. Whereas prospect theory can be potentially reconciled with this
behavior, prospective reference theory predicts ex ante that this and
similar anomalies will occur. For example, it also resolves the Russian
roulette paradox. As a consequence, the theory has much greater predictive power. Moreover, the situation in which people treat the lottery
information as being fully informative reduces to the standard expected
utility model so that the analysis need not contradict rational choice.
Consequently, it is possible to reconcile the expected utility model and
the literature on the psychology of choices under uncertainty. The main
deficiency in the literature is that the reference point used is that of a
classical statistician rather than a Bayesian decision maker. The classroom
experiments and hypothetical studies that have been conducted may not
have been treated as fully informative by the participants. This behavior in
and of itself can account for the anomalies that have resulted.
In short, psychologists often have used the wrong reference point to
assess rationality. Rather than assuming that people are perfectly informed, a more realistic reference point for comparison would be one
in which people have incomplete information but act on this information
in a rational manner. Imposing this Bayesian learning structure greatly
enhances the predictive power that our theories can provide.

ECONOMIC THEORIES OF DECISION MAKING UNDER UNCERTAINTY

103

6. Toward a sensible Basis for Risk Policies

Perhaps the final caveat with respect to the literature on rational choice
is that substantial judgment is required to assess which aberrations are
important. Tversky and Kahneman (1974) frequently recount the example
of assessing the contents of two urns. In particular, the task is to identify
which urn has two- thirds red balls and one-third white balls as opposed to
the reverse. Under one situation, we are told that the draws from an urn
led to four red balls and one white ball being selected. Under a second
scenario, we have drawn twelve red balls and eight white balls. Which of
the two sets of draws provides stronger confidence that the urn is twothirds red and one-third white?
Not surprisingly, experimental subjects presented with such lotteries
(when recruited for the study at a local shopping mall) give incorrect
answers. What is particularly note-worthy is that a swing of only one ball in
terms of the draws from the urns will lead to a reversal in the assessment of
the correct response. (It is left as an exercise for the reader to determine
which of the answers is correct. It should be noted that the reader has much
stronger probabilistic training than the typical experimental subject.)
Determining whether people can intuit applications of the laws of
probability dealing with nontransparent choice situations about which they
may care very little is not the best test of the rationality of individual
behavior. What needs to be identified are the situations in which actual
choices diverge from full rationality. Which deviations are important, and
which are not? What systematic patterns of errors can be observed? What
policies suggest themselves to address these difficulties? These are the truly
fundamental issues that are often ignored in the literature's self-sustaining
search for the new and often unimportant behavioral anomaly.
The inadequacies in choice also affect the degree to which providing risk
information can potentially eliminate the market failure. The principal
assumption necessary for information to be effective is that there must be a
choice that the individual can make. In the case of risks traded in the
market, such as job risks and consumer product hazards, there is an
element of discretion that makes programs providing risk information
meaningful. In contrast, in situations in which there is no discretion,
particularly in the short run, risk information will be less effective. For
example, the market response may be slow to incorporate information
about broadly based environmental hazards ranging from air pollution
exposures to toxic waste leakage into the water supply. Eventually such
risks will influence long-run mobility patterns, but in most cases the
individual response will not be sufficient to ensure an efficient distribution
of risks because the risks are not being traded in the marketplace.

104

POLICY ANALYSIS AND ECONOMICS

Informing individuals in these contexts may be important so that, for

example, people can choose to relocate or to create safety incentives
through tort liability. Residents in the Love Canal, New York, area exited
quite quickly once learning of the the toxic contaminants there. Moreover,
the risk information also facilitated their legal suits. Yet such information
alone does not provide incentives for generators of the risks to reduce their
magnitudes. There must be either a market response or action by some
other social institution.
Informational policies alone are not sufficient to address environmental
risks. Even in a case of risks traded in the market, provision of risk
information may not always be sufficient. The major weak link in such
programs is that the same kinds of impediments to sound decision making
that lead to a variety or market failures in situations of risky decisions also
may limit the efficacy of risk communication efforts. If we could provide
full information about the risks and be assured of a rational response,
market outcomes would be ideal. Human cognition, however, limits the
kinds and amounts of information that public programs can effectively
convey. As a result, informational efforts, correctly, do not attempt to
provide comprehensive risk information; instead, they generally indicate in
a succinct way the character of the risk and the pertinent precautionary
behavior.
Consider, for example, the well-known case of cigarette warnings. The
purpose of the warnings is to alert consumers to classes of risks posed by
cigarettes, but in no instance is there any effort to convey the particular
probabilities involved. Is the risk of cancer 0.01, 1, or 0.9 over a lifetime?
In situations in which we convey risk information that is not specific, there
should be continuing efforts to monitor the impacts of these programs on
risk perceptions. Are the risk perceptions induced by the broad wording of
these warnings biased in any particular direction? If so, can we select
different phraseology to convey the risk more accurately?
A striking case in point is an example that took place in California. In an
effort to provide consumers with pertinent risk information, California
instituted a comprehensive hazard warning program under Proposition 65.
That statute requires that all producers warn consumers of potential
carcinogenic risks. Any product posing a lifetime cancer risk from daily
consumption of at least one chance in 100,000 must be accompanied by a
pertinent warning. The specific wording of the warning is as follows:
'WARNING: This product contains a chemical known to the state of
California to cause cancer." Although the implementation of this act is not
yet complete, experimental studies of the implications of this wording for
consumers in other states indicate that the risk perceptions induced differ

105

ECONOMIC THEORIES OF DECISION MAKING UNDER UNCERTAINTY

Table 5-5.

Effects of Bleach Labels on Precaution-Taking (percentages)

Precaution

1.

2.

3.

Do not mix with

toilet bowl
cleaner (if toilet
is badly stained)
Do not add to
ammonia-based
cleaners (for
particularly
dirty jobs)
Store in childproof location

No
Warning
(n = 51)

Clorox
(n = 59)

Bright
(n = 42)

Test
(n = 44)

Maximum
Incremental
Effect

16

23

36

40

24

69

68

69

84

16

43

63

50

76

33

Source: Viscusi and Magal (1987: 66).

quite markedly from the trace carcinogens that are intended to be captured
by this program. In particular, the average risk assessment of adult
consumers seeing this warning is that the product poses a lifetime risk of
12/100, a risk assessment that dwarfs the risk threshold for labeling and
greatly exceeds the risks that are believed to be associated with any of the
products that will be captured by this program. lO
This example suggests that, in evaluating programs intended to provide
information about risks, we should not ignore the impediments to rationality that often led to a rationale for government intervention initially. There
must be a continuing effort to monitor how the risk information is
processed and to determine whether it is leading to sound decisions.
The potential efficacy of informational efforts that take the form of
hazard warnings is illustrated by the results in table 5-5. This study
pertained to the different precautions induced by a variety of warning
labels for bleach. The table's columns provide information for four alternative labels, the first a Clorox label that has been purged of all warning
information. The second label is the Clorox label currently on the market,
and the third the label for Bright bleach, a brand of bleach marketed by the
Kroger Company. The final label is a test label that has been redesigned
based on cognitive principles for the effective design of hazard warnings. 11
What these results indicate is that properly designed warnings can have
an incremental effect. The main risk from bleach is that it will be mixed
with a product such as ammonia and form chloramine gas, which is the

106

POLICY ANALYSIS AND ECONOMICS

leading source of poisoning among adults. The hazard warning would

increase precautions to avoid such mixing by 16 percent to 24 percent.
Similarly, the hazard warning would induce greater placement of the
product in a child-proof location in 33 percent of the instances. Moreover,
for families with young children vulnerable to poisoning, the efficacy is
almost 100 percent.
These results suggest that warnings can influence behavior in the
intended manner. Having an impact, however, does not always indicate
efficacy. In the usual instance, hazard warnings are adopted in situations
where precautions are usualy desirable, but that may not always be the
case. Consumers, for example, may differ in the disutility they attach to
wearing rubber gloves, and some individuals may therefore rationally
choose not to follow this precaution when using household cleaners. Such a
decision may be quite plausible based on the benefits and costs to the
particular individual. What is important is that the warning alerts the
consumer to the potential benefits of taking the precaution so that
the individual can make an informed decision.
In many important cases, however, we do not permit such consumer
sovereignty to reign. There are some precautions that we choose to make
more mandatory, thus overriding individual preferences. One notable
example is motorcycle helmets. Yet, many cyclists choose not to wear
these helmets and several Hollywood movie stars, such as Sylvester
Stallone, have visibly proclaimed that wearing helmets is not in keeping
with the appropriate image that motorcyclists wish to convey. If only the
motorcyclist were at risk, society might wish to condone such deliberate
self-destruction. However, motorcyclists also impose substantial risks on
others, notably through the accident costs imposed on automobile drivers
involved in collisions with motorcyclists. In particular, if the auto driver
is at fault, the cost of hitting a motorcyclist is considerably higher when
cyclists do not wear helmets and thereby sustain severe injuries. Many
states consequently make wearing helmets mandatory, largely in an effort
to protect the well-being of nonmotorcyclists.
It may also be the case that some requirements are needed to protect the
well-being of the participants themselves. Hockey helmets were long
viewed with disfavor by professional hockey players who, much like the
motorcyclists, believe that wearing such protective equipment would
threaten their rugged image. Yet, once wearing a helmet was mandatory,
complaints of this type were fewer, perhaps because players no longer
risked being singled out for the apparent weakness of relying on protective
equipment.

ECONOMIC THEORIES OF DECISION MAKING UNDER UNCERTAINTY

107

7. Conclusion
Analyses of choices under uncertainty have been a particularly active topic
in the economics literature during the past decade. Perhaps the main
theme of this work is that these choices are flawed in ,a number of ways.
Individuals often do not process information well, and the decisions that
result often are in substantial error.
One potential pitfall in addressing this literature is that analysts frequently attempt to summarize its implications by simply noting that people
are often irrational. Such an observation is correct as far as it goes, but it
does not convey the rich detail of the results found in the literature. What
is most noteworthy is not that people are irrational, but rather that they
have displayed systematic patterns of irrationality. Some mistakes indicate
a failure to recognize properly the implications of the risk to one's
well-being, whereas in other cases errors arise when people overreact to
the risks they face. Thus a belief that there is an inadequacy in the way that
individuals respond to risk does not necessarily mean that more governmental intervention is needed. Indeed, the situation may be that the
level of risk individuals are choosing to bear is below the efficient amount.
Although one would be hard-pressed to argue that government should
encourage additional risk taking in these instances, at the very least
government should not be engaged in efforts to further discourage such
behavior by providing risk information that inflates risk perceptions that
may already be too high. Rather, the objective of public policy should be
to promote informed decisions and efficient bearing of risks.
The challenge for public policy is that we often need to know a great
deal about the particular context before we know which action is appropriate. Thus we need to assess the extent to which people err in their risk
assessments and in their subsequent behavior. Moreover, if we were going
to pursue a government policy, such as a hazard warning program, we
should ascertain the ways in which it will influence risk perceptions and
behavior. Gaps in rationality consequently complicate the role of the
policy-maker in much the same manner as they create difficulties for the
individual decision maker.

Acknowledgment
The author wishes to thank Howard Kunreuther for detailed and helpful
comments.

108

POLICY ANALYSIS AND ECONOMICS

Notes
1. For alternative reviews of these issues, see Camerer and Kunreuther (1989), Fischhoff
et aI. (1981), Fishburn (1988), and Machina (1987). More generally, an entire journal, the
Journal of Risk and Uncertainty, is devoted to these issues.
2. The potential for learning is particularly great when there are opportunities for search
and experience. See Vining and Weimer (1988).
3. See Combs and Siovic (1974).
4. See, for example, Svenson (1981).
5. See Viscusi and Magat (1987).
6. See the discussion in Allais (1953), Camerer and Kunreuther (1989), or Viscusi (1989).
7. See, for example, the survey by Machina (1987).
8. Detailed reviews of regret theory appear in Camerer and Kunreuther (1989), Machina
(1987), and Viscusi (1989).
9. See, for example, the discussion of risk taking by Arrow (1971).
10. See Viscusi (1988).
11. See, especially, chapter 2 of Viscusi and Magat (1987) for a review of these principles.

References
Allais, Maurice, "Le Comportement de L'homme Rationel Devant Ie Risque,
Critique des Postulates et Axiomes de L'ecole Americaine," Econometrica 21
(1953), 503-546.
Arrow, Kenneth, Essays in the Theory of Risk Bearing (Chicago: Markham
Publishers, 1971).
Camerer, Colin, F., and Howard Kunreuther, "Decision Processes for Low
Probability Risks: Policy Implications," Journal of Policy Analysis and Management 8:4 (1989), 565-592.
Combs, Barbara and Paul Siovic, "Causes of Death: Biased Newspaper Coverage
and Biased Judgments," Journalism Quarterly, 56 (1974), 837-843, 849.
Fishchhoff, Baruch, et aI., Acceptable Risk (Cambridge: Cambridge University
Press, 1981).
Fishburn, Peter c., "Expected Utility: An Anniversary and a New Era," Journal of
Risk and Uncertainty 1:3 (1988), 267-284.
Kahneman, Daniel, and Amos Tversky, "Prospect Theory: An Analysis of Decision under Risk," Econometrica 47:2 (1979) 263-281.
Kunreuther, Howard, "Limited Knowledge and Insurance Protection," Public
Policy 24:2 (1976), 227-261.
Kunreuther, Howard, et aI., Disaster Insurance Protection: Public Policy Lessons
(New York: Wiley, 1978).
Machina, Mark J., "Choice under uncertainty: Problems Solved and Unsolved,"
Journal of Economic Perspectives 1:1 (1987), 121-154.
Svenson, ala, "Are We All Less Risky and More Skillful Than Our Fellow Drivers
Are? Acta Psychologica 47 (1981), 143-148.

ECONOMIC THEORIES OF DECISION MAKING UNDER UNCERTAINTY

109

Tversky, Amos, and Kahneman, Daniel, "Judgment under Uncertainty: Heuristics

and Biases," Science 185 (1976), 1124-1131.
Vining, Aidan, and David L. Weimer, "Information Asymmetry Favoring Sellers:
A Policy Framework," Policy Sciences 21:4 (1988), 281-303.
Viscusi, W. Kip, "Prospective Reference Theory: Toward an Explanation of the
Paradoxes," Journal of Risk and Uncertainty 2:3 (1989), 235-264.
- - , "Predicting the Effect of Food Carncer Risk Warnings on Consumers,"
Food Drug Cosmetic Law Journal 43:2 (1988), 283-307.
Viscusi, W. Kip, and Wesley A. Magat, Learning About Risk: Consumer and
Worker Responses to Hazard Information (Cambridge, Mass.: Harvard University Press, 1987).
Viscusi, W. Kips, Wesley A. Magat, and Joel Huber, "An Investigation of the
Rationality of Consumer Valuations of Multiple Health Risks, Rand Journal of
Economics 18:4 (1987), 465-479.

THE NEW INSTITUTIONAL

ECONOMICS: IMPLICATIONS
FOR POLICY ANALYSIS
Howard Frant

1. Introduction

Imagine a typical policy analyst-or better, an ideal policy analyst-asked

to evaluate the question of whether the federal government should provide
financial assistance in reconstructing deteriorated municipal water delivery
systems. How might such an analyst proceed?
He might begin by trying to evaluate the extent of the problem:
What the economically optimal rate of repair or replacement really was,
and how much of a shortfall in capital spending that implied, if any. Next,
the analyst might consider whether there were economic and philosophical
justifications for a federal role, or whether the policy was one more
appropriately left to states and municipalities. Then the analyst might take
up the optimal design of a federal assistance program: A simple indifference curve analysis, for example, shows that a matching grant will produce
a greater increase in output than a flat grant. Finally, one might
examine organizational and political issues: What agency should administer the program, and what conflicts might arise between this program and
other aspects of its mission? Who should receive the funds? How could
such a program be made most acceptable to Congress?

111

112

POLICY ANALYSIS AND ECONOMICS

When this problem was posed to the General Accounting Office (1980),
their approach was quite different. The GAO looked at water systems in
several cities and concluded that, while the problem was serious in
some cities, it was not serious in those cities whose water systems were
required to be self-financing. A clear implication of this conclusion (though
not one the GAO explicitly drew) is that changing the organizational
structure of public water departments is an alternative to a federal
spending initiative.
Such an approach is not one that comes naturally to policy analysts. In
a market-oriented approach to policy design, we tend to think of policy
as various ways of altering individual incentives, with organizational
and institutional considerations operating as constraints on policy design.
Organizational issues, that is, affect implementation, and as such may need
to be foreseen by the policy analyst, but only after consideration of the
appropriate policy tools. We are much less accustomed to thinking of
organizational structure as a policy variable affecting individual incentives,
just like tax structure or grant structure or regulatory structure. This is
surely due in part to the lack of a coherent framework for thinking about
the effect of organizational structure on individual incentives. It is this lack
that the "new institutional" economics tries to address.
2. The New Institutional Economics: Precursors
What exactly is the new institutional economics? Drawing the boundaries
of the field is difficult because is has come from so many sources and has
gone in so many directions. Indeed, the mainstream of economics today is
far more "institutional" than it was a few years ago. The brief summary
attempted here cannot possibly do justice to all the important contributors.
The goal is simply to sketch out some main ideas and then suggest how
those ideas could improve the practice of policy analysis.
One part of the new institutional economics traces back to Ronald
Coase (1937), who posed the simple yet knotty question, Why do organizations exist? Economists are quick to extol the virtues of the price system as
a coordinating mechanism. "Yet, having regard to the fact that if production is regulated by price movements, production could be carried on
without any organisation at all, well might we ask, why is there any
organisation?" (p. 388). Conversely, if there is a need for non-price
organization, "why is not all production carried on by one big firm?"
(p. 394).
Coase's proposal was to focus on the transaction as the unit of analysis:
Firms will compare the costs associated with carrying out a given transac-

THE NEW INSTITUll0NAL ECONOMICS

113

tion in a market with the costs of organizing it internally. Coase suggested

that the costs of market transactions included first, costs of using the
price mechanism, such as discovering what prices are; second, and more
importantly for the later literature, the costs of negotiating a separate
contract for each transaction. Putting some transactions under the umbrella of the firm, whose workers are covered by a single comprehensive
contract, is more economical. At some point, however, it becomes too
costly to organize everything within one firm because of "decreasing
returns to the entrepreneur function" (p. 394). The efficient size of the
organization, then, is one where the marginal costs of organizational and
market transactions are equal. It remained to later writers, however, to try
to identify the factors that made the efficient size larger or smaller.
Another precursor of modern institutional economics is the literature on
the economics of property rights. This literature points to the allocation of
property rights as a key factor in determining individual incentives. Coase
(1960) is again a key figure. The famous "Coase theorem," stating that the
allocation of property rights does not affect efficiency in the absence of
transaction costs, is widely misunderstood to this day. It is often taken to
mean that property rights do not matter, while Coase's point was just the
opposite. As he remarks (1988: 174): "The world of zero transaction costs
has often been described as a Coasian world. Nothing could be further
from the truth. It is the world of modern economic theory, one which I was
hoping to persuade economists to leave." Because transaction costs do
exist, the allocation of property rights does have efficiency consequences.
Harold Demsetz (1967) follows Coase in emphasizing the way in which the
appropriate allocation of property rights can reduce the costs of negotiating and enforcing contracts.
Armen Alchian and Demsetz (1972) made the seminal attempt to apply
property rights theory to organizations. They focus on what they call "team
production." Team production is group production that has inherent
nonseparabilities, so that individual productivities are hard to estimate.
(Their example is two men jointly lifting heavy loads into a truck.) Because
monitoring true productivities is costly, it will not be efficient to monitor
perfectly, and therefore each member of the team will have an incentive to
"shirk" to some degree. This is a simple Prisoner's Dilemma problem-all
members of the team would prefer to minimize shirking, but shirking still
takes place. What to do? Alchian and Demsetz suggest that the workers
could agree (obviously, this is an "as if" abstraction rather than a historical
statement) to hire one worker to monitor others, giving the monitor the
power to fire shirking workers. We then have the problem of who monitors
the monitor. This problem can be overcome, they suggest, if the workers

114

POLICY ANALYSIS AND ECONOMICS

agree to fixed wages and give the monitor the right to residual profits, so
that the monitor has no incentive to shirk. This arrangement of property
rights is beneficial to all because it maximizes the size of the total pie to be
divided among workers and monitor. Thus results the classical capitalist
firm.
The idea that team production was the essence of the problem was, by
and large, rejected by later writers. But Alchian and Demsetz did set the
agenda in other respects. One key idea was the emphasis on choosing an
efficient level of monitoring to reduce shirking. Another was the forceful
statement of the firm as a set of contractual relations. Their views on the
latter were quite radical:
It is common to see the firm characterized by the power to settle issues by fiat,
by authority, or by disciplinary action superior to that available in the conventional market. This is delusion .... To speak of managing, directing, or assigning workers to various tasks is a deceptive way of noting that the employer
continuaIly is involved in renegotiation of contracts on terms that must be
acceptable to both parties. Telling an employee to type this letter rather than file
that document is like my telling a grocer to sell me this brand of tuna rather than
that brand of bread. [Alchian and Demsetz, 1972: 777]

3. Transaction-Cost Theory
At this point the literature separates into two largely parallel streams. One
stream returns to Coase's focus on transaction costs and tries to operationalize it. The leading contemporary exponent of such an approach to
understanding the boundary between organization and market is Oliver
Williamson (1975, 1985). Williamson attempts to identify the general
factors that determine transaction costs, particularly contracting costs,
and relate them to specific organizational structures. He assigns a key role
to Herbert Simon's (1957) idea of "bounded rationality": the idea that
both information-gathering and information-processing are costly, and
hence it is impossible for human beings to identify and evaluate all possible
alternatives in making decisions. For Simon, organization then emerges
as a way to process information. But Williamson points to a different
problem: that bounded rationality makes it infeasible to respond to
environmental uncertainty by writing complete contracts that cover all
possible contingencies. The absence of complete contracts in turn creates
scope for "opportunism"-Williamson's term for behavior that is deceitfully or dishonestly self-serving. As Williamson (1990: 12) puts it: "Given
bounded rationality, aLL complex contracts are unavoidably incomplete.

THE NEW INSTITUTIONAL ECONOMICS

115

Given opportunism, contract-as-promise unsupported by credible commitments is hopelessly naive" (emphasis in original).
How does this approach lead us back to the issue of organization versus
markets? Williamson emphasizes the role of specialized commitments,
leading to a "small numbers" bargaining situation. The initial transaction
may be made in a competitive market. But once a contract is agreed on,
and costly commitments are made, the relationship is transformed into one
of bilateral bargaining. If one party must make a specialized investment
that does not transfer well to other uses-in Williamson's terminology, if
"asset specificity" is high-that party is vulnerable to opportunism, since
the other party can threaten to break off the relationship and so render the
investment worthless. The other party may thus be able to extort a higher
payment.
The alternative is to substitute some sort of governance mechanism so
that continual contract renegotiation is not required. For example, we
might move the transaction out of the market and into the organization.
But this entails costs as well. Williamson (1985) contrasts the "highpowered" incentives that usually exist in markets-the ability to appropriate the gains from a transaction for one's personal use-with the relatively
"low-powered" incentives in organizations. High-powered incentives have
the advantage of encouraging efficiency but the disadvantage of encouraging opportunism. The choice between organization and market involves a
weighting of these two factors.
Consider the labor market, for example. An employee may acquire
specialized job knowledge that would be expensive to replace; this gives
the employee more bargaining power than a new hire would have. Williamson, Wachter, and Harris (1975) discuss how many features of so-called
"internal labor markets" are effective in reducing the incentive of
employees with job-specific knowledge to extort a higher ex-post payment
from employers. A specific empirical prediction is implied: Many jobs are
not characterized by acquisition of highly specialized knowledge and thus
should have less elaborate governance mechanisms.
Note that although this view follows the Alchian and Demsetz view of
the organization as a set of contracts, it disagrees sharply with the claim
that organizations are characterized by continual renegotiation of contracts. People may have entered an organizational relation precisely to
avoid this. In particular, transactions requiring long-term advance commitments may be more viable in organizational contexts. If contracts are the
essence of the firm, then some of the contracts are long-term ones; thus
there is a difference between my employee and my grocer.
An obvious application of these ideas is to the question of vertical

116

POLICY ANALYSIS AND ECONOMICS

integration. The decision whether to integrate vertically or not is, after

all, simply the organization-or-market decision. The issue of vertical integration has never reposed very comfortably in the "old neoclassical"
framework, under which there is no theory of the organization-or-market
decision. In a competitive market, the opportunity cost of an input ought
to be the same whether or not it is produced in a plant owned by the firm.
On the other hand, a firm with market power might use vertical integration
in a strategic way; this leads one to view all cases of vertical integration with suspicion as being potentially anticompetitive. Williamson's
approach, in contrast, explains why, in some cases, a firm might want to be
vertically integrated independent of any aspirations to market power.
4. Agency Theory
The second stream in institutional economics is agency theory. An important antecedent of this work is the literature on the economics of insurance,
especially that dealing with "moral hazard." The moral hazard problem
arises because an insurer cannot perfectly observe the behavior of insurees.
Since insurees do not bear the costs of their own risky behavior, they will
engage in more of it than if uninsured, and more than would be efficient if
their behavior could be monitored. The optimal contract under these
conditions will in general sacrifice perfect efficiency in risk-spreading,
making insurees bear some of the risk in order to improve incentives
(Spence and Zeckhauser, 1971). Thus insurance policies commonly include
such features as deductibles and co-payments.
This idea was quickly formulated more generally as the problem of an
"agent" (analogous to the insuree) who has information not accessible to
the "principal" with whom she is transacting in an environment with
exogenous uncertainty. The information asymmetry is typically portrayed
as relating to "effort." Obviously this is easy to relate to the Alchian and
Demsetz emphasis on shirking, but in a more general sense effort in these
models is simply something that has a negative effect on the utility of the
agent but a positive effect on the total product of the transaction. (As will
be noted later, the significance of the term "effort" is often misunderstood.) The key problem is the difficulty in monitoring the agent's effortperhaps because of exogenous uncertainty, perhaps because of the sort
of nonseparabilities in production that concern Alchian and Demsetz,
perhaps because of informational barriers. A series of theorems were
developed about the forms of optimal contracts between principal and
agent, under various conditions of uncertainty and risk aversion (see Ross,
1973; Harris and Raviv, 1978; Holmstrom, 1979; and Shavell, 1979).

THE NEW INSTITUTIONAL ECONOMICS

117

Michael Jensen and William Meckling (1976) were apparently the first
to use the term "agency" in a specifically organizational context. Their
account starts with the classical capitalist firm owned by a single entrepreneur. Jensen and Meckling, however, focus on what happens as the
entrepreneur begins to sell off shares of the firm to outsiders (to raise
capital, diversify, and so on). At this point governance problems arise, for
the entrepreneur-manager now shares the cost of his actions with others
(much as in the insurance case) and thus has increased incentive to
shirk-by granting himself too many perquisites, for example. Jensen and
Meckling point out that if investors are rational, entrepreneur-managers
will bear these costs when they sell shares, and therefore it is in their
interest to put in place mechanisms to keep shirking to a minimum. They
use the term "agency costs" for the combination of the costs of the
mechanisms and the unavoidable efficiency losses that arise because it is
too costly to control all actions.
Jensen and Meckling hark back to Alchian and Demsetz in their view of
organizations as contractual relations; for them organizations are "simply
legal fictions which serve as a nexus for a set of contracting relationships
among individuals" (p. 310; emphasis in original). Eugene Fama (1980)
goes further, arguing that ownership of the firm is largely a meaningless
concept. He too views the firm as a nexus of contracts between individuals,
but he insists on the absence of a privileged position for either management
or the shareholders. The latter he views as merely individuals who specialize in risk bearing. This idea-we could call it "ownership iIIusion"-was
extended by Fama and Jensen (1983a, b). If shareholders are just residual
risk bearers, they ask, why should they also be those with ultimate control
over the decisions of the organization? Alchian and Demsetz would have
ascribed this to the need to motivate monitors; Fama and Jensen instead
point to the agency problems that could arise in defining what the residual
would be if other parties had control. But, they add, in complex orgainzations there is a need to specialize the management and risk-bearing
functions, and this requires the development of organizational structures
(such as boards of directors) to oversee the decisions of managers. They
also discuss why the possibility of agency problems means that certain
organizational forms have advantages in certain situations--for example,
that nonprofit status (meaning the absence of a residual claimant) enables
one to solicit donations by reassuring donors that their gifts will not simply
be diverted to the residual claimant. They make a selectional argument:
The organizational forms that survive will be those that have the lowest
costs, including agency costs.
The parallels between the agency approach and the transaction cost
approach should be apparent. The agency problem is just the problem of

118

POLICY ANALYSIS AND ECONOMICS

opportunism: Those whose behavior cannot be monitored with precision

have an incentive to behave ex post in ways that they would have agreed to
exclude ex ante if such an agreement were possible. If people are rational,
the primary impact of opportunistic behavior will be on efficiency rather
than equity; that is, some potentially beneficial transactions will not take
place because of the threat of opportunism. The focus is then on determining what sort of institutional structures-<>rganizational, legal, perhaps
even cultural~an be developed to reduce this threat.
The differences between the two streams are less significant than are
the similarities. But the transaction-cost theorists tend to differ from the
agency theorists, at least those already cited, with respect to the issue of
long-term versus spot contracting. As we have seen, the transaction-cost
theorists place more weight on the long-term nature of many relationships
and on the scope for opportunism that is created by the need to make
long-term commitments.
5. Institutional Economics in the Public sector

As should be apparent from the discussion so far, the new institutional

economics has been almost exclusively concerned with the private sector
firm. As Terry Moe (1984) has pointed out, there are dangers in a
too-facile adaptation of this framework to the public sector. To take only
one example, there is an implicit or explicit assumption in much of the
literature that economic Darwinism will lead to selection of the organizational form that is relatively, if not absolutely, efficient. It is by no means
obvious that any similar mechanism operates in the public sector. But
efficiency arguments are not the essence of a new institutional approach
(statements to the contrary by some of its leading practitioners notwithstanding). A new institutional approach is one that opens up the black box
of the organization by analyzing the incentives facing rational individuals.
There is nothing intrinsic in such an approach that requires us to maintain
that outcomes will necessarily be efficient. 1
Of course, the application of models of rational choice to the public
sector is an idea now in use in a number of existing literatures. One is
political scientists' mathematical approach to "positive political theory."
This literature, well summarized by Peter Ordeshook (1986), uses game
theory, social choice theory, and other fruits of mathematical economics to
investigate the effects of different institutional forms on outcomes. In
general, however, this literature has focused much more on voting and
legislatures than on bureaucracies. (A fascinating exception is Gary Miller

THE NEW INSTITUTIONAL ECONOMICS

119

and Terry Moe's [1986] application of some social-choice theorems to

decision making in organizations.)
Another related literature, going back to James Buchanan and Gordon
Tullock (1962), now goes under the name of "public choice." Though this
literature had no discernible impact on the institutional-economics literature, it has many parallel features. Buchanan and Tullock's problem of
designing a constitution that commands unanimous consent is analogous to
the problem of designing an efficient contract or organization, and their
idea that finding the optimal structure involves weighing the costs
of organization versus the costs of forgoing beneficial transactions foreshadows some of the issues that arise in the agency and transactioncost literature.
In general, though, this literature, like the political science literature,
has been more concerned with legislatures and voting than with public
bureaucracy. There is, however, a line of public choice work specifically
concerned with bureaucracy. This line traces back to William Niskanen's
(1971, 1975) theory of the budget-maximizing bureaucrat, which is by now
well established in the public finance literature. Niskanen's model was a
major advance over the then-prevailing (perhaps still-prevailing) assumption implicit in most public finance literature that a benevolent government
could costlessly remedy all sorts of market failures. Niskanen sharpened
the general public choice point that we need to look at the incentives of
individuals in the public sector as much as in the private sector.
A Niskanen-type approach, however, is limited by its high level of
abstraction-we have "bureaucrats" and "legislators" playing a very stylized budget game. Models that consider other players as well (for example,
Fred Thompson and L. R. Jones [1986]) may yield quite different conclusions about likely outcomes. Moreover, in Niskanen's framework, only
one type of agency problem, the misrepresentation of budgetary needs, is
assumed to exist. But this is by no means the only agency problem in
bureaucracies, nor even obviously the most significant one.
Consider, for example, the question of policy implementation. In the
classical agency model, as we saw, "effort" is anything that increases
output but has a negative effect on utility. If output is the degree of
implementation of policy, it is not hard to see that "effort" could be, for
instance, a bureaucrat's degree of compliance with a policy with which she
disagrees. Failure to appreciate this point has led some political scientists
to assert that agency models are inapplicable to the public sector. 2

120

POLICY ANALYSIS AND ECONOMICS

6. The New Institutional Economics and Policy Analysis

What, then, do policy analysts have to learn from the new institutional
economics? Certainly, an agency approach to public policy includes issues
already familiar from other conceptual frameworks, such as the literature
on "accountability." Accountability, however, captures only half the
story-the half that in agency theory would be called "monitoring." The
other half is incentives. A new institutional focus, unlike an accountability
focus, highlights the fact that how much monitoring one needs depends on
how well incentives are aligned in the first place.
One area where a new institutional approach seems particularly promising is the analysis of public-private interactions-privatization and government contracting. The non-economic work in this area tends to make
appeals to ineffable qualities of communal action that are lost by going
from public to private. Although there are certainly areas where this is
a legitimate concern-the criminal justice system, for instance-the difficulty seems greatly overstated for workaday services such as garbage
collection or electricity.
Among economists, on the other hand, the theoretical and empirical
work on private versus public provision generally relies on an extremely
simple property rights schema: Private providers have the right to capture
efficiency gains; public providers don't; so private providers should be
more efficient than public providers. Like the Niskanen bureaucrat, this
idea was at one time a major advance over the conventional public finance
view of a neutral and benevolent government efficiently curing market
failures, but it is no longer sufficient.
A more sophisticated approach would begin by recognizing that the
issue of privatization is, to a first approximation, simply the issue of vertical
integration: Should a service be "made" or "bought"? And the issue of
vertical integration is, as we have seen, the issue of firm versus market that
is so central in the transaction cost literature. Viewed in this light, the issue
becomes whether the possible efficiency gains from a market transaction
outweigh the increased risk of opportunism.
The answer will depend on characteristics of the transaction, such as
ease of monitoring and asset specificity. When complete contracts are very
difficult to write because outcomes are difficult to define, or when one side
must make a substantial advance commitment, the possibilities for opportunism increase: Even if a private provider is more efficient, the efficiency
gains will not necessarily be passed on to the public. Moe (1984) sketches
out the basic theory of an incomplete-contracts approach to privatization,
but to my knowledge John Donahue (1989) is the first author actually to

THE NEW INSTITUTIONAL ECONOMICS

121

apply such an approach to a study of a variety of public services. 3 Donahue

does not engage in formal modeling, but he identifies salient features of
programs that would make them more amenable or less amenable to
market provision, and uses this framework to explain outcomes. This is
a promising area for more research.
Even this approach can be criticized as insufficiently "discriminating,"
to use Williamson's word. For the framework here is still the simple
property-rights dichotomy between the private entrepreneurial firm and
the public bureaucracy. As Burton Weisbrod (1989: 541) remarks dryly,
"Other institutional forms exist." For example, a substantial amount of
government contracting, particularly in the human services area, is done
with nonprofits. One would expect that problems of opportunism might be
substantially different when contracting with nonprofits than when contracting with for-profits (and that the contractor's incentive for efficiency
might be substantially different as well). Yet until very recently there has
been little explicit consideration of this possibility. Weisbrod notes that a
1987 issue of the Journal of Policy Analysis and Management devoted
entirely to privatization contained only passing mention of the existence of
the private nonprofit sector. There is increasing academic interest in this
area, of such recent vintage that much of the work is not yet published.
Similarly, the basic property-rights hypothesis about relative public and
private efficiency has often been tested by comparing public and private
utilities. This research tends to find no effect from the difference in
ownership, or even an advantage to public provision. On the other hand,
studies of such services as garbage collection have found that private
provision is significantly cheaper. One possible explanation is that in
monopoly contexts regulation of private providers may have its own
intrinsic inefficiencies, such as the Averch-Johnson (1962) effect.
But there is another factor worthy of consideration: Many public
utilities are not government bureaus but independent public corporations
or authorities. 4 Such organizations have governance structures--and, in
a certain sense, property-rights structures--quite different from those of
bureaus. One might suspect, then, that their efficiency properties would be
quite different. Yet researchers studying public utilities have not distinguished between these two types of organizations. (Admittedly, the task
would be time-consuming because most data sources do not make the
distinction.) Would the generally negative findings of researchers looking
for superior efficiency from private sector utilities be different if the
comparison group were bureaus rather than all public agencies? We do
not know.
This brings us to another area where the need for contact between

122

POLICY ANALYSIS AND ECONOMICS

institutional economists and policy analysts is pressing: quite simply, the

development of a more nuanced understanding of the differences among
different types of not-for-profit organizations and how those differences
affect policy outcomes.
In the case of private nonprofits, the situation is relatively good. As
already noted, until recently the empirical research in public policy gave
rather little thought to the nonprofit alternative (an exception being in
health care, where the co-existence of public, private nonprofit, and
private for-profit institutions is so strikingly apparent). But theory building
toward an economics of nonprofit institutions is well underway. Henry
Hansmann (1980) pioneered the application of new institutional economics
to private nonprofits. 5 He noted that what distinguished nonprofits was a
"nondistribution constraint"-not that nonprofits could not earn profits,
but that there were severe restrictions on the distribution of those profits to
individuals. He argued that where monitoring of output (specifically,
quality of output) was difficult, the nondistribution constraint might make
customers more confident. Note the similarity to a Williamson-style argument for an intraorganizational rather than a market transaction. Spurred
on by Hansmann, as well as the forceful advocacy of Weisbrod (1977),
research interest in nonprofits has surged.
The picture is less sanguine when it comes to understanding the different sorts of public institutions. Ideas from agency theory and transactioncost economics have begun to diffuse into the study of the public sector,
but one is hard-pressed to find the sort of differentiation among different
public organizational forms that one finds for the private sector in, say,
Fama and Jensen. With the possible exception of the considerable literature on state-owned enterprises, there has been little effort among economists (as opposed to political scientists) to consider the policy impacts of
alternative forms of public organization. I have examined two such
questions (Frant, 1989). First, does a city's choice of governance structure~lected mayor or city manager-by creating different incentives,
affect its degree of bureaucratization? This question has considerable
importance for public management, and there is indeed evidence of such an
effect. Second, can we explain the fact that agencies organized as public
authorities seem at times to perform better than government bureaus in
maintaining infrastructure, and at other times to perform worse? This
question is relevant to a variety of policy fields: transportation, water
resources, housing, and so on. It appears that predictable differences in
the ability and incentive to monitor different types of services may help
account for differences in performance.
Considerable work remains to be done on these two research questions,

THE NEW INSTITUTIONAL ECONOMICS

123

and many more suggest themselves as well: Does an agency whose budget
comes from an enterprise fund more nearly resemble a bureau (which it
resembles in terms of "control") or an authority (which it resembles in
terms of "ownership")? Can we gain any insight into the effects of pure
"publicness" by comparing policy outcomes in intermediate public and
private organizational forms, such as private nonprofits versus public
authorities? To what extent do public for-profit corporations in the United
States (Comsat, Amtrak) differ from state-owned enterprises in other
countries? Does the strength of the link between employee compensation
and performance in public agencies vary with the monitoring or contracting
characteristics of the output?
Asking such questions requires some changes in our thinking. We are
used to viewing a variety of policy variables-taxes, subsidies, regulations,
budgets-as determining policy outcomes within the boundaries of a set of
fixed parameters: the intrinsic natures of the private and public sectors.
Now we need to consider that those apparently fixed parameters may in
fact be policy variables as well. Institutional structure is a dimension of
policy choice, one that may be critical to understanding a policy's success
or failure.

Notes
1. In one sense, it is almost tautological to say that rational individuals will seek the
efficient outcome, since to do otherwise would potentially make them worse off. In this sense
one could say that monopoly pricing is efficient because there is no way to make both sides
better off given the monopolist's inability to price discriminate. But this is not what people
generally mean when they speak of efficiency.
2. This point is, as noted earlier, often misunderstood. See, for instance, the exchange
between Cook and Wood (1989) about political control of the EPA. Wood asserts (and Cook
implicitly agrees) that "principal-agent models assume bureaucrats are passive, lazy, and
calculating only to the extent they want to avoid work." Wood goes on, "EPA bureaucrats
engaged in strategic manipulation at crucial times to move outputs in unexpected directions.
They acted independently and sometimes in opposition to political principals to maintain
established policy" (p. 971). What Wood takes to be a refutation of principal-agent models is
of course a confirmation of them (and a refutation of, say, a Weberian model).
3. John Vickers and George Yarrow (1988) apply some formal modeling, including
agency theory, to a study of British-style privatization (denationalization), but their approach
is quite different from the incomplete-contracts approach of Donahue.
4. Similarly, Caves and Christensen's (1980) study of Canadian railroads involved a
comparison of a private corporation with a Crown corporation.
5. Both Alchian and Demsetz and Fama and Jensen (1983b) discuss nonprofits, but in
much less detail.

124

POLICY ANALYSIS AND ECONOMICS

References
Alehian, Armen A., and Harold Demsetz, "Production, Information Costs, and
Economic Organization," American Economic Review 62:5 (December 1972),
777-795.
Averch, H., and L. Johnson, "Behavior of the Firm Under Regulatory Constraints," American Economic Review 53:5 (December 1962), 1052-1069.
Buchanan, James M., and Gordon Tullock, The Calculus of Consent: Logical
Foundations of Constitutional Democracy (Ann Arbor: University of Michigan
Press, 1962).
Caves, D. W., and L. R. Christensen, "The Relative Efficiency of Public and
Private Firms in a Competitive Environment: The Case of Canadian Railroads,"
Journal of Political Economy 88:5 (October 1980), 958-976.
Coase, Ronald H., "Notes on the Problem of Social Cost," in The Firm, the
Market, and the Law (Chicago: University of Chicago Press, 1988).
- - , "The Problem of Social Cost," Journal of Law and Economics 3 (October
1960), 1-44.
- - , "The Nature of the Firm," Economica 4 N.S. (1937), 386-405.
Cook, Brian J., and B. Dan Wood, "Principal-Agent Models of Political Control
of Bureaucracy," American Political Science Review 83:3 (September 1989),
965-978.
Demsetz, Harold, "Toward a Theory of Property Rights," American Economic
Review 57:2 (May 1967), 347-359.
Donahue, John D., The Privatization Decision: Public Ends, Private Means (New
York: Basic Books, 1989).
Fama, Eugene F., "Agency Problems and the Theory of the Firm," Journal of
Political Economy 88:2 (April 1980) 288-307.
Fama, Eugene F., and Michael C. Jensen, "Separation of Ownership and Control," Journal of Law and Economics 26:2 (June 1983a), 301-325.
- - , "Agency Problems and Residual Claims," Journal of Law and Economics
26:2 (June 1983b), 327-349.
Frant, Howard L., Incentive and Structure: On the Control of Managerial Opportunism (Ann Arbor: University Microfilms International, 1989).
Hansmann, Henry B., "The Role of Nonprofit Enterprise," Yale Law Journal,
89:5 (April 1980), 835-898, reprinted in Susan Rose-Ackerman, ed., The
Economics of Nonprofit Institutions: Studies in Structure and Policy (New York:
Oxford University Press, 1986).
Harris, Milton, and Artur Raviv, "Some Results on Incentive Contracts with
Applications to Education and Employment, Health Insurance and Law
Enforcement," American Economic Review 68:1 (March 1978), 20-30.
Holmstrom, Bengt, "Moral Hazard and Observability," Bell Journal of Economics
10:1 (Spring 1979), 74-91.
Jensen, Michael C., and William Meckling, "Theory of the Firm: Managerial
Behavior, Agency Costs and Ownership Structure," Journal of Financial Economics 3:4 (October 1976), 305-360.

THE NEW INSTITUTIONAL ECONOMICS

125

Miller, Gary J., and Terry M. Moe, "The Positive Theory of Hierarchies," in
Herbert F. Weisberg, ed., Political Science: The Science of Politics (New York:
Agathon Press, 1986).
Moe, Terry M., "The New Economics of Organization," American Journal of
Political Science 28:4 (1984), 739-777.
Niskanen, William, "Bureaucrats and Politicians," Journal of Law and Economics
18:3 (December 1975), 617-644.
- - , Bureaucracy and Representative Government (Chicago: Aldine Atherton,
1971).
Ordeshook, Peter c., Game Theory and Political Theory: An Introduction (Cambridge: Cambridge University Press, 1986).
Ross, Stephen A., "The Economic Theory of Agency: The Principal's Problem,"
American Economic Review 63:2 (May 1973), 134-139.
ShaveII, Steven, "Risk Sharing and Incentives in the Principal and Agent Relationship," Bell Journal of Economics 10:1 (Spring 1979), 55-73.
Simon, Herbert A., Administrative Behavior, 2nd ed. (New York: Free Press,
1957).
Spence, Michael, and Richard Zeckhauser, "Insurance, Information, and Individual Action," American Economic Review 61:2 (May 1971), 380-387.
Thompson, Fred, and L. R. Jones, "Controllership in the Public Sector," Journal
of Policy Analysis and Management 5:3 (Summer 1986), 547-571.
U.S. General Accounting Office, Additional Federal Aid for Urban Water Systems
Should Wait Until Needs Are Clearly Established, Report #CED-81-17
(Washington, D.C.: U.S. General Accounting Office, 1980).
Vickers, John, and George Yarrow, Privatization: An Economic Analysis (Cambridge, Mass.: MIT Press, 1988).
Weisbrod, Burton A., "Rewarding Performance That Is Hard to Measure: The
Private Nonprofit Sector," Science 244 (5 May 1989), 541-546.
- - , The Voluntary Nonprofit Sector: An Economic Analysis (Lexington, Mass.:
Lexington Books, 1977).
Williamson, Oliver E., "The Firm as a Nexus of Treaties: An Introduction," in
Masahiko Aoki, Bo Gustafsson, and Oliver E. Williamson, eds., The Firm as a
Nexus of Treaties (Beverly Hills: Sage, 1990).
- - , The Economic Institutions of Capitalism: Firm, Markets, Relational Contracting (New York: Free Press, 1985).
- - , Markets and Hierarchies: Analysis and Antitrust Implications (New York:
Free Press, 1975).
Williamson, Oliver E., Michael L. Wachter, and Jeffrey E. Harris, "Understanding the Employment Relation: The Analysis of Idiosyncratic Exchange," Bell
Journal of Economics 6:1 (Spring 1975), 250-280.

MACROECONOMICS AND
MACROECONOMISTS AS
INSTRUMENTS OF POLICY
George Horwich

1. Introduction

Macroeconomic theory has evolved for several centuries into a powerful

tool for policy analysts who understand both its strengths and its limitations. Its strengths lie in its highly developed and more or less empirically
validated logical framework. Besides a number of unresolved theoretical
issues, its limitations involve the difficulty of applying the theorydetermining the relevant time frame and level of aggregation, measuring
accurately the underlying variables, and estimating the parameters and
dynamic processes relevant to particular policy problems. The dynamic
issues include the oft-mentioned leads and lags of policy interventions and
the role of expectations.
This chapter summarizes the main features of modern macrotheory as
they apply, and have been applied, to U.S. policy experience. The discussion begins with a brief survey of the monetary equation of exchange and
the textbook aggregate supply and demand framework and then examines
the major policy episodes since 1929. The chapter ends with a review of the
episodes both as historical events and as objects of macroeconomic interpretation by professional economist-analysts and policy decision makers.
127

128

POLICY ANALYSIS AND ECONOMICS

2. The Monetary Equation of Exchange

The earliest attempts at modeling the macroeconomy involve the aggregate
exchange relationship between money and goods: Total money expenditures equal total receipts from the sale of goods and services. Dividing each
side of the equation into two underlying components, we write

MxV=PxQ
where, on the expenditures or demand side, M is the total existing quantity
of money and V is its velocity or rate of turnover; on the receipts or supply
side, Q is the total quantity of output and P is its average price.
There are, of course, many alternative definitions of the four variables,
depending on the purpose or data at hand. Generally, M is defined as the
medium of exchange (currency plus demand deposits) held by the public,
and Q is final output of goods and services or, simply, the real GNP.
As a simple statement equating expenditures and receipts, the equation
of exchange (EOE) is an undeniable identity. The variables are subject,
however, to numerous underlying behavioral hypotheses and statements of
causality that have broad relevance to policy analysis. These relationships
are derived in terms of traditional supply and demand microanalysis, as
in the supply and demand for bank reserves (a determinant of M), the
demand for money (a determinant of V), and the supply and demand for
labor (a determinant of Q). One aggregate relationship that falls out of
the equation is the causal connection between changes in the quantity
of money and resulting changes in the price level-the so-called "quantity theory"-given fixed or otherwise explainable values of the other
variables.
In skillful hands, the EOE is as useful today as it was before 1936,
when Keynes published his General Theory of Employment, Interest and
Money. The General Theory, however, opened the way to a more complex
classification of expenditure and wealth variables beyond the quantity
of money, the modern formulation of which-the "aggregate supply
and demand" framework-is a durable bastion of macro understanding.
Nevertheless, despite its detractors, including Keynes, the EOE stands as a
useful summary statement of the monetary/price-Ievel/aggregate-output
relationships generated by the more complex Keynesian system. For
example, in terms of EOE, a sudden reduction in the supply of output Q
caused by a reduction of worldwide energy supplies must, ceteris paribus,
cause the general price level P to rise. This rather straightforward link
between the oil crises of the 19708 and the inflation rate was totally missed
by early modelers of those events, including whole segments of the
macropolicy fraternity.

MACROECONOMICS AND MACROECONOMISTS

129

3. Aggregate Supply and Demand

The Keynesian system divides money expenditures MV into consumption,
investment, and government outlays, which sum to total money income or,
equivalently, the value of output PQ. Tax revenues, and thus the government deficit, are also identified in the expenditure stream, as is saving,
defined as the difference between disposable income (total income minus
taxes) and consumption. The stock of money is treated as a form of wealth,
the demand for which depends on income and the rate of interest.
This basic model, expressed in constant prices, formed the demand side
of a general equilibrium framework that was gradually enlarged to include
a labor market and an explicit production or supply sector. Keynes'
original formulation was cast purely in terms of the expenditure or demand
components; implicitly, supply was infinitely elastic and the price level was
constant. Even the rate of interest was generally assumed fixed by depression circumstances. But Hicks (1937), in his famous IS-LM generalization,
allowed the expenditure functions, in combination with the supply and
demand for money, to determine the level of both national income and
interest, both of which could vary. Marschak (1951), in his lectures,
Income, Employment, and the Price Level, introduced the production side
of the economy, which, together with expenditure equations, determined
an explicit equilibrium price level.
In Marschak's model it is possible to derive two independent equilibrium relationships between total income or output and the general price
level. One equilibrates the demand side (expenditure and tax functions and
money market), and the other the supply side (labor market and production function), of the system. Plotted on a diagram with the price level P on
the vertical axis and aggregate output Q on the horizontal, the resulting
aggregate demand AD and aggregate supply AS curves bear an obvious
and appealing parallel to the familiar demand and supply schedules of
microanalysis.
Although the aggregate schedules look, and in many ways behave, like
their micro counterparts, they are really quite different (see Barron,
Loewenstein, and Lynch, 1989: chap. 7). They are not the simple sum
of the micro demand and supply curves, which are related to indivdual
relative prices and not to the general price level. In fact, the existence of a
relationship between the general price level and the aggregate supply or
demand for output is neither obvious nor inevitable. On the demand side,
a rise in the price level reduces the real value of cash balances held in
private portfolios. In the absence of offsetting responses by those to whom
the balances are a liability (ultimately, on net, the government), consumption spending, which depends positively on wealth, will fall. If, at the same

130

POLICY ANALYSIS AND ECONOMICS

time, the demand for real balances varies inversely with the rate of
interest, wealthholders will try to regain some portion of the lost balances
by selling securities. This action will raise the rate of interest and reduce
the level of investment spending. With both consumption and investment reduced, the level of income (the aggregate quantity of output
"demanded") at which the expenditure components sum exactly to income is itself lower at the higher price level.
On the supply side, a positive relationship between the price level and
output rests, for a given technology and productivity schedule of labor, on
an incomplete response by labor to changes in real wages caused by
changes in the price level. For a given money wage, a rise in the price level
reduces the real wage, which will raise the quantity of labor demanded by
firms. If, because it is in a state of excess supply, labor accepts the loss of
real wages by allowing its money wage to remain fixed or to rise less than
prices, the induced increase in labor demand will raise employment and
output. Labor's acquiescence in its loss of real wages, sparked by the
higher price level, thereby imparts an upward slope to aggregate supply in
the P-Q space. We assume, for simplicity, that all points along a given AS
schedule are generated by a fixed money wage, as well as underlying
productivity function. In the event that labor insists on a full adjustment of
the money wage to the price level, real wages remaining constant, AS is
simply a vertical line.
The aggregate supply and demand framework is easily enlarged to
include a foreign sector-net exports balanced by a net capital flow of
opposite sign. The foreign transactions will make aggregate demand a
flatter schedule. A rise in the price level, for example, and the associated
rise in the interest rate reduce not only the domestic spending components
but net exports as well. Exports fall and imports rise both as a direct result
of the higher domestic price level and as an indirect consequence of the
higher domestic interest rate. The latter attracts foreign capital, which
drives up the exchange rate.
Expectations of price level change or other shocks to the system can also
be incorporated in one or more schedule shifts, which will be illustrated by
applying the model to an examination of the major macropolicy issues that
have arisen since 1929.

4. The Depression of the 1930s

More than 60 years after the event, there is still no consensus among
economists as to the direct cause of the catastrophic downturn that began

MACROECONOMICS AND MACROECONOMISTS

131

in September 1929. There is widespread agreement, however, that the

failure of the Federal Reserve to maintain the money supply helped turn
what might otherwise have been a severe, but normal, fluctuation into a
macroeconomic disaster.
4. 1. Monetary Policy: The Early Years

One can readily make the case that policy-makers and their advisers should
have understood the importance of defending the money supply immediately against the forces reducing it. 1 The causative role of money in
determining prices and output-including the havoc that a sudden contraction of money can wreak-was well understood by Hume (1955: 40) almost
two centuries earlier. The quantity theory had, in fact, received its most
detailed and elegant treatment at the hands of Keynes (1950) in his Treatise
on Money. Though published in 1930, the ideas had been discussed and
widely circulated in the 1920s. At the same time, Charles Hardy (1932), of
the Brookings Institution, was only one of a number of American economists who thoroughly understood the nature of central bank operations.
Another American, Chester Phillips (1921), had earlier produced the
definitive statement on the relation between bank reserves and the money
stock.
Nevertheless, the Federal Reserve Board, distracted by alternative
indicators and policy goals, permitted a currency drain and other forces
to decimate bank reserves and the stock of money. For some decision
makers, causality in the equation of exchange moved primarily from the
right side of the equation to the left: The quantity of money was passively
determined by the forces of commerce and the state of economic activity,
as expressed in the demand for loanable funds. From this perspective, the
adequacy of the money supply was gauged by its price, a low interest rate
indicating that the supply, relative to demand, was ample. Because shortterm rates dropped steadily despite the loss of reserves following the
October 1929 stock market crash and the collapse of investment, one not
inconsequential view was that the Federal Reserve needed to take no
aggressive expansionary action. At the same time, monetary officials were
eager to protect the gold reserve by preventing short-term rates from
falling too low-a goal that aggressive monetary injections might defeat.
The Federal Reserve eventually conducted open market purchases
sufficient to create a net rise in Federal Reserve credit, but only after the
economy had experienced a full two years of its steepest decline on record.
At that point the ability of the central bank to create money was severely
constrained by the unwillingness of banks or their customers to borrow.

132

POLICY ANALYSIS AND ECONOMICS

4.2. Fiscal Policy

An alternative means of stimulating the economy was through Treasury

fiscal policy, either by increased government spending, if one wanted to
take that route, or via tax reduction. Both measures raise disposable
incomes and create a deficit, which raises the quantity or velocity of
money. But in 1930 that approach was still in its prenatal stage. It was, in
fact, expecting too much of U.S. policy-makers to appreciate the value of
fiscal deficits or to understand the utter perversity of the tax increase of
1932. 2 On the other hand, the Hoover administration in 1931 created a
prototypical New Deal agency, the Reconstruction Finance Corporation,
which extended loans to distressed state and local governments and private
businesses.
Although much discussed and maligned by the media and social commentators, the federal government's deficit in the 1930s peaked at 5.4
percent of the GNP in 1934 and averaged only 2.9 percent for the
remainder of the decade. 3 Although high by historical peacetime standards, the deficits were hardly adequate to the task of stimulating an
economy whose output had dropped 30 percent between 1929 and 1933. By
comparison, in 1983, in a much healthier economy, the federal deficit
(based on a unified budget) was 6.8 percent of GNP. But it should be noted
that that deficit was 21 percent of 1983's federal expenditures, whereas the
deficits of the 1930s were often 40 or 50 percent of the total budget-as
much, perhaps, as could reasonably be expected, given the limited understanding of fiscal policy as an economic stabilizer.
Between 1933 and 1937 the economy grew 46 percent, essentially
restoring the 1929 level of output. Whereas 9 million jobs had been lost in
the downturn, generating a 1933 unemployment rate of 24.9 percent, only
7.5 million were added in the upswing. 4 At the same time, the working-age
population grew between 1929 and 1937, leaving 7.7 million, or 14.3
percent of the labor force, unemployed. The growth of real GNP during
the upswing averaged 9.8 percent per year, a more rapid increase than had
previously been seen in the recorded data for other than a single year. But
as impressive as the recovery was, it was not sufficient to achieve full
employment.
4.3. The Problem of Excess Bank Reserves

The monetary authorities, meanwhile, focused on the huge accumulation

of excess reserves in the banking system, due mainly to the inflow of gold,

MACROECONOMICS AND MACROECONOMISTS

133

whose price had almost doubled in 1934. By January 1936 the excess
reserves had reached $3 billion-ten to fifteen times their normal level.
The Board of Governors feared that these reserves placed the banking
system outside their control; the banks might decide spontaneously to use
the reserves to extend vast amounts of credit (Friedman and Schwartz,
1963: 520). The economic point of view held that the excess reserves were
reflective of the increased demand for liquidity characteristic of all sectors
in the low interest rate economy of the 19308 and that they would not be
disgorged precipitously. One economic interpretation saw the banks' marginal propensity to invest reserves as low but positive; the excess reserves
were desired and their removal by the Federal Reserve would force the
banks to contract their earning assets and the money supply in an effort
to replace them (Friedman and Schwartz, 1963: 527). An alternative
(Keynesian) view held that the marginal propensity to invest reserves was
zero (reserves and earning assets were perfect substitutes) and that only
increased demands for funds by borrowers could activate them; removal of
the excess reserves by the authorities would be passively accepted without
affecting the banks' holdings of earning assets or the money supply
(Horwich, 1963).
In no event was there empirical or theoretical evidence supporting the
concerns of the Board of Governors. Nevertheless, between August 1936
and May 1937, the excess reserves were sharply lowered by a phased
doubling of reserve requirements. During this period, but with a considerable lag behind the policy moves, the banks reduced their earning assets
and the money supply. In May the economy entered a sharp decline.
Though deep, the recession lasted only a year. Real GNP at the end of the
decade (1939) was only slightly above its 1929 level, and the unemployment rate was a staggering 17.2 percent.
Whether the increase in reserve requirements precipitated the banks'
response and triggered the recession or whether an independently declining external demand for loanable funds induced banks to liquidate their
assets will never be known with certainty. There is evidence on both sides.
In retrospect, it is probably fair to say that the Federal Reserve should
have waited for the excess reserves to cause trouble, not anticipate that
they might do so by taking an inherently risky neutralizing action. 5 In a
decade plagued by insufficient aggregate demand and no net rise in the
price level, fears of aggressive use by banks of their excess reserves seem
grossly misplaced. Indeed, the reemergence of excess reserves in 19381941 to almost triple their mid-thirties levels further undercut the board's
view that excess reserves in this period were nonfunctional and might be
spontaneously drawn down.

134

POLICY ANALYSIS AND ECONOMICS

4.4. The 19305 in Perspective

The 1930s, America's worst economic decade, suffered not only from the
limitations of macrotheory-fiscal policy was only dimly understood during
the downturn and later-but from a failure to act seriously on the theory
that was known: the quantity theory of money. Similarly, one of the most
egregious policy errors of the decade outside of pure macropolicy was the
Smoot-Hawley tariff of 1930, which imposed tariffs equal to an average of
60 percent of the value of U.S. imports. There was never any doubt that
the overwhelming majority of American economists understood this measure to be destructive of trade and welfare; 1,028 of them signed a petition
to President Hoover urging him to veto it (Kindleberger, 1986: 124). The
tariff predictably invited global retaliation and contributed to the twothirds emasculation of world trade that occurred in 1929-1933 just when
aggregate demand was most in need of shoring up.
Most of the subsequent New Deal measures had little to do with
macropolicy and involved instead attempts to protect selected sectors by
price supports or legalized cartelization. Their impact was relevant more to
redistribution and perceived social equity than to economic efficiency or
business (in particular, investment) confidence.
Today the continuing search for the causes of the depression is yielding
new insights into the more microeconomic dimensions of the collapse. A
recent contribution suggests that the 1930s ushered in an era of remarkably
extensive and rapid change-from manufacturing to services, and within
manufacturing, to larger scale and technologically innovative techniques of
production (Bernstein, 1987). The extraordinary magnitude of the change
in the optimal mix of output required an economic adjustment that would
have been painful and prolonged even in the presence of sound monetary
and fiscal policies and that was catastrophic in their absence.
In terms of the aggregate supply and demand framework, the new
technology and character of output caused a temporary leftward shift of
the aggregate supply schedule: Output fell in declining industries but could
not rise much, if at all, in expanding industries until resources underwent
the massive reallocation required of them. At any average level of prices,
total output was thus less, implying a leftward shift of the entire AS
schedule. In the 1930s this shift of AS was, of course, overwhelmed by
an even greater leftward shift of AD, reflecting the contraction of money
and resulting in a decline of both the general price level and aggregate
output.

MACROECONOMICS AND MACROECONOMISTS

135

5. The 19408: War and Peace

Both the conversion to a wartime economy in World War II and the return
to civilian output after the war were unique, all-encompassing episodes in
U.S. economic history. We examine some of their policy implications.

5. 1. Mobilization6
Between 1939 and 1944 U.S. real GNP grew at a remarkable average rate of
11.2 percent per year, a cumulative increase of 70 percent. This increment,
roughly equal to the additional output of war materiel and services,
absorbed not only the remaining pool of unemployed workers but drew
millions of additional workers, primarily women, into the work force.
The demand changes were thus government-driven, and the industrial
response was also to a large extent mandated by centralized authorities.
General price controls were imposed early in 1942, severely limiting free
market activity. Nevertheless, price adjustments in response to market
forces were occasionally made and a fair degree of wage flexibility was
permitted to draw millions of workers to new and geographically disparate
employments.
The price controls were widely accepted as an inevitable wartime
measure, necessary to contain inflation. In an interesting reappraisal of this
policy, Paul Evans (1982) finds that while the controls reduced the price
level by 30 percent, they reduced aggregate output 7 percent and employment almost 12 percent. These reductions resulted from the inability of
households to acquire additional goods with their increased incomes,
thereby reducing their work incentives. As well as the wartime economy
performed, it thus could have produced an even greater output (or, more
importantly, the same output more rapidly) under free market prices and
an accommodating monetary policy that permitted the same degree of
measured inflation.
As a contribution to social harmony, particularly in a national emergency, a defense of controls can perhaps be made. But they are economically tolerable only for limited periods. If the war had lasted longer, the
inefficiencies resulting from the absence of relative price signals would
have cut deeper and deeper into aggregate output. Fortunately, as the war
wound down in 1945, enforcement of the controls became increasingly lax.
By September 1946 they were totally removed.

136

POLICY ANALYSIS AND ECONOMICS

5.2. Demobilization
The freeing of prices set the stage for the most extensive restructuring of
the economy to occur in any single year of U.S. history. In 1946 government defense spending fell by 80 percent, from 37 percent of 1945's GNP
to 8 percent of 1946's GNP. Eight million people were released from the
armed forces, and 2.5 million civilian jobs were added. The difference,
equal to 5.6 million people, left the labor force or joined the unemployment rolls. As many as a third of the 65 million persons employed by the
private or public sectors (including the military) changed jobs or returned
to the household and nonprofit sector.
The real GNP (in 1947 prices) fell 11.1 percent in 1946, but unemployment rose by only 1.3 million from a wartime low of 1.9 percent to a
sustainable 3.9 percent. Most remarkable is that all this occurred without a
trace of centralized direction or control. The adjustment was a purely
price-directed market phenomenon.
Many, if not most macroeconomists, however, were fearful that demobilization would ravage aggregate demand and result in a deep postwar
depression (see Copeland, 1944; Hagen and Kirkpatrick, 1944). What they
overlooked was the enormous sums of money in the hands of the public
built up by wartime deficit spending and which the public, facing wartime
shortages of goods, had retained (velocity had fallen sharply). This was the
very opposite of the monetary situation in the early 1930s. Between 1939
and 1946, real M1 balances (the public's currency plus demand deposits)
rose 129 percent. 7 By comparison, real balances in World War I fell 8
percent, which may have contributed to the downturns of both 1918
and 1920.
American economists, like others, were not aware of the importance
of real balances or, more generally, wealth as a determinant of expenditures. Nor were they likely to believe, after the 1930s, that a capitalist
economy was capable of undergoing a vast peacetime transformation
without a high degree of government planning or, at least, centralized
coordination. Nineteen forty-six was instructive on both counts. That the
demobilization proceeded in a postwar frenzy of laissez-faire without
serious intervention by government or macroeconomists is one of those
fortunate accidents of economic history.

6. 1950-1953: The Korean War

Before turning to the cyclical behavior of the postwar economy, the
Korean War will be examined for its particular blend of macropolicies.

MACROECONOMICS AND MACROECONOMISTS

137

Prices rose almost immediately following the June 1950 outbreak of

hostilities. The cause was an increase in expenditures--govemment outlays
for war materiel and consumer spending in anticipation of higher prices.
Both money and velocity rose, but the rise in velocity was especially
pronounced, as might be expected in these circumstances.
A tax increase was quickly enacted, and the money supply gradually
tightened. In January 1951 price controls were imposed. Unlike the price
ceilings of World War II, the removal of the controls in 1953 was not
followed by an increase of prices that had been only temporarily repressed.
The Korean controls had not been binding. Nevertheless, they may have
played a useful role in defusing inflationary expectations and imposing
pricing discipline on the part of both suppliers and demanders in the
context of general stabilizing policies. This, at least, is the econometric
finding of George Perry (1973). Others, such as Herbert Stein (1984:
119-120), consider the monetary and fiscal measures to have been sufficient and the price controls superfluous.
Inflationary expectations shift aggregate demand upward (to the right)
and aggregate supply upward (to the left). The supply shift is now
characteristic of recessionary periods, as will be argued in the next
section. In fact, it was the belief (though not articulated as such) that such a
shift was taking place in the summer of 1971 that led in August to the only
other postwar imposition of general wage and price controls. But the
evidence for the gathering inflation was weak, and the controls proved to
be binding and disruptive and positively disastrous when the oil price
shocks occurred in the fall of 1973.
Still, the distinction is worth making between price controls of the
World War II variety that serve only to repress inflation caused by
underlying monetary or fiscal forces versus controls that defuse inflationary
expectations that need not be realized and need not reappear at a later
date. There is a difference, in principle, at least. In practice, recognizing
perverse expectations, imposing controls in a timely and effective manner
to neutralize them, and then removing the controls before their interference with relative prices causes too much damage is almost certainly
undoable-both economically and politically-as the controls of 19711973 demonstrated. But there is a difference.
By contrast with 1950, policy-makers in 1971 focused on unionmanagement negotiations and price indices whose trends bearing on
inflation and inflationary expectations were not clear. The controls were
binding from the start in that inflationary forces due not to expectations but
to excessive growth of the money supply were at work. In the following two
years the controls masked continued expansionary monetary and fiscal

138

POLICY ANALYSIS AND ECONOMICS

policies and were still in force in 1973 in the energy markets, when world
oil markets erupted and imposed an economy wide reallocation of resources that the controls severely hampered (see Stein, 1984: chap. 5).

7. The Post-1949 Business Cycle

The first postwar recession occurred in 1948-1949 and, though a minor
event by comparison, was like the downturn of 1929-1933 in that both
output and the price level fell. In later recessions a new pattern was
established in which output fell but prices rose. These subsequent fluctuations are described in terms of the aggregate supply and demand
framework. What follows is schematic description of more or less ideal
states, not necessarily a complete account of an actual sequence of events.
Starting at a trough of economic activity, recovery is characterized by a
rightward shifting AD schedule in response to an accelerated growth of
money or autonomous increases in the expenditure functions. AD initially
shifts along a relatively flat AS schedule, with output rising relatively more
than the price level. Two factors contribute to this pattern: To regain their
jobs unemployed workers are willing to allow money wages to rise less than
the price level, reducing the real wage; at the same time, productivity
gains for labor are likely to be substantial in the earlier stages of recovery,
so that even with an increase in real wages, marginal production costs do
not necessarily rise or rise very much. 8 As full employment of resources is
approached, continued increases in prices tend to evoke equiproportionate
increases in money wages. Since AS is constructed on the assumption of a
fixed money wage, the increase of money wages gradually pushes real
wages above productivity and causes AS to shift leftward.
Eventually the excess supply of labor disappears and a point of full
employment is reached in which output (relative to trend) is at capacity and
can rise no more. AD continues shifting to the right, however, causing
prices, wages, and other costs to rise equally. Accompanying every rightward (or, equivalently, upward) movement of AD is thus an equal leftward
(or upward) movement of AS. Output is constant and prices continue to
increase.
Although what constitutes an acceptable rate of inflation will differ
in time and place, a continuing inflation will sooner or later become politically intolerable to the administration, whichever party is in power. The
money supply is tightened and AD decelerates in its rightward shifts;
relative to trend, or absolutely, it may cease shifting altogether. AS,
possibly for a variety of reasons, fails to adapt to the change of policy and

MACROECONOMICS AND MACROECONOMISTS

139

continues to shift upward along the relatively more stable AD schedule,

causing prices to continue to rise but output to fall. In effect, the supply
side of the economy prices its product (AS shifts upward) in anticipation
of demand increases (upward shifts of AD) that fail to materialize. At the
higher prices (traced along a given AD schedule), sales decline, as do
output and employment. The economy is in a typical post-1949 recession.
The failure of aggregate supply to adapt to the deceleration of demand
may result from an informational gap: suppliers and owners of resources
do not immediately know what is happening to aggregate demand or,
if they do, how it will affect their individual markets. Continuing
supply-price increases may also reflect the skepticism of suppliers that the
monetary authority has the will or the ability to pursue and successfully
implement a policy of restraint.
The lack of coordination between demand and supply pricing will
gradually correct itself, however, causing the upward shifts of AS to
diminish. When they cease (generally after about 11 months) ,9 the decline
of output and employment ends, and the economy is poised for a renewed
period of economic growth with little or no inflation.
Table 7-1 summarizes the price and output changes of eight postwar
recessions and shows the net decline of demand of the 1948-1949 downturn occurring in the first two quarters of 1949 (that is, output and the price
level simultaneously decline). In the last two quarters, output successively
increased and then decreased while prices remained constant. We can infer
that AS and AD shifted equally-to the right in 1949:3 (the third quarter of
1949) and to the left in 1949:4.
In the 1953-1954 recession, however, a net leftward shift of AD
occurred only in 1953:4. In the other three quarters, output fell while
prices rose by varying degrees, indicating that the inflation was supply
induced in the sense that it was caused by a net leftward shift of AS. In
1957-1958 output declined sharply in two quarters, during which prices
rose, though only moderately. A good guess would be that the upward
movement of AS was being effectively restrained by significant decreases
of AD, an expression of the anti-inflationary resolve of the monetary
authority. Indeed, in the 1960-1961 recession prices were virtually flat,
indicating that in the two quarters in which output fell, the inflationary
tendency in supply (the leftward shift of AS) was only strong enough to
neutralize the price effect of the decline (the leftward shift) of AD.
It is a plausible conclusion that three closely spaced Eisenhower recessions, all of which served to restrain inflation, laid the groundwork for the
next nine years of steady growth at low rates of inflation. By the late 196Os,
however, price increases were beginning to occur, and in those quarters of

140

POLICY ANALYSIS AND ECONOMICS

Table 7-1. Annual Percentage Changes in Real GNP (0) and the GNP Deflator
(P) in Eight Post-World War II Recessions.
11148-10/49

1948:4
1949:1
2
3
4

7/53-5/54

-1.7
-3.4
-3.4
0.0
0.0

3.3
-4.7
-2.3
3.0
-3.9

4/60-2161
P

1960:2
3
4
1961:1

0.0
1.3
0.0
0.0

1980:1
2
3

8.4
9.2
9.0

1.5
-3.1
6.2

-3.1
-3.2
-5.6
-1.6

1.5

Q
-1.1
0.4
-3.4
4.1

1969:4
1970:1
2
3
4

5.0
6.9
5.8
2.9
4.7

1957:3
4
1958:1
2

4.2
1.4
2.8
1.4

2.3
-6.2
-8.1
2.2

11/73-3/75

12169-11/70

1/80-7/80
P

1953:3
4
1954:1
2

8157-4158

Q
-1.6
-2.5
-0.3
4.9
-3.6

1973:4
1974:1
2
3
4
1975:1

9.6
5.5
8.5
13.6
10.9
9.9

3.6
-2.2
1.1
-5.2
-3.5
-7.8

7/81-11182
p

Q
4.0
-9.5
0.3

1981:3
4
1982:1
2
3
4

9.1
7.6
6.2
4.9
5.6
3.6

Q
1.8
-5.6
-6.0
1.2
-3.2
0.6

Note: 1948:4 denotes 1948, fourth quarter.

Source: Recessions and upswings are designated by the National Bureau of Economic
Research as reported in the U.S. Department of Commerce periodical, Business Conditions
Digest (January 1987: 104). All percentage changes are computed from data in U.S.

Department of Commerce, Bureau of Economic Analysis (1986: 37-39 and 335-337).

the 1969-1970 downturn in which output fell, the inflation was pure supplypush (caused by a net leftward shift of AS), as table 7-1 indicates. A
similar pattern appears in the recessions of 1973-1975, 1980, and 19811982, in which all reductions in output were accompanied by significant
price increases, the result of leftward shifts of AS. As it happened, these

MACROECONOMICS AND MACROECONOMISTS

141

shifts were exacerbated by the world oil disruptions then occurring. The oil
price shocks will be discussed in a later section.
Like the 19605, the postrecession 19805 experienced good growth at
tolerable rates of inflation, inflationary expectations having been more or
less wrung out of the economy in the Carter/Reagan recessions. The tax
cuts of 1964 and 1981 also appear to have contributed to the prosperity, as
did the absence of destabilizing shocks and a monetary policy that on
hindsight was roughly stabilizing in the upswings of both decades.
As noted, the foregoing sketch of the postwar business cycle is schematic, not necessarily historical. The pattern of events was in fact irregular, an
upswing sometimes being interrupted by a temporary faltering of the
economy (1952: 2, 1962: 4, 1966: 2, 1968: 4), the recession perhaps ending
with inflation above acceptable levels (1970: 4, 1975: 1), and the supply
side shifting for reasons other than its delayed reaction to a policy-driven
slowdown of demand (for example, as a result of the energy shocks).

8. The Phillips Curve

One of the empirical generalizations drawn from the postwar cyclical
experience was the Phillips Curve, an inverse relation between the rate of
change of money wages (or the price level) and unemployment (Phillips,
1958). The curve enjoyed considerable currency among macroeconomists
and policy-makers in the 19605 and 19705. In his 1967 presidential address,
Milton Friedman (1968) questioned the basis for it, arguing that inflation
could not lower real wages and unemployment below their "natural" level
unless it was unexpected, and even then, only temporarily (see also Phelps,
1968).
In terms of our analysis, increases in the price level, whether anticipated
or not, are effective in lowering unemployment when there is excess supply
in the labor market. In that circumstance inflation may be the only way to
bring about the reduction in real wages that will clear the market. But once
the excess supply disappears, Friedman is surely correct that further
increases in the price level will not permanently raise employment.
In general, both the labor market and financial markets appear to have
seriously embraced inflationary expectations for the first time in American
history in the late 19605. Prior to that time U.S. markets seem to have
regarded peacetime inflation as a rare, nonrepeatable event-which, in
large measure, it was.
The anticyclical pattern of real wages implied by the Keynesian analysis
of the labor market is often challenged by claims that real wages are in fact

142

POLICY ANALYSIS AND ECONOMICS

acyclical or even procyclical (Mankiw, 1988: 446). The total wage bill,
however, as a percentage of GNP is markedly anticyclical in every cycle
since 1929, with the possible exception of the 1937-1938 downturn. 10 This
could indicate that changing technology and overall factor productivity
alter the net burden of real wages-however they alone are changing-in a
direction consistent with the Keynesian hypothesis.

9. The Oil Price Shocks

Macroeconomists had a difficult time interpreting the macro impacts of the
world oil supply disruptions of 1973-1974 and 1979-1980. The ongoing
postwar debates between Keynesians and monetarists focused on the
superiority of one set of demand-side variables (Keynesian expenditure
and asset-holding functions) over another (money and velocity). Neither
camp was particularly adept at analyzing disruptions originating on the
supply side of the system, although world food shortfalls due to weather
and natural disturbances had occurred in the early 19705.

9. 1. The Oil Supply Disruption as a Macro Event

As noted earlier, the macroanalysis of the oil shocks is straightforward.
The worldwide reduction in supply and increase in price of the most widely
used energy resource raised the marginal cost of producing almost every
component of the world's GNP. In response, every country's aggregate
supply curve (including those of OPEC and other oil exporting countries)
shifted up or to the left, reflecting those higher costs and, simultaneously,
lower factor productivities. The upward shift of AS along a given AD curve
reduced output and raised the general price levelY In the equation of
exchange, Q fell autonomously and P rose to balance it.
The large econometric models, such as the Data Resources Incorporated and the Wharton, were based on the Keynesian demand-side-only
model of the 1930s and early 1940s. In the absence of an independent
production sector, which, together with the demand equations determined
output and the price level, there was no feasible way to introduce the oil
shocks and come out with the correct answers. Manipulators of the models,
knowing that the shocks entailed a loss of income, tended to treat that
as the disturbance. Entering a reduced income into the model reduced
consumption and, through various steps, investment. In effect, aggregate
demand, rather than supply, shifted to the left. The impact on the price

MACROECONOMICS AND MACROECONOMISTS

143

level was problematical and varied with the model and the operator.
Usually prices rose at first but then, as might be expected in a demand-side
shock, fell (see, for example, Curtis, 1979).
A number of government agencies, most notably the Department of
Energy (DOE), relied on the big econometric models. DOE, under
Secretary Schlesinger, mandated that all its departments use DRI for their
analyses. But some agencies, such as the Council of Economic Advisers
(CEA), had more flexibility and were able to employ a wider view of
macroeconomic relationships. The policy analysis, however, was complicated by the fact that both oil price shocks of the 1970s occurred just as the
economy was about to enter a downturn in which aggregate supply, driven
by perverse inflationary expectations, was already drifting upward along
a stalled (monetary-restricted) aggregate demand. The energy shocks
accelerated the leftward movement of supply, intensifying the inflation and
the decline of output and employment.
Few, if any, macroeconomists saw the first energy shock or the simultaneous cyclical downturn in these stark supply-demand terms. In their
February 1974 report, the CEA, though recognizing that productive capacity might be reduced by the oil curtailments, felt that capacity was as yet
unaffected (CEA, 1974: 24). They discussed the weaknesses in real income
and demand, but without distinguishing between two possible causes: (1)
an independent reduction of demand-side variables, such as the quantity of
money, and (2) an independent reduction in aggregate supply, whether
due to uncoordinated supply-side pricing or the energy shock. In the first
case demand is weak in the sense of a leftward shift of the aggregate
demand schedule due, in this instance, to a restrictive monetary policy. In
the second case demand is weak because of an induced upward movement
along aggregate demand. Only the latter, of course, explains the simultaneous reduction of real income and rise in the price level. The 1974
Economic Report explains the inflation by reference to the importance of
energy and food prices in the overall inflation rate (CEA, 1974: 28). But
this is tautologous: Any price index is an identity with respect to its
components, none of which can themselves explain anything as to the
index's overall level or underlying determinants-for example, whether
the price changes originated in autonomous supply or demand forces.
In this connection a curious assertion, commonly made by monetarists
but others as well, is that, in the absence of monetary accommodation, the
rise in energy prices tends to be offset by an equal decline in other prices. 12
Herbert Stein, who was chairman of the CEA that produced the 1974
report but is not a monetarist, remarks in his recent policy book (1984:
185): "If demand had been rigorously controlled by tight fiscal and

144

POLICY ANALYSIS AND ECONOMICS

monetary policy, the rises of food and petroleum prices might have been
absorbed by an offsetting decline of other prices." There is, however, no
basis for this belief. It makes sense only if the increase in energy prices is
purely a demand phenomenon resulting from a transfer of demand to
energy from nonenergy markets. Moreover, in the context of the oil supply
disruption and aggregate supply shift, with all the likely attendant controls
and natural bottlenecks, it is doubtful that monetary or fiscal tightening
would reduce absolute or relative prices very much. Most of the demand
curtailment would fall on real output and employment.
In the second disruption, 1979-1980, the CEA did not offer a more
logical or internally consistent macro interpretation of the events than the
council had in 1974. The Carter CEA seemed equally and alternately
concerned with restraining the inflationary impact of the oil shock and
dealing with a possibly massive "oil price drag." The latter phenomenon is
a deflationary decline (leftward shift) of aggregate demand owing to a
diversion and piling up of funds in the hands of domestic and foreign oil
producers. It could be a particular problem in oil markets because the
short-run demand for oil and oil products is very inelastic-a rise in price
causes a relatively much smaller reduction in consumption, resulting in an
increase in the total dollar outlay on oil. In the 1970s, with crude oil prices
first quintupling and then, later, more than doubling, the increased flow of
funds to oil producers, domestic and foreign, was enormous. The possibility of drag was addressed with growing concern by the CEA's Economic
Report of 1974 (p. 29), 1975 (pp. 30,42), and 1980 (pp. 29, 51), all of which
feared that the companies would be unable to recycle their receipts fast
enough to avoid a serious loss of purchasing power from the economy at
large. The empirical evidence, in this author's opinion, does not support
their concern. 13

9.2. The Macropolicy Response

Meanwhile, the 1980 report did not see the increase in oil prices as directly
raising production costs or, in our terms, shifting AS to the left. Instead,
it viewed the oil price hikes as contributing to general inflation and to
heightened inflationary expectations, particularly by wage earners. The
cure for this, it declared, was strong monetary and fiscal restraint (CEA,
1980: 79). But if the report had distinguished between leftward shifts of
AS due to inflationary expectations and those caused by a shrinkage of
capacity output due to higher oil prices, it could not have taken quite so
unqualified a stand for demand tightening.

MACROECONOMICS AND MACROECONOMISTS

145

There is in fact little that macropolicy can do for an oil-constricted

economy. Demand restraint, as noted, will likely reduce output much
more than prices-to no apparent net gain. The rise in prices due to an oil
shock is, moreover, a one-time effect, not a component of an ongoing
inflation. Considering only the oil shock, some monetary or fiscal ease
would be a far more appropriate policy response. It would facilitate the
reallocation of resources dictated by the higher price of energy and it
would counter any possible oil price drag. In the context of upward shifting
AS schedules due to perverse inflationary expectations, macropolicy,
which has directly triggered such supply behavior by restraining aggregate
demand, might pursue a middle course-restraining demand less than if it
had only the business cycle to deal with.
The monetary authorities, however, possessing even less sophistication
than others as to aggregate supply, demand, and their interaction, treated
the inflations of the 1970s as purely demand-driven and calling for unqualified restraint. 14 Though occurring on a much smaller scale, the situation in
the 1970s was comparable to that of 1929, 1939, or 1946, when a huge
adjustment in the composition of output loomed, and for which monetary
ease and stability were necessary catalysts. Instead, as in 1929, the monetary response in the 1970s was contractionary and destabilizing. 15
The central banks of all the countries of the Organization for Economic
Cooperation and Development plunged ahead with monetary restriction
whose harmful effects were magnified by the energy-induced leftward
shifts of aggregate supply. Several researchers have argued that tight
money, both in 1974 and 1979-1981, was responsible for a greater reduction
in output and employment than was caused directly by the rise in oil prices
(Bohi, 1989: chap. 7).
9.3. Policy with Respect to Oil Market Controls

If macropolicy cannot do much to protect or restore an oil-disrupted

economy, particularly in the stagflation phase of the cycle, other varieties

of public policy may be more helpful. The construction and drawdown of
national petroleum stockpiles is an obvious direct offset to the loss of oil in
the world economy (Horwich and Weimer, 1984: chap. 4; 1988: chaps. 5,
9). Public policy can make a further contribution to economic welfare by
dismantling any prevailing oil price ceilings or mandatory fuel allocations,
which almost invariably prevent resources from going to their most valued
uses. The 1974 report gingerly suggested that higher prices might accomplish much of what needed to be done in energy markets; it did not criticize

146

POLICY ANALYSIS AND ECONOMICS

the controls or discuss their ability to damage the economy (CEA, 1974:
33). The 1975 report, after more than a full year's experience with controls,
called forthrightly for their removal from oil and natural gas, for construction of a strategic petroleum reserve, and for general reliance on free
market prices (CEA, 1975: 21, 30).
In June 1979 the oil decontrol process initiated by President Carter
began, and the 1980 report endorsed it (CEA, 1980: 107-108). But its
approval was qualified. While applauding more rational prices, it urged
establishment of a standby fuel rationing plan and, by implication, price
ceilings on gasoline, diesel fuel, and heating oil for use in a major
disruption (p. 106). It warned that decontrol would bring producer windfalls that would do nothing to increase supplies and that decontrol generally would reduce the real incomes of consumers (p. 109). The latter
assertion is diametrically at odds with the conventional demonstration of
the gain of consumer surplus from removal of price ceilings (Arrow and
Kalt, 1979; Ford Foundation, 1979: chap. 5). Price ceilings and product
rationing make internal sense only if decontrol and free-market prices
are seen as increasing consumer vulnerability to oil price drag: The
ceilings may limit the drag by limiting the transfer from consumers to the
oil industry. 16 But when one considers that U.S. controls, by discouraging
production and subsidizing imports, resulted in increased world oil prices
(Horwich and Weimer, 1984: 70, 72, 107), it is unclear that drag would
be less under domestic controls than in uncontrolled markets.
10. The 1980s: The Fiscal Deficits

Macroeconomists frequently lament the absence of a systematic theory of

the optimal mix of monetary and fiscal stimulus. An even larger problem,
in this writer's opinion, is our inability to know what the deficit is. We do
not have a widely accepted standard procedure for separating endogenous
from exogenous components or for looking at the deficit as part of the
government's total income and wealth position. The measurement of
monetary stimulus suffers from the same failure to distinguish between the
change in money that is endogenous from that which is exogenous-that
is, attributable to the monetary authority. A measure of the exogenous
money supply would certainly be worth having (see Hendershott, 1968).
But it is less critical than for fiscal policy because money is more easily and
readily altered and more subject to government control than are the
federal budget and its determinants.

MACROECONOMICS AND MACROECONOMISTS

147

In the 1980s, as the measured deficit reached historic peacetime highs

even as the economy flourished, the criticism of established government
accounting practice mounted. The complaints came from an unusual
mixture of ideological voices in the economics profession, all in agreement
that to continue lumping everything government does into a single current
consumption account was highly misleading (Eisner and Pieper, 1985;
Darby, 1987). The inability of recent econometric studies to find a correlation between the deficit and interest rates (Bailey et aI., 1984; Evans, 1985)
suggests at least the possibility that the measured deficit is not a very
meaningful magnitude.
In retrospect, we should have realized the inadequacy of the conventional measure. Some of the simplest suggested adjustments make for
enormous changes in the deficit:
The observed deficit as a fraction of the GNP is a scale adjustment
that obviously should be, and in fact usually is, made routinely.
2. The high-employment deficit, a measure of the exogenous (noncyclical) component, is often significantly different from the
observed deficit even on the basis of a small difference between the
prevailing and an assumed full-employment level of unemployment.
John Tatom (1984) reports on the Commerce Department highemployment deficit series, which is based on a 5 percent level of
frictional unemployment. Expressing the Commerce series as a
fraction of GNP reduces the relatively traumatic deficits of 1958:2
and 1958:3 (they "curled" the hair of Treasury Secretary Humphrey) from almost 3 percent observed to 0.0 percent and 0.5
percent adjusted, respectively; the 1975:2 deficit from 6.5 percent
observed to 3.5 percent adjusted; the 1982:2 deficit from almost 4
percent observed to 0.2 percent adjusted; and the 1982:4 deficit
from almost 7 percent observed to 2 percent adjusted. On this
procedure alone, there do not appear to be any really serious
deficits after World War II and before 1975 and the mid-1980s. All
the pre-19708 deficits and most of those of the 19708 reflected the
operation of the automatic stabilizers and little else.
3. If one takes the first step in a total balance sheet approach to the
government's fiscal position and adjusts the current account for
the change in the inflation-adjusted value of outstanding liabilities
(the federal debt), the effect on the current deficit can be overwhelming. The public holds over $2 trillion in government bonds; a
5 percent inflation reduces their value more than $100 billion,
1.

148

POLICY ANALYSIS AND ECONOMICS

which, when subtracted from recent deficits, reduces them by half

or more. Increases in interest rates can also lower the value of
outstanding debt and current deficits significantly. From a policy
perspective, these adjustments for the change in the real value of
debt are especially important. There are both theoretical and
empirical reasons for believing that the interest-rate inflation premium on government bonds compensates the public for losses of its
government-bond portfolio due to inflation (Cagan, 1981: 3). The
public makes up these losses by applying revenues from the inflation
premium to the purchase of an equal additional amount of government bonds. To that extent, the current deficit is neutral in its
impact on the capital market.
4. So-called Ricardian equivalence, whereby the public increases its
saving in anticipation of future tax liabilities, amplifies the reaction
described in item 3. A long-held belief, first voiced (but not endorsed) by Ricardo, is that households will save more in response to
increased government deficits in anticipation of taxes that must be
levied to pay the interest on the debt (Bailey, 1962: 71-80; 1971:
chap.9; Barro, 1974). There are some dramatic examples of this
possible tendency, such as the 50 percent and 79 percent increases
in the ratio of saving to GNP in World Wars I and II, respectively
(Horwich and Bjornstad, 1991). A government capital account
would help, of course, to distinguish between government expenditures that are pure consumption and those that add to social
capital. The latter will provide its own saving and funding to pay
the interest on the debt (Kormendi, 1983: 1005-1(07).
The larger step of classifying government expenditures into consumption and investment categories is more difficult. But the political thicket
this effort would encounter should not preclude such an attempt. It could
be done to a limited degree (as other countries do), including in government investment, at an initial stage, only physical assets such as buildings,
roads, and dams. Almost certainly the general task should be delegated
to a private nonprofit agency, such as the National Bureau of Economic
Research (which is also the organization that determines the business
cycle turning points).
Alternative measures of the deficit, including various degrees of consolidation with state and local government (which ran enormous surpluses
in the 1980s) and various degrees of disaggregation with respect to the
Social Security and other pension funds, would also be useful.

MACROECONOMICS AND MACROECONOMISTS

149

11. Summary and Conclusion: The Major Polley Issues

During most of the period covered by this survey, macrotheory focused on
demand-side variables, moving from the relatively simple world of money
and its demand (or velocity) to the more complex Keynesian expenditure
models. By the end of World War II, the convulsive economic events of the
1930s and 1940s, including the phenomenal wartime expansion, had made
Keynesianism the preeminent doctrine among macroeconomists. A large
government fiscal presence, particularly on the side of spending, was
widely regarded as necessary to compensate for the weakness of private
investment in a mature economy. At the same time an automatic countercyclical role was created by the increased level and progressivity of income
taxes and various government social programs enacted in the 1930s and
1940s.
Although Keynesianism was the predominant macro framework, it was
not ensconsed in the highest policy circles until Walter Heller became
chairman of the Council of Economic Advisers in 1961. And while the
early Keynesians had emphasized government spending as the primary
stabilizer, a change in tax rates was used effectively in the 1960s in the
celebrated tax cut of 1964. The use of a surcharge to contain Vietnam-era
spending in 1968 was widely judged to be ineffective because the public
viewed the policy as only temporary, which it was.
11. 1. The Revival of Money

Independent monetary policy suffered doctrinal eclipse following its failure

in the 1929-1933 downturn and its de facto paralysis under the 1941
Federal Reserve commitment to support interest rates at designated low
levels. At the same time Keynesians had undermined the theoretical
underpinnings of monetary policy by arguing that, in any case, it was
ineffective in the low interest rate, low investment world that had emerged
from the 1930s and 1940s.
As observed earlier, the successful World War II demobilization and
peacetime conversion lent renewed credence to the importance of real
balances--the centrality of money-in economic activity. But it was not
until April 1951 that the Federal Reserve-Treasury Accord freed the
Federal Reserve from its hand-tying policy of automatically supporting
interest rates at low levels. As we saw, three recessions followed in the next
nine years, all preceded by monetary restraint of one degree or another.

150

POLICY ANALYSIS AND ECONOMICS

But the restraint was compatible with overall GNP growth in the 1950s that
was quite satisfactory (3.3 percent compounded annually) and, as noted,
created a noninflationary environment for the long upswing of the 1960s.
11.2. Aggregate Supply

Serious macroeconomic attention to the production or supply side of the

system did not appear until the 1960s, when it surfaced as discussion of the
Phillips Curve. This inverse relationship between the rate of change of
money wages or the inflation rate and unemployment was one of the
implications of the upward-sloping aggregate supply curve-an increase
in the price level evokes an increase in employment and a decrease in
unemployment. According to the Keynesian interpretation of the labor
market, employment rises only if the real wage falls; thus the accompanying rise in money wages cannot be as great as the inflation rate. This
remains the most widely accepted explanation of labor market behavior,
but it has not been verified empirically. Rather than moving countercyclically, as the theory predicts, real wages tend to be procyclical or even
acyclical. An internally consistent, fully empirically validated explanation
of the aggregate supply/price level relationship remains elusive.
Meanwhile, the Phillips Curve came under attack by those who claimed
that, in general, workers would not allow inflation to reduce their real
wages unless it was unexpected, and even then only temporarily. But the
Phillips Curve and the aggregate supply curve that underlies it are valid if
there is excess supply in the labor market; that is, unemployment is above
the natural or frictional level. As inflation gradually rose in the late 1960s
and in the upswing of the late 1970s, plots of the Phillips relation indeed
kept shifting to the right, suggesting that workers were not allowing
inflation to reduce their real wages very much and that the positive
response of employment to inflation was growing weaker.
Some of the rightward shifts, notably in 1970, were so great as to trace a
positive inflation/unemployment path. This was the Phillips counterpart to
the upward shifting aggregate supply schedules as they lagged the deceleration of upward shifting aggregate demand, catapulting the economy into a
typical post-1949 inflationary recession. Adding to the upward shifts of
aggregate supply (and the rightward shifts of the Phillips Curve) were,
of course, the oil price shocks, which increased inflation by decreasing
productive capacity.

MACROECONOMICS AND MACROECONOMISTS

151

11.3. The Decline of Countercyclical Fiscal Policy

The United States enters the 1990s with monetary policy as the single
flexible tool of demand management. The only certain role of fiscal policy
in the short term is the operation of the automatic stabilizers. In the 1980s
a tax cut (1981) and reduction of marginal rates (1986) provided shortterm and, in the latter case, longer-term stimulus, but such measures are
extremely cumbersome politically and cannot be counted on for routine
countercyclical duty. Moreover, the measured deficit component of the
budget has been criticized for (1) being a meaningless end product of the
crude and unsatisfactory way in which government keeps its accountseverything goes into a current statement (there is no capital account);
and (2) the very real possibility that the economic impact of the deficit may
be partially or totally offset by increases in household saving that anticipate
the future taxes required to service the additional debt. At this moment,
fiscal policy as a short-term tool is in a highly unsettled state: in particular,
the deficit, accurately measured, is probably neither as large nor as
harmful as popularly supposed.
11.4. A Monetary Rule?
A useful way to conclude this chapter would be to enunciate a policy rule
for the stock of money or the monetary base. In his historical survey, Stein
(1984) described the failures of policies that, at various times, focused on
narrow targets: unemployment, always a sluggish indicator of the state of
the economy; the price level, as often as not poorly measured by the
various indices; the market rate of interest, determined by many forces,
many of them international, and thereby an unreliable target of stabilization without detailed knowledge of other variables. To this list we can add
the exchange rate, which, like the interest rate, is too endogenous to be
manipulated by the authorities with impunity.
It is tempting to entertain the use of a monetary policy rule, such as
constant growth of the monetary base. Such a rule might at least have
avoided the monetary debacles of the 1930s. It might also have provided
greater stability and softened the recessions of 1948-1949, the 1950s,
1960-1961, and 1969-1970, none of which were associated with a severe
nonmonetary shock. In the face of such shocks, it is probably neither
realistic nor desirable to adhere rigidly to any rule. For wars such as World

152

POLICY ANALYSIS AND ECONOMICS

War II, in which government defense spending rose from 1.4 percent of
GNP in 1939 to 42.8 percent in 1944, it is unrealistic and almost certainly
suboptimal not to allow additional money and some inflation to reduce the
necessary increases in taxes and real interest rates. But even for smaller
mobilizations, an increased monetary or fiscal stimulus may be desirable.
For supply-side shocks, such as the adjustments imposed by the energy
disturbances of the 19708 and 1990 or, even more, the vast transformation
required of the 1930s, as described by Bernstein (1987), it is unlikely that a
fixed rule would or should be tolerated.
None of the preceding, however, should be construed as a case for fine
tuning if by that we mean the mindless, almost automatic support of
interest rates--a policy we have grown accustomed to; the response to
changes in price indices before a trend has been established; or the focus
on unemployment rates that may contain a large structural-in addition to
a cyclical~omponent.
Finally, we need to underscore the fact that monetary policy is not a
panacea. If monetary restraint has triggered the usual inflationary recession, the only effective response may be resolute determination by the
authorities to stay the course, thereby defusing the public's inflationary
expectations. In future upswings the Federal Reserve may want to exercise
tighter control of money to limit the need for later reversal. In response to
supply-side shocks, monetary policy should avoid compounding the adjustment with a monetary shock. In these cases the most useful policy moves
are likely to lie outside the macro realm-the drawdown of strategic
stockpiles and the removal of price ceilings, mandatory allocations, and
other impediments to the free operation of markets. Although price
controls make no economic sense as a response to supply shocks, they may,
in principle, be imposed effectively to dampen general inflationary expectations triggered by demand forces. In practice, however, they are
unlikely to be used expeditiously-they are the ultimate in fine tuning. The
worst part of the 1970 recession was the impatience with the subsequent
recovery and the decision to impose price controls soon after it began.
11.5. Beyond Supply and Demand?

This chapter concludes by repeating the claim that aggregate supply and
demand is a trustworthy tool for understanding what macro-policy-makers
need to know. Although it is a barebones model and needs constant
refinement and elaboration, it is a great organizer. At the same time, new
theoretical developments lie ahead, including the exciting possibilities of

MACROECONOMICS AND MACROECONOMISTS

153

rational expectations. I7 When and if they reach the policy stage, however,
they will not render supply and demand, whether micro or macro, obsolete. On the contrary, one would expect to know more at that point about
the aggregate schedules and how and where they will shift in response to
policy initiatives. Properly armed, macroeconomists can thereby playa
meaningful role in the effort to mitigate the impact of future disturbances.
Acknowledgment

For many helpful comments and cntIcisms, I thank John Carlson, Stan
Engerman, Glenn Hueckel, Dan Kovenock, and Steve Sullivan.
Notes
1. See the exhaustive and scholarly treatment of the monetary events of the 1930s in
Friedman and Schwartz (1%3), chapter 7.
2. See Stein (1984) for a definitive treatment of this and other central episodes in
post-World War II macroeconomic policy. On the material in this paragraph, see p. 33.
3. National income and government expenditure and receipts data throughout this study
are taken from various issues of CEA, Economic Report of the President and V.S. Department of Commerce (1975, 1986).
4. Michael Darby (1976) discovered that the official employment statistics of the 1930s
counted as unemployed some 3.5 million workers employed in various government work
programs: the Civilian Conservation Corps, the Works Progress Administration, and the
Public Works Administration. Counting these individuals as employed would reduce the
official unemployment rate by over 6 percentage points.
5. This was the view expressed by William McChesney Martin of the St. Louis Federal
Reserve Bank, as reported in Friedman and Schwartz (1963: 522).
6. This and the section on demobilization draw on Horwich and Bjornstad (1991).
7. By mid-1946 prices were essentially free of controls and reflected with reasonable
accuracy the real costs of goods and services. The 1946 level of real balances can therefore
also be regarded as accurately measured.
8. Firms, of course, make substantial investments in the training of their personnel and
often find that it pays to retain workers throughout the business cycle, avoiding the costs of
rehiring and retraining. In view of this, measured productivity per worker will tend to fall with
the decline of output in downturns, when the work force is above the short-run optimum, and
rise in the subsequent upswing when output expands. On long-term contracts and labor
"stockpiling," see Samuel Morley (1979: 45-59).
9. This is the average length of recessions in the post-World War II period. See V.S.
Department of Commerce (1987: 104).
10. Author's worksheets are available on request. The relevant time series can also be
expressed as the ratio of real wages to average labor productivity: letting w be the money
wage rate and N the level of employment, the ratio of real wages (wIP) to average labor
productivity (QIN) reduces to wNlPQ, which is labor's share of the national product.

154

POLICY ANALYSIS AND ECONOMICS

11. The earliest clear statement of the supply-side nature of an oil-supply disruption is by
Phelps (1975: 53). See also Phelps (1978.) It is possible, and indeed probable, that the oil
shocks will cause investment to decline in the sense of a schedule shift. In response, AD itself
would shift to the left. There is no indication, however, that any such shift of AD, which
exerts downward pressure on the price level, was as great as the leftward shift of AS, which
raises prices and was unquestionably the dominant shift in 1973-1974 and 1979-1980.
12. Phelps (1975: 52) also criticizes the assertion.
13. The Carter CEA (CEA, 1980: 65) defined gross oil price drag as the "gross value of
changes in cost [exceeding the rate of inflation) that users of oil and refined products must
pay." Net drag was obtained by subtracting from gross drag expenditures of oil producers on
goods in U.S. markets in the same year: an assumed 30 percent of after-tax revenues by
domestic producers and 20 percent by foreign producers (spent on U.S. exports). On this
procedure, the CEA estimated drag for 1979 at 2.1 percent of GNP and projected it at 3
percent for 1980. An alternative measurement of drag based on the actual balance sheets and
income statements of U.S. oil companies, and the assumption, following the CEA, that
foreign oil companies spend only 20 percent of their receipts on U.S. exports of goods in the
current year, yielded much lower estimates for 1973, 1974, 1979, and 1980: 0.83, 1.86, 1.39,
and 0.45 percent of GNP. See Horwich and Weimer (1984: 29-30, 152-165, 200-13).
14. Simulations by John Tatom (1981) of the St. Louis Federal Reserve Bank and Mork
and Hall (1980), employing versions of the aggregate supply-demand framework, attribute
one-fifth to one-fourth of the measured inflations to the rise in energy prices. But that fraction
does not necessarily capture the full extent of the resource reallocation ultimately imposed
by the oil shocks.
15. The 1970s and the early 1930s are alike in that both experienced leftward drifting AS
schedules-the 1970s because of the oil shocks and the 1930s probably as a result of the
adjustments described by Bernstein (1987). Both decades also suffered policy-induced
leftward shifts of demand, which, in the 1930s but not the 1970s, were dominant. The Carter
CEA, however, fully expected oil price drag to mushroom and create a deficiency of demand
that would be the major consequence of the oil shocks. See the next section.
16. The desire to retain or impose oil price ceilings in a disruption in order to stave off oil
price drag is made explicit in numerous contemporaneous CEA memoranda and correspondence. For example, a memorandum of April 21, 1980, from George Eads, a
member of the CEA, to various OMB, DOE, and White House staff personnel asserts:
"the potential for oil price drag would be reduced with coupons" (p. 14), "Using a new
currency [i.e., rationing coupons) automatically neutralizes much of the effect of potential
oil price drag" (p. 20).
17. One possible way to anticipate the impact of expectations on macro behavior is to
employ techniques of experimental economics (Smith, 1980), feeding subjects realistic (that
is, not necessarily perfectly adjusted) measures of macroeconomic variables such as the
federal deficit, the price level, the stock of money, the rate of unemployment, and the
Commerce Department's index of leading indicators. There is, moreover, no reason to limit
the analysis of macropolicies to the aggregate supply and demand framework. Growth
models, for example, may be more appropriate for analyzing longer-run phenomena, such as
the rate of saving or the ultimate impact of a merchandise trade deficit (see Gordon, 1990).

References
Arrow, K. J., and J. P. Kalt, Petroleum Price Regulation: Should We Decontrol?
(Washington, D.C.: American Enterprise Institute, 1979).

MACROECONOMICS AND MACROECONOMISTS

155

Bailey, M. J., National Income and the Price Level (New York: McGraw-Hill,
1962, 1971).
Bailey, M. J., G. P. Balabanis, G. Tavlas, and M. Ulan, "U.S. Deficits and Interest
Rates," manuscript, U.S. Department of State, Office ofthe Under Secretary of
State for Economic Affairs (May 11, 1984).
Barro, R.J., "Are Government Bonds Net Wealth?" Journal of Political Economy
82:6 (November-December 1974), 1095-1117.
Barron, J. M., M. A. Loewenstein, and G. J. Lynch, Macroeconomics (Reading,
Mass.: Addison-Wesley, 1989).
Bernstein, M. A., The Great Depression (Cambridge: Cambridge University Press,
1987).
Bohi, D. R., Energy Price Shocks and Macroeconomic Performance (Washington,
D.C.: Resources for the Future, 1989).
Cagan, P., "The Real Federal Deficit and Financial Markets," The AEI Economist
(November, 1981).
Copeland, M. A., "How [to] Achieve Full and Stable Employment," American
Economic Review 34:1 Supplement, Part 2 (March 1944), 134-147.
Council of Economic Advisers, Economic Report of the President (Washington,
D.C.: Superintendent of Documents, 1974, 1975, 1980).
Curtis, W., Analysis Memorandum: Macroeconomic Effects of Petroleum Supply
Interruptions, vol. I (DOE/EIA-0102I48/1) (March 1979).
Darby, M. R., "Accounting for the Deficit: An Analysis of Sources of Change in
the Federal and Total Government Deficits," Research Paper No. 8704, U.S.
Department of Commerce (October 2, 1987).
- - , "Three-and-a-Half Million U.S. Employees Have Been Mislaid: Or an
Explanation of Unemployment," Journal of Political Economy 84:1 (February
1976), 1-16.
Eisner, R., and P. J. Pieper, "How to Make Sense of the Federal Deficit," The
Public Interest 78 (Winter 1985), 101-118.
Evans, P., "Do Large Deficits Produce High Interest Rates?" American Economic
Review 75:1 (March 1985), 68-87.
- - , "The Effects of General Price Controls in the United States During World
War II," Journal of Political Economy 90:5 (October 1982), 944-966.
Ford Foundation, Energy: The Next Twenty Years (Cambridge, Mass: Balinger,
1979).
Friedman, M., "The Role of Monetary Policy," American Economic Review 58:1
(March 1968), 1-17.
Friedman, M., and A. J. Schwartz, A Monetary History of the United States
(Princeton: Princeton University Press, 1963).
Gordon, R. J., "The Gordon Update," Supplemental Newsletter for Use with
Gordon's Macroeconomics, Fifth Edition (Spring, 1990).
Hagen, E. E., and N. B. Kirkpatrick, "The National Output at Full Employment in
1950," American Economic Review 34:1 (September 1944), 472-500.
Hardy, Charles 0., Credit Policies of the Federal Reserve System (Washington,
D.C.: The Brookings Institution, 1932).

156

POLICY ANALYSIS AND ECONOMICS

Hendershott, P. H., The Neutralized Money Stock (Homewood, III.: Richard D.

Irwin, Inc., 1968).
Hicks, J. R., "Mr. Keynes and the 'Classics'; A Suggested Interpretation,"
Econometrica 5:2 (April 1937), 147-59.
Horwich, G., "Effective Reserves, Credit, and Causality in the Banking System of
the Thirties," in D. Carson (ed.), Banking and Monetary Studies (Homewood,
III.: Richard D. Irwin, Inc., 1963).
- - and D. J. Bjornstad, "Spending and Manpower in Four U.S. Mobilizations--A MacrolPolicy Perspective," Journal of Policy History 3:2 (Spring
1971), 173-202.
- - and D. L. Weimer, eds. Responding to International Oil Crises (Washington, D.C.: American Enterprise Institute, 1988).
- - , Oil Price Shocks, Market Response, and Contingency Planning (Washington, D.C.: American Enterprise Institute, 1984).
Hume, D., Writings on Economics, ed. E. Rotwein (Madison: University of
Wisconsin Press, 1955).
Keynes, J. M., A Treatise on Money, vols. I and II (London: Macmillan & Co.,
Ltd., 1950).
- - - , The General Theory of Employment, Interest and Money (New York:
Harcourt, Brace & Co., 1936).
Kindleberger, C. P., The World in Depression (Berkeley: University of California
Press, 1986).
Kormendi, R. c., "Government Debt, Government Spending, and Private Sector
Behavior," American Economic Review 73:5 (December 1983), 994-1010.
Mankiw, N. G., "Recent Developments in Macroeconomics: A Very Quick
Refresher Course," Journal of Money, Credit, and Banking 20:3 Part II (August
1988), 436-449.
Marschak, J., Income, Employment, and the Price Level (New York: Augustus M.
Kelley, Inc., 1951).
Mork, K. A., and R. E. Hall, "Energy Prices, Inflation, and Recession,
1974-75," The Energy Journal 1:3 (July 1980), 31-63.
Morley, S. A., Inflation and Unemployment (Hinsdale, III.: Dryden Press, 1979).
Perry, G. L., "The Success of Anti-Inflation Policies in the United States," Journal
of Money, Credit, and Banking 5:1 Part II (February 1973), 569-593.
Phelps, E. S., "Commodity-Supply Shock and Full-Employment Monetary
Policy," Journal of Money, Credit, and Banking 10:2 (May 1978), 206-221.
- - , "Stopover Monetarism: Supply and Demand Factors in the 1972-74
Inflation," The Japan-U.S. Assembly (Washington, D.C.: American Enterprise
Institute, 1975), pp. 51-68.
- - , "Money Wage Dynamics and Labor Market Equilibrium," Journal of
Political Economy 76:4 (August 1968), 687-711.
Phillips, A. W., "The Relation Between Unemployment and the Rate of Change of
Money Wage Rates in the U.K., 1861-1957," Economica 25 N.S. (November
1958), 283-299.
Phillips, C. A., Bank Credit (New York: Macmillan and Co, 1921).

MACROECONOMICS AND MACROECONOMISTS

157

Smith, V. L., "Relevance of Laboratory Experiments to Testing Resource Allocation Theory," in J. Kmenta and J. Ramsey, eds., Evaluating Econometric
Models (New York: Academic Press, 1980).
Stein, H., Presidential Economics (New York: Simon and Schuster, 1984).
Tatom, J. A., "A Perspective on the Federal Deficit Problem," Federal Reserve
Bank of St. Louis Review (June-July 1984), 5-17.
- - , "Energy Prices and Short-Run Economic Performance," Federal Reserve
Bank of St. Louis Review (January 1981), 3-17.
U.S. Department of Commerce, Bureau of the Census, Historical Statistics of the
United States: Colonial Times to 1970, Part I (Washington, D.C.: Superintendent of Documents, 1975).
U.S. Department of Commerce, Bureau of Economic Analysis, The National
Income and Product Accounts 1929-82. Statistical Tables (Washington, D.C.:
Superintendent of Documents, 1986).
U.S. Department of Commerce, Bureau of Economic Analysis, Business Conditions Digest (Washington, D.C.: Superintendent of Documents, January 1987).

THE NEW TRADE THEORY:

IMPLICATIONS FOR POLICY
ANALYSIS
John Pomery

1. Overview
Changes in the way economists perceive international trade have led to a
greater variety of trade policy prescriptions, each given with muted confidence. A selective sample gives some hint of the diversity. Paul Krugman
(1987: 132) puts the case for free trade:
free trade is not passe, but it is an idea which has irretrievably lost its innocence.
Its status has shifted from optimum to reasonable rule of thumb. There is still a
case for free trade as a good policy, and as a useful target in the practical world
of politics, but it can never again be asserted as the policy that economic theory
tells us is always right.

Reviewing a book by Elhanan Helpman and Krugman (1989), Robert Lucas

(1990: 666-667) questions the value of this "rule of thumb":
Helpman and Krugman seem not so much to be defending the validity of what
they call the "central economic tenet" of free trade as trying to avoid the blame
for being the first to expose its emptiness!

Lucas then goes on to assert that

159

160

POLICY ANALYSIS AND ECONOMICS

theory does not provide us a blanket vindication of any single, universally

applicable policy conclusion, but it does provide a coherent framework for
examining specific policy interventions on their merits, case by case. I think it is
a mistake to ask more than this from policy analysis.

Later Lucas concludes, concerning some of the "surprising implications"

emerging from the "Smithian" models in the Helpman-Krugman monograph, "They are, I think, outstanding illustrations of why we work to
construct useful, explicit theories rather than being content with good rules
of thumb."
Robert Baldwin (1989: 132-133) expresses a concern for understanding
the political process within which trade policy decisions are made:
If economists wish their advice to carry more weight in the political process, they

must be more willing to examine the influences of institutions and procedures on

policies .... If they are to make policy recommendations, economists need
better empirical and experimental studies of the economic and political feasibility of well-known policy measures.

Later, in the spirit of Krugman's position, Baldwin remarks:

Even if further research ... confirms that there are some real world situations
where protection would be the best available measure for maximizing social
welfare, economists should be very cautious about surrendering the efficiency
argument for free trade.

Judith Goldstein and Stephen Krasner (1984: 284-285) have a different

emphasis again, focusing on the possibility of using strategic policies to
induce greater cooperation among national governments:
The United States is in an ideal position to play Tit for Tat because of its large
domestic market, provided that both parties are clear about the values in the
matrix and the classification of behaviour as cooperation (fair) and defection
(unfair) .... The defense of liberalism by the most liberal solution to trade
distortions has not worked in practice for it does not work in theory. The United
States should meet protectionism in kind. We cannot defend liberalism unilaterally; without pressure, our trading partners will not act in accordance with
GAIT norms. [Emphasis in original.)

Goldstein and Krasner's focus on repeated games echoes a criticism of

Lucas (1990: 666) concerning Helpman and Krugman's narrow concern
with static, noncooperative models of interaction. Going further into the
"political" dimension of the debate, Robert Gilpin (1987: 44-5) asserts:
As a means to understand society and especially its dynamics, economics is
limited; it cannot serve as a comprehensive approach to political economy.
. . . The first of these limitations is that economics artificially separates the

THE NEW TRADE THEORY: IMPLICATIONS FOR POLICY ANALYSIS

161

economy from other aspects of society and accepts the existing sociopolitical
framework as given, including the distribution of power and property rights;
resource and other endowments of individuals, groups, and national societies;
and the framework of social, political, and cultural institutions. The liberal
world is viewed as one of homogeneous, rational, and equal individuals living in
a world free from political boundaries and social constraints.

Jagdish Bhagwati (1988: 17), almost anticipating a Gilpin-style criticism,

adopts a broader view (for most economists) of the policy arena:
Profound commitments to policies are generally due to a mix of ideological
factors (in the form of ideas and example), interests (as defined by politics and
economics), and institutions (as they shape constraints and opportunities).

Bhagwati (1988: xiii), perhaps even more so than Baldwin or Krugman,

has no doubt as to where the enemy is:
To assist the hand of history, so that it does not falter, we will have to reform
and strengthen the institutional framework, national and international, to
harness these pro-trade interests and contain the forces of protectionism more
effectively.

These quotes are selective but not unrepresentative of differing perspectives found in the literature. This chapter will argue that the heterogeneous
views of economists concerning the appropriate trade policy are not, to any
significant extent, empirical issues or normative issues. Rather they reflect
more basic opinions about the interaction of theory, empirics, and policy.
In a way, the difficulty arises from our tendency to seek unambiguous
policy conclusions from economic science. Such unambiguity would seem
to require a strong form of "foundationalism," but the message from
philosophers for the past half-century has been that we cannot locate a
satisfactory foundationalist philosophy. For this reason, the ambivalence
about trade policy, which Lucas so keenly identifies, holds not only
between economists but seems to have a hold on many individuals. There
is a temptation to be pulled a little way toward each of the views already
outlined. The goal here is not to resolve the debate concerning trade policy
but instead to argue that much of the discussion is at cross purposes and
that the attempt to resolve the debate is in itself partially misconceived.
The ambivalence in the policy debate can be summarized in simple
fashion, as partially captured by the previous quotes. A large number,
possibly the majority, of economists seem committed to the view that free
trade is the desired policy stance, at least as a "rule of thumb." Recent
advances in trade theory, incorporating models of imperfect competition,
have cast increased doubt as to the relevance of such a rule of thumb.

162

POLICY ANALYSIS AND ECONOMICS

Many of the models of imperfect competition of the last decade do not

invite a laissez-faire approach to trade policy. Furthermore, the properties
of these models tend to be very sensitive to apparently minor changes in
specification, thus raising a suspicion that there can be no general "rule of
thumb" for trade policy in an imperfectly competitive environment.
The situation is further clouded by the parallel development of classes
of models, involving (for example) strategic policy interaction between
national governments or explicit roles for domestic interest groups, where
the model endogenously predicts trade policy for each country. In such
models it is not clear where an exogenous policy recommendation gains its
leverage to alter behavior. In a related, but always interconnected literature, specialists in political and economic dimensions of international
relations use a different language, not immediately commensurate with
that of many economists.
To these concerns should be added the observation that, despite the
persistence of academic advocacy of free trade, the level of trade barriers
(particularly nontariff barriers) seems to remain obstinately high. Also add
considerable debate as to the appropriate way to compare and interpret,
for example, the recent histories of the U.S. and Japanese economies: Is
the more rapid Japanese growth sustainable? Does it constitute a case for
more active industrial policy? And, if so, can the experience be translated
to the United States? The comparison with Japan has to be melded with
interpretation of recent experiences in Eastern Europe, and of the comparative track records of export-oriented newly industrialized countries
compared to other developing countries. These latter data can be interpreted as offering presumptive evidence of the value of liberal, marketoriented policies with respect to trade and other aspects of resource
allocation.
With all this as background, even the staunchest advocates of free trade
are not completely confident, nor in unity, about the basis for, or the
extent of relevance of, their advocacy. A significant minority in the debate
wonder whether it is time to abandon the traditional loyalty to free trade
once and for all.
2. The Nature of the Dominant Belief in Free Trade
For many years, mainstream economists in the United States have had a
strong tendency to be advocates of free trade (at minimum, as a good "rule
of thumb"!). Even in the past there have been reservations about the
justification of this position in the literature. The impact of the "New

THE NEW TRADE THEORY: IMPLICATIONS FOR POLICY ANALYSIS

163

Trade Theory" has been to alter the grounds, and to amplify the extent, of
this ambivalence. A discussion of the impact of the New Trade Theory
on policy debates should consider the reasons for the earlier ambivalence
and the ways in which recent theoretical developments have altered the
situation.
The case for free trade originates with the study of the international
economy as a system of perfectly competitive markets. The simple textbook versions of traditional models of international trade (such as the basic
Ricardian or Heckscher-Ohlin models) assume price-taking behavior by
the private sector, free entry and exit, constant returns to scale in production, fixed supplies of "nation-specific" factors of production, no
nonpecuniary externalities (that is, complete markets), no uncertainty, no
intertemporal effects, no transportation costs, and more. In this environment, free trade is efficient for the world as a whole (relative to the
assumption of nation-specific factors) and leaves each country being at
least as well off as it would be under autarky (the absence of international
trade).
That the country as a whole benefits from trade does not imply that
everyone in the country will be better off from free trade than from
autarky. As relative prices alter in the move between autarky and free
trade, income redistributes between the net buyers and the net sellers of a
good within a country. In particular there is a presumption that the net
sellers of goods (or factors of production), which become less scarce as a
result of the opportunity to trade in international markets, will be hurt by
the change in relative prices. What might be termed "the surplus revenue
result" says that, given a world price different from national autarkic price
and given the ability to substitute in production or consumption, it is
possible at world prices under free trade to allocate each individual in a
nation enough revenue to leave him or her as well off as at autarky, and for
the nation to have additional revenue left over (Alan Woodland, 1982: sec.
9.2; Wilfred Ethier, 1988: app. J). In this sense, the winners in the move
from autarky to free trade are capable of compensating the losers so that
everyone is better off than at autarky-although whether such compensation takes place is another matter.
There are two traditional qualifications to the argument for free trade.
First, the optimal tariff argument is basically an argument for exertion of
market power by the tariff-imposing country (John Pomery, 1987). From
the national viewpoint, an optimizing country should equalize its trade-off
in consumption (the marginal rate of substitution in consumption), in
production (the marginal rate of product transformation), and in trade.
The world price represents the average rate of transformation through

164

POLICY ANALYSIS AND ECONOMICS

trade and is not equal to the marginal rate l unless the country has no power
to influence price in world markets. If it does possess such power, and if it
can assume that the rest of the world will not retaliate, the marginal rate of
transformation through trade is given by the slope of the foreign offer
curve. The optimal tariff requires the country to reduce its own volume of
trade (in response to the tariff) so that the market-clearing equilibrium
shifts around the rest of the world's offer curve until that offer curve is just
tangent to the highest attainable trade indifference curve for the tariffimposing country.
Second, an argument for trade intervention arises in the case of domestic market distortions. For example, suppose there is a minimum wage
constraint in a Heckscher-Ohlin economy, such that a low relative price of
the labor-intensive good triggers the minimum wage constraint and leads
to unemployment. Autarky can occur at full employment while free trade
leads to a fall in the price of the labor-intensive good and to a dramatic
surge in unemployment. In this case, autarky could be better than free
trade for the nation as a whole, and a prohibitive tariff, which takes the
country back to autarky, would be better than free trade (Brecher, 1974).
With such claims of distortions in domestic resource allocation, it is
desirable to counter the distortion as close to the source as possible. A
wage subsidy would put the country back on its undistorted production
possibility frontier and lead to free trade which is superior to autarky. Thus
for policy recommendations it is important to identify which policy alternatives are feasible and, among those, which works most directly at the
source of distortion. (At an even more basic level, our underlying philosophy of science may influence what is to count as a "genuine" externality.)
Nevertheless, the traditional advocacy of free trade had an apparently
strong foundation arising from the efficiency properties of free trade in
a world of perfectly competitive markets. Perfect competition had an
entrenched "benchmark" status for empirics, theory, and policy, although
qualifications to the intuition of a world of a complete set of perfectly
competitive markets were clearly understood. 2 For example, economists
were well aware of the optimum tariff argument, but they tended to view it
as of limited importance because of its implications for global efficiency,
the likelihood of retaliation, and even its lack of an appropriate moral
commitment. Economists were also aware of the central implication of
second-best theory, namely, that first-best policies may not be optimal in
the presence of existing distortions (Bhagwati, 1971). The possibility that
certain interest groups, such as labor or capital, might prefer restricted
trade to free trade was firmly emphasized by Wolfgang Stolper and Paul
Samuelson (1941).3 Harry Johnson (1953-54) had shown that it was possible
for one country to benefit from a tariff war.

THE NEW TRADE THEORY: IMPLICATIONS FOR POLICY ANALYSIS

165

Advocacy of free trade was not predicated on ignorance of possible

counterexamples nor on a presupposition that free trade reflected the
narrow self-interests of all participants in the global economy. Rather, it
seemed to arise from a combination of beliefs that it was appropriate, as a
first approximation descriptively, to treat the world economy as a system
of perfectly competitive markets and that the goal of free trade, with
its efficiency properties given the perfectly competitive description of the
world, was, normatively, an appropriate ideal. 4
There was also the ever-present concern among those interested in
policy that any admission that free trade was not the appropriate policy
would be seized by special interests or the economically illiterate and used
to subvert the efficiency of global resource allocation. The knowledge that
trade taxes or subsidies were often not the most appropriate form of
intervention, and the prominence of "fallacious arguments" for trade
barriers, gave added strength to this concern.

3. Newer Perspectives on Trade

The results and implications outlined in the previous section were largely
understood 20 or 30 years ago (although the depth of our understanding or
the manner of presenting results often has been modified by subsequent
developments). There is some ambiguity over the extent of the more
recent changes. One possible interpretation of the significance of the
changes in the perspectives of international economists is that the profession is on the brink of totally abandoning the perfectly competitive view of
international trade. A more conservative interpretation treats the recent
developments as leading to greatly increased emphasis on the exceptions
and qualifications of the free trade argument while still preserving the
central insights of the perfectly-competitive view. Thus the changes can be
viewed either as fairly revolutionary or as a significant shift in emphasis
while conserving the original framework. At the same time, sensitivity to
the potential endogeneity of policy decisions, and to the conspicuous
failure of policy-makers to follow the advice of economists, reinforces the
suggestion of an urgent need for a greater attempt to describe the political
process.
It is not possible here to give an exhaustive list of topics and models that
have influenced the way the international economy is viewed, but it is
worth touching on some examples. The coverage here is selective; see
Ethier (1987) for a more extended treatment.
Consideration of national economies of scale (that is, economies of scale
determined by the level of output of the national industry and affecting

166

POLICY ANALYSIS AND ECONOMICS

costs only for firms in that nation's industry) introduces the possibility that
a country that is led to abandon an industry with increasing returns under
free trade can be worse off at free trade than at autarky (Helpman, 1984;
Ethier, 1979). If industries with strong national economies of scale are
associated with positive economic rents, each national government has an
incentive to try to preempt the pattern of world production by subsidizing
its own industry. A parallel result is to be expected if there are dynamic
learning effects that are specific at the national level.
Consideration of international economies of scale (where scale effects
depend on the level of world output for the industry and affect costs in all
firms in the industry worldwide) gives a different intuition, namely, that
countries will receive all the standard gains from trade plus an additional
benefit arising from pooling all production in an international industry that
creates the scale economies. Though the case of international scale economies is less destructive to the intuitions of traditional trade theory than
the case of national scale economies, it does suggest that the international
coordination problem is more complicated than envisaged by the models of
perfect competition under constant returns to scale.
Models of imperfect competition, and particularly of international oligopolies, also undercut any narrow intuition from perfect competition.
Cournot oligopolists fail to set price equal to marginal cost (except in the
limit), and so introduce a domestic distortion. They also achieve economic
rents in foreign markets. The poor coordination aspects of Cournot
oligopolists allow such "strange" phenomena as two-way trade in the same
physical commodity, even in the presence of nonzero transportation costs
(James Brander, 1981). Moreover, the presence of rents creates a case
for increasing national welfare by shifting profits from foreign firms. In
an example that has become a standard in the literature (Brander and
Spencer, 1985), if two Cournot duopolists from different countries sell in a
third country, an export subsidy by the government of one of the firms can
lead to capture of profits (in excess of the amount of the subsidy) at the
expense of the foreign firm. In traditional perfectly competitive trade
theory, export subsidies, when viewed as a potential first-best instrument,
tend to expand trade beyond the optimal volume while creating excess
supply of the exportable good in the world market. Given the lack of
convincing explanation within the perfectly competitive framework about
the prevalence of export subsidies, this type of result contributed to
increased interest in models of imperfect competition.
The results under imperfect competition have to be used with caution
due to lack of robustness of conclusions. For example, the conclusion that
adopting export subsidies is in the national interest has to be significantly

THE NEW TRADE THEORY: IMPLICATIONS FOR POLICY ANALYSIS

167

modified if the producing country whose government intervenes also

consumes the good, or if there is more than one producer in that country,
or if the duopolists interact through prices rather than quantities (Eaton
and Grossman, 1986), or if there are several such duopolistic industries
with all the domestic firms drawing on the same fixed pool of factors of
production (Dixit and Grossman, 1986). (All this without even considering the impact of repetition, retaliation, or the implications for global
efficiency.) In general, situations of imperfect competition may lead to
behavior that may look paradoxical to an intuition honed on models
of perfect competition (Helpman and Krugman, 1989; Richardson, 1989;
Grossman and Richardson, 1985). Trade barriers may lead a "domestic
monopolist" to produce more output or less output for the domestic market;
trade intervention mayor may not preempt entry into a global market by
foreign rivals; trade intervention may increase or decrease tacit collusion
between domestic and foreign firms.
Other areas of economic theory also complicate the possibilities for
international trade policy. For example, if the theory of search is applied to
national labor markets, another source of externalities is introduced
(Davidson, Martin, and Matusz, 1988). The presence of an additional
participant on either side of the labor market alters the probabilities of a
match, and hence the incentive for search, by all participants. The presence of uncertainly with incomplete markets can also generate externalities (Newbery and Stiglitz, 1981; Grinols, 1987; Dixit, 1987). Such externalities again may give rise to justification for intervention (although not
necessarily trade intervention as a first-best policy).
Policy issues have also become hemmed in by the tendency to endogenize the political decision-making process. One of the first hints of the
potential importance of this development was given by Anne Krueger's
(1974) application of the concept of rent seeking to developing countries. In
Krueger's model the presence of an import quota created economic rents;
resources shifted from agricultural production to compete for the fixed
number of jobs in the distributional sector, generating unemployment such
that the expected wage in distribution is equalized to the certain wage in
agriculture. Whereas the natural policy recommendation might be to avoid
import quotas in the first place (and particularly for a small country), the
Krueger application emphasized the possible implications of applying the
paradigm of narrow self-interest to political aspects of resource allocation.
Significant work by authors such as Bhagwati (1982) (summarized in the
notion of DUP, or directly unproductive profit-seeking activities), Ronald
Findlay and Stanislaw Wellisz (1982), Stephen Magee, William Brock,
and Leslie Young (1989) (using game-theoretic models of interest groups),

168

POLICY ANALYSIS AND ECONOMICS

and of Wolfgang Mayer (1984) (applying median voter theory) suggested

that self-centered actions can lead to socially unfavorable outcomes. It is
important to note that these political forms of intervention and redistribution are typically not "first-best" ways to redistribute income. This raises
questions about the appropriate policy instruments to model as vehicles for
DUP-style activities, and also about the level of "social rationality" that
can be expected from the combination of political and economic interaction. (Baldwin, 1989, provides an introduction to a much wider literature
concerning government behavior.)
Another strand of endogenization of trade policy arises from treating
governments as actors in a game between governments (McMillan, 1986;
Bagwell and Staiger, 1990). Typically, each government is assumed to act
as representative of national interests (appropriately aggregated) and to
optimize given its perception of the intergovernmental game being played
(examples include Marie Thursby and Richard Jensen, 1983; Dan Bernhardt and Alice Enders, 1989; John Kennan and Raymond Riezman,
1988). Such models are sensitive to the solution concept used and to the
way that perceptions feed into reputations. Equally important, the fact that
policies are endogenously determined suggests that any policy recommendation may be irrelevant. (If not, policy advocacy would seem to
change the structure of the game.) Policy analysis becomes essentially
descriptive.
These game-theoretic models overlap with the literature on international relations. Some of the concepts in the latter literature have natural
interpretations in the economists' conversation; for example, "hegemony"
has a natural, albeit partial, analog in the familiar concepts of the dominant
firm of industrial organization, or the dominant contributor to a public
good. On the other hand, discussion of national goals and considerations,
such as stability, security, identity, power, or growth (Krasner, 1976;
Gilpin, 1987; Rangarajan, 1984), sometimes involves a flavor of "group
identity" or "group preferences" or of perceptual relationships that are
less individualistic than traditional fare for economists.
4. Changing Perceptions of the International Economy

Over the past two or three decades, many things have changed concerning the international economy. Certainly, the global economy itself has
changed in many ways, with increased openness of most economies,
decolonialization, lowered real costs of transportation, communication and
information processing, areas of deregulation, increased internationaliza-

THE NEW TRADE THEORY: IMPLICATIONS FOR POLICY ANALYSIS

169

tion of financial markets and of the banking system, changes in exchange

rate regimes, increased literacy, and more. Economists' knowledge of the
international economy has increased: The presence of intraindustry trade,
the presence of scale economies, the presence of imperfect competition,
the sensitivity of neo-factoral explanations of trade to the description and
level of disaggregation of factors are all better understood.
Yet an even more important change has come in how economists tend to
perceive perfect competition in relation to other forms of social interaction. The roles of game theory (and certain very special examples such as
the one-shot Prisoner's Dilemma), and of applications of game theory to
industrial organization and to political behavior that influences the economic system, have altered the way economists think about international
trade.
As we have seen, these changes create an awkward situation. Game
theoretic models-with oligopolistic firms interacting at the private level,
or national governments interacting strategically at the level of trade
intervention, or interest groups interacting strategically to influence
national governments--seldom led to an unambiguous conclusion that free
trade is both attainable and in the best interests of either global efficiency
or national welfare. Thus the policy advocacy of free trade is stripped of
the implicit support arising from the presumption that the world is essentially a system of perfectly competitive markets.
The awkwardness runs even deeper than the possible failure of the
"natural" policy prescription of free trade. In many cases, the answer to
the question whether a trade tax or a trade subsidy is a better policy from
the national viewpoint is acutely sensitive to the specification of the model.
It is no longer obvious that there is any single best policy prescription;
rather each situation may have to be evaluated in great detail before any
recommendation can be made.
We have identified a third level of difficulty, which arises because
models of strategic interactions between governments, or of endogenous
optimizing choices by domestic interest groups, suggest that policy decisions are not unrestricted choice variables for the purposes of policy
advocates but may already be determined by "rational" decisions in realms
beyond the narrow market context.
Despite these difficulties, many (but not all) economists are reluctant to
abandon the presumption that free trade is the recommendation of first
resort. What has changed is less the policy prescription than the confidence
with which the prescription is offered, and the grounds used to support the
prescription.
Much of what has been said so far should be relatively uncontroversial,

170

POLICY ANALYSIS AND ECONOMICS

albeit highly selective and at times oversimplified. The potential for

controversy arises in deriving implications from these changes in modeling
the international economy for policy purposes. The range of views is fairly
wide. Some economists see a justification for an actively interventionist
policy, perhaps coupled with an explicit industrial policy (Brander, 1986).
Others deplore the sense of confusion and reaffirm the dominant role of
free trade as the preferred policy. Still others welcome the diversity of
possibilities implicit in a more eclectic approach, viewing this diversity as a
symptom of a richer understanding of the world and as a reflection of the
limits of very sweeping policy recommendations, but they warn of the
likely limited applicability of some of the less traditional implications
of the new theory (Dixit, 1986). Several authors have tried to draw a set of
principles out of the literature on imperfect competition, either with an eye
to showing the feasibility of a carefully constructed trade policy or with an
eye to showing how difficult to implement, and limited in scope, such a
policy might be (Brander, 1986; Krugman, 1986: 15-20, give somewhat
opposed conclusions). Often there is a plea to be cautious and to await
further developments. Other economists suggest a need for increased
research into the description of political decision making and the role of
lobbies (Baldwin, 1989). At times there is a concern that any abandonment
of advocacy of free trade may give comfort to special interest groups
(Krugman, 1987; Bhagwati, 1988). There has been some attempt to get
empirical estimates of the nature and importance of the various effects,
although typically comparing free trade and autarky rather than comparing
free trade with trade under selective interventions (Richardson, 1989).
From a more "management-oriented" perspective, the MIT Commission believes it found a role for a much more interventionist policy, based
on some perceived successes of the Japanese economy and on some
observations on productivity failings of many U.S. firms (Michael Dertouzos et al. 1989; also see William Baumol et al. 1989, where a more
traditional perspective for economists is used to invite a less alarming
conclusion). From another direction, discussions of international relations
by authors with a background in political science seem to assume the
almost-complete demise of the perfectly competitive view of international
trade and the irrelevance of much of the perspective of liberal economists
(Gilpin, 1987: 215).
It would be convenient at this point to pick one of the perspectives
found in the literature and to declare it the winner. Such an approach
would be satisfying since policy debates, being action-oriented, invite an
unequivocal outcome. However, closer inspection suggests doubts whether
it is reasonable to expect such a strong conclusion.

THE NEW TRADE THEORY: IMPLICATIONS FOR POLICY ANALYSIS

171

For example, if one looks at the defense of free trade as a "rule of

thumb" (identified here with such authors as Krugman, Baldwin, and
Helpman), it seems clear the there is a market for policy advice in terms of
broad rules of thumb and that there is room for legitimate concern about
the way interest groups selectively use ideas emanating from academia. At
the same time, the extent to which research has drifted from the perfectly
competitive paradigm raises doubts as to how weight can be put on such a
rule of thumb.
Lucas's view, that we drop rules of thumb and turn to explicit, useful
theories seems to be the most "rational" approach. However, as will be
discussed later, underlying a Lucas-style position appears to be a commitment to a very heroic view of the philosophy of science.
As we move to broader views of the arena of policy debate, the implicit
criticisms of the political economy models and of the multi-faceted views of
Bhagwati (not to mention the explicit criticisms of authors such as Gilpin),
the expanding domain of "positive" description seems to squeeze out an
effective role for policy advocacy.
The remainder of this chapter will raise the following questions: Why
are we driven so strongly to seek an unequivocal resolution of policy
debates? What reasons are there for believing the policy debates are
incapable of complete closure? How does our view of the underlying
philosopy of science affect our stance in this policy debate? Why does the
policy debate always appear to be pushed back by expanding domains of
scientific description? How can we make something constructive out of
policy debate that seems to fracture into sometimes overlapping, sometimes competing perspectives?

5. Policy Debate in an Intellectual Environment of

Antifoundationalism

The way one describes the current situation in trade theory depends on the
philosophy of science that one espouses. For many contemporary economists, it would be natural to consider the debate as a signal of an imminent
paradigm shift. As will be explained later, this way of looking at things is
probably of limited interest. Instead, the discussion is couched in terms of a
clash between "foundationalism" and "antifoundationalism."
In a very crude form, the problem of foundationalism can be summarized as follows: (a) as practical beings, we need to make decisions (about
beliefs, actions, values, et cetera); (b) we do not know how to make
decisions without introducing "foundational" beliefs to underpin those

172

POLICY ANALYSIS AND ECONOMICS

decisions; (c) all attempts to build a credible position to justify those

foundations, whether in mathematics, analytic philosophy, or science,
appear to be seriously inadequate; (d) yet we have difficulty making sense
of a pure "antifoundationalist" world. 5 The upshot, of about 60 years of
debate since the crest of foundationalist thinking, is a curious milieu of
foundationalist and antifoundationalist ideas.
As we evaluate alternative positions with respect to trade policy, part of
the debate is over the role of scientific models. In turn, our differing
conceptions of that role carry different levels of foundationalist presuppositions. One can see the way that an implicit philosophy of science
directs our perception of policy issues. For example, writing almost two
decades ago, Corden naturally follows a traditional distinction between
positive and normative economics. 6
Such a two-stage procedure suggests that, first, the world is described in
a neutral, value-free manner (albeit, following Popper, in a fallible way);
then, given the description, a value judgment is imposed as to how the
world should be. But an adequate neutral, value-free scientific methodology is precisely what foundationalists have been unable to deliver in
this area. 7
The positive-normative distinction is often accompanied by an emphasis
on critical empirical tests. Yet it is not clear whether it is more useful to
think in terms of competing theories or of complementary perspectives.
Even simple descriptions tend to be norm-driven. The competing-theories
view seems to see the descriptive choice as analogous to choosing between
the proposition "It is raining outside" and the proposition "It is not raining
outside." Surely we simply look out the window and record the facts?8
Contrast this decision with the question, "Is this table top flat?" From a
very macro perspective, the table top may look smooth and level; within
that perspective the table top can legitimately be described as flat. But
from a more micro perspective, using say a high-powered microscrope, we
would expect the table top to look extremely uneven; within that perspective, the table-top is clearly not flat. Worse still, if we talk to a particle
physicist, we may learn that, when the table is conceived as a bunch of
elementary particles, we have no way to describe the table top, and indeed
there is a sense that, within this perspective, the table top does not exist.
If the question concerning the flatness of the table is a better analogy for
policy-relevant descriptions than the question as to whether it is raining
outside, description has to be evaluated relative to a perspective. Only if
we can find a privileged perspective among the alternatives can we hope to
resolve the policy debate unambiguously. 9 (Furthermore, results in social
choice theory suggest that our tendency to equivocate over the use of
concepts will only exacerbate any barrier to consensus. Consider Graciela

THE NEW TRADE THEORY: IMPLICATIONS FOR POLICY ANALYSIS

173

Chichilnisky and Geoffrey Heal [1983], for example, and their focus on the
role of common regions of excluded preferences.)
Shifting to a policy prescription involves more than adding a value
judgment to pure description; it implies the much stronger claim that the
chosen perspective has a privileged status.
To give a simple example, if a particle physicist were to claim that the
table is a bunch of neutrons and orbiting electrons such that the table top
cannot be defined meaningfully, and then used that to conclude that one
should not put books on the table, we would rightly regard him (or her) as
very confused. The confusion arises because of moving from an assertion
that is legitimate within a particular (scientific) perspective to a claimin this case, an absurd claim-that this has implications outside that perspective. In the example of the table top, the foolishness of this leap
is transparent. Yet there seem to be times when economists make an
analogous jump without qualms. The moral is that within-perspective
conclusions cannot automatically be assumed to carryover into the
broader realm of policy.
If one drops the emphasis on science as neutral description, it appears
more useful to consider the policy prescription process as involving four
stages. The first stage is the choice of perspective. The second is a choice of
description within that perspective. The third is an assertion of how the
world should be, given the perspective-driven description. The fourth step,
for effective policy prescription, is an assertion of how the world could be.
Notice that whereas empirical testing occurs at the second stage, much of
the policy thrust will come from the first stage; for economists with a strong
foundationalist outlook, however, the two steps tend to be compressed
into one. Under a simple positive-normative dichotomy, there are only two
steps: the combined first step and the value judgment entering at stage
three.
Even though there will always be a role for additional data, the issues
involved in the current trade policy debate are less empirical or normative
in the traditional sense. The disagreements come in terms of the significance
of the observed data and of the various formal models, that is, a difference
driven by choice of perspective, and (in part) by implications concerning
the effectiveness of changing the system. to
6. Possible Ways Out?

The previous section, though necessarily briefly stated, should raise serious
doubts about whether a closure of policy debates is possible. It was
suggested that to think of the current debate as a potential paradigm shift is

174

POLICY ANALYSIS AND ECONOMICS

not very helpful. Not only does ambiguity over the scope of a paradigm
leave the question hard to resolve, but the confusions in the current debate
are not to be resolved simply by switching paradigms. From the point
of view of a practitioner of normal science, the notion of paradigm is
conservative and protective, even though for a "scientific revolutionary" it
may appear to have radical connotations. The danger of characterizing the
shift in the policy debate as a paradigm shift is that the "within-paradigm"
myopia may simply reappear in a new context. The problem is not just
which paradigm to use (however the boundaries of a paradigm are defined), but how to use information acquired within any given paradigm.
It seems that the problem of finding a balance between what lies within a
perspective and what lies outside is at the heart of the policy debate.
Before turning to that topic, the discussion turns to a line that may, in the
long run, be a false avenue for resolving the policy debate.
There is a strong temptation to believe that expansion in the knowledge
base of economists will, by increasing their power of description (and,
hopefully, prediction), result in unambiguous policy conclusions. If the
problem were simply one of inadequate description, this might work. Yet,
to the extent that the difficulties arise from the constraints imposed by
partially conflicting and partially overlapping perspectives, and to the
extent that these are inevitable constraints in a process characterized by
abstraction and conceptualization, descriptive progress within any given
perspective will not be sufficient.
This issue, of the power of improved description to resolve policy
disputes, is particularly important because of the role of repeated games
in the current thinking of economists. Thus Lucas' (1990) querying of
Helpman and Krugman's focus on static non-cooperative models of interaction seems to carry with it a suggestion that a broader approach,
possibly allowing more cooperative outcomes, would be more insightful.
Authors such as Goldstein and Krasner (1984) emphasize the role of
repeated games, and of particular strategies within those games, to influence outcomes.
If there is any hope for a "better description" of the trade arena, it
would seem to come from repeated games. There are a number of reasons
for being extremely cautions in expecting unequivocal policy pronouncements from this kind of analysis. It is hard to know which game we should
focus on or, more accurately, how we should abstract from the world to
obtain a well-defined game. Do we focus on interactions between national
governments? Do we include interactions between interest groups and
governments? Do we include military-type links between countries that
may influence the application of trade policies? Do we include the way that

THE NEW TRADE THEORY: IMPLICATIONS FOR POLICY ANALYSIS

175

social factors, such as education, religious beliefs, family structure, and the
like, feed into perceptions of how individuals fit into their economic and
political system? Any notion that there is, out there in the "real world," the
correct repeated game just waiting to be discovered seems a hopelessly
heroic leap of faith.
Although repeated games promise to add major insights to the way
economists interpret the world, they also have potential weaknesses.
Within the perspective of repeated games, it is well known that "almost
anything" can be a possible outcome and that attempts to refine solution
concepts have not succeeded in clarifying the possibilities as unequivocally
as might be hoped (Ledyard, 1986). "Outside" the perspective of repeated
games seems to be a deeper problem, namely, whether situations can be
constrained to be viewed as repeated games without any loss of generality.
(One might note the "foundationalist" aspects of John Harsanyi's (19671968) ingenious, and path-breaking, method of dealing with this issue.)
Thus one would expect to see weaknesses emerging in any "game-theoretic
paradigm" (however defined), just as are found now in the "perfectlycompetitive paradigm"; equivocation within the paradigm and inability to
think of what may lie outside the paradigm may just reinstate confusion.

7. A Bottom Line?
Thus, while one might naturally hope that the ambivalence exhibited in the
trade policy debate will be resolved by increased knowledge, to the extent
that the ambivalence emanates in part from the process of conceptualization and abstraction, one is likely to be disappointed. Equivocal use of
concepts, differing perspectives, and underlying that, different beliefs
about the role of economic models, all should act as barriers to consensus.
Perhaps the best move, not to resolving the debate but to making it
more productive, is to recognize the different presuppositions we bring to
the debate-and the different conversations in which we think we are
involved. 11 And a first step may be to discuss more carefully the distinction
between what lies within a perspective and what lies outside it, and why
this matters.
One place where the distinction plays a role is in a clear trend for a shift
in focus for the "leverage" of policy. Twenty-five years ago the political
system was viewed as essentially disjoint from economists' explanatory
domain. Since then, economists have devoted much effort to bringing both
the political system and aspects of information within the bounds of
economic description. By endogenizing these channels for policy leverage,

176

POLICY ANALYSIS AND ECONOMICS

prescient authors, such as Bhagwati (1988), have been forced to bring into
the discussion the notions of ideologies, ideas, and institutions.
Policy analysis suffers a parallel problem to that of philosophy. In
philosophy, once an area becomes clarified and well defined, it is likely to
be spun off into a separate discipline; as a consequence philosophy is left
with its most intractable problems. In policy analysis, if a channel (such as
the political system) is a particularly effective way to alter behavior, it is
likely to be considered as another locus of optimizing behavior. But once
the channel has been endogenized within a model, questions arise as to
how, within the perspective of that model, it can be simultaneously
assumed that policy advocacy can alter behavior through the channel. The
move to endogenize both political behavior and information acquisition
has indeed pushed potential policy "leverage points" to areas such as
ideology, ideas, and institutions.
The tendency of economists to expand the domain of economics leaves
some awkward questions. One might argue that the progress of natural
science should not be viewed as one of unqualified expansion; though the
natural sciences have tended to increase the number of applications and
the extent of their "within-discipline" understanding, that progress has
sometimes been accompanied by a narrowing of the goals of the discipline
in terms of the questions the discipline views as within its purview. Thus
while one would expect economists to find insightful things to say about
what were once "noneconomic areas," such as political decisions or
institutional design, progress in economics may also entail a narrowing of
the objectives of the science. Any such narrowing clearly has implications
for policy advocacy because it invites questions as to how the perspectives
of economists fit with perspectives that lie outside the discipline. Moreover, the trend to moving potential policy channels within the domain of
economics suggests a possible limiting outcome: If economics is the study
of conditioned behavior (in terms of optimization subject to constraints
and to the incentives implied by those constraints), the complete dominance of such economic analysis would imply the only way policy changes
could be effected is by unconditional changes in people's beliefs, perceptions, preferences, or the like. We are reminded that, although within the
perspective of such a science there would be no role for "moral (or
unconditional) behavior," it could playa dominant role in policy debates.
Put another way, what is missing from a model, or theory, may be as
important as what is included. Traditionally, the way to present a model is
to list its assumptions. (What else can one do?) However, from a policy
perspective this may be inadequate. It is impossible to understand the
significance of assumptions without knowing not only what is assumed, but

THE NEW TRADE THEORY: IMPLICATIONS FOR POLICY ANALYSIS

177

what is excluded. Indeed it is arguable that, from the point of view of a

policy-maker, what the model cannot comprehend may be crucial in terms
of evaluating the usefulness of the policy prescription.
All this suggests that we focus on what is inside each policy advocate's
perspective, what is outside it, and how the advocate views the relation
between the inside and the outside. Rather than view the foundationalist
approach as an algorithm for rational decision making, it may be more
helpful to view it as a way of understanding how stringent are the
requirements for achieving consensus--even in the scientific arena. To
bring some order to the debate, it is useful to recognize the nature of the
conversation that each author is involved in. For example, Helpman and
Krugman, along with many other economists, are involved in a conversation where it is accepted that policy pronouncements of economists are
potential influences on special interest groups and that this is a relevant
consideration. Underlying this is a view of the theory-empirics-policy
nexus that assumes that, in order to discuss policy, one must move back
somewhat from the narrow focus of the theory. (At the same time, they do
not move back enough to introduce concerns about whether the presence
of special interest groups may make moot the whole discussion of optimal
policy, either at the national or the global level. They certainly do not go as
"far back" as Bhagwati does in his monograph, nor as far back as some of
the authors of the 19408 and early 19508, who allowed themselves the
luxury of talking of a moral dimension to trade policy.)
Lucas is clearly not in the same conversation. For example, for him,
there is a strong distinction between what is learned from economic theory
and personal stances about how one expresses one's opinions before the
general public. There is no doubt that the difference between Lucas, on the
one hand, and Helpman and Krugman, on the other, is driven largely by a
different perception of the role of economic theory. Whether one could
flesh out Lucas's view of science without resorting to "rules of thumb" is an
open question. Moreover, with such a "hard-line" conception of economic
theory, one wonders how policy analysis can (or is intended to) break free
of mere description and what criterion is used to determine which theories
are useful-and for what purposes. This view is so focused on being
"inside" economics that there seems to be no place for policy to get
leverage.
Baldwin, to take another example, seems to be advocating expanding
the boundaries of economics to include modeling of the political system.
Baldwin is more clearly trying to stand both inside the broader conception
of economics (by modeling and testing political behavior in relation
to trade policy) and outside economics (by asserting a role for policy

178

POLICY ANALYSIS AND ECONOMICS

advocacy despite this endogenization of the political mechanism). This

seems again a different conversation from that of Helpman and Krugman
(who do not focus on this dimension of decision making) and of Lucas
(who seems determined to avoid acknowledging anything of value
"outside" economics).
Goldstein and Krasner emphasize the lack of commitment to liberal
trade policies among U.S. trading partners and the value of a U.S. strategy
to induce more cooperative behavior by the rest of the world. Although the
"inside" of their perspective seems broader in some ways than that of
Helpman and Krugman or of Lucas, it carries the twin presumptions that
liberal trade is the appropriate goal and that the U.S. internal political
structure would permit a well-defined, and well-oriented, tit-for-tat policy.
Again, a different conversation, and again a case where what is "outside"
the perspective seems as important as what is inside.
One can read "moral crusades" in many of the writings. Both Helpman
and Krugman and Baldwin, not to mention Bhagwati, are fighting to keep
the faith in free trade as the presumptive, or default, policy. Lucas' implicit
crusade seems to involve keeping the faith in a philosophy of science that
allows theory to be given a privileged status, compared to mere personal
stances and other subjective aspects of the policy-making environment.
Crusades are usually about rules, about perceiving the world "in the right
way." And rules are notoriously hard to evaluate, in part because of the
difficulty of locating a correct perspective from which to do the evaluation.
One way to understand the wedge between Lucas on the one hand, and
Helpman, Krugman, Baldwin, and Bhagwati on the other, is that Lucas
conceives a world where each case can (costlessly, neutrally, and unambiguously) be evaluated on its merits, whereas the others have a picture of
a more constrained situation. In other words, it is not (just) different
perceptions of the international economy, but different perceptions of the
way the international economy is evaluated, that are affecting the policy
stances. Obtaining a consensus on policy involves not just an empirical and
a normative consensus, but deeper levels of consensus too.
This chapter has argued that the absence of an adequate theory of
empirical confirmation (or verification, or refutation), which allows us to
get beyond the theory being tested, implies that policy debates are rarely
going to be resolved purely by more information. Stepping back outside
the theory can be done in a relative sense, but not (in an antifoundationalist world) to the point of allowing a neutral starting point. Leaving the
debate to be characterized as a battle over a paradigm shift ignores the
balance between the "inside" and the "outside" of alternative perspectives. Philosophers are seeking to develop a workable theory of ideas

THE NEW TRADE THEORY: IMPLICATIONS FOR POLICY ANALYSIS

179

related to our notion of being inside versus outside a theory or a perspective; but it is unclear that there is a way to reconcile completely the
conflicting demands on such a theory (Nicholas Rescher, 1985; Thomas
Nagel, 1986; Hilary Putnam, 1987).
In a sense we have been exploring the bite of the conflict between a
desire to be practical and effective and the desire to be neutral and free of
cultural contingency. Policy needs pull us toward seeking an unambiguous
policy stance. Practicality requires we adopt a particular position and hence
(at least temporarily) ignore the unquestioned presuppositions. If we could
find acceptable foundations in which to ground our knowledge, there
would be no conflict between being practical and being neutral; unfortunately no one has a watertight foundationalist position, and even antifoundationalists are unsure how to proceed without foundations. If we
appeal to communal consensus as a substitute source of grounding, we
need to be aware of the "bias" implicit in the presuppositions of the
community.
It is often suggested that we can do no better than allow the experts to
make the decisions. But if we look, for example, at the conflict between
Lucas and Helpman and Krugman, both sides have claims to expert status.
Neither empirical criteria, nor logical criteria, nor expertise may be
sufficient to adjudicate between alternative policy recommendations. In
addition, we should remember that expertise brings with it not only
positive connotations of skills and communal respect but also negative
connotations of self-interest and selective myopia. (For a gentle suggestion
of such a role for self-interest, see Robert Solow [1988]; Rom Harre [1986]
gives a very different emphasis. The potential ubiquity of network externalities in the market for ideas raises doubts about the efficacy of competitive
forces as an antidote to self-interest; John Haltiwanger and Michael
Waldman [1985] and Michael Katz and Carl Shapiro [1986].) There is also
a suspicion that many policy pronouncements of economists reflect in
significant part the presuppositions of their "world-view," while claiming
the cloak of scientific respectability in terms of empirical accountability.
It would be nice to unveil a bottom line to the discussion. However,
there should be enough evidence so far to suggest that the cost of achieving
any bottom line is to ignore some potentially important aspects of the
situation. There is no useful sense in which to talk about an unambiguous
trade policy, and there are good reasons to expect failure of closure in such
a debate. The trade policy debate seems to have some of the flavor of an
argument, as to whether the table top is flat, between proponents of
different perspectives. The stances in the debate are driven largely by
differences in perspective, both immediate differences about how best to

180

POLICY ANALYSIS AND ECONOMICS

view the international economy and by more indirect differences about

what we learn from economic models and how this learning feeds into
policy discussion.
In essence, economists resolve the "foundationalist dilemma" by exiting
with some, at least temporarily, unchallenged presupposition. But different economists choose different points of exit. Foundationalist issues
occur at multiple levels: the basis for decisions of economic agents within
the model; the basis for choosing one model over another; the basis for
interpreting what we learn from the chosen model and how it is to be seen
in the policy setting. Not surprisingly, it is hard to reach a consensus, and
tendencies toward consensus seem to reflect common presuppositions
above all else.
It may be time to recognize that trade policy operates in an environment
that cannot usefully be comprehended within a single perspective. A
broadly defined paradigm is likely to result in equivocal, and still blinkered, recommendations, whereas a narrowly defined paradigm seems
certain to miss important dimensions of the situation. Our foundationalist
heritage leaves us prone to two forms of questionable belief. One belief is
that, if our descriptive abilities are refined enough, then the policy debate
can be closed. The other belief is that, if we analyze our different
presuppositions, we can fully understand each other and can agree to
disagree. Both these beliefs, one is tempted to call them "mirages," seem
to require stronger foundations than can reasonably be assumed. While we
should continue to seek better descriptions (however we choose to define
"better") and greater understanding of the presuppositions of various
competing advocates, nevertheless the "trade policy conversation" is likely
to go on, and on, and on ...

Notes
1. As always, "marginal" is defined relative to the perceptions of the decision maker. If
our philosophy of science allowed us to believe that we can describe the world objectively, that
is, in some sense as it "really is," we would note that these perceptions may be mistaken.
However, as will be argued later, it is not clear that we can achieve the required objectivity.
2. In this respect, Krugman's talk, of free trade having lost its innocence in the light of
recent developments, seems a trifle heroic in its interpretation of the situation of the 1950s
and 19605. Max Corden (1974: 7-8) has a cautionary approach: "Theory is vital, but it is not
enough. Theory does not 'say'-as is often asserted by the ill-informed or the badly
taught-that 'free trade is best.' It says that, given certain assumptions, it is 'best.' "
(Emphasis in original.)
3. The prepublication history of this article gives a fascinating insight into some of the
norms of the profession 50 years ago.

THE NEW TRADE THEORY: IMPLICATIONS FOR POLICY ANALYSIS

181

4. These beliefs were probably reinforced by a number of considerations. The development of proofs of optimality and existence of abstract general equilibrium models of complete
sets of perfectly competitive markets paralleled the growth of formal models of international
trade and suggested the intellectual respectability of competitive models of trade. (See
McKenzie, 1954, for example.) The central descriptive issue, the pattern of trade, appeared
to be handled in a robust way by focusing on the role of competitive markets. Major
alternatives to models of perfectly competitive markets were viewed at that time as either
essentially disjoint approaches (as with Keynesian macroeconomic models) or as dominated
by the capabilities of an ideal market system (as with centrally planned command economies).
5. See Evan Simpson (1987) for a good discussion of some of the current debate between
"foundationalists" and "antifoundationalists". Richard Bernstein (1983) gives a flavor of the
issues at stake, while Alfred MacKay (1980: 112-117) gives an account of the type of
impossibility theorem that drives the problem.
6. See Corden (1974: 5-8). As noted in an earlier quote from this text, Corden
rightly-given his positive-normative dichotomy-emphasizes the contingent nature of policy
prescriptions.
7. Here we are talking not just of the rapid demise of verificationism, and the weaknesses
of refutationism as emphasized by Thomas Kuhn (1970). Even one of the stronger hopes of
the foundationalist tradition in the philosophy of science, namely the confirmation criterion
espoused, for example, by Clark Glymour (1980), is still an incompletely developed theory of
confirmation and, more importantly, aspires at best to provide a criterion of confirmation
within a theory.
8. Even this example is not as simple as to looks. There should be little doubt that, had
human physiology been different so that, for example, any contact between drops of rain and
the human skin were fatal, our criterion for deciding whether it is, or is not, raining would be
much different, and far more refined, than it is now.
9. This may explain some of the strong feelings concerning foundationalism in the
economics profession. On a perspectival view, the assertion that there is a privileged, neutral
description of the phenomena under observation amounts to a claim that there is only one
"correct" perspective to adopt. How do we locate that correct perspective? Maybe from
critical empirical tests that distinguish between competing hypotheses. But this is exactly what
the foundationalists have failed to deliver in the philosophy of science, a criterion that
discriminates between different perspectives in a neutral way. At best we can seek confirmation within a theory, not in any absolute sense. As a consequence, from a "nonfoundationalist" perspective, the committed foundationalist seems to have introduced a spurious sense of
objectivity into the discussion, while to the committed foundationalist, his or her critics simply
don't understand the nature of "good science."
10. A concern with a position such as that of Lucas (1990) would be whether one can flesh
out the notion of "useful, explicit theories" (and other implicit commitments, concerning the
philosophy of science, made by Lucas) in a way that meets criticisms from an antifoundationalist perspective. An instrumentalist interpretation would still leave a question as to which
useful, explicit theories should be accessed for policy recommendations. On the other hand,
turning to a strong critic of traditional Western philosophy such as Richard Rorty (1989) has
its own set of problems; see, for example, the discussions of Rorty's position in Simpson
(1987).
11. The (related) issues of incommensurability and of indeterminacy of translation have a
major place in twentieth-century philosophy. Donald Davidson (1984) makes an argument
that the notion of a conceptual scheme, so central to relativistic approaches, is incoherent
because we do, in fact, communicate to some extent-and this would not be possible with

182

POLICY ANALYSIS AND ECONOMICS

entirely distinct conceptual schemes. Nonetheless, given the fluidity of concepts envisaged in
the later works of Ludwig Wittgenstein (1958), there seems little doubt we can get at cross
purposes without always being aware of it. (The contrast between the later Wittgenstein and
his earlier works [see Wittgenstein, 1974) seems to capture well the tension between
anti-foundationalist and foundationalist views of philosophy.)

References
Bagwell, Kyle, and Robert W. Staiger, "A Theory of Managed Trade," The
American Economic Review, 80:4 (September 1990), 779-795.
Baldwin, Robert E., "The Political Economy of Trade Protection," The Journal of
Economic Perspectives, 3:4 (Fall 1989), 119-135.
Baumol, William J., Sue Anne Batey Blackman, and Edward N. Wolff, Productivity and American Leadership: The Long View (Cambridge, Mass.: The MIT
Press, 1989).
Bernhardt, Dan, and Alice Enders, "Free Trade Equilibria to Multi-Country
Quota Games," Journal of International Economics 27:3/4 (November 1989),
319-333.
Bernstein, Richard J., Beyond Objectivism and Relativism (Philadelphia: University of Pennsylvania Press, 1983).
Bhagwati, Jagdish, Protectionism (Cambridge, Mass.: The MIT Press, 1988).
- - - , "Directly Unproductive, Profit-Seeking (DUP) Activities," Journal of
Political Economy 90:5 (October 1982), 988-1002.
- - - , "The Generalized Theory of Distortions and Welfare," in Jagdish N.
Bhagwati, Ronald W. Jones, Robert A. Mundell, and Jaroslav Vanek, eds.,
Trade, Balance of Payments, and Growth: Papers in International Economics
in Honor of Charles P. Kindleberger (Amsterdam: North-Holland Publishing
Co., 1971).
Brander, James A., "Rationales for Strategic Trade and Industrial Policy," in Paul
R. Krugman, ed., Strategic Trade Policy and the New International Economics
(Cambridge Mass.: The MIT Press, 1986).
Brander, James A., "Intra-industry Trade in Identical Commodities," Journal of
International Economics 11:1 (February 1981), 1-14.
Brander, James A., and Barbara J. Spencer, "Export Subsidies and International
Market Rivalry," Journal of International Economics 18:1/2 (February 1985),
83-100.
Brecher, Richard A., "Minimum Wage Rates and the Pure Theory of International
Trade," The Quarterly Journal of Economics 88:1 (February 1974), 98-116.
Chichilnisky, Graciela, and Geoffrey Heal, "Necessary and Sufficient Conditions
for a Resolution of the Social Choice Paradox," Journal of Economic Theory
31:3 (October 1983), 68-87.
Corden, W. M., Trade Policy and Economic Welfare (Oxford: Clarendon Press,
1974).
Davidson, Carl, Lawrence Martin and Steven Matusz, "The Structure of Simple

THE NEW TRADE THEORY: IMPLICATIONS FOR POLICY ANALYSIS

183

General Equilibrium Models with Frictional Unemployment," Journal of Political Economy 96:6 (December 1988), 1267-1293.
Davidson, Donald, "On the Very Idea of a Conceptual Scheme," reprinted as
Essay 13 of Inquiries into Truth and Interpretation, D. Davidson (Oxford:
Oxford University Press, 1984).
Dertouzos, Michael L., Richard K. Lester, Robert M. Solow and The MIT
Commission on Industrial Productivity, Made in America (Cambridge Mass.:
The MIT Press, 1989).
Dixit, Avinash, "Trade and Insurance with Moral Hazard," Journal of International Econmics 23:3/4 (November 1987), 201-220.
- - - , "Trade Policy: An Agenda for Research" in Paul R. Krugman, ed.,
Strategic Trade Policy and the New International Economics (Cambridge, Mass.:
The MIT Press, 1986).
Dixit, Avinash K., and Gene M. Grossman, "Targeted Export Promotion with
Several Oligopolistic Industries," Journal of International Economics 21 :3/4
(November 1986), 233-250.
Eaton, Jonathan, and Gene M., Grossman "Optimal Trade and Industrial Policy
Under Oligopoly," The Quarterly Journal Of Economics 101:2 (May 1986),
383-406.
Ethier, Wilfred J., Modern International Economics, 2nd. Edition (New York:
W. W. Norton, 1988).
- - - , "The Theory of International Trade," in L. H. Officer, ed., International
Economics (Boston: Kluwer Academic Publishers, 1987).
- - - , "Internationally Decreasing Costs and World Trade," Journal of International Economics 9:1 (February 1979), 1-24.
Findlay, Ronald, and Stanislaw Wellisz, "Endogenous Tariffs, the Political Economy of Trade Restrictions, and Welfare," in J. N. Bhagwati, ed., Import
Competition and Response (Chicago: University of Chicago Press, 1982).
Gilpin, Robert, The Political Economy of International Relations (Princeton:
Princeton University Press, 1987).
Glymour, Clark, Theory and Evidence (Princeton: Princeton University Press,
1980).
Goldstein, Judith L., and Stephen D. Krasner, "Unfair Trade Practices: The Case
for a Differential Response," The American Economic Review, Papers and
Proceedings 74:2 (May 1984), 282-287.
Grinols, Earl L, Uncertainty and the Theory of International Trade (Chur, Switzerland: Harwood Academic Publishers, 1987).
Grossman, Gene M., and J. David Richardson, Strategic Trade Policy: A Survey of
Issues and Early Analysis, Special Papers in International Economics, No. 15
(April 1985).
Haltiwanger, John, and Michael Waldman, "Rational Expectations and the Limits
of Optimality," The American Economic Review 75:3 (June 1985), 326-340.
Harre, Rom, The Varieties of Realism: A Rationale for the Natural Sciences
(Oxford: Blackwell, 1986).
Harsanyi, John c., "Games with Incomplete Information Played by 'Bayesian'

184

POLICY ANALYSIS AND ECONOMICS

Players, I-III," Management Science 14:3, 5, 7 (November 1967, January and

March 1968), 159-182, 320-334, 486-502.
Helpman, Elhanan, "Increasing Returns, Imperfect Markets, and Trade Theory,"
Ronald W. Jones and Peter B. Kenen, eds., Handbook of International Economics, volume I, (Amsterdam: North-Holland, 1984).
Helpman, Elhanan, and Paul R. Krugman, Trade Policy and Market Structure
(Cambridge, Mass.: The MIT Press, 1989).
Johnson, Harry G., "Optimum Tariffs and Retaliation," The Review of Economic
Studies 21:2 (1953-4), 142-153.
Katz, Michael L., and Carl Shapiro, "Technology Adoption in the Presence of
Network Externalities," Journal of Political Economy 94:4 (August 1986),
822-841.
Kennan, John, and Raymond Riezman, "Do Big Countries Win Tariff Wars?"
International Economic Review 29:1 (February 1988),81-85.
Krasner, Stephen D., "State Power and the Structure of International Trade,"
World Politics 28:3 (April 1976), 317-47.
Kreuger, Anne 0., "The Political Economy of the Rent-Seeking Society," The
American Economic Review 64:3 (June 1974), 291-303.
Krugman, Paul R., "Is Free Trade Passe?" The Journal of Economic Perspectives
1:2 (Fall 1987), 131-144.
- - , "Introduction: New Thinking About Trade Policy," in Paul R. Krugman,
ed., Strategic Trade Policy and the New International Economics (Cambridge,
Mass.: The MIT Press, 1986).
Kuhn, Thomas S., The Structure of Scientific Revolutions, 2nd ed., enlarged,
International Encyclopedia of Unified Science, Foundations of the Unity of
Science, Volume II, Number 2, (Chicago: University of Chicago Press, 1970).
Ledyard, John 0., "The Scope of the Hypothesis of Bayesian Equilibrium,"
Journal of Economic Theory 39:1 (June 1986), 59-82.
Lucas, Robert E., Jr., "Book Review of 'Trade Policy and Market Structure," by
Elhanan Helpman and Paul R. Krugman" Journal of Political Economy 98:3
(June 1990), 664-667.
MacKay, Alfred F., Arrow's Theorem: The Paradox of Social Choice (New Haven:
Yale University Press, 1980).
McKenzie, Lionel, "On Equilibrium in Graham's Model of World Trade and
Other Competitive Systems," Econometrica 22:2 (April 1954), 147-161.
McMillan, John, Game Theory in International Economics (Chur, Switzerland:
Harwood Academic Publishers, 1986).
Magee, Stephen P., William A. Brock, and Leslie Young, Black Hole Tariffs and
Endogenous Policy Theory (Cambridge: Cambridge University Press, 1989).
Mayer, Wolfgang, "Endogenous Tariff Formation," The American Economic
Review, 74:5 (December 1984), 970-985.
Nagel, Thomas, The View from Nowhere (New York: Oxford University Press,
1986).
Newbery, David M. G. and Joseph E. Stiglitz, The Theory of Commodity Price
Stabilization (Oxford: Oxford University Press, 1981).

THE NEW TRADE THEORY: IMPLICATIONS FOR POLICY ANALYSIS

185

Pomery, John, "Optimal Tariffs," in John Eatwell, Murray Milgate, and Peter
Newman eds., The New Palgrave (London: The MacMillan Press Ltd, 1987).
Putnam, Hilary, The Many Faces of Realism (LaSalle IL: Open Court Publishing
Co., 1987).
Rangarajan, L., "The Politics of International Trade," in Susan Strange, ed., Paths
to International Political Economy (London: George Allen and Unwin, 1984).
Rescher, Nicholas, The Strife of Systems (Pittsburgh: University of Pittsburgh
Press, 1985).
Richardson, J. David, "Empirical Research on Trade Liberalization with Imperfect
Competition: A Survey," OECD Economic Studies (Spring 1989), 7-50.
Rorty, Richard, Contingency, Irony, and Solidarity (Cambridge: Cambridge University Press, 1989).
Simpson, Evan, ed., Anti-foundationalism and Practical Reasoning: Conversations
between Hermeneutics and Analysis (Edmonton, AB: Academic Print. & Pub.,
1987).
Solow, Robert M., "Comments from Inside Economics," Chapter 3 of Arjo
Klamer, Donald N. McCloskey, and Robert M. Solow, eds., The Consequences
of Economic Rhetoric (Cambridge: Cambridge University Press, 1988).
Stolper, Wolfgang F., and Paul A. Samuelson, "Protection and Real Wages," The
Review of Economic Studies 9:1 (November 1941), 58-73.
Thursby, Marie, and Richard Jensen, "A Conjectural Variation Approach to
Strategic Tariff Equilibria," Journal of International Economics, 14:112 (February 1983), 145-161.
Wittgenstein, Ludwig, Tractatus Logico-Philosophicus, trans. D. F. Pears and
B. F. McGuiness, with introduction by Bertrand Russell (London: Routledge
& Kegan Paul, 1974).
- - , Philosophical Investigations, the English text of the third edition, trans.
G. E. M. Anscombe (New York: MacMillan, 1958).
Woodland, A. D., International Trade and Resource Allocation (Amsterdam:
North-Holland Publishing Co., 1982).

POLICY RESEARCH:
A WITHERING BRANCH
OF ECONOMICS?

w. Lee Hansen

1. Introduction

What are the prospects that policy research will remain an important
preoccupation of economists and that new economics Ph.D.s will continue
to be trained for, interested in, and rewarded for doing policy research?
This question is timely as policy schools and economics departments
reexamine their respective missions and clarify their goals and objectives.
Debate within public policy and public administration schools continues
over the relative emphasis they should place on policy research and
analysis versus public administration or management (Wildavsky, 1985;
KettI, 1990).
Meanwhile, the American Economic Association, responding to criticism about the profession's increasingly narrow focus, has undertaken a
major study of how new Ph.D.s are educated and how their education
shapes the direction offuture research in economics (Hansen, 1990; Hansen,
1991; and Commission on Graduate Education in Economics, 1991). The
implications of this study are important for both fields because economics
and economists playa central role in public policy programs, and public
187

188

POLICY ANALYSIS AND ECONOMICS

policy issues have important dimensions that require the expertise of

economists.
This chapter explores the linkages between economics and policy research, beginning with a brief review of the changing relationships between
economics and policy research. The chapter will then examine the nature
of the knowledge and skills required to do policy research, the extent to
which economics graduate students are acquiring such knowledge and
skills, and the implications of these findings for the linkages between
economics and policy research.
Before proceeding, readers should realize that because the distinctions
between policy research and policy analysis are blurred, these terms will be
used interchangeably. In principle, they are separable, with policy analysis
being more specifically client oriented (Weimer and Vining, 1989). As a
consequence, academics doing policy research are less likely to be involved
in policy analysis than their nonacademic colleagues who, even if they
describe themselves as policy researchers, are typically more heavily
engaged in policy analysis. The focus of the comments here is on this
author's area of specialty in policy research, on human resource issues such
as those in the areas of labor and welfare, as contrasted to health and the
environment where conditions may be somewhat different.
2. The setting
Understanding the linkages between economics and policy research is
complicated by several developments. These developments include: the
economic profession's apparently declining interest in, and attention to,
public policy concerns; the current disillusionment of economists with
policy analysis and its accomplishments; the weakening of interest in, and
funding of, policy-oriented research; the growing presence of policy
analysts in government; and finally the emergence of public policy schools
which seek to educate and train policy researchers and analysts.
The waning of interest in policy issues among younger academic
economists has raised concern about the scope and content of graduate
training. Such training, it is argued, has become ever more narrowly
focused on theoretical and technical matters, and gives less attention to
empirical and problem-oriented research questions and to policy issues
(Leontief, 1982; Morgan, 1988; Colander and Klamer, 1987). There seems
to be a growing belief that talented undergraduates interested in a more
policy-oriented approach to economics are discouraged from beginning
graduate study in this subject (Kasper, 1991). Those students who do
retain strong policy-oriented interests must either put them aside while

POLICY RESEARCH: A WITHERING BRANCH OF ECONOMICS?

189

doing graduate work in economics or try to satisfy them by moving into

some related graduate program, such as a public policy school, an agricultural economics department, industrial relations institute, or a business
school. Of course, many economics graduate students still concentrate on
policy research, but there is a sense that opportunities to build on these
interests are eroding. In any case, because the training provided in
graduate programs outside economics is by definition different, it remains
unclear whether the existing stock of expertise in policy research and policy
analysis, which has come to be dominated by economists, can continue to
reproduce itself.
Traditionally, economists have demonstrated a keen interest in policy
issues. Indeed, as the federal government broadened the scope of its
activities, and particularly its social programs in the 19605 and the early
19705, it relied increasingly on economists. They were called to offer advice
on programs, to analyze program effectiveness and in some cases to help
operate these programs. Substantial numbers of academic economists were
attracted to the policy field because of its challenges, the availability of
substantial funding to carry out research, and the possibility of having
some effect on public policy (Nelson, 1987). Moreover, in the 19705 many
new Ph.D.s in economics took positions in government agencies and
policy research organizations to apply their knowledge and skills by
carrying out policy research and analysis and by helping to formulate and
implement policy.
As time passed, economists began to look back in an effort to assess
their role in the arena of public policy. Early indications of the disillusionment about the prospects for policy research and analysis emerge from
Henry Aaron's influential book, Politics and The Professors: The Great
Society in Perspective (Aaron, 1978). In a wide-ranging analysis of the
Great Society programs of the 19605 and early 19705, he concludes that
social science research on these programs, much of it undertaken by
economists, often did not, and could not, answer the questions that needed
to be answered. Even when it did, their results did not necessarily influence
decision makers. And because many of these social programs were attacking complex problems, the results fell short of earlier expectations, as
demonstrated by social science research. Aaron's sobering critique no
doubt dampened the enthusiasm of economists for the future of policy
research and analysis.
It is not clear that matters have improved as we enter the 1990s. Nothing
demonstrates this more clearly than a recent discussion of social science
research in the Journal of Human Resources (1990). The occasion was a
broad-based review of two recent books by economists on policy research

190

POLICY ANALYSIS AND ECONOMICS

and analysis: Robert H. Haveman's Poverty Policy and Poverty Research:

The Great Society and the Social Sciences and Richard P. Nathan's Social
Science in Government: Uses and Misuses. Haveman and Nathan offer
their assessments of the state of social science research in poverty and in
the evaluation of employment and training programs. Both indicate that
important progress has been made in policy analysis and evaluation
research while at the same time they highlight continuing problems. After
reading these volumes, one comes away with a more hopeful view than that
expressed by Aaron.
Aaron, who participated in the JHR discussion, states his views even
more bluntly, noting that economists have too often proposed policies
that, though designed to improve welfare, gave far too little attention to
whether such policies can be implemented and sustained. He suggests the
existence of an inherent conflict between the values of the academic and
policy worlds, where "the former praises technical virtuosity and elegance
and rewards analytical innovation" and "the latter insists on simplicity and
transparency and is indifferent to analytical innovation" (Aaron, 1990;
279).1

Another veteran policy researcher and analyst, Edward M. Gramlich,

highlights the different incentives driving academics and policy-makers. As
a result, policy-makers are often confused by policy research because
"academic researchers do not exactly speak with one voice" (Gramlich,
1990: 288). Indeed, he goes on to question why politicians listen to
researchers at all.
Eric A. Hanushek offers a different perspective, arguing that greater
emphasis is needed on disciplinary research that is "motivated by the
challenges perceived within the separate disciplines" (Hanushek, 1990:
291). Such research in economics, for example, may address some
issues relating to public policy, but that is not the central purpose of disciplinary research. The objective of policy research, by contrast, is "to
produce policy implications that have some hope or expectation of being
taken seriously" (Hanushek, 1990: 291). Policy research differs from
policy analysis, the latter being highly specific, client-oriented, and with a
short time horizon. Based on his review, Hanushek suggests that resources
should be concentrated on disciplinary research because such an allocation
will have a stronger and more definite impact, including important effects
on the beliefs of future generations of policy makers.
A harsher view is expressed by James J. Heckman who, focusing on the
low level of existing social science knowledge, argues that the "pretence to
knowledge undermines the acquisition of knowledge" (Heckman, 1990:
303). He goes on to say that "the failure of speculative social science theory

POLICY RESEARCH: A WITHERING BRANCH OF ECONOMICS?

191

and policy advocacy research has hurt empirical social science" (Heckman,
1990: 303) and accounts for diminished funding for all social science
research. He laments the low quality of empirical advocacy research, which
he asserts has harmed the academic credability of all empirical research. 2
Unfortunately, Heckman does not amplify on these intriguing and provocative comments, thus leaving readers to elaborate for themselves.
Though it might be argued that these economists do not constitute a fair
sampling of the economics profession, they are nevertheless an influential
group whose judgments about the relationships between disciplinary research, policy research, and policy analysis cannot easily be dismissed.
None of the four reviewers express much optimism about the likelihood
that policy research can have strong positive effects. Without saying so
directly, though Hanushek and Heckman come closest, they suggest that
the "real pay dirt" lies in disciplinary research.
The decline of interest and certainly funding of policy research in the
labor-welfare area is quite apparent to economists even though comprehensive supporting data are not readily available. It is true that research
funding grew rapidly from the mid-1960s to the late 19708. Since then, real
levels of funding for research on employment and training programs have
plummeted, as have those for poverty research. The declining availability
of federal research funds has been even more pervasive, with federal
obligations for basic and applied research in economics falling by more
than 40 percent from 1977 to 1986. 3 These declines are reflected in reduced
research output. Haveman documents the falling number of published
papers on poverty and poverty-related research that began in the late
19708. Casual inspection of the journals indicates a similar decline in
published papers in the human resources field.
Perhaps associated with declining funding are the sharply diminished
opportunities for academic economists to gain policy experience. Unlike
the situation in the early 1960s when the number of doctoral economists in
government was minuscule, government agencies now support substantial
staffs of economists as a result of being able for more than a decade to
hire capable Ph.D. economists in permanent positions. The presence of
staff economists in key agencies reduces, if not eliminates, the need for
short-term expertise of the kind academics could and did provide; it seems
that fewer young academic economists go off to work for the federal
government in Washington, D.C., either during the summer or on
academic year leaves. It is possible, of course, that government agencies
have discovered that academics who might be interested in temporary
employment in Washington would be of only marginal value because of the
nature of their training and the requirements of the available jobs.

192

POLICY ANALYSIS AND ECONOMICS

Meanwhile, a wide array of private for profit and nonprofit organizations employ Ph.D. economists to carry out research studies that academic
economists might have done in the 1960s and early 1970s. These developments mean that fewer young academic economists have direct policy
experience, and as a consequence they are less well equipped to convey
what policy research and analysis means to their students.
The rise of public policy schools and the expansion of graduate
programs in related fields are also causing some uneasiness. With the help
of Ford Foundation funds, a number of public policy schools expanded in
the 1970s, and they continue to be vigorous and productive. Whether their
limited numbers of Ph.D. graduates and M.A. recipients can compete with
Ph.D. economists is not clear. It is quite apparent that their production of
Ph.D.s never was large, and it shows little or no change in recent years.
These schools have produced a steady flow of master's degree students who
may be at least partial substitutes for economics Ph.D.s in policy analysis.
On the other hand, the market for public policy analysts, whether they
hold master's or doctorate degrees, did not grow as rapidly in the 1980s as
had been expected (Conant, 1991).
Less can be said about the production of policy researchers and analysts
in economics because few if any graduate departments offer a special track
in policy research and analysis. Nor is there a research field with such a
designation. Rather, there are ten major fields, divided along traditional
lines: economic theory, public finance, labor, international, and so on.
Within these fields there is no easy way of knowing the methodological
orientation of published research, that is, whether it is theoretical, empirical, institutional, or policy oriented. 4
How the relative attractiveness of economics and public policy programs
will change in the 1990s remains unclear. Much depends on what happens
within the two fields and how the labor market evaluates the products of
programs in these two fields.

3. What It Takes To Do Polley Research and AnalysIs

What is it that distinguishes the training of policy researchers and policy
analysts? Several attempts have been made to describe the kinds of
knowledge and skills required in these positions. Weimer and Vining
(1989) list five areas of preparation: basic information-acquiring skills,
a framework for considering the appropriateness of public policy interventions, the technical skills acquired through disciplinary study in fields
such as economics and statistics, an understanding of political and organ-

POLICY RESEARCH: A WITHERING BRANCH OF ECONOMICS?

193

izational behavior, and some ethical framework for dealing with clients.
Patton and Sawicki (1986) list three broad goals: learning to apply basic
methods of analysis quickly and appropriately as the situation demands,
becoming competent in using the methods of analysis and designing
approaches to policy issues, and communicating the results. They also list
12 practical principles, ranging from learning how to focus quickly on the
central features of the problem, to deciding how much time and effort to
devote to a problem.
Nelson (1987) offers the fullest description of what it takes for an
economist to participate effectively in the making of public policy. The
knowledge and skills Nelson distills from his research are best summarized
in his own words:
Many economists would need to invest greater effort in improving writing skills,
facility in reasoning by analogy, command of institutional details, knowledge of
legal processes and reasoning, and political awareness and savvy. They might
need to devote more time and effort to investigations of history, law, politics,
and institutions, and their bearing on the economic topics of policy concern.
Advocates of economic policies would need to tailor their policy proposals
to reflect an accurate understanding of how these policies will be publicly
perceived-in terms of social equity, the public sense of "fairness," impact on
personal liberties, infringements on private property rights and other such
public concerns. Proponents of economic policies need to be able to defend
these policies, not only on narrow technical grounds, but also in broader
ideological and philosophical terms. Sensitivity and knowledge in these areas
are needed to establish one's standing and influence with policy makers, as well
as to shape persuasive policy arguments. Understanding that reaches beyond the
confines of economics is also likely to yield more promising economic policy
proposals in the first place.
To be more effective policy advocates, many economists would generally
need to give more attention to "big-picture" skills. The ability to "tell a story"
that makes sense, to "paint a picture," is at a premium in government. Top
policy makers are often confronted with overwhelming amounts of information
and data. Their greatest concern is to organize this diverse material in some
meaningful way. The biggest asset of economists is the conceptual equipment
that enables them to impose sense and order on an immensely confusing world
of employment, industry, commerce, finance, and administration. Within the
recently established profession of public policy analysis, a number of its members have described the practice of policy analysis as more an "art" than a
"science" (Wildavsky, 1979). The skills of the "craftsman," rather than the
"scientist," are most in demand in professional roles in government [Po 86J

The knowledge and skills used in policy research and analysis, as

already described, appear to differ from the knowledge and skills of

194

POLICY ANALYSIS AND ECONOMICS

current mainline economists. Aside from differences in the substance of

their interests, there appear to be important differences in the relative
emphasis given to various types of knowledge and skills. To understand
the commonalities and differences between economists and people in the
policy field, it is essential to identify the special knowledge and skills
possessed by economists as a result of their education and subsequent
experience.
4. The Knowledge and Skills Taught to Economists
What knowledge and skills are being imparted to new economics Ph.D.s
through their graduate training that might equip them to do policy research
and analysis? An important source of information is the research undertaken for the American Economic Association's Commission on Graduate
Education in Economics (COGEE), which recently completed a major
examination of graduate education. That study focused on the structure
and content of graduate training, in the realization that how new economists are trained helps shape the direction of the profession's future
research. Based on this research and its deliberations, the commission
recently issued its recommendations (Commission on Graduate Education
in Economics, 1991). The presentation that follows draws on selected
results from the COGEE study (Hansen, 1990; Hansen, 1991), principally
surveys of faculty members and recent Ph.D.s. s
Before proceeding, it may be helpful to point out that no formal
statement exists describing the knowledge and skills that are required of,
or used by, economists. In an effort to capture what seem to be the
distinguishing differences among economists, a classification of knowledge
and skills was devised for use in the surveys. Interestingly, the survey
responses of faculty members and graduate students about which knowledge and skills are being taught and which should be taught proved to be
amazingly similar. Moreover, the responses were surprisingly similar among
programs of different reputation (e.g., the top six ranked departments, the
next nine ranked departments). These similarities no doubt reflect the
socialization process that occurs in graduate training; students with different
views either do not enter economics Ph.D. programs or they leave for other
pursuits.
Knowledge. To find out what graduate students are learning, faculty (and
also graduate students) were asked to rank the relative importance now
given, and that should be given, to each of six types of knowledge:

POLICY RESEARCH: A WITHERING BRANCH OF ECONOMICS?

195

Table 9-1. Ranking of Skills and Knowledge of Ph.D. Graduates, Based on

Faculty Responses
Panel A: Ranking of Emphasis on Different Types of Knowledge That Is and
Should Be Given in Graduate Education
Ranking
1st
2nd
3rd
4th
5th
6th

Currently Emphasized
Theory
Econometrics
Empirical
Applications
Institutions
Literature

Should Be Emphasized
Theory
Econometrics
Empirical
Applications
Institutions
Literature

Panel B: Ranking of Importance That Is and Should Be Given to Different Skills

in Graduate Education
Ranking
1st
2nd
3rd
4th
5th
6th
7th

Currently Important
Analytics
Mathematics
Critical judgment
Applications
Computation
Creativity
Communication

Should Be Important
Analytics
Creativity
Critical judgment
Applications
Communication
Mathematics
Computation

Source: Hansen (1991).

economic theory, econometrics, institutions and history, literature, applications, and empirical economics. 6
The results in table 9-1, panel A, indicate general satisfaction of faculty
members with the current emphasis in graduate training, as evidenced by
the similar rankings for "what is"and "what should be." The order of the
rankings of knowledge is about what most economists would have expected. Economic theory leads the way, followed by econometrics; well
behind come empirial economics, and applications, institutions, and literature are at the bottom. These results suggest that faculty members at Ph.D.
granting departments see nothing seriously wrong with the relative emphasis given to the six different types of knowledge.
Skills. Graduate education also imparts special skills utilized by economists. Thus respondents were asked to rank the importance that is given and
should be given to seven different skills: analytics, mathematics, critical
judgment, applications, computation, creativity, and communications. 7

196

POLICY ANALYSIS AND ECONOMICS

As shown in panel B of table 9-1, faculty members indicate that

analytics and mathematics are the most important skills, with creativity and
communication at the bottom of the list and critical judgment, applications, and computation ranking in the middle. The rankings change
considerably when the focus is on which skills should be important. While
analytics remains at the top the list, creativity vaults to second place, and
communication moves up several notches. Meanwhile, mathematics and
computation decline sharply in importance, moving to the bottom of the
list of skills. Interestingly, the skill of application remains in the middle.
Summary. Several conclusions emerge. First, there appears to be general
agreement about what knowledge is, and what knowledge should be,
emphasized in graduate education. Second, there is considerable divergence between the importance that is, and should be, given in graduate
training to the various skills required of economists. Third, the differences
in the skill rankings support the popular view that mathematics is overemphasized and creativity is undervalued in graduate education in
economics.
5. The KnOWledge and Skills Required In Public Polley
Because policy researchers and policy analysts could not be identified in
the COGEE surveys, there was no direct way of finding out what knowledge and skills they thought were, and should be, emphasized. Nor was
there time to survey public policy professionals to find out what knowledge
and skills they gained in their graduate education and later used in their
jobs. It might be possible to divine this from a detailed examination of the
already-cited works by Weimer and Vining, Patton and Sawicki, and
Nelson. In principle, they are all describing the relative importance of
which knowledge and skills are learned, which should be learned, and
which are used in practice.
A next best alternative is to examine what knowledge and skills recent
economics Ph.D.s say they use in their jobs, with particular attention to
nonacademic economists who are most heavily involved in policy research
and analysis. The results presented here are for 1977-1978 economics
Ph.D.s who by 1989 had extensive and varied experience on the job and
knew what their work required.
Knowledge. While recent Ph.D.s ranked the emphasis given to the six
types of knowledge in their own graduate training quite similarly to the
rankings of current faculty members, their rankings based on what they use
in their current jobs proved to be quite different, as shown in table 9-2,

POLICY RESEARCH: A WITHERING BRANCH OF ECONOMICS?

197

panel A. Recent Ph.D.s ranked applications and empirical economics first

and second by a wide margin, with theory and econometrics lagging behind
and institutions and literature at the bottom. This emphasis given to
applications and empirical economics contrasts sharply with what goes on
in graduate school, where emphasis is given to economic theory and
econometrics.
It is important to separate recent Ph.D.s by type of employment. Those
in nonacademic jobs rate applications first and as strongly as they rated
theory in their own Ph.D. training, with empirical economics a close
second. These rankings are not much different from those for recent
Ph.D.s employed in departments that educate graduate students.
Skills. The rankings of skills used on the job by recent Ph.D.s also
differ considerably from those of current faculty members, being closer
to the skills faculty members thought should be important relative to the
skills faculty members thought actually were important in graduate training. Topping the rankings of skills used on the job (table 9-2, panel B) was
communication, followed by critical judgment. Applications, analytics,
and creativity were closely ranked; computation and mathematics trailed
at the bottom of the rankings. These results indicate that the skills
used by, and hence needed by, nonacademics differ considerably from
what is emphasized in graduate education. These skills also differ from
those used by recent Ph.D.s who are in Ph.D.-granting departments. They
give first place to creativity, with communication, analytics, applications,
and critical judgment ranked closely behind, followed by computation and
mathematics at the bottom. Thus, the rankings of skills provided in
graduate training differ considerably both for academic researchers at
Ph.D.-granting institutions and for practicing economists in the nonacademic sector who most closely approximate policy researchers and
analysts.
Summary. To the extent the rankings of the knowledge and skills used by
recent Ph.D.s in nonacademic jobs can be taken as a reflection of the
knowledge and skills used by policy researchers-analysts, the emphasis and
importance given to the different types of knowledge and skills in graduate
training differ markedly from what policy researchers-analysts use on the
job.
6. Confirming Evidence from Economics Faculty
What do faculty members have to say about the capacity of graduate
students to use their knowledge and skills? How proficient are they
as future economists, not necessarily as policy researchers or analysts?

198

POLICY ANALYSIS AND ECONOMICS

Table 9-2. Ranking of Knowledge and Skills of Recent Ph.D. Graduates,

Based on Their Own Responses

Panel A: Emphasis Given to Different Types of Knowledge in Own Ph.D.

Education, in Current Job, and in Current Nonacademic Job

Ranking
1st
2nd
3rd
4th
5th
6th

In Own Ph.D.
Education
Theory
Econometrics
Empirical
ApplicationPolicy
Literature
InstitutionsHistory

In Current Job
All Jobs
ApplicationPolicy
Empirical
Theory
Econometrics

Nonacademic Jobs
ApplicationPolicy
Empirical
Theory
Econometrics

Literature

InstitutionsHistory
Literature

InstitutionsHistory

Panel B: Importance Given to Different Skills in Own Ph.D. Training, in Current

Job, and in Current Nonacademic Job

Ranking
1st
2nd
3rd
4th
5th
6th
7th

In Own Ph.D.
Education
Analytics
Mathematics
Critical judgment
Applications
Creativity
Computation
Communication

In Current Job
All Jobs
Communication
Critical judgment
Applications
Analytics
Creativity
Computation
Mathematics

Nonacademic Jobs
Communication
Critical judgment
Analytics
Applications
Creativity
Computation
Mathematics

Source: Hansen (1991).

Faculty members indicated that by the time of their comprehensive examinations many graduate students had not gained proficiency in using the
knowledge and skills developed in their course work. At the high extreme,
58 percent of faculty members indicated that all or most students (as
compared with "some" and "few or none") were well-grounded in economic theory. In sharp contrast, only 14 percent of faculty members could
say that by the time students completed their comprehensive examinations
most or all of them were good at using theory in empirical applications or
at applying theory to the real world.
Nor is the quality of Ph.D. dissertations viewed as particularly impres-

POLICY RESEARCH: A WITHERING BRANCH OF ECONOMICS?

199

sive by faculty dissertation supervisors. Asked to characterize the dissertations they had recently supervised, 68 percent of faculty members agreed
that most or all dissertations were good training instruments. No more than
half thought that all or most dissertations applied economics to real-world
problems or were well-grounded in economic theory, and only 40 percent
could say that most or all dissertations offered good empirical evidence to
support economic theory. Finally, less than a third of the faculty supervisors indicated that all or most dissertations were well written.
The importance of these selected findings should be apparent. To the
extent that economists have long played a major role in public policy
programs, it would appear that the training currently provided in graduate
programs may limit the interest and ability of new Ph.D.s to do public
policy research and to engage in public policy analysis. In the absence of
comparable data from an earlier period, however, this conclusion rests on
the assumption that the rankings of knowledge and skills from a decade or
two ago would have been different. Unfortunately, this possibility cannot
be checked because no comparable information is available, either for the
19705 when policy research and analysis received greater attention or for
the early 1960s when policy research and analysis had not yet taken shape
as a special area of expertise.
7. Economists and Public Polley

In light of these findings, what is the likely future role of economics in the
policy research and analysis? It is clear that currently economists play a
. major role, with substantial numbers of faculty in public policy schools
holding doctorates in economics (Friedman, 1991). Though exact numbers
are difficult to obtain, the preponderance of economists emerge from
examining the membership list of the Association for Public Policy Analysis and Management. Thus, it is not surprising to learn that public policy
programs give extensive attention to economics in their master's and Ph.D.
curricula. Even a cursory acquaintance with policy analysis in federal
government agencies suggests that economists playa major role in carrying
out public policy research and in shaping public policy analysis.
The commanding position of economists in the public policy arena
is hardly surprising. Each year, approximately 40 percent of new Ph.D.s
enter nonacademic jobs, with most of them (about two thirds) going into
government positions, many of which involve policy research and analysis.
This is an underestimate because other new Ph.D.s who initially opt for
academic positions eventually gravitate to the nonacademic sector, where
they utilize their knowledge and skills in ways demanded by their positions
rather than by their graduate education. What is perhaps surprising is that

200

POLICY ANALYSIS AND ECONOMICS

graduate training does not seem to serve well the educational needs of this
large fraction of graduates.
Of those economics Ph.D.s who take academic jobs, a considerable
number will require, because of the orientation of their own research,
teaching, and professional interests, the mix of knowledge and skills that
public policy programs seek to develop even though these academic Ph.D.s
may not be labeled as policy researchers and analysts. Some will work in
these areas in parallel with their disciplinary research. Others will utilize
this intellectual equipment in their daily lives as teachers and members
of their departments and university communities. Though not engaged in
policy research or analysis, they will be discussing policy issues in their
courses and demonstrating to their students how such issues are examined. In addition, developing new course materials, supervising student
projects, and participating in study groups and university administrative
activities all require expertise in the five areas of preparation mentioned
by Weimer and Vining, in practicing the 12 principles presented by Patton
and Sawicki, and in following the multifaceted approach described by
Nelson. The many new Ph.D. economists who do not take academic
jobs will benefit even more directly because of the nature of the work they
perform.

8. Implications for Economics and for Public

Polley Programs

What are the implications of the results presented here? Several emerge.
In economics, graduate training is not well structured to prepare new
Ph.D.s for the range of work that they will do, particularly for those
graduate students interested in policy research and policy analysis. This
development may indicate a turning away by the economics profession
from its long-standing interest in economic policy issues. As a result, we
can expect to find a diminishing proportion of economists with active
interest in policy research and analysis.
It is conceivable that these patterns will be reversed. The creation of the
AEA Commission may signal the end of a more than decade-long phase in
economics that has emphasized abstract and technical issues to the exclusion of real-world, empirical policy issues. If so, possibilities are stronger
that the Commission's findings and recommendations may produce
changes in how graduate students are educated and in that sense gradually
redirect the profession's interests. Whether the combined effects of these

POLICY RESEARCH: A WITHERING BRANCH OF ECONOMICS?

201

two forces can accelerate the pace of change and help policy research
regain its position in economics is uncertain.
Should economics move away from policy research and analysis, an
opportunity will arise for public policy schools to expand to fill the void.
Whether public policy schools will be able to, or choose to, do so remains
unclear. They labor under several constraints, one being the move to place
greater emphasis on public management. This move appears to be motivated by the view that producing better managers is socially more important, particularly when neither the supply of students nor the demand for
graduates can support large programs with a heavy emphasis on policy
analysis. Even if there is no redirection of these programs toward public
management, public policy schools may not be able to recruit talented
young economists to their faculties. This difficulty may not be serious if
public policy schools seek out experienced economists who have made
their mark in policy research and analysis. On the other hand, if the
number of new economists interested in, and prepared to move into, policy
research and analysis does decline, the supply of eligible economists will be
smaller than it has been in the past.
These developments suggest that public policy schools may want to
think about "training their own" faculty members while at the same time
attempting to meet the presumably growing nonacademic demand for their
graduates, including the demand from employers who in the past may have
hired economists. The major difficulty with this approach is the matter of
scale. Gearing up to produce any appreciable number of Ph.D.s in public
policy is both a formidable and costly undertaking. Moreover, such an
approach could end up producing lower-quality substitutes whose presence
would undermine the long-run prospects for the success of public policy
programs. A further difficulty is that the demand for public policy Ph.D.s is
apparently not that strong right now; certainly the flow of Ph.D.s is still
quite small. And if interest in public policy is reversed in economics, public
policy schools might see their Ph.D. market evaporate.
Whatever happens, it is apparent that the role of economics and
economists in public policy research and analysis helps to bridge the gap
between disciplinary research and the need by policy-makers for improved
knowledge, specific studies of policy issues, and policy advice. Let us hope
that economists will renew their interest and commitment to policy research and analysis, particularly in graduate economics but also in public
policy schools. Let us hope that in doing so they can also be responsive to
what has been learned in the past decade about how policy research and
analysis can be strengthened and how the capacity of new economics
Ph.D.s to contribute to the study of policy issues can be enhanced.

202

POLICY ANALYSIS AND ECONOMICS

Acknowledgments

The author appreciates the comments and suggestions of Edward Bird,

James Conant, Robert Haveman, and Robert Lampman.
The author is indebted to the Andrew W. Mellon Foundation, the
Alfred P. Sloan Foundation, and the National Science Foundation for its
funding of the work of the American Economic Association's Commission
on Graduate Education in Economics, from which this chapter is derived.
The author also thanks Edward Bird, Thomas Buchmueller, and Jeffrey
Dominitz for their research assistance. Suzanne Vinmans for her secretarial support, and Elaine Moran for her editorial advice.

Notes
1. Another statement of Aaron's current views is contained in his recent Ely Lecture to
the American Economic Association (Aaron, 1989).
2. In a throw-away line, Heckman notes that "it is easier to prove theorems than to
establish enduring facts."
3. Based on data for federal obligations for basic and applied research, deflated by the
GNP deflator; see National Science Foundation, Profiles-Economics: Human Resources and
Funding, Special Report NSF 88-333, table 31 and 33.
4. It might be useful for more fields to adopt key word identifiers to be supplied with
paper abstracts, thereby permitting identification of papers by both subject and methodological orientation.
5. The samples of faculty members, graduate students, and recent Ph.D.s were drawn
from a sample of 91 economics departments whose graduate programs were among those
ranked in the 1982 study by the National Academy of Sciences (Jones, Lindzey, and
Coggeshall, 1982). Further details on the surveys and other work for the Commission will be
available in forthcoming publications by the author, who served as Executive Secretary of the
Commission.
6. The knowledge categories reflect an effort to classify the major types of knowledge that
are used by economists in their work and are taught to graduate students as part of their
preparation as economists. Six categories or types of knowledge emerged:

a.
b.
c.
d.
e.
f.

Economic theory (e.g., assumptions and theorems of economic behavior).

Econometrics (e.g., statistical theorems in economics, properties of models, and
distribution theory).
Economic institutions and history (e.g., forms of economic association, historical
economic forces).
Economic literature (e.g., recent or comprehensive histories of economic ideas and
approaches) .
Economic applications and policy issues (e.g., current topics of concern in business,
government, and society).
Empirical economics (e.g., testing implications of theoretical models, estimating
behavioral responses, practical analysis of data, experience with economic databases).

POLICY RESEARCH: A WITHERING BRANCH OF ECONOMICS?

203

7. The seven categories of skill used follow:

a.

Critical judgment (e.g., analyzing ideas, reviewing literature, formulating pertinent

comments).
b. Analytics (e.g., understanding and solving problems, making and analyzing logical
arguments).
c. Applications (e.g., seeing practical implications of abstract ideas, analyzing realworld policies and processes).
d. Mathematics (e.g., constructing and analyzing proofs, manipulating mathematical
abstractions) .
e. Computation (e.g., effectively and quickly finding and manipulating relevant data
sources, translating statistical theory into functioning programs).
f. Communication (e.g., speaking and writting effectively and with good style, quickly
understanding spoken and written ideas of others).
g. Creativity (e.g., conceiving interesting research questions, finding new ways of
analyzing topics).

References
Aaron, Henry, "Social Science Research and Policy: Review Essay," Journal of
Human Resources 25 (Spring 1990), 276-280.
- - - , "Politics and the Professors Revisited," Richard T. Ely Lecture, in
American Economic Association: Papers and Proceedings 79 (May 1989),1-15.
- - - , Politics and the Professors: The Great Society in Perspective (Washington,
D.C.: Brookings Institution, 1978).
Aaron, Henry, et aI., "Review Essays of Poverty Policy and Policy Research by
Robert Haveman and Social Science Research in Government: Uses and Misuses
by Richard Nathan," Journal of Human Resources 25 (Spring 1990), 275-311.
Colander, David, and Arjo Klamer, "The Making of an Economist," Journal of
Economic Perspectives 1 (Fall 1987), 95-112.
Commission on Graduate Education in Economics, "Report of the Commission
on Graduate Education in Economics," Journal of Economic Literature (Forthcoming September 1991).
Conant, James, "The Enrollment Crash in Schools of Public Affairs and
Administration-And the Aftermath: A Search for Winners and Losers,"
Robert M. La Follette Institute of Public Affairs, University of WisconsinMadison, 1991, 18 pp. mimeo., plus tables and appendices.
Jones, Lyle V., Gardner Lindzey, and Porter E. Coggeshall, editors, An Assessment of Research-Doctorate Programs in the United States: Social and Behavioral
Sciences (Washington, D.C.: National Academy Press, 1982).
Friedman, Lee S., "Economists and Public Policy Programs," Journal of Policy
Analysis and Management 10 (Spring 1991), 343-359.
Gramlich, Edward M., "Social Science Research and Policy: Review Essay,"
Journal of Human Resources 25 (Spring 1990), 281-289.
Hansen, W. Lee, "Educating and Training New Economics Ph.D.s: How Good a
Job Are We Doing? American Economic Review: Papers and Proceedings 80

204

POLICY ANALYSIS AND ECONOMICS

(May 1990), 437-444; and the related" Discussion" by Alan S. Blinder, Claudia
Goldin, T. Paul Schultz, and Robert M. Solow, 445-450.
- - - , "The Education and Training of Economics Doctorates: Major Findings of
the Executive Secretary of the American Economic Association's Commission
on Graduate Education in Economics," JourTUJI of Economic Literature (Forthcoming, September 1991).
Hanushek, Eric A., "Social Science Research and Policy: Review Essay," Journal
of Human Resources 25 (Spring 1990), 290-295.
Haveman, Robert H., Poverty Policy and Poverty Research: The Great Society and
the Social Sciences (Madison: University of Wisconsin Press, 1987).
Heckman, James J., "Social Science Research and Policy: Review Essay," Journal
of Human Resources 25 (Spring 1990), 297-304.
Kasper, Hirschel, "The Education of Economists, From Undergraduate to
Graduate Study: A Report of the Committee of College Faculty," Journal of
Economic Literature (Forthcoming, September 1991).
KettI, Donald F., "The Perils-and Prospects-of Public Administration,' 50 PAR
(July/August 1990), 411-419.
Leontief, Wassily, "Academic Economics," Science 217 (2 July 1982), 106-107.
Morgan, Theodore, "Theory Versus Empiricism in Academic Economics: Update
and Comparisons," Journal of Economic Perspectives 2 (Fall 1988), 159-164.
Nathan, Richard P., Social Science in Government: Uses and Misuses (New York:
Basic Books, Inc., 1988).
National Science Foundation, Profiles-Economics: Human Resources and Funding, Special Report NSF 88-333, 1989.
Nelson, Robert H., "The Economics Profession and the Making of Economic
Policy," Journal of Economic Literature 25 (March 1987), 49-91.
Patton, Carl V., and David S. Sawicki, Basic Methods of Policy Analysis and
Planning (New York: Pentice-Hall, 1986).
Wildavsky, Aaron, "Social Science Research and Policy: Review Essay," Journal
of Human Resources 25 (Spring 1979), 305-311.
- - , "The Once and Future School of Public Policy," The Public Interest 79
(Spring 1985), 25-41.
Weimer, David L., and Aidan R. Vining, Policy Analysis: Concepts and Practice
(Englewood Cliffs, N.J.: Prentice-Hall, 1989).

Index

Aaron, Henry, 189

Accountability, 120
Ackley, Gardner, 7
AEA, see American Economic Association
Agency costs, 117
Agency theory, 116-118, 119, 120, 122
Aggregate supply and demand, xv, 128,
129-130, 136
going beyond, 152-153
in Great Depression, 134
in Korean War, 137
oil prices and, 142, 143-144, 145
recessions and, 138-140
Aggregate supply curve, 150
Agricultural economics, 4
Agricultural Economics, Bureau of, 4
Agriculture Department, 9
Alchian, Armen, 113-114, 115, 116, 117
Allais Paradox, 95, 96,101,102
Allocation branch of public finance, 49, 50
Altruism, xii-xiii, 65
misunderstanding of, 67
in private choices, 72-73, 74
in public choices, 75
American Economic Association (AEA),
2,187
COGEE of, see Commission on
Graduate Education in
Economics
American Enterprise Institute, 13
Antifoundationalism, 171-173, 178, 179,
see also Foundationalism

Antitrust Division of Justice Department,

4

Area Redevelopment program, 52

Army, Department of the, 9
Arrow's possibility theorem, 53-54
Asset specificity, xiv, 115
Association for Public Policy Analysis and
~anagement, 199
Autarky, 164, 166
defined, 163
Average rate of transformation, 163-164
Averch-Johnson effect, 121
Banks, in Great Depression, 131, 132-133
Barone, Enrico, 47
Bayesian theory, xiv, 90, 91,101,102
Behavioral anomalies, in decision making,
94-96,98'
Benefit-cost analysis, x, xi-xii, 12
criticism of, 29, 30, 32, 33, 38, 39
in distributional analysis, 55, 62
morality vS., 68
in pollution issues, 11, 14
in progressive model, 9
self-interest paradigm in, 77
Bentham, Jeremy, 24-25, 26
Benthamite welfare functions, 54
Bergson, Abram, 47
Blueprints for Basic Tax Reform
(Bradford), 14
Board of Governors, 133

205

206
BOB (Bureau of the Budget), 4,8,9
Bounded rationality, xiv, 114
Bradford, David, 14
Brookings Institution, 13, 14, 131
Brown, E. Cary, 5
Budget, Bureau of the (BOB), 4, 8, 9
Budget-maximizing bureaucrat theory, 119
Bush administration, 79

Carter administration, 52, 141, 144, 146

CEA, see Council of Economic Advisers
Charitable contributions, 72-73
Child care industry, 51, 60
Civil Aeronatics Board, 13
Civil Service Commission, 4
Clean Air Act, 14
Coase, Ronald, xiv, 112-113, 114
Coase theorem, 113
COGEE (Commission on Graduate
Education in Economics), xvii, 194,
196,200
Command-and-control methods, 14
Commerce Department, 5, 9, 10
Commission on Graduate Education in
Economics (COGEE), xvii, 194, 1%,
200
Commons, John R., 3
Communications industry, deregulation of, 14
Communism, decline of, 17
Compensatory behavior, 70-71
Complementary goods, 25-26
Congressional Budget Office, x, 11
Congressional Research Service, x
Consumer sovereignty, 58, 77, 106
Contracts
morality and, 70
new institutional economics and, xiv,
114-115,117,120-121
Controversy, 23-39
hard vs. soft values in, 33-34, 38
institutional transactions and, xii, 34-37
normative ciritiques in, xi-xii, 38-30
public vs. private values in, 30-31
standing problem in, 34
utility theory and, 24-28, 32-33, 38
Corps of Engineers, 12
Cost-benefit analysis, see Benefit-cost
analysis

POLICY ANALYSIS AND ECONOMICS

Cost-Benefit Analysis and the Economics of
Investment in Human Resources, 51
Cost-effectiveness analysis, in progressive
model,9
Council of Economic Advisers (CEA), x,
10,16,149
creation of, 5
influence of, 6, 7
oil prices and, xv, 143, 144
Countercyclical fiscal policy, decline of,
151
Credit, in Great Depression, 133
Crime deterrence, self-interest paradigm in
77-78
Currie, Lauchlin, 4

Data Resources Incorporated, 142

Decision making under uncertainty,
85-107
behavioral anomalies in, 94-96, 98
insurance and, 86-90, 97
policies concerning, 103-106
prospect theory in, xiii-xiv, 96-102
risk perception and, 85-86, 90-94, 95,
97,98-99,101,104-105,107
Defense Department, 15
influence of economists in, 11
Defense spending
in Korean War, 137
in WW II, 136, 152
Delaney clause, 91
Demsetz, Harold, 113-114, 115, 116, 117
Deontological ethics, 65, 71
Depressions, 129, see also Great
Depression; Recessions
Deregulation, 13-14, 17
Diminishing marginal utility, theory of, 25,
26
Disciplinary research, 190, 191
Distributional analysis, 45-62
equity/efficiency issue, see Equity/
efficiency issue
Great Society and, 51-53
human capital in, 49-51
organizational efficiency in, 59-60
positive economics in, xii, 46, 48, 58-59
in welfare economics, 45-46, 47-49, 50,
58,59,60

INDEX
Distribution branch of public finance, 49,
50
DOE (Department of Energy), 143
Domestic market distortions, 164
Duopolists, 166, 167
Durkheim, Emile, 68
Eastern Europe
communism decline in, 17
trade and, 162
Economic Analysis, Bureau of (Commerce
Department),5
Economic Darwinism, 118
Economic rents, see Rent seeking
Economic Report (CEA), 143, 144
Economics training, x-xi
changes needed in, 18-19
in policy research, xvii-xviii, 187,
188-189,192,194-199,200
progressive movement and, 9
in traditional vs. nontraditional agencies,
13
Economies of scale, 49
trade and, 165-166, 169
Economists, x-xi
impact of, 13-17
numbers of in government, 9-10
policy research and, 191, 192, 194-196,
199-200, see also Economics
training
politics and, 6-7
progressive movement and, 3, 4, 8-10
Education Summit of 1989, 79
Education system, 17,50, see also Voucher
system
Efficiency theory, 29, 34-37,38, see also
Equity/efficiency issue
institutional transactions and, 34-37
in microeconomics, 23
standing problem in, 34
utility theory in, 24-28
Egoistic incentives, 75
Eisenhower administration, 139
Ely, Richard, 2
Empirical advocacy research, 191
Employment Act of 1946, 5
Employment research, 191
Energy, Department of (DOE), 143

207
Energy industry, deregulation of, 14
Energy policy, 17
Environmental issues, 11
lack of success in, 14
Environmental Protection Agency, x, 11
EOE (Equation of exchange, monetary),
128
Equity/efficiency issue, 45, 48, 50, 53-56,
58, see also Fairness
static nature of, 60-61
Exchange rates, 130, 151
Exchange theories, 72
Exports
aggregate supply and demand and, 130
new trade theory on, 162, 166-167
Ezekiel, Mordecai, 4

Fairness, 56-58, 59-60, see also Equity/

efficiency issue
Federal government, social programs of,
51-52, see also specific programs
Federal Reserve System, x
economists' influence on, 10
in Great Depression, 131, 133
monetary policy and, 152
money revival and, 149
in progressive era, 4
Federal Reserve-Treasury Accord, 149
Federal Trade Commission, 4
Fiscal deficits, xv, 17, 146-148
Food and Drug Administration, 4
Food stamps, 58
Ford administration, 52
Forest Service, 4, 12
For-profit organizations
efficiency of, 60
new institutional economics and, 121,
123
policy research and, 192
Foundationalism, 161, 175, 177, 179, 180
antifoundationalism vs., 171-172, 173
Foundations of Economic Analysis
(Samuelson),47
France, 53
Free trade, xvi, 159-160,161, 166, 169,
171,178
nature of belief in, 162-165
Friedman, Milton, 5, 141

208
Game theory
in new institutional economics, 118
in trade, xvi, 167, 168, 169, 174-175
Gaskins, Darius, 13
General Accounting Office (GAO), 112
General Theory of Employment, Interest,
and Money (Keynes), 128
GNP, see Gross national product
Gold reserves, in Great Depression, 131,
132-133
Goode, Richard, 5
Gordon, Kermit, 9
Government bonds, 147-148
Government contracting, 120
Graaff, J. de V., 47
Great Britain, 53
Great Depression
government spread to economics and, 4
macroeconomics of, xv, 130-134, 138
Great Society, 51-53
Gross national product (GNP), 5
EOEin, 128
fiscal deficits and, 147, 148
in Great Depression, 132, 133
money revival and, 150
morality and, 70
oil prices and, 142
WW II and, 135, 136, 152

Hansen, W. Lee, 16
Hard vs. soft values, 33-34, 38
Hardy, Charles, 131
Harriss, C. Lowell, 5
Haveman, Robert H., 190
Hazard warnings, 104-106
Head Start program, 51, 52
Health, Education and Welfare
Department, 11
Health care issues, 11, 122
Heckscher-Ohlin economy, 163, 164
Heller, Walter, 5, 7,149
Helmet laws, 106
Hicks, J.R., 129
High-powered incentives, 115
History, 1-19
of economic policy development, 10-13
of government in economics, 4-5
of impact of economists, 13-17

POLICY ANALYSIS AND ECONOMICS

of progressive movement, x, 1, 2-4, 6,
7-10
Hobbes, Thomas, 65
Hoover administration, 132, 134
Hotelling, Harold, 47
Housing and Urban Development
Department, 11
Human capital, 49-51
Hume, D., 131
Humphrey, George, 147

Imperfect competition model, in trade

theory, 161-162, 166-167, 169, 170
Implicit contracts, 70
Import quotas, 167
Impulse control movements, 78
Income, Employment, and the Price Level
(Marschak),129
Income redistribution, see Redistribution
of income
Incomplete-contracts approach to
privatization, 120-121
Indifference curve analysis, 26, III
Inflation, 6, 11
fiscal deficits and, 147-148
Korean War and, 137
monetary policy and, 152
oil prices and, 143, 144, 145
Phillips curve on, 141, 150
recessions and, 138-141
WW II and, 135
In-kind transfer programs, 52-53, 59
Institutional school of economics, 3
Institutional transactions, xii, 34-37
Insurance
agency theory in, 116
decision making and, 86-90,97
Interdependent utility, 67
Interest-group liberalism, 8
Interest groups, see Special interest groups
Interest rates
aggregate supply and demand in, 129,
130
fiscal deficits and, 147, 148
in Great Depression, 131
monetary policy and, lSI, 152
Internalization, 69
Internal Revenue Service, 81

209

INDEX
International economies of scale, 166
International economy, 168-171
Interstate Commerce Commission, 4, 13
Interventionism, in trade, 170
Intraindustry trade, 169
IS-LM generalization, 129
Isoelastic welfare functions, 54, 56

Japan, trade and, 162, 170

Jevons, Stanley, 25
Johnson administration, 51, 52

Journal of Human Resources 189-190

Justice Department, 4

Kahn, Alfred, 13, 15

Karl, Barry, 6
Kennedy administration, 7, 9, 51
Keynes, John Maynard, 128, 131
Keynesian theory, 4, 5,149
aggregate supply and demand in, xv,
128,129
aggregate supply curve in, 150
on oil price shocks, 142
on wages, 141-142
Keyserling, Leon, 6
Korean War, 136-138
Krueger, Anne, 167
Kuznets, Simon, 5
Labor Department, 9,10
Labor market
aggregate supply and demand and, 130
history of, 2
transaction-cost theory in, 115
Labor Statistics, Bureau of, 4
Labor unions, 81, 137
Laissez-faire theory
demise of, 5
new trade thoery vs., 162
WW II and, 136
Land Management, Bureau, 13
Lerner, Abba, 47
Lindblom, Charles, 8
Lotteries, decision making in, xiii-xiv, 95,
97,101-102
Love Canal, N.Y., 104

Low-powered incentives, 115

Lump-sum income transfers, 47,48, 50

Macroeconomics, xv-xvi, 127-153

aggregate supply and demand in, see
Aggregate supply and demand
aggregate supply curve in, 150
countercyclical fiscal policy decline and,
151
EOEin,l28
of fiscal deficits, xv, 146-148
of Great Depression, xv, 130-134, 138
historical factors in, 5
in Korean War, 136-138
monetary rule in, 151-152
money revival and, 149-150
oil price shocks and, xv, 128, 137, 138,
141, 142-146, 150 see also Oil price
shocks
Phillips curve in, 141-142, 150
post-1949,138-141
in WW I, 136, 148
in WW 11,135-136,137,147,148,149,
151-152, see also World War II
Manpower Development and Training
program, 52
Marginal costs, 138, 166
Marginal rate of product transformation,
163
Marginal rate of substitution in
consumption, 163
Market permit system, for pollution, 14
Marschak, J., 129
Median voter theory, 168
Menger, Carl, 25
Metapreferences, 71
Microeconomics
efficiency theory in, 23
of Great Depression, 134
supply and demand in, 128, 129
Middle East conflicts, 17
Mitchell, John Wesley, 3, 5
MIT Commission, 170
Monetary equation of exchange (EOE),
128
Money
policy rule for, 151-152
quantity theory of, 128, 134

210
revival of, 149-150
tight, 145
in utility theory, 25, 32
Money transfer programs, 559
Money wages, 141, 150
Monopoly, 49
Moral hazards, 70,116
Morality, 76, 77, 79
criteria for, 68-69
personality theory of, 69-72
in private choices, 72-74
in public choices, 75
utility theory vs., 66-67
Musgrave, Richard, 49-51
Nathan, Richard P., 190
National Bureau of Economic Research
(NBER), 5, 148
National economies of scale, 165-166
National income statistics, 5
National Science Foundation (NSF), 9-10
NBER (National Bureau of Economic
Research), 5, 148
Negative income taxes, 15
Neoclassical economics, xiv
progressive model vs., 3
self-interest paradigm in, xii, xiii, 65, 66,
67,72,73,74,76,77-78,80
vertical integration and, 116
New Deal, 4-5, 52, 132, 134
New institutional economics, ix, xiv-xv,
111-123
agency theory in, 116-118, 119, 120, 122
policy analysis and, 120-123
precursors of, 112-114
in public sector, 118-119, 122
transaction-cost theory in, xiv, 112-113,
114-116,117-118,119,122
New trade theory, xvi-xvii, 159-180
antifoundationalism and, 171-173, 178,
179, see also Foundationalism
free trade vs., xvi, 159-160, 161,
162-165,166,169,171,178
international economy and, 168-171
possible end to debate about, 173-175
Nixon administration, 11, 52
Nondistribution constraint, 122
Nonprofit organizations

POLlCY ANALYSIS AND ECONOMICS

efficiency of, 60
new institutional economics and, 117,
121,122,123
policy research and, 192
Nonzero transportation costs, 166
Normative economics, ix, xi-xii, 28-30
Notch-effect, 70
NSF (National Science Foundation), 9-10
Nursing homes, 60
Office of Personnel Management (OPM),
9,10
Office of Tax Analysis, 4-5,14
Oil price drag, 144, 146
Oil price shocks, xv, 137, 138
aggregate supply curve and, 150
EOEin, 128
as macro event, 142-144
macropolicy response to, 144-145
market controls and, 145-146
recessions and, 141
Oligopolists, 166
OPEC, 142
OPM (Office of Personnel Management),
9,10
Opportunism, xiv, 114-115, 118, 120, 121
Optimal tariff argument, 163-164
Organization for Economic Cooperation
and Development, 145
Ownership illusion, 117
Pareto, Vilfredo, 47
Pareto optimality, xi, xii, 27, see also
Potential Pareto optimality
in distributional analysis, 46-47, 48, 49,
53,60
fairness and, 57, 58
Great Society and, 52
human capital and, 51
in institutional transactions, 36, 37
Pechman, Joseph, 5,14
Perfect competition model, in trade
theory, 164, 165, 166, 167, 169, 171,
175
Personality theory, 69-72
Ph.D. dissertations, 198-199
Phillips, Chester, 131

211

INDEX
Phillips curve, 141-142, 150
Planning, Programming, and Budgeting
System (PPBS), 9,11
Policy research, xvii-xviii, 187-201
faculty input on, 197-199
funding for, xvii, 191
implications of, 200-201
knowledge and skills in, 194-197
requirements for, 192-194
role of economics in, 199-200
setting for, 188-192
Polio vaccine, 61
Politics, economists and, 6-7
Politics and The Professors: The Great
Society in Perspective (Aaron), 189
Pollution issues, x, 11, 14,61
Popper, Karl, 172
Positive economics, ix
in distributional analysis, xii, 46, 48,
58-59
Positive political theory, 118
Potential Pareto optimality, xi, 27-28, 37,
62, see also Pareto optimality
Poverty
Great Society and, 52, 53
human capital investments and, 51
research on, 191
Poverty Policy and Poverty Research: The
Great Society and Social Sciences
(Haveman), 190
PPBS (Planning, Programming, and
Budgeting System), 9,11
Price controls, 152
in Great Depression, 134
in Korean War, 137-138
on oil, 146
in WW II, 5, 135, 137
Price indexes, 11, 143
Prices, 47, 48, 129, see also Price controls
controversy over, 30-31
decision making and, 89
in distributional analysis, 58
monetary policy and, 151
new institutional economics and, 112
Phillips curve on, 141
Principal/agent model, 74
Prisoner's Dilemma problems, 113
Private vs. public values, 30-31
Privatization, xiv, 120-121

Progressive movement, x, 1,2-4,6

demise of, 7-10
Project evaluation, 55, 62
Property rights, 33
in new institutional economics, 113-114,
120, 121
Proposition 65,104-105
Prospective reference theory, xiv, 101-102
Prospect theory, xiii-xiv, 96-102
Public choice literature, 119
Public goods, 74
Public policy schools, 13
policy research and, xviii, 192,201
Public sector, new institutional economics
in, 118-119, 122
Public vs. private values, 30-31
Pure transfer programs, 45, 56

Quality circles, 81
Quantity theory of money, 128, 134
Quotas, import, 167
Racism, morality vS., 68
Rationing of oil, 146
Rawlsian welfare functions, 54, 55, 56
Reagan administration, 52,141
Real wages
aggregate supply and demand in, 130
Phillips curve on, 141-142, 150
recessions and, 138
Recessions
macroeconomics of, 138-141, 149, 151,
152
self-interest paradigm in, 78
Reconstruction Finance Corporation, 132
Redistribution of income, see also Pareto
optimality; Transfer programs
in distributional analysis, 55-56, 58-59
trade and, 165, 168
Reflection affect, 97
Regular preferences, 71
Rent seeking, 36, 38, 166, 167
Repeated games, 174-175
Ricardian equivalence, 148, 163
Risk perception, decision making and,
85-86,90-94,95,97,98-99,101,
104-105,107

212
Roosevelt administration, 53
Ruggles, Nancy, 5

Samuelson, Paul, 5, 47, 48, 74,164

Savings and loan industry, 14-15
Schlesinger, James, 143
Schultze, Charles, 8-9
Second-best theory, 164
Self-interest paradigm, ix, xii-xiii, xiv,
65-82
morality vs., see Morality
in personality thoery, 69-72
in private choices, 72-74
in public choices, 74-75
public policy implications of, 76-86
in utility theory, 33, 66-67, 74
Shadow prices, 29, 30
Shoup, Carl, 5
Simon, Herbert, 114
Small numbers bargaining, 115
Smith, Adam, 1,47
Smith-Welch analysis, 59
Smoking
institutional transactions on, 35
self-interest paradigm on, 76-77
warnings against, 104
Smoot-Hawley tariff, 134
Social choice theory
in new institutional economics, 118, 119
in trade, 172
Social gospel movement, 2

Social Science in Government: Uses and

Misuses (Nathan), 190
Social sciences, 1,6,8,10, see also
Socioeconomics
Social Security Administration, 81
Social security programs, 52, 61, 148
Social welfare functions, see Welfare
functions
Socioeconomics, xiii, 76, see also Social
sciences
Soft vs. hard values, 33-34, 38
Special interest groups
in nontraditional agencies, 12
progressive movement and, 8
in trade, 164, 165, 169, 170, 171,177
Stabilization branch of public finance, 50
Stallone, Sylvester, 106

POLICY ANALYSIS AND ECONOMICS

Standing problem, 34
State Department
growth of, 5
number of economists in, 9
Stein, Herbert, 143-144
Steuerle, Eugene, 14
Stock market crash of 1929, 131
Stolper, Wolfgang, 164
Subsidies
lump-sum, 47, 48
trade, 165, 166-167, 169
wage, 56
Substituting goods, 25-26
Substitution axiom, 102
Supply and demand, 128, 129, 141, 148,
152, see also Aggregate supply and
demand
Surplus revenue result, 163
Surveys, criticisms of, 31

Tariffs, 134, 164

Tax cuts, 7,141,151
Taxes, 30, 129
in distributional analysis, 46
in Great Depression, 132
Korean War and, 137
lump-sum, 47, 48
negative, 15
pollution, x, 14
self-interest paradigm and, 81
special-interest groups and, 12
trade, 165, 169
in welfare economics, 47, 48
WWIIand,5
Taxpayer sovereignty, 58
Tax reform, 13, 14, 17
Tax Reform Act of 1986,12,14
effect of on charity, 72-73
Team production, 113-114
Theoretical Welfare Economics (Graaf), 47
The Theory of Public Finance (Musgrave),
49-51
Tight money, 145
Tort liability system, 85, 104
Trade barriers, 162, 165, 167
Trade subsidies, 165, 166-167, 169
Trade taxes, 165, 169
Training programs, research on, 191

213

INDEX
Transaction-cost theory
in new institutional economics, xiv,
112-113,114-116,117-118,119,
122
self-interest paradigm in, 70, 74
Transfer programs
in-kind, 52-53, 59
lump-sum, 47, 48, 50
money, 59
pure, 45, 56
Transportation costs, nonzero, 166
Transportation Department, 11
Transportation industry, deregulation of,
13-14
Treasury Deapartment, x
in Great Depression, 132
influence of economists in, 10
number of economists in, 9
Office of Tax Analysis in, see Office of
Tax Analysis
Treatise on Money (Keynes), 131
Truman, David, 8

Unemployment, 6
fiscal deficits and, 147
in Great Depression, 132, 133
monetary policy and, 151, 152
oil prices and, 143
Phillips curve on, 141, 150
post-WW 11,136
trade and, 164, 167
Upward Bound program, 52
Utility theory
decision making and, 87,94-96,97,98,
99-100,102
development of, 24-28
dimensions of, 24
inadequacies of, 32-33, 38
self-interest paradigm in, 33, 66-67,74

Veblen, Thorstein, 3
Veil of ignorance, 33
Vertical integration, 115-116, 120
Vietnam War, 149
Voting, 48, 76, 168
Voucher system, 11
self-interest paradigm in, xiii, 79-80

Wages
money, 141, 150
real, see Real wages
Wage subsidies, 56
Waldo, Dwight, 7
Walras, Leon, 25
War on Poverty, 52, 53
Waugh,Frederick,4
Weisbrod, Burton, 7,54-55,56,60,61,
121,122
Welfare economics, ix, xi-xii
controversy over, see Controversy
distributional analysis in, 45-46,
47-49,50,58,59,60
history of, 3, 10
Welfare functions, 48, 53-56, 60, 61
Benthamite, 54
isoelastic, 54, 56
Rawlsian, 54, 55, 56
Wharton model, 142
Williamson, Oliver, 114, 116, 121, 122
Willingness-to-pay analysis, 29
Working, E. J., 4
World War I, 136, 148
World War II, 135-136, 137, 151-152
demobilization of, 135, 149
fiscal deficits following, 147, 148
government spread to economics and, 4,
5
mobilization of, 135