econstor

A Service of

zbw
Leibniz-Informationszentrum
Wirtschaft

Make Your Publications Visible.

Leibniz Information Centre
for Economics

Dalgaard, Carl-Johan; Strulik, Holger

Working Paper
The history augmented Solow model

cege Discussion Papers, No. 151

Provided in Cooperation with:

Georg August University of Göttingen, cege - Center for European,
Governance and Economic Development Research

Suggested Citation: Dalgaard, Carl-Johan; Strulik, Holger (2013) : The history augmented
Solow model, cege Discussion Papers, No. 151, University of Göttingen, Center for European,
Governance and Economic Development Research (cege), Göttingen

This Version is available at:

http://hdl.handle.net/10419/71085

Standard-Nutzungsbedingungen: Terms of use:

Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Documents in EconStor may be saved and copied for your
Zwecken und zum Privatgebrauch gespeichert und kopiert werden. personal and scholarly purposes.

Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle You are not to copy documents for public or commercial
Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich purposes, to exhibit the documents publicly, to make them
machen, vertreiben oder anderweitig nutzen. publicly available on the internet, or to distribute or otherwise
use the documents in public.
Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen
(insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, If the documents have been made available under an Open
gelten abweichend von diesen Nutzungsbedingungen die in der dort Content Licence (especially Creative Commons Licences), you
genannten Lizenz gewährten Nutzungsrechte. may exercise further usage rights as specified in the indicated
licence.

www.econstor.eu
 Number 151 – March 2013

THE HISTORY AUGMENTED SOLOW

MODEL

Carl-Johan Dalgaard
Holger Strulik

ISSN: 1439-2305
 The History Augmented Solow Model
Carl-Johan Dalgaard∗ and Holger Strulik†

First version: November 2010. This Version: March 2013

Abstract. Unified growth theory predicts that the timing of the fertility tran-
sition is a key determinant of contemporary comparative development, as it
marks the onset of the take-off to sustained growth. Neoclassical growth theory
presupposes a take-off, and explains comparative development by variations
in (subsequent) investment rates. The present analysis integrates these two
perspectives empirically, and shows that they together constitute a powerful
predictive tool vis-a-vis contemporary income differences.

Keywords: Comparative Development; Unified Growth Theory; Neoclassical

Growth Theory.
JEL: O11, O57.

∗
University of Copenhagen, Department of Economics, Øster Farimagsgade 5, building 26, DK-1353 Copenhagen K,
Denmark; email: [email protected].
†
University of Goettingen, Department of Economics, Platz der Goettinger Sieben 3, 37073 Goettingen, Germany;
email: [email protected].
 1. Introduction

One of the biggest questions in social science is why some countries are more developed than
others. A good place to start thinking about an answer, from an economics angle, remains the
seminal Solow (1956) model. As shown by Mankiw, Romer and Weil (1992), and by many since, the
basic Solow model holds considerable explanatory power vis-a-vis contemporary income differences.
Yet, it leaves much income variation unexplained, which has led to various “augmentations” involving
human capital (Mankiw, Romer and Weil, 1992; Ram, 2007), R&D (Nunnemann and Vanhout, 1996)
and health and longevity (Knowles and Owen, 1995).
Another strand of literature argues, however, that in order to better understand comparative
development one must also understand the take-off to sustained growth, which separates an epoch
of stagnation from the current regime where (many) countries grow persistently in terms of income
per capita (Galor and Weil, 2000; Lucas, 2002; Hansen and Prescott, 2002; Galor and Moav, 2002).
Hence, this literature – often referred to as “unified growth theory” – provides the mechanics of
economic development before and after the take-off to growth (see Galor, 2011 for a survey). From
the perspective of unified growth theory, neoclassical growth theory, which only analyzes the growth
process once it has begun, is therefore an incomplete basis for an understanding of comparative
development. A fuller understanding is arguably achieved by recognizing that a large fraction of
contemporary income differences is due to the differential timing of the take-off itself (See Galor,
2010a). Theoretically, this critical historical turning point is hypothesized to be narrowly connected
to the demographic transition; the fertility transition, specifically.1 In theory, the onset of the fertility
transition stimulates subsequent growth in labor productivity for two major reasons: (i ) by lowering
population (labor force) growth it reduces the extent of capital dilution, (ii ) by stimulating human
capital accumulation it drives growth directly, but also indirectly via technological innovation.2
The present paper demonstrates empirically that the combination of these two literatures delivers
a powerful predictive tool. In its simplest version, the basic Solow model along with a control for
the timing of the fertility transition accounts for about 2/3 of the cross-country variation in GDP
per worker in 2000; up from about 40% without the control for the timing of the fertility decline.

1 The demographic transition generally involves two separate transitions: a mortality transition and a fertility tran-
sition. The mortality transition often (but not always) precedes the fertility transition generating the familiar hump
shaped trajectory for population growth.
2 The fertility transition also leads to temporarily faster growth of the labor force relative to the total population
which stimulates growth in income per capita (but not per worker ). See Bloom and Williamson (1998) for more on
this effect.

1
Moreover, the well known “anomaly” when estimating the parameters of the empirical Solow model
– that the capital share is estimated to be exceedingly high – is eliminated.
The link between the year of the fertility transition and contemporary labor productivity is robust
to a range of controls for early institutional developments, cultural variables, geography and climate
as well as early levels of technological sophistication. However, we also demonstrate that the partial
correlation between the year of the onset of fertility transition and labor productivity is eliminated
if we control for the channels through which the transition (in theory) should affect the growth
process: post take-off capital dilution and human capital accumulation.
The paper proceeds as follows. In the next Section we lay out the empirical framework, and
explain how we integrate insights from unified growth theory with the basic neoclassical growth
framework in a simple way. Section 3 provides the empirical analysis, and Section 4 concludes.

2. Empirical Framework

Figure 1 provides a stylized account of contemporary income differences. One reason why income
differences may arise is found in the timing of the take-off, as described by unified growth theory.
The figure illustrates a situation were one country experiences an early take-off at time τ1 , which
puts it on a positive growth trajectory. Meanwhile another country has not taken off at all (as
yet); the “no take-off” scenario. Naturally, if we compare the income difference between these two
countries at (say) time τ2 the difference in income would mainly be due to the differential timing of
the take-off.
But the timing of the take-off is probably not the full story. For instance, Japan did not move
onto a sustained growth trajectory until sometime during the 20th century, and yet it is today
richer than many Western European countries that went through the transition in the 19th century.
The combination of high investment rates, and the process of catch-up is a compelling explanation.
Indeed, it is the standard textbook account, which involves the use of the neoclassical growth model.
Figure 1 illustrates this possibility; a country that takes off at τ2 > τ1 may either partially catch
up with the early take-off group, or even overtake them. The overtaking phenomenon would reflect
different (investment) behavior post take-off.
Hence, whereas unified growth theory explains how comparative development is affected by the
difference in timing of take-off (i.e., τ1 − τ2 ), neoclassical growth theory provides a complementary
explanation as to the growth process after the take-off, which relates to the location of the steady

2
 Figure 1: Stylized Comparative Development

log y
delayed take-off, overtaking
log y(τ2 , ŷ2∗ )

log y(τ1 , ŷ1∗ ) early take-off

log y(τ2 , ŷ1∗ ) delayed take-off

log ȳ no take-off

τ1 τ2 t time

The figure illustrates how the relative timing of the take-off (τ1 ,τ2 or “no-take-off”) and post
take-off investment behavior, implicit in relative steady state paths (ŷ1∗ , ŷ2∗ ), jointly explain
GDP per worker differences at time t.

state growth trajectory.3 One might expect that the combination of the two theories would be a
powerful vehicle for an understanding of comparative development.
To put these considerations slightly more formally, suppose observed GDP per worker stagnated
prior to some point in time τ , after which the growth process takes off.4 Observed GDP per worker
at time t > τ can then be written (as a matter of pure accounting, in a continuous time setting) as

y(t) = y(τ ) · egy (t−τ ) , (1)

in which gy it the observed average growth rate between τ and t. In order to say something about
how yt is determined, we need to make statements about how the average growth rate between τ
and t, gy , is determined. This requires the use of theory.
Consider a standard continuous time Solow (1956) model, where production is Cobb-Douglas.
Accordingly, output is given by Y = K α (AL)1−α , where Y is GDP, K is physical capital, A

3 Of course, unified growth theory also speaks to the post take-off period. But from an empirical angle the Solow
model (so far) provides more structure, which is useful for present purposes.
4 Naturally, in practice, there is probably no such thing as “the year” of the take-off. The process is gradual and
initially accelerating (e.g., Phillips and Sul, 2009). Hence, this is a simplistic representation designed to deliver a
regression equation, the results from which are amendable to interpretation.

3
represents technology and L is labor; α, is the capital share in production and is bounded between
zero and one. Total savings in the economy is determined as a constant share, s, of total income,
and total investments, I, equal total savings. Physical capital accumulates in accordance with
K̇ = I − δK, where δ is the rate of capital depreciation. Finally, technology grows at a constant
rate, g, and the labor force at the rate n. In the long-run steady state of the model, GDP per
efficiency unit of labor, ŷ ≡ Y /AL, is given by
  α
∗ s 1−α
ŷ = . (2)
n+δ+g

In order to use the model to think about how observed GDP per worker is determined, we need to
work out the predicted average growth rate of GDP per worker between τ and t (i.e, gy in equation
(1)). Seen through the lens of the Solow model the average growth rate, gy , is approximately given
5
by
1
gy ≈ g + [log ŷ ∗ − log ŷ(τ )] . (3)
t−τ
That is, the average growth rate depends on the underlying - constant - trend growth rate, g, and
on the contribution from “convergence”, which is given by the distance between the level of GDP
per worker in efficiency units at time τ , ŷ(τ ) ≡ Y (τ )/(L(τ )A(τ )), and the steady state level of GDP
per efficiency unit of labor, as predicted by the Solow model, ŷ ∗ .
If we insert equation (3) into the accounting equation (1), we obtain:
  α
s 1−α
y(t) ≈ Aτ · eg(t−τ ) , (4)
n+δ+g

in which Aτ is the level of technology at time τ , A(τ ) = Aτ . Taking logs we are thus left with

α α
log(y(t)) ≈ log (Aτ ) + g (t − τ ) + log (s) − log (n + δ + g) . (5)
1−α 1−α

The only difference between this expression and the one emanating from a standard Solow model
is that the initial level of technology is given as of time τ rather than at some arbitrary “time
zero”.6 While this difference superficially seems minute, it has a substantive implication in that it

5 See Appendix for details on derivations. The adopted approximation assumes that the period t − τ is sufficiently long
to admit the country in question to approximately reach its steady state path. This means our regressions presume
that GDP per worker at the time of observation (in practise, the year 2000) is a point on the steady state path. This
is a common assumption in the empirical literature, going back to the work of Mankiw et al (1992). For a recent
contribution invoking this assumption, though in the context of studying historical levels of economic activity, see
Ashraf and Galor (2011).
6 See e.g. Mankiw et al (1992, equation 6), for the standard expression.

4
demonstrates how neoclassical growth theory as well as unified growth theory jointly contribute to
an understanding of contemporary comparative development.
To see this, note that the last two terms on the right hand side of equation (5) are the neoclassical
contribution to the understanding of contemporary levels of income, reflecting the level of steady
state income. That is, these are the post take-off growth determinants.
The contribution from “history” is found in the two other terms on the right hand side of equation
(5): time passed since the take-off, t − τ , and the initial level of technological sophistication at the
eve of take-off, Aτ (rather than at “time 0”). Accordingly, conditional on post take-off growth
determinants (s and so forth) we expect countries to be richer the longer its been since the economy
underwent the “industrial revolution”, and the more sophisticated the country was technologically
at the time.7
Before we move on, it is useful to observe how equation (5) speaks to the conceptualization of
contemporary income differences in Figure 1. In the depicted illustration a collection of the countries
are on their - parallel - balanced growth paths when t becomes “large”, whereas another country
(group of countries) is stuck in stagnation. It it clear that the implied income variation at this
juncture is attributable to both factors that influence the location of the steady state growth path
and factors that influence whether the process has begun at all. That is, the observed cross sectional
variation in income (log (y [τ2 , ŷ2∗ ]) − log (ȳ)) is partly due to “neoclassical factors” (that influence
ŷ2∗ ) and partly due to the difference in timing of the take-off τ2 . More generally, the fact that Aτ
enters into equation (5) implies that even though two countries may have identical ŷ ∗ they do not
necessarily attain the same growth path; if one were more technologically sophisticated initially this
effect will linger, as shown in Figure 1. This is the “decomposition” that equation (5) captures.
In order to take equation (5) to the data a few issues need to be addressed up front. The first issue
is how to date the “take-off”; τ in the derivations above. Based on the literature on unified growth
theory, the year of the onset of the fertility transition should be a good proxy for τ (Galor, 2010b).
While our preferred measure of τ thus is the year of the onset of the fertility transition, we also
explore an alternative indicator of the take-off: the “year of industrialization”, which is determined

7 Considerable variation in technological sophistication, prior to the take-off, is not inconsistent with limited pre take-
off variation in GDP per worker. In a pre-take off environment changes in technology would be expected to manifest
themselves in changes in population density, but not in sustained increases in living standards; see Ashraf and Galor
(2011) for theory and evidence; see Comin el al. (2010) for evidence that early technological sophistication (pre 1500)
seems correlated with contemporary income differences.

5
as the year in which employment in industry exceeds that in agriculture.8 These data are described
below.
The second issue is how to think about “g”; the underlying rate of technology growth. We will
assume that once a take-off is achieved the country taps into the world technology frontier. Hence,
g can be viewed as exogenously given, from the perspective of the individual country.9 As a result,
g is common across countries in the sample which means that we can think of it as a parameter to
be estimated (though see footnote 13 below).
The third issue relates to the neoclassical controls: n in particular. Empirically, peak population
growth rates, during the demographic transition, vary considerably across countries (e.g., Reher,
2004; Strulik, 2008). As a result, even conditional on the transition occurring, we expect differences
in average growth rates of the labor force, post transition. We proxy post transition labor force
growth by 1960-2000 average growth rates; similarly, s (also an 1960-2000 average) is to be thought
of as the post transition investment rate in physical capital.
A final issue is how to deal with log (Aτ ) . The simplest approach is to follow Mankiw et al.
(1992) and assume it is subject to random variation. Indexing countries by i this would imply


log (Aiτ ) = log A¯τ + ǫi . For our baseline regressions we also invoke an arguably slightly more
reasonable approach by assuming log (Aτ ) is subject to systematic regional variation, but random
variation within regions.10 Denoting regions by R this would mean log (ARiτ ) = log (ARτ ) + ǫi .
This can be handled in a cross-sectional context by adding regional fixed effects to the right hand
side of the regression equation.11 For the robustness checks we push the matter a little further by
controlling for plausible correlates with Aτ (such as early levels of schooling), as well as (fundamental)
determinants of Aτ : geography, culture and institutions (Acemoglu, 2009, Ch. 4).
In sum the regression equation to be estimated below is:

log (yi ) = β0 + β1 log (si ) + β2 log (ni + δ + g) + β3 τi + X′i γ + ǫi . (6)

8 See Strulik and Weisdorf (2008) for a unified growth theory that explicitly studies both the fertility transition and
the process of structural change, thereby providing theoretical foundations for the use of either of the two adopted
indicators as proxies for the timing of the take-off in GDP per worker.
9 Globally, of course, g is undoubtedly endogenous. See e.g., Howitt (2000).
10 The celebrated book by Diamond (1997) provides historical foundations for systematic differences in historical levels
of technological sophistication across regions or continents.
11 The regions included in our regressions are: East Asia and the Pacific; Europe and Central Asia; Latin America
and the Caribbean; Middle East and North Africa; South Asia and Sub-Saharan Africa.

6
where the matrix Xi contains regional fixed effects and alternative confounders to be discussed
below; γ is the associated vector of regression coefficients.12
The history augmented Solow model holds several strong predictions, which we test below.

• The predictions that β1 > 0, β2 < 0 and β3 < 0.

• The restriction that the coefficients on log (s) and log (n + δ + g) are numerically identical:
β1 + β2 =0.
• The prediction that β1 = −β2 is between 0.5 and 2/3; this mirrors that the capital share α
is expected to be between 1/3 and 0.4.
• The prediction that β3 should capture the frontier productivity growth rate in the sample;
a reasonable prior for this rate is about 2 percent on account of the growth trajectory of the
US economy (arguably the technological leader) during the 20th century.13

Hence, the model offers predictions regarding the signs of the parameters (1); the relative size of
the parameters (2); as well as their numerical value (3), (4). In particular, β3 is negative as each
year of delay of the fertility transition lowers GDP per worker relative to countries that have taken
off. The extent of the shortfall in GDP per worker depends on trend growth, g, which we expect is
about two percent per year.
Before we turn to the empirical analysis it is worth considering what the prospects are for identi-
fication by way of ordinary least squares (OLS). In order for OLS to deliver unbiased estimates the
right hand side variables need to be orthogonal to the residual in equation (6), which structurally
should contain information about the level of technology at the time of take-off.
As discussed in Mankiw et al. (1992), the independence between s, n and the level of A is plausible.
The fact that the potential covariance involves the level of technology lagged by up to a century
would not appear to make the case any worse. Based on unified growth theory it is also plausible
to assume that the onset of the fertility transition is largely orthogonal to the (pre take-off) level of
A; rather, it is the (expected) growth rate that might trigger the fertility decline (Galor and Weil,
2000). In these respects the key orthogonality conditions are plausible.
12 Comparing equation (6) with equation (5) it follows immediately that, structurally, β = A + g · t, and that ǫ
0 τ i
therefore captures unobserved country variation in Aτ .
13 Alternatively, one could assume that g varies across countries. For example, suppose individual countries are
visualized as gradually converging towards their technological steady state, as in the model by Nelson and Phelps
(1966). Then g would exhibit country specific variation in transition; in the long run, however, all countries would
converge to the frontier rate. In this case the interpretation of β3 changes from the unique “frontier rate” to the average
growth rate in the country sample (see e.g., Zellner, 1969). Again one would think that 2 % might be a reasonable prior
as the global distribution of income seems to have remained fairly stable during the second half of the 20th century,
which suggests that the frontier and average expanded at roughly similar rates (Acemoglu and Ventura, 2003).

7
 Nevertheless, identification is obviously not ensured. In theory, the (unobservable) Aτ impacts
on GDP per worker. Moreover, Aτ might be determined by factors that are correlated with “con-
temporary” savings, labor force growth as well as τi . If so, identification is jeopardized. In Section
3.3 we try to gauge the likely severity of this concern by controlling as best as we can for plausible
correlates with Aτ (such as early levels of schooling) as well as fundamental determinants of Aτ :
geography, culture and institutions (Acemoglu, 2009, Ch. 4).

3. Regression Analysis

3.1. Data. We estimate the history augmented Solow model in two samples, following the approach
of Mankiw et al. (1992). The first sample only excludes oil producing nations, whereas the second
sample excludes both oil producers and “small nations” with less than two million inhabitants.
Mankiw et al. defend these sample restrictions by arguing that labor productivity likely is poorly
understood by the mechanisms emphasized in the Solow model in both oil nations (where resource
extraction dominates) and small nations (where “idiosyncratic factors” might dominate).14
The Appendix provides details on our data sources, and Table 1 reports summary statistics for
our main variables. Investment rates and growth in the labor force are calculated as averages for
the period 1960-2000, as explained above.
In measuring the year of the take-off, τi , we employ three variables. Our main variable records
the year of the fertility decline (YFD), and is taken from Reher (2004) who determines the “date”
of the decline as (p. 21): “...the first quinquennium after a peak, where fertility declines by at least
8% over two quinquennia and never increases again to levels approximating the original take-off
point.” The fertility measure underlying Reher’s analysis is the crude birth rate, which is available
for extended periods of time in most countries around the world. A minor issue with the resulting
data is that Reher codes countries that, as of the time of his research, had not undergone the
transition as having experienced it in “2000”. Obviously, including these countries in our sample
would introduce considerable measurement error, for which reason we limit the sample to countries
that had undergone the transition before 2000.15 With the YFD indicator we are able to test the
14 We follow Mankiw et al. in the interest of comparability with earlier studies; nothing substantive changes if we
use a “full” sample of countries that only is limited by data availability. The difference in sample size between an
unrestricted sample, and our non-oil sample, is only eight observations. Specifically, the countries that drop out in the
non-oil sample are Bahrain, Benin, Brunei, Iran, Kuwait, Oman, Qatar and Saudi Arabia.
15 By focusing on countries in which the take-off has (in theory) occurred, we also limit the risk of misspecification
error, as it implies that equation (6) can legitimately be said to characterize all countries in our sample as of the year
2000, where GDP per worker is measured. Before the take-off a different equation would likely be appropriate for GDP
per worker, see Ashraf and Galor (2011), Dalgaard and Strulik (2013).

8
model on a sample involving up to 110 countries. In light of the discussion above, we expect to see
YFD enter the regression with a negative sign, and with a coefficient in the vicinity of 0.02.

Table 1: Summary Statistics of Main Variables

variable obs mean stdev min max

y 110 18136 18373 1052 67078
s 110 0.15 0.08 0.02 0.45
n 110 0.02 0.01 -0.01 0.05
YFD 110 1961 29 1865 1995
YFD2 91 1955 36 1800 1993
YIT 56 1956 37 1801 1998

The table reports summary statistics for GDP per worker in 2000 (y),
the average investment rate 1960-2000 (s), the average rate of labor
force growth (n), the year of the fertility transition dated by Reher
(2004) (YFD) and Caldwell and Caldwell (2001) (YFD2) as well as the
year of the industrial transition (YIT) from Bentzen et al. (2013).

Our second measure of the fertility transition, which we employ in our robustness checks, is taken
from Caldwell and Caldwell (2001). These data speak to the year of the fertility transition, judged
from total fertility rates rather than the crude birth rate, which could make a difference. This
variable is labeled YFD2, and it allows for a test of the model on a sample of up to 91 countries.
Finally, our third measure of the year of the take-off captures structural change: “The year of
the industrial transition” (YIT). This variable records the year in which employment in industry
exceeds employment in agriculture, and is taken from the work of Bentzen et al. (2013). As with the
year of the fertility decline from Reher (2004) we ignore countries that experienced the transition
after 2000. With this sample limitation we can estimate the model on a sample involving up to 56
countries.
Descriptively it is interesting to note that the three measures of the “take-off” have fairly similar
characteristics (albeit they are available for different samples of countries): On average the transition
occurs around the same time (late 1950s), exhibits fairly similar cross-country variation (i.e., a
standard deviation of roughly 35 years) and is observed within a broadly similar time frame (19th
century onwards). Figure 2 provides a visual sense of the overall concordance between the three
indicators.

3.2. Baseline results. Table 2 reports our baseline results. The first six columns refer to the “non
oil” sample, whereas the last six columns concern the more limited sample where we also exclude
countries with less than 2 million inhabitants.

9
 Figure 2: Correlation Between Alternative Indicators for Year of take-off
A: Fertility Decline: Reher vs. Caldwell and Caldwell B: Fertility Decline vs. Industrialization
2000

2000
RWA
KHM OMN
PRY
HND
IRN
NIC
GTM
DJILSO
SYR
GHA
BGD
VNM
KEN
MRT
LBY
SEN
SAU
SDN SYR IRN
GTM
ZAF
BOL
KWT
DZA
BWA
JOR
MNG
SWZ LBY
BHR
CHN
ECU
IDN
PRK
PAN
MEX
ZWE JOR
ZAFDZA BOL
BWA
Year of fertility decline (Reher)

ALB
AZE
BRA
CRI
MYS
TTO
VEN
COL
DOM
SLV
EGY
GUY
LBN
PER
SUR
THA
TUN
MAR
TKM
UZB
BLZ MEX PAN

Year of fertility decline (Reher)

KOR
LKA
BRN
CHL
MUS
IND VENTUNSUR
MYS
SLV
BRA
CRI
MAR
CHL MUSKOR
1950

SGP
PHL QAT

1950
JPN TWN
PRI
JPN

ROM
ROM
ITA PRT JAM USA ITA PRT
BGR JAM
CUB
AUT
FIN CAN
AUT FIN
GBRNLD CHE
DNK
ARG ESP GBR CHE
NLD
ARG DNK ESP
1900

BEL NOR

1900
BEL NOR
FRA DEU DEU FRA
URY
HUN HUN

SWE SWE
1850

1850
1800 1850 1900 1950 2000 1800 1850 1900 1950 2000
Year of fertility decline (Caldwell & Caldwell) Year of industrial transition

The figure depicts the correlation between year of the The figure depicts the correlation between year of the
fertility decline according to Reher (2004) and Cald- fertility decline (Reher version) vs. year of industrial
well and Caldwell (2001), respectively. The pure cor- transition (Bentzen et al, 2013). The pure correlation
relation is 0.88; 0.93 if France is omitted. is 0.65; 0.72 if England is omitted.

In column 1 and 7 we test the unrestricted baseline Solow model in the “full sample”. The sign
of the included variables (average investment rate, s, and average growth in the labor force, n) are
as expected, and the basic specification accounts for some 40% of the cross-country variation in
GDP per worker in 2000. Moreover, the structural prediction of the coefficients being the same in
absolute value cannot be rejected on neither the full sample nor the more limited sample (see bottom
of column 1 and 7, respectively). Hence, in the remaining we employ the restricted model where the
coefficients are required to be the same in absolute value.
The restricted model is estimated in column 2 and 8, from which we can elicit information about
the size of the structural parameter, α. Theoretically α should reflect the capital share in national
accounts and thus fall somewhere in an interval from 1/3 and 0.4. However, as is familiar from the
Mankiw et al. (1992) study, the parameter is estimated to be much higher at about 0.6.
In columns 3 and 9 we include the year of the fertility transition; in column 4 and 10 we further
include regional fixed effects (without substantive implications beyond a slightly higher R2 ), and
in columns 5-6 and 11-12 we limit attention to countries that went through the fertility transition
before 1970. There are two reasons why the latter sample restriction is of interest. First, our “Solow
controls” are measured as averages from 1960–2000, and are thought to proxy for post transition
developments. By restricting the sample to “early transition countries” this approximation should
improve. Second, when we derived the equation for estimation we assumed countries are reasonably

10
 Table 2: The History Augmented Solow Model

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
log GDP per worker, 2000

log(s) 1.215*** 1.239***

(0.148) (0.144)
log(n + d + g) -1.707*** -1.701***
(0.435) (0.443)
log(s) − log(n + d + g) 1.255*** 0.595*** 0.594*** 0.564*** 0.662*** 1.278*** 0.583*** 0.523*** 0.662*** 0.729***
(0.135) (0.162) (0.153) (0.198) (0.207) (0.133) (0.157) (0.141) (0.190) (0.200)
YFD -0.0222*** -0.0186*** -0.0111*** -0.0128*** -0.0227*** -0.0177*** -0.0119*** -0.0128***
11

(0.0034) (0.0038) (0.0030) (0.0037) (0.0033) (0.0035) (0.0031) (0.0038)

regional FE no no no yes no yes no no no yes no yes
β1 + β2 = 0 (p-value) 0.31 0.34
implied α 0.56 0.37 0.37 0.36 0.40 0.56 0.37 0.34 0.40 0.42
α = 1/3 (p-value) 0.00 0.53 0.51 0.73 0.38 0.00 0.58 0.87 0.34 0.19
observations 110 110 110 110 58 58 99 99 99 99 52 52
R2 0.448 0.445 0.634 0.735 0.462 0.589 0.481 0.478 0.681 0.792 0.546 0.643

i) log(s) refers to the 1960-2000 average investment rate, log(n + delta + g) is the sum of 1960-2000 labor force growth and 0.05, YFD is the year of the fertility decline, all models
contain a constant term; six regional fixed effects. (ii) Robust standard errors in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1. (iii) Cols. 1-4 represents the ”full” sample
where only oil countries are omitted; cols. 5 and 6 further removes non-oil countries with YFD in 1970 or later; cols. 7-10 considers non-oil and non-small countries, and cols. 11-12
further drop non-small and non-oil countries with YFD in 1970 or later.
close to their steady state path by 2000, which seems more plausible in the case of early transition
countries. As is clear from Table 2, however, we obtain very similar results in either of the three
samples; i.e., whether we examine non-oil countries, non-oil and non-small and non-oil (and non-
small) that went through the fertility transition before 1970.16
Several results from columns 3-6 and 9-12 are worth noting. First, the year of the fertility transition
is highly significant in statistical terms and contributes significantly to the overall R2 , which increases
to between 0.6 and 0.7 (column 3, 9). If regional fixed effect are added, the model can account for
up to 80% of the variation in GDP per worker in 2000 (column 10).
Second, the parameter estimate is reasonable on a priori grounds; we obtain estimates in the
vicinity of -0.02. Hence, each year the transition is delayed comes (on average) at a cost of roughly
2% lower GDP per worker in 2000. While this number matches expectations on structural grounds
it is worth spelling out its implications. From Table 1 we know that there is a lot of variation
in the timing of the fertility decline; the mean is 1961, but the standard deviation is roughly a
generation (i.e., about 30 years). If each year “costs” 2 percent lost output per worker, a one
standard deviation’s worth of delay lowers GDP per worker in 2000 by nearly 50%.
The third noteworthy result is that the “anomaly” detected originally by Mankiw et al. (1992),
and that is re-confirmed in column 2 and 8, is eliminated: the estimated share of capital is around
1/3 (see bottom of column 3-6 and 9-12 for a formal test). It is worth pausing to reflect on why
this result emerges. Naturally, the mechanical explanation is that the year of the fertility decline is
negatively correlated with savings (at -0.45 in the non-oil sample), and positively correlated with the
growth rate of the labor force (at 0.63 in the non-oil sample). As a result, once the fertility decline
is included in the regression, the parameter estimate for the savings rate and for the term involving
n will shrink in absolute value. But what is a plausible economic interpretation?
The positive correlation between the year of the fertility decline and observed growth in the labor
force 1960–2000 is relatively easy to rationalize. In the aftermath of the fertility decline, growth in
the labor force gradually declines. Ceteris paribus, one would therefore expect to see faster growth
in the labor force 1960-2000 in countries for which the transition occurred relatively recently.17

16 More formally, on statistical grounds one cannot reject the hypothesis of equality between the point estimates for
YFD in columns 4 and 6, and columns 10 and 12, respectively.
17 Moreover, late transition countries tend to attain higher peak-fertility rates compared with early transition countries
(Reher, 2004; Strulik, 2008), which will serve to reinforce the positive correlation between YFD and average (post-
transition) labor force growth.

12
 The observed negative correlation between the year of the take-off and the average savings rate
might also have a relatively straightforward explanation. It has long been known that the average
propensity to save tends to increase in the process of industrialization. Indeed, Lewis (1954, p. 155)
viewed this fact as the central problem of development economics to explain. In more recent work
Laitner (2000) provides a theory, according to which gradual structural change during the take-off
shifts savings from largely consisting of capital gains on land holdings to savings in the form of
physical capital; during development (NIPA) savings therefore increase. Since the fertility transition
is the critical watershed beyond which the growth process takes off, according to unified growth
theory, one would expect the savings rate to be higher in places farther away from the transition
(i.e., where the YFD is a “low” number), ceteris paribus.18 This mechanism can therefore account
for a negative correlation between the year of the fertility decline, and the average savings rate
1960-2000.19 Taken together, the year of the fertility transition should reduce (in absolute value)
both the coefficient on log(s) and log(n + δ + g), leading to a lower estimate for α.
The baseline results leave some clue as to the relative importance of the timing of the take-
off and the forces emphasized by neoclassical growth theory (cf. the discussion in Section 2), in
accounting for cross-country inequality in labor productivity. In order to compare the estimate
for YFD and the “Solow term” on an equal footing one can calculate the standardized regressions
coefficients, which then reflect the impact from the two controls on GDP per worker in 2000, if either
variable is changed by one standard deviation. Accordingly, conditional on regional fixed effects
the standardized regression coefficient associated with the “neoclassical term” is 0.31, whereas it
is 0.46 for YFD in the non-oil sample; the corresponding numbers for the sample that excludes
small countries as well are 0.28 and 0.45. Taken at face value these findings suggests that the
forces emphasized by unified growth theory are somewhat more powerful than those traditionally
emphasized by neoclassical growth theory in explaining GDP per worker in 2000.

3.3. The History Augmented Solow Model: Robustness of Baseline Results.

3.3.1. Alternative Measures of the “Take-off ”. Table 3 explores whether our results are sensitive
to the choice of indicator for the “year of the take-off”. The first check consists in employing an

18 In the theory developed by Strulik and Weisdorf (2008) it is demonstrated that the fertility decline is associated
with rapid growth in relative employment in industry during the take-of.
19 Complementary explanations include the lowered relative importance of subsistence consumption in the aftermath
of the take-off (Strulik, 2010) and that the development process instigates rising patience, which elevates savings
(Strulik, 2012).

13
alternative indicator for the year of the onset of the fertility decline due to Caldwell and Caldwell
(2001) (C&C). As seen from Figure 2 there are instances where the two measures give rise to widely
different transition years. The extreme example is France for which C&C suggest the year 1800,
whereas Reher estimates the onset of the decline to take place around 1900. But overall the difference
between the two measures is actually not all that large: On average the fertility transition occurs
only 3.5 years (median: 2.5 years) later according to C&C’s estimates compared to those from Reher.
In this light it is perhaps unsurprising that our baseline results are robust to the change in the
indicator for the year of the fertility decline (see Columns 1-4 of Table 3). The only change is that
the implied estimate for α edges slightly higher, and in some cases rises significantly above 1/3. Still,
it is never significantly different from 0.4 (see bottom of Table 3 for formal tests).
The second check consists of using an altogether different indicator for the year of the take-off,
namely the “year of the industrial transition” (YIT). In most respects the results remain unaffected.
The year of the take-off is highly significant (along with the “Solow controls”), and we also obtain
similar results with respect to α, which is estimated around 0.4. However, the point estimate for
the take-off declines, when the YIT indicator is used.
A possible explanation is that the regressions involve considerably fewer observations when YIT
is used, compared with those involving YFD ad YFD2. Moreover, most of the countries that drop
out are found among the currently poorest countries in the world thus limiting the variation in the
data appreciably. To check this account we re-ran the regressions using YIT and YFD2 on the
same sample.20 Conditional on regional fixed effects and the “Solow term”, we find -0.0059 for YIT
(p-value of 0.007), which is insignificantly different from the result in Table 3 column 6, and -0.0086
(p-value of 0.047) for YFD2, which is significantly different (at a 10% level of significance) from the
result for YFD2 reported in Table 3, column 2. This suggests that the difference in results, when
moving from YFD2 to YIT, probably is driven by the change in sample.
The overall impression is that our baseline results are reasonably robust to the use of alternative
indicators. In particular, the amount of the variation explained by the history augmented Solow
model, the strong statistical significance of the year of the take-off, and the lowered estimate for the
capital share all carry over with little or no qualifications to the use of YFD2 or YIT.

20 We use YFD2 because it allows for the maximum common sample, N= 45. Naturally, this involves the “non-oil”
sample.

14
 Table 3: The History Augmented Solow Model: Alternative Indicators for Take-off

(1) (2) (3) (4) (5) (6) (7) (8)

log GDP per worker, 2000

log(s) − log(n + d + g) 0.721*** 0.728*** 0.691*** 0.663*** 0.742*** 0.693*** 0.751*** 0.682***
(0.180) (0.152) (0.187) (0.163) (0.131) (0.156) (0.134) (0.169)
YFD2 -0.0121*** -0.0119*** -0.0134*** -0.0129***
(0.0026) (0.0031) (0.0030) (0.0039)
YIT -0.0068*** -0.0055*** -0.0069*** -0.0056***
15

(0.0017) (0.0016) (0.0017) (0.0016)

regional fe No Yes No Yes No Yes No Yes

implied α 0.42 0.42 0.41 0.40 0.43 0.41 0.43 0.41

α = 1/3 (p-value) 0.16 0.08 0.25 0.27 0.03 0.16 0.03 0.23
α = 0.4 (p-value) 0.75 0.67 0.89 0.98 0.55 0.86 0.51 0.93
observations 91 91 81 81 56 56 50 50
R2 0.615 0.684 0.650 0.745 0.617 0.696 0.633 0.713

(i) All regressions contain a constant. (ii) Robust standard errors in parentheses, *** p < 0.01, ** p < 0.05, *p < 0.1. (iii) YFD2
is the year of the fertility transition according to Caldwell and Caldwell (2001), whereas YIT is the year of the industrial transition
from Bentzen et al. (2013). (iv) Cols. 1-2, 5-6 estimate the model in the non-oil sample, cols. 3-4, 7-8 also exclude small countries.
3.3.2. Confounders: Additional Controls for Aτ . A lingering concern is that omitted variables might
be affecting the significance of the year of the fertility transition in the regressions above. In theory
our regression model involves the level of technological sophistication at the time of the take-off,
Aτ , which we cannot measure directly. Some of this unobserved variation is likely picked up by our
regional fixed effects, but this naturally leaves out unobserved variation in Aτ within regions.
In brief the concern is then the following. If we denote by Ãτ unobserved variation (within regions)
in the pre-take off level of technology our OLS estimates for our parameter of main interest, β̂1 can
be written as

!
cov(Y F D, Ãτ )
β̂1 = β1 + . (7)
var(Y F D)
 
cov(Y F D,Ãτ )
As discussed in Section 2, the key orthogonality condition, var(Y F D) = 0, is theoretically
plausible: in Galor and Weil (2000) the onset of the fertility transition is unleashed when the rate
of technological change attains a critical level; it is not triggered by a sufficiently high level of A.
Similar arguments can be made regarding covariances pertaining to log(s) and log(n + δ + g) (see
Mankiw et al (1992), or Section 2 above). Nevertheless, at this stage the orthogonality conditions
amounts to an assumption, which seems worth testing to the greatest extent possible.
Since we cannot measure Aτ directly we have to take an indirect route, which involves a two
pronged strategy. First, we control for outcomes that might be correlated with Aτ : primary school
enrollment rates in 1900, and GDP per capita in 1900. The motivation for the former control is
simple: If education is important for technological change (either via adoption or innovation), early
levels of education might be a marker for the initial level of technology.21 The main problem with
this control is that unified growth theory predicts that the fertility transition unleashes a process of
human capital accumulation (the quantity-quality trade-off is triggered), which stimulates growth
and technological change. If we control for post-take off human capital then the year of the fertility
transition should turn insignificant, which makes the proposed robustness check hard to interpret.
We return to this issue in the context of the results below, and at greater length in Section 3.4. The

21 In early states of development primary schooling might be highly important for technology diffusion along the
intensive margin. That is, technology usage lags might be shorter (Comin et al, 2008) in societies where the citizen’s
have at least basic education. In addition, primary school enrolment rates might well be highly correlated with
secondary and tertiary enrolment rates. Observe also that Murtin (2012) documents a clear link between fertility and
schooling over the period 1870-2000; in accordance with unified growth theory, Murtin finds that changes in schooling
are negatively correlated with changes in fertility during this period.

16
motivation for the use of GDP per capita in 1900 is that technology enters the production function
for which reason GDP per capita in 1900 could act as a stand-in for early technology.22
The second strategy is to control for fundamental determinants of productivity: institutions,
culture and geography (Acemoglu, 2009, Ch. 4). If these factors influence productivity, they might
well have helped shape Aτ making them relevant indirect controls for Aτ . Moreover, they might
also have worked to influence the timing of the fertility transition, as argued in several recent papers
(Andersen et al, 2010; Basso and Cuberes, 2011), for which reason these checks are non-trivial on a
priori grounds.
In controlling for institutions we employ direct measurement of executive constraints in 1900,
as well as possible historical controls for the development of institutions: population density in
1500, fraction of population speaking a European language as well as a full set of legal origins
dummies (Acemoglu et al., 2001, 2002; Hall and Jones, 1999; Glaeser and Shleifer, 2002).23 In
controlling for “culture” we deploy three sets of determinants: religious affiliations (see e.g. Barro and
McCleary, 2003), ethno-linguistic fractionalization (e.g., Devleeschauwer et al, 2003; Michalopoulos,
2012), and finally the genetic distance between country citizens and the population in the UK anno
1500. The latter control is inspired by Spolaore and Wacziag (2009) and might be interpreted
as capturing cultural distance between individual countries and the population that saw the first
industrial revolution. Finally, our geography controls are distance to the equator (absolute latitude),
tropical area and distance to coast or river, in addition to a full set of regional fixed effects (e.g.,
Hall and Jones, 1999; Gallup and Sachs, 2000; Rappaport and Sachs, 2003). Data sources are given
in the Appendix.
Table 4 reports our results. In column 1 and 3 we check for the impact on the point estimate
of interest from including early education and income alongside the “standard” controls. As seen,
this does seem to leave an imprint in the point estimate for YFD, which could be a sign of omitted
variable bias. Another possibility, however, is that enrollment and GDP per capita in 1900 represent
an outcome from the fertility transition, as discussed above.

22 But it might not; in a Malthusian steady state environment A and GDP per capita are uncorrelated (e.g., Ashraf
and Galor, 2011; Dalgaard and Strulik, 2013).
23 Why not use measures of contemporary institutions? The reason is that institutional quality likely is determined by
human capital (at least in part), which in theory is determined by the timing of the fertility decline as noted above (see
for instance, Glaeser and Saks (2006) who document that human capital accumulation has worked to lower corruption
in the US and, more generally, Glaeser et al. (2004)). Consequently, adding measures of contemporary institutions
would not allow for a discriminatory test since these (in part) are an outcome from a process which may have begun
with the fertility transition.

17
 Table 4: The History Augmented Solow model and Potential Confounders

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
log GDP per worker, 2000

log(s) − log(n + d + g) 1.106** 0.835 0.497** 0.477** 0.354** 0.537*** 0.485*** 0.550*** 0.580*** 0.525***
(0.420) (0.488) (0.207) (0.210) (0.162) (0.173) (0.155) (0.159) (0.156) (0.133)
YFD -0.007* -0.016** -0.007** -0.011* -0.011*** -0.011** -0.015*** -0.0169*** -0.0189*** -0.0162***
(0.004) (0.006) (0.003) (0.006) (0.003) (0.005) (0.003) (0.004) (0.004) (0.004)
log GDP per capita 1900 -0.059 -0.238
(0.173) (0.212)
log prim enr. rate 1900 0.054 0.030
(0.114) (0.110)
Executive constraints (1900) 0.140***
(0.0349)
log Population density 1500 -0.053
(0.068)
Fraction speaking european language 0.461**
(0.184)
Legal origin: British 0.192
(0.245)
Legal origin: French 0.133
(0.201)
Legal origin: Socialist -0.574**
(0.243)
Legal origin: German 0.589*
(0.307)
Fraction Catholic (1980) 0.004
(0.003)
Fraction Muslim (1980) -3.84e-05
(0.003)
Fraction Protestant (1980) 0.005
(0.004)
Ethnic linguistic fractional (1985) -0.001
(0.003)
Genetic distance to UK, 1500 0.080
(0.245)
Absolute latitude -0.001
(0.011)
Tropical area -0.536*
(0.313)
Distance to coast or river -0.305*
(0.164)
Regional FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

p-value (added controls, joint) . . . . . 0.04 0.00 0.39 . 0.11

observations 30 28 58 55 43 94 109 107 110 104
R2 0.777 0.794 0.739 0.751 0.897 0.768 0.768 0.739 0.735 0.772

(i) Robust standard errors in parentheses, *** p < 0.01, ** p < 0.05, * p < 0.1. (ii) All models contain a constant term. (iii) The reported p-value refers to
the joint exclusion of added controls (if more than one). (iv) In cols. 2 and 4 countries with transition before 1900 are excluded. (iv) See Data appendix for
sources of controls.

A simple way to check which of the two possibilities is the viable one consists of focusing on
countries in which the fertility transition had yet to occur as of 1900. As it turns out, in the sample
underlying column 1 only two countries experienced the transition before 1900, implying GDP per
capita in 1900 might have been affected by the fertility transition: Sweden and Uruguay. In Column
2 we therefore omit these two countries, after which the point estimate rises (in absolute value)
to a level reasonably comparable to that reported in Table 2 (see Table 4, column 2). Similarly,

18
in the (slightly larger) sample involving primary school enrolment rates, the list of pre-1900 take-
off countries is Hungary, Sweden and Uruguay. Column 4 omits them, which also produces a
numerically greater point estimate for YFD. In general, these results seem to provide some credence
to the orthogonality condition discussed above.24
The remaining columns of Table 4 report the results from the “fundamental determinants” check.
In some cases the added controls appear significant (individually or collectively), yet they seem to
have little influence on the coefficient for YFD. This suggests that while (say) early institutions
might have impacted on the level of technological sophistication, the orthogonality between the level
of technology and the timing of the transition ensures consistent OLS estimates.25 In other cases
the controls turn out insignificant. That should not be taken as evidence of “lack of importance”.
Rather, it suggests that the variables in question likely have impacted on contemporary economic
development through one or more proximate sources of income: the year of the take-off, or post
take-off investment and fertility.26

3.4. The Fertility Transition and Current Labor Productivity: Exploring the Channels
of Influence. In the regressions above we did not control for current human capital. The reason
is that human capital accumulation, in theory, is the key mechanism by which the fertility decline
influences contemporary income. If true, it follows that once contemporary human capital levels are
included in the regression, the impact from the fertility decline should evaporate.
To see the connection more clearly, suppose the evolution of average years of schooling in the labor
force is described by the following linear first order differential equation (Aghion and Howitt, 2008,

24
If we employ YFD2, the list of pre-1900 transition countries grows: Belgium, France, United Kingdom, Iceland,
Netherlands, New Zealand and Ireland. Nevertheless, all results in Table 4 carry over to this alternative indicator.
The point estimate for YFD2 is around -0.01 in all specifications, with all countries included. If we then drop the pre-
demographic transition countries we once again see an increase in economic significance for YFD2. In the regression
involving GDP per capita 1900 we obtain -0.012,t = 2.32; in the human capital setting the point estimate is -.013,
t = 2.46. These results are available upon request. We have also explored an alternative data source for human capital
(average years of schooling), which recently has been compiled by Morrison and Murtin (2009). The basic results
are very similar. A difference is, however, that when we use the YFD indicator both human capital and YFD turn
insignificant; this is true whether we omit countries that went through the fertility decline before 1900 or not. When
we use YFD2 as take-off indicator along side Morrison and Murtin (2009)’s data for human capital, only YFD2 remains
significant. This is true whether or not we exclude countries that observed a fertility decline before 1900, albeit the
economic significance of the transition year is increased if we exclude countries that experienced the fertility transition
in the 19th century.
25
The significance of institutions in the regression is consistent with the analysis in Eicher et al. (2006), which
establishes empirically a causal impact of institutions on productivity, conditional on the Mankiw et al. set of controls,
that is, institutions seem to raise productivity chiefly via A.
26
A case in point would be genetic distance to the UK. Basso and Cuberes (2011) document that a greater distance
to the US leads to a delayed fertility decline, which (as shown above) leads to lower GDP per worker today.

19
Ch. 13):
ṡ(t) = u − µs(t)

where ṡ is change in years of schooling, u is the fraction of time spend per year accumulating skills,
and µs(t) is the loss of schooling years originating from people exiting the labor force. For simplicity
u and µ are treated as constants. The solution to this differential equation, assuming s(τ ) = 0, is

uh i
s(t) = 1 − e−µ(t−τ ) .
µ

Accordingly, the level of schooling at some point in time t (say, the year 2000) depends on how
much time is invested on net per period (u/µ), but also on the calendar date at which investments
commenced, τ .
The influence from the latter is intuitive: even at a 100% enrollment rate, average years of schooling
will only gradually rise when well educated cohorts replace less well educated ones. For investments
given, average years of schooling must be – in part – determined by the time at which such invest-
ments started being undertaken. Now, according to the bulk of the literature on growth in the long
run, the year of the fertility decline should be a good marker for the time at which human capital
investments start being made in earnest; a good proxy for τ in other words (e.g., Galor and Weil,
2000; Lucas, 2002). As a result, the fertility transition should matter to current labor productivity
via its impact on current human capital.
Of course, the influence from human capital on growth may involve multiple channels. The
influence could be direct, as in Lucas (1988). Alternatively, or in addition, it could influence growth
via faster technology adoption, as in e.g. Nelson and Phelps (1966). Moreover, one may hypothesize
that a higher level of schooling in the population at large gradually works to improve institutional
quality (e.g., Glaeser et al, 2004). Hence, for any of these reasons one would expect that current
human capital levels should appear as a determinant of contemporary labor productivity. If the
fertility transition is influencing productivity via human capital, its significance should diminish
once current human capital is controlled for in the regression.27
In order to assess whether this is the case or not, we invoke three measures of outcomes from
(past and present) “quality investments”: Average years of schooling in the population (+15) in
27 As noted in the introduction, the fertility decline also influences post-transition growth via reduced capital dilution;
with (gradually) slower population growth the capital-labor ratio is increased. Notice, however, that the regressions
above already controls for this channel, as the specification includes a control for the rate of labor force growth – post
transition, proxied by labor force growth 1960-2000. The significance of the timing of the fertility transition above
cannot therefore be attributed to reduced capital dilution, leaving only the human capital channel.

20
2000; the average IQ score in the population, and life expectancy at birth in 2000. The results from
re-estimating our baseline model – human capital augmented – are reported in Table 5.
All three “child quality variables” are highly significant correlates with GDP per worker, as ex-
pected. However, whereas average years of schooling and IQ scores eliminate the influence from the
fertility transition (column 2 and 4) the same is not true for life expectancy (column 1). As can
be seen from column 3 and 5, years of schooling and IQ render the fertility transition insignificant
mainly due to a marked reduction in the point estimate on the latter; the coefficient associated with
the fertility decline is reduced by 2/3 or more when either the schooling variable or IQ is introduced.
These results are consistent with the main prediction of unified growth theory: the differential timing
of the fertility transition is a strong determinant of current income differences due to its influence
on the path of human capital accumulation, which in turns stimulates growth in multiple ways.
At the same time, some of the results in Columns 2-4 appear problematic. Specifically, as seen
from Column 2, the estimated return to schooling is large: about 20%. This return estimate is
much too large to be attributed to the direct impact of schooling on individual-level productivity;
in practise, micro estimates are typically “only” about 10% (e.g., Caselli, 2005). An interpretation
of this result is that the estimates from Table 5 convolute the direct effect of human capital on
productivity (which is what micro studies estimate) and the indirect effects, via innovation. Such
indirect effects may either result from the production of new ideas, as emphasized in the R&D
growth literature (e.g., Romer, 1990; Aghion and Howitt, 1992), or be the consequence of more
rapid technology adoption (e.g., Nelson and Phelps, 1966; Benhabib and Spiegel, 2005).
In order to check the viability of this explanation for the elevated return estimate, we introduce
the R&D investment rate. If indeed the large return on skills (recovered above) is attributable to the
indirect effects of human capital on technology innovation or adoption, we expect to see it decline if
the R&D investment rate is introduced, as the latter should pick up the innovation channel.
As seen from column 6, when R&D investment (as a fraction of GDP) is added to the model
the estimated return on schooling shrinks to about 10%. To isolate the pure effect of introducing
R&D investments we also provide the results from estimating the model without R&D, but in the
same sample (column 7); in this setting the estimated return is considerably higher at 15%, as
expected. Interestingly, however, the estimate for IQ does not change much; when the average IQ
score increases by 1 it is associated with an increase in labor productivity of about 4% whether we
control for R&D or not.

21
 Table 5. Channeling the Effect of the Fertility Decline

1 2 3 4 5 6 7 8 9
log(y) log(y) log(y) log(y) log(y) log(y) log(y) log(y) log(y)

log(s) − log(n + d + g) 0.464*** 0.3 0.573*** 0.08 0.471* 0.642*** 0.800*** 0.31 0.596**
(0.14) (0.19) (0.19) (0.28) (0.28) (0.21) (0.21) (0.26) (0.22)
YFD -0.0123*** -0.0052 -0.0149*** -0.0022 -0.00927**
0.0038 0.0042 0.0045 0.0036 0.0045
life expectancy (at birth, 2000) 0.0549***
(0.01)
avg. years of schooling (above 15, 2000) 0.211*** 0.103* 0.143***
(0.05) (0.06) (0.04)
IQ score 0.0548*** 0.0482** 0.0495***
22

(0.01) (0.02) (0.02)

log(sR&D ) 0.197* 0.277**
(0.11) (0.12)

regions Yes Yes Yes Yes Yes Yes Yes Yes Yes

constant 30.04*** 18.27** 38.98*** 9.94 28.39*** 8.988*** 8.390*** 5.657*** 5.415***
(8.09) (8.29) (8.85) (7.23) (8.73) (0.65) (0.38) (1.75) (1.55)
observations 110 80 80 49 49 54 54 39 39
R2 0.78 0.78 0.72 0.78 0.73 0.67 0.63 0.76 0.72
(i) Robust standard errors in parentheses. (ii) ***: p < 0.01, **: p < 0.05, *: p < 0.1. (iii) The sample excludes oil producers. Column 3 re-estimates the
model on the same sample as in Column 2 to assess the importance of the sample to the insignificance of YFD; the same exercise is conducted in Column
4 in the context of the alternative “child quality indicator”: IQ. Columns 6-9 probe the consequences of adding R&D investment rates for the return on
skills and IQ.
 Overall, the results from Table 5 paint a coherent picture. Early transition countries managed to
start accumulating human capital sooner than late take-off countries. The resulting human capital
divergence instigated divergence in livings standards. Hence, while the timing of the fertility decline
is a determinant of contemporary comparative development, its effect is operating through human
capital and ideas; the more proximate sources of growth.
Finally, one may observe that the above results are fully consistent with the various augmentations
of the Solow model that have been proposed (Mankiw et al, 1992; Ram, 2007; Knowles and Owen,
1995; Nunnemann and Vanhout, 1996). Yet it adds one important additional piece of information.
Namely, that the origin of observed differences in these additional “capital stocks” (be it schooling,
IQ, life expectancy or technological “ideas”) is found in the differential timing of the take-off to
sustained growth, which is connected to the onset of the fertility transition.

4. Concluding Remarks

The present paper has augmented the empirical Solow model by taking into account the timing
of the take-off to sustained growth. According to unified growth theory, the timing of the fertility
decline should be a good proxy for the timing of the take-off to growth. We have shown theoretically
how one can combine key insights from unified growth theory with the neoclassical model in a simple
way.
Our empirical results support the notion that contemporary income differences are highly influ-
enced by both the year of the take-off and post take-off influences captured by the neoclassical growth
model. We find that adding the onset of the fertility transition to the neoclassical growth model
improves the empirical fit considerably. Taken together the onset of the fertility decline and (post
take-off) investment and labor force growth can account for about 70% of the variation in GDP per
worker; nearly 80% if one on top adds regional fixed effects. When we examine the standardized
regression coefficients we find that the influence from the fertility transition is slightly greater than
the influence from the “neoclassical determinants” of productivity, as of the year 2000.
Another key finding is that the obtained parameters square well with priors. The share of capital is
estimated to be between 1/3 and 0.4, and the point estimate for the fertility decline suggests that each
year the transition is delayed lowers GDP per worker by about two percent. The latter is consistent
with the view that once the take-off has occurred individual countries (on average) manage to tap

23
into the “world technology frontier”, thus experiencing a rate of underlying productivity growth
similar to “leader nations” like the US, conditional on physical capital investments etc.
The influence from the fertility transition is robust to the inclusion of controls for institutions,
culture, geography, as well as correlates with early levels of technological sophistication, like early
education. But the partial correlation between the year of the fertility decline and current GDP per
worker is eliminated if we control for the channels through which the fertility transition is supposed to
affect contemporary comparative development: reduced post take-off capital dilution and increased
human capital accumulation.
Admittedly, our regressions do not necessarily establish causality, although the obtained point
estimates square well with priors. Hence, an important topic for future work is to obtain a clearer
empirical understanding of the determinants of the fertility transition. This will eventually pave
the way for a study where the causal impact from the fertility transition on global cross-country
inequality in labor productivity can be assessed.

24
 Appendix A. Derivation of equation (5)

The Solow mode consists of the aggregate production function Y = K α (AL)1−α , the equation of
motion of capital K̇ = I/K −δK and the investment equation I = sY . Production in efficiency units
is defined as ŷ = Y /(AL) and capital stock in efficiency units is defined as k̂ ≡ K/(AL), implying
ŷ = k̂ α . Log-differentiating k̂ we obtain
˙
k̂ K̇ Ȧ L̇ I
= − − = − δ − g − n, (A.1)
k̂ K A L K
in which the last equality makes use of the equation of motion for K and the definitions of pro-
ductivity growth Ȧ/A ≡ g and population growth L̇/L ≡ n. Rewriting the investment rate we
get
I Y ŷ
= s = s = sk̂ α−1 .
K K k̂
Inserting this information into (A.1) we get the fundamental equation of the Solow growth model in
efficiency units (A.2).
˙
k̂
= sk̂ α−1 − δ − g − n. (A.2)
k̂
˙
At the steady-state k̂/k̂ = 0 and thus solving for k̂
  1   α
∗ s 1−α
∗ s 1−α
k̂ = ⇒ ŷ = , (A.3)
δ+g+n δ+g+n
where a star (∗) denotes a steady-state value. Taking the first order Taylor approximation of (A.2)
evaluated at the steady state we obtain
!
˙ ∗
k̂ k̂ − k̂
≈ s(α − 1)(k̂ ∗ )α−2 · (k̂ − k̂ ∗ ) = − (1 − α)(δ + n + g) · ,
k̂ | {z } k̂ ∗
=λ

where the right hand side follows after inserting k̂ ∗ from (A.3). The parameter λ denotes the speed
of convergence. To see this clearly consider another Taylor approximation at k̂ ∗ :
! !
k̂ k̂ ∗ 1 k̂ − k̂ ∗
log ≈ log + (k̂ − k̂ ∗ ) =
k̂ ∗ k̂ ∗ k̂ ∗ k̂ ∗
Use this information in
! !
˙
∂ log(k̂/k̂ ∗ ) k̂ k̂ − k̂ ∗ k̂
= = −λ = −λ log
∂t k̂ k̂ ∗ k̂ ∗
and observe that this equation constitutes a linear ordinary differential equation for the variable
log(k̂/k̂ ∗ ). The solution is given by log(k̂(t)/k̂ ∗ ) = log(k̂(τ )/k̂ ∗ ) · e−λ·(t−τ ) that is
 
log(k̂(t)) ≈ log(k̂ ∗ ) · 1 − e−λ(t−τ ) + e−λ(t−τ ) log(k̂(τ )). (A.4)

Recall that ŷ(t) = k̂(t)α and thus log(ŷ(t)) = α log(k̂(t). Inserting log(k̂(t)) from (A.4) provides
(A.5).  
log(ŷ(t)) ≈ log(ŷ ∗ ) · 1 − e−λ(t−τ ) + e−λ(t−τ ) log(ŷ(τ )). (A.5)

25
 Now consider income per worker defined as y ≡ Y /L = Aŷ. Let gy denote the average growth
rate of y between time t and τ and gŷ the respective average growth rate of ŷ. Then from log-
differentiating y
log(ŷ(t)) − log(ŷ(τ ))
gy = g + gŷ = g +
t−τ
Inserting log(ŷ(t)) from (A.5) this can be written as
1 − e−λ(t−τ )
gy ≈ g + (log(ŷ ∗ ) − log(ŷ(τ ))) . (A.6)
t−τ
From this follows for sufficiently large t − τ the approximation (3) in the main text.
Finally note that (A.6) can equivalently be written as gy (t − τ ) = g(t − τ ) = log [ŷ ∗ /ŷ(τ )] that is
e(gy −g)(t−τ ) = ŷ ∗ /ŷ(τ ). This information can be used to rewrite the growth equation for income per
capita:
y(t) = y(τ )egy (t−τ ) = A(τ )ŷ(τ )egy (t−τ ) = A(τ )ŷ ∗ eg(t−τ ) .
Inserting ŷ ∗ from (A.3) provides (4) in the main text and taking logs provides (5) in the main text.

Appendix B. Data Sources

Absolute latitude:
Source: Gallup et al. (1999).
Distance to coast of river:
Source: Gallup et al. (1999).
Executive constraints in 1900:
Source: Polity IV., http://www.systemicpeace.org/polity/polity4.htm.
Fraction speaking a major European language:
Source: Hall and Jones (1999).
Fraction of area in tropics:
Source: Gallup et al. (1999).
GDP per worker in 2000:
Source: Penn World Tables, 6.3. (Heston et al., 2009),
GDP per capita in 1900:
Source: Barro–Ursúa macroeconomic data (2010), downloaded from http://rbarro.com/data-sets/.
Genetic distance to England in 1500:
Source: Spolaore and Wacziarg (2009).
Growth in the Labor Force:
Source: Penn World Tables 6.3. (Heston et al., 2009).
Investment rates 1960-2000:
Source: Penn World Tabels 6.3 (Heston et al., 2009).
IQ score:
Source: Jones and Schneider (2006).
Legal origin dummy’s:
Source: Treisman (2007).
Life expectancy at birth in 2000:

26
 Source. World Development Indicators, available at: http://data.worldbank.org/.
Population density in 1500:
Source: McEvedy and Jones (1978).
Primary school enrollment rates in 1900:
Source: Benavot and Riddle (1988).
R& D expenditures as a fraction of GDP, 2000:
Source: World Development Indicators, available at http://data.worldbank.org/
Religion: Measures of fraction of population muslim, catholic and protestant, respectively. All
measured in 1980.
Source: Treisman (2007)
Schooling in 2000 (average years):
Source: Barro and Lee (2001).
Year of fertility decline, YFD:
Source: Reher (2004).
Year of fertility decline, YFD2:
Source: Caldwell and Caldwell (2001). In cases where Caldwell and Caldwell report an interval
(say, 1990–1995) we date the decline to the middle of the interval. All observations without precise
indications of the decline (reported in Caldwell and Caldwell as “ transition before year X”) are
coded as missing observation.
Year of Industrialization, YIT, which measures the first year that industry accounts for a
greater proportion of total employment than agriculture.
Source: Bentzen et al. (2013).

27
References
Acemoglu, D., 2009. Introduction to modern economic growth. Princeton University Press.

Acemoglu, D. and J. Ventura, 2003. The world income distribution, Quarterly Journal of Economics
117, 659-694.

Acemoglu, D., S. Johnson, and J.A. Robinson, 2001. The colonial origins of comparative develop-
ment: An empirical investigation, American Economic Review 91, 1369–1401.

Acemoglu, D., S. Johnson, and J.A. Robinson, 2002. Reversal of fortune: geography and institutions
in the making of the modern world income distribution, Quarterly Journal of Economics 117,
1231–1294.

Aghion, P., and P. Howitt, 1992. A model of growth through creative destruction, Econometrica 60,
323–351.

Aghion, P. and P. Howitt, 2008. The Economics of Growth, MIT Press.

Andersen, T.B., C-J. Dalgaard and P. Selaya, 2010. Eye disease and development. Mimeo (Univer-
sity of Copenhagen).

Ashraf, Q, and O. Galor, 2011. Dynamics and stagnation in the Malthusian epoch, American
Economic Review 101, 2003-2041.

Basso A. and D. Cuberes, 2011. Institutions, culture and the onset of the demographic transition.
Mimeo (University of Sheffield).

Barro, R.J. and J.W. Lee, 2001. International data on educational attainment: updates and impli-
cations, Oxford Economic Papers 53, 541-563.

Barro, R. and R.M. McCleary, 2003. Religion and economic growth across countries, American
Sociological Review 68, 760–781

Benavot, A. and P. Riddle, 1988. The expansion of primary education, 1870-1940: Trends and
Issues, Sociology of Education 61, 191–210.

Benhabib, J. and M.M. Spiegel, 2005. Human capital and technology diffusion, in (P. Aghion and
S. Durlauf eds.) Handbook of Economic Growth, Amsterdam: North Holland, 935–966.

Bentzen, J., N. Kaarsen and A. M. Wingender, 2013. Industrialization: When and How Across 114
Countries. Mimeo (University of Copenhagen).

Bloom, D. E and J.G. Williamson, 1998. Demographic transitions and economic miracles in emerging
Asia, World Bank Economic Review, 12, 419–455.

Caldwell, J and P. Caldwell, 2001. Regional paths to fertility transition, Journal of Population
Research 18, pp. 91–117.

Caselli, F., 2005. Accounting for cross-country income differences. In: Aghion P. and S. Durlauf
(eds.) Handbook of Economic Growth.

28
Comin, D., B. Hobijn and E. Rovito, 2008. Technology usage lags, Journal of Economic Growth,
13, 237–256.

Comin, D., W. Easterly and E. Gong, 2010. Was the wealth of nations determined in 1000 B.C.?
American Economic Journal: Macroeconomics 2, 65-97.

Dalgaard, C-J and H. Strulik, 2013. The Physiological Foundations of the Wealth of Nations. Mimeo
(University of Copenhagen).

Devleeschauwer, A., R. Wacziarg, S. Kurlat, W. Easterly, and A. Alesina, Alberto, 2003. Fraction-
alization, Journal of Economic Growth 8, 155–194.

Diamond, J., 1997. Guns, Germs and Steel: The Fates of Human Societies, Penguin.

Eicher, T., C. Garcia-Pealosa and U. Teksoz, 2006. How do institutions lead some countries To
produce so much more output per worker than others?, in Eicher and Garcia-Pealosa Institutions,
Development and Economic Growth, MIT Press.

Galor, O., 2010a. The 2008 Lawrence R. Klein Lecture – Comparative economic development:
insights from unified growth theory, International Economic Review 51, 1–44.

Galor, O., 2010b. The demographic transition: causes and consequences, Working Paper (Brown
University).

Galor, O., 2011. Unified Growth Theory. Princeton University Press.

Galor, O. and D. Weil, 2000, Population, technology and growth: from Malthusian stagnation to
the demographic transition and beyond, American Economic Review 90, 806-828.

Galor, O. and O. Moav, 2002. Natural selection and the origin of economic growth, Quarterly
Journal of Economics 117, 1133–1191.

Gallup, J.L. and J. Sachs, 2000. Agriculture, climate, and technology: why are the tropics falling
behind? American Journal of Agricultural Economics 82, 731–737.

Gallup, J.L, J. Sachs, and A. Mellinger. 1999. Geography and Economic Development, CID Harvard
University Working Paper No. 1.

Glaeser, E., R. La Porta, F. Lopez-de-Silanes and A. Schleifer, 2004. Do institutions cause growth?
Journal of Economic Growth 9, 271-303.

Glaeser, E. and R. Saks, 2006. Corruption in America, Journal of Public Economics, 90, 1053-1072.

Glaeser, E. and A. Shleifer, 2002. Legal origins, Quarterly Journal of Economics 117, 1193–1229.

Hall, R. and C.I Jones, 1999. Why do some countries produce so much more output per worker than
others? Quarterly Journal of Economics 114, 83–116.

Hansen, G. and E. Prescott. 2002. Malthus to Solow. American Economic Review 92, 1205-17.

Heston, A. R. Summers and B. Aten, 2009. Penn World Table Version 6.3. Center for International
Comparisons of Production, Income and Prices at the University of Pennsylvania.

29
Howitt, P., 2000. Endogenous growth and cross-country income differences, American Economic
Review 90, 829–846.
Jones, G. and W.J. Schneider, 2006. Intelligence, human capital, and economic growth: A bayesian
averaging of classical estimates (BACE) approach, Journal of Economic Growth 11, 71–93.
Knowles, S. and P.D. Owen, 1995. Health capital and cross-country variation in income per capita
in the Mankiw–Romer–Weil model, Economics Letters 48, 99–106.
Laitner, J., 2000. Structural change and economic growth. Review of Economic Studies, 67, 545–561.
Lewis, W.A., 1954. Economic development with unlimited supplies of labour, Manchester School of
Economic and Social Studies, 22, 139-191.
Lucas, R. E. Jr., 2002, “The industrial revolution: past and future”, in: Lucas, R. E. Jr., Lectures
on Economic Growth, Cambridge, Massachusetts: Harvard University. Press.
Mankiw .N.G., D. Romer, and D.N. Weil, 1992. A contribution to the empirics of economic growth,
Quarterly Journal of Economics 107 , 407–437.
McEvedy, C. and R. Jones, 1978. Atlas of World Population History. Penguin Books Ltd., New
York, NY.
Michalopoulos, S., 2012. The origins of ethnolinguistic diversity, American Economic Review 102,
1508-1539.
Morrison, C. F. Murtin, 2009. The century of education. Journal of Human Capital, 3, 1–42.
Murtin, F., 2012. Long-term determinants of the demographic transition: 1870-2000, forthcoming:
Review of Economics and Statistics.
Nelson, R.R. and E.S. Phelps, 1966. Investment in humans, technological diffusion, and economic
growth, American Economic Review 56, 69-75.
Nunneman W. and P. Vanhout, 1996. A further augmentation of the Solow model and the empirics
of economic growth for the OECD. Quarterly Journal of Economics 111, 943-53.
Phillips, P.C.B. and D. Sul, 2009. Economic transition and growth, Journal of Applied Econometrics
24, 1153–1185.
Ram, R., 2007. IQ and economic growth: further augmentation of Mankiw-Romer-Weil model,
Economics Letters 94, 7–11.
Rappaport, J. and J. Sachs, 2003. The United States as a coastal nation, Journal of Economic
growth,8, 5–46.
Reher, D. S., 2004. The demographic transition revisited as a global process, Population Space and
Place 10, 19–42.
Romer, P., 1990. Endogenous technological change, Journal of political Economy 98, 71–102.
Solow, R.M, 1956. A contribution to the theory of economic growth, Quarterly Journal of Economics
70, 65–94.

30
Spolaore, E. and R. Wacziarg, 2009. Diffusion of development, Quarterly Journal of Economics
124, 469–529.
Strulik, H., 2008. Geography, health, and the pace of demo-economic development, Journal of
Development Economics 86, 61–75.
Strulik, H., 2010. A note on economic growth with subsistence consumption, Macroeconomic Dy-
namics 14, 763–771.
Strulik, H., 2012. Patience and prosperity, Journal of Economic Theory 147, 336–352.
Strulik, H. and J. Weisdorf, 2008. Population, food, and knowledge: a simple unified growth theory,
Journal of Economic Growth 13, 195–216.
Treisman, D., 2007, What have we learned about the causes of corruption from ten years of cross-
national empirical research? Annual Review of Political Science, 2007.
Zellner, A., 1969. On the aggregation problem: A new approach to a troublesome problem. In
M. Beckmann and H.P. Künzi (eds.) : Economic Models, Estimation and Risk Programming —
Essays in honor of Gerhard Tintner. Springer, Berlin.

31
Bisher erschienene Diskussionspapiere

Nr. 151: Dalgaard, Carl-Johan; Strulik, Holger: The History Augmented Solow Model, March 2013
Nr. 150: Strulik, Holger; Trimborn, Timo: The Dark Side of Fiscal Stimulus, Januar 2013
Nr. 149: Prettner, Klaus: Public education, technological change and economic prosperity, Januar
2013
Nr. 148: Lankau, Matthias; Bicskei, Marianna; Bizer, Kilian: Cooperation Preferences in the
Provision of Public Goods: An Experimental Study on the Effects of Social Identity,
Dezember 2012
Nr. 147: Krenz, Astrid: Modeling Services Sectors' Agglomeration within a New Economic
Geography Model, Dezember 2012
Nr. 146: Krenz, Astrid: A Panel Co-integration Analysis of Industrial and Services Sectors'
Agglomeration in the European Union, Dezember 2012
Nr. 145: Strulik, Holger: Knowledge and Growth in the Very Long Run, November 2012
Nr. 144: Baskaran, Thushyanthan: Ideology and fiscal policy: quasi-experimental evidence from the
German States, Oktober 2012
Nr. 143: Ehlers, Tim; Schwager, Robert: Honest Grading, Grade Inflation and Reputation, Oktober
2012
Nr. 142: Gehringer, Agnieszka: Another look at the determinants of current account imbalances in
the European Union: An empirical assessment, Oktober 2012
Nr. 141: Strulik, Holger; Werner, Katharina: Life Expectancy, Labor Supply, and Long-Run Growth:
Reconciling Theory and Evidence, September 2012
Nr. 140: Strulik, Holger; Prettner, Klaus; Prskawetz, Alexia: The Past and Future of Knowledge-
based Growth, September 2012
Nr. 139: Prettner, Klaus; Trimborn, Timo: Demographic change and R&D-based economic growth:
reconciling theory and evidence, September 2012
Nr. 138: König, Jörg; Ohr, Renate: Homogeneous groups within a heterogeneous community -
Evidence from an index measuring European economic integration, August 2012
Nr. 137: Schwager, Robert: Student Loans in a Tiebout Model of Higher Education, Juli 2012
Nr. 136: Martínez-Zarzoso, Inmaculada: Exporting and Productivity: Evidence for Egypt and
Morocco, April 2012
Nr. 135: König, Jörg; Ohr, Renate: Messung ökonomischer Integration in der Europäischen Union –
Entwicklung eines EU-Integrationsindexes -, April 2012
Nr. 134: Gehringer, Agnieszka: Financial liberalization, growth, productivity and capital
accumulation: The case of European integration, März 2012
Nr. 133: Berner, Eike; Birg, Laura: Retailers and Consumers. The pass-through of import price
changes, März 2012
Nr. 132: Gehringer, Angnieszka: Current accounts in Europe: implications of the external imbalances
for the future of the common monetary policy, März 2012
Nr. 131: Ohr, Renate; Özalbayrak, Mehmet: The Euro – A „MUST“ for Small European States?,
Januar 2012
Nr. 130: Zeddies, Götz: Der Euro als Triebfeder des deutschen Exports?, November 2011
Nr. 129: Geishecker, Ingo; Siedler, Thomas: Job Loss Fears and (Extreme) Party Identification: First
Evidence from Panel Data, Oktober 2011
Nr. 128: König, Jörg; Ohr, Renate: Small but Beautiful? Economic Impacts of the Size of Nations in
the European Union, August 2011
Nr. 128: König, Jörg; Ohr, Renate: Small but Beautiful? Economic Impacts of the Size of Nations in
the European Union, August 2011
Nr. 127: Schüder, Stefan: Monetary Policy Trade-Offs in a Portfolio Model with Endogenous Asset
Supply, Juni 2011
Nr. 126: Hiller, Sanne: The Export Promoting Effect of Emigration: Evidence from Denmark, Juni
2011
Nr. 125: Martínez-Zarzoso, Inmaculada; Voicu, Anca M.; Vidovic, Martina: CEECs Integration into
Regional and Global Production Networks, Mai 2011
Nr. 124: Roth, Felix; Gros, Daniel; Nowak-Lehmann D., Felicitas: Has the Financial Crisis eroded
Citizens‟ Trust in the European Central Bank? Panel Data Evidence for the Euro Area,
1999-2011, Mai 2011, Revised Version März 2012
Nr. 123 Dreher, Axel; Vreeland, James Raymond : Buying Votes and International Organizations,
Mai 2011
Nr. 122: Schürenberg-Frosch, Hannah: One Model fits all? Determinants of Transport Costs across
Sectors and Country Groups, April 2011
Nr. 121: Verheyen, Florian: Bilateral Exports from Euro Zone Countries to the US - Does Exchange
Rate Variability Play a Role?, April 2011
Nr. 120: Ehlers, Tim: University Graduation Dependent on Family‟s Wealth, Ability and Social
Status, April 2011
Nr. 119: Cho, Seo-Young; Dreher, Axel; Neumayer, Eric: The Spread of Anti-trafficking Policies –
Evidence from a New Index, März 2011
Nr. 118: Cho, Seo-Young; Vadlamannati, Krishna Chaitanya: Compliance for Big Brothers: An
Empirical Analysis on the Impact of the Anti-trafficking Protocol, Februar 2011
Nr. 117: Nunnenkamp, Peter; Öhler, Hannes: Donations to US based NGOs in International
Development Cooperation: How (Un-)Informed Are Private Donors?, Februar 2011
Nr. 116: Geishecker, Ingo; Riedl, Maximilian: Ordered Response Models and Non-Random
Personality Traits: Monte Carlo Simulations and a Practical Guide, Revised Version Februar
2012
Nr. 115: Dreher, Axel; Gassebner, Martin; Siemers, Lars-H. R.: Globalization, Economic Freedom
and Human Rights, Oktober 2010
Nr. 114: Dreher, Axel; Mikosch, Heiner; Voigt, Stefan: Membership has its Privileges – The Effect
of Membership in International Organizations on FDI, Oktober 2010
Nr. 113: Fuchs, Andreas; Klann, Nils-Hendrik: Paying a Visit: The Dalai Lama Effect on
International Trade, Oktober 2010
Nr. 112: Freitag, Stephan: Choosing an Anchor Currency for the Pacific, Oktober 2010
Nr. 111: Nunnenkamp, Peter; Öhler, Hannes: Throwing Foreign Aid at HIV/AIDS in Developing
Countries: Missing the Target?, August 2010
Nr. 110: Ohr, Renate; Zeddies, Götz: „Geschäftsmodell Deutschland“ und außenwirtschaftliche
Ungleichgewichte in der EU, Juli 2010
Nr. 109: Nunnenkamp, Peter; Öhler, Hannes: Funding, Competition and the Efficiency of NGOs:
An Empirical Analysis of Non-charitable Expenditure of US NGOs Engaged in Foreign
Aid, Juli 2010
Nr. 108: Krenz, Astrid: La Distinction reloaded: Returns to Education, Family Background, Cultural
and Social Capital in Germany, Juli 2010
Nr. 107: Krenz, Astrid: Services sectors' agglomeration and its interdependence with industrial
agglomeration in the European Union, Juli 2010
Nr. 106: Krenz, Astrid; Rübel, Gerhard: Industrial Localization and Countries' Specialization in the
European Union: An Empirical Investigation, Juli 2010
Nr. 105: Schinke, Jan Christian: Follow the Sun! How investments in solar power plants in Sicily can
generate high returns of investments and help to prevent global warming, Juni 2010
Nr. 104: Dreher, Axel; Sturm, Jan-Egbert; Vreeland, James Raymon: Does membership on the
Security Council influence IMF conditionality?, Juni 2010
Nr. 103: Öhler, Hannes; Nunnenkamp, Peter; Dreher, Axel: Does Conditionality Work? A Test for an
Innovative US Aid Scheme, Juni 2010
Nr. 102: Gehringer, Agnieszka: Pecuniary Knowledge Externalities in a New Taxonomy: Knowledge
Interactions in a Vertically Integrated System, Juni 2010
Nr. 101: Gehringer, Agnieszka: Pecuniary Knowledge Externalities across European Countries – are
there leading Sectors?, Juni 2010
Nr. 100: Gehringer, Agnieszka: Pecuniary Knowledge Externalities and Innovation:
Intersectoral Linkages and their Effects beyond Technological Spillovers, Juni 2010
Nr. 99: Dreher, Axel; Nunnenkamp, Peter; Öhler, Hannes: Why it pays for aid recipients to take
note of the Millennium Challenge Corporation: Other donors do!, April 2010
Nr. 98: Baumgarten, Daniel; Geishecker, Ingo; Görg, Holger: Offshoring, tasks, and the skill-wage
pattern, März 2010
Nr. 97: Dreher, Axel; Klasen, Stephan; Raymond, James; Werker, Eric: The costs of favoritism: Is
politically-driven aid less effective?, März 2010
Nr. 96: Dreher, Axel; Nunnenkamp, Peter; Thiele, Rainer: Are „New‟ Donors Different? Comparing
the Allocation of Bilateral Aid between Non-DAC and DAC Donor Countries, März 2010
Nr. 95: Lurweg, Maren; Westermeier, Andreas: Jobs Gained and Lost through Trade – The Case of
Germany, März 2010
Nr. 94: Bernauer, Thomas; Kalbhenn, Anna; Koubi, Vally; Ruoff, Gabi: On Commitment Levels
and Compliance Mechanisms – Determinants of Participation in Global Environmental
Agreements, Januar 2010
Nr. 93: Cho, Seo-Young: International Human Rights Treaty to Change Social Patterns – The
Convention on the Elimination of All Forms of Discrimination against Women, Januar 2010
Nr. 92: Dreher, Axel; Nunnenkamp, Peter; Thiel, Susann; Thiele, Rainer: Aid Allocation by
German NGOs: Does the Degree of Public Refinancing Matter?, Januar 2010
Nr. 91: Bjørnskov, Christian; Dreher, Axel; Fischer, Justina A. V.; Schnellenbach, Jan: On the
relation between income inequality and happiness: Do fairness perceptions matter?,
Dezember 2009
Nr. 90: Geishecker, Ingo: Perceived Job Insecurity and Well-Being Revisited: Towards Conceptual
Clarity, Dezember 2009
Nr. 89: Kühl, Michael: Excess Comovements between the Euro/US dollar and British pound/US
dollar exchange rates, November 2009
Nr. 88: Mourmouras, Alex, Russel, Steven H.: Financial Crises, Capital Liquidation and the
Demand for International Reserves, November 2009
Nr. 87: Goerke, Laszlo, Pannenberg, Markus: An Analysis of Dismissal Legislation: Determinants
of Severance Pay in West Germany, November 2009
Nr. 86: Marchesi, Silvia, Sabani, Laura, Dreher, Axel: Read my lips: the role of information
transmission in multilateral reform design, Juni 2009
Nr. 85: Heinig, Hans Michael: Sind Referenden eine Antwort auf das Demokratiedilemma der EU?,
Juni 2009
Nr. 84: El-Shagi, Makram: The Impact of Fixed Exchange Rates on Fiscal Discipline, Juni 2009
Nr. 83: Schneider, Friedrich: Is a Federal European Constitution for an Enlarged European Union
Necessary? Some Preliminary Suggestions using Public Choice Analysis, Mai 2009
Nr. 82: Vaubel, Roland: Nie sollst Du mich befragen? Weshalb Referenden in bestimmten
Politikbereichen – auch in der Europapolitik – möglich sein sollten, Mai 2009
Nr. 81: Williamson, Jeffrey G.: History without Evidence: Latin American Inequality since 1491,
Mai 2009
Nr. 80: Erdogan, Burcu: How does the European Integration affect the European Stock Markets?,
April 2009
Nr. 79: Oelgemöller, Jens; Westermeier, Andreas: RCAs within Western Europe, März 2009
Nr. 78: Blonski, Matthias; Lilienfeld-Toal, Ulf von: Excess Returns and the Distinguished Player
Paradox, Oktober 2008
Nr. 77: Lechner, Susanne; Ohr, Renate: The Right of Withdrawal in the Treaty of Lisbon: A game
theoretic reflection on different decision processes in the EU, Oktober 2008
Nr. 76: Kühl, Michael: Strong comovements of exchange rates: Theoretical and empirical cases
when currencies become the same asset, Juli 2008
Nr. 75: Höhenberger, Nicole; Schmiedeberg, Claudia: Structural Convergence of European
Countries, Juli 2008
Nr. 74: Nowak-Lehmann D., Felicitas; Vollmer, Sebastian; Martinez-Zarzoso, Inmaculada: Does
Comparative Advantage Make Countries Competitive? A Comparison of China and Mexico,
Juli 2008
Nr. 73: Fendel, Ralf; Lis, Eliza M.; Rülke, Jan-Christoph: Does the Financial Market Believe in the
Phillips Curve? – Evidence from the G7 countries, Mai 2008
Nr. 72: Hafner, Kurt A.: Agglomeration Economies and Clustering – Evidence from German Firms,
Mai 2008
Nr. 71: Pegels, Anna: Die Rolle des Humankapitals bei der Technologieübertragung in
Entwicklungsländer, April 2008
Nr. 70: Grimm, Michael; Klasen, Stephan: Geography vs. Institutions at the Village Level, Februar
2008
Nr. 69: Van der Berg, Servaas: How effective are poor schools? Poverty and educational outcomes
in South Africa, Januar 2008
Nr. 68: Kühl, Michael: Cointegration in the Foreign Exchange Market and Market Efficiency since
the Introduction of the Euro: Evidence based on bivariate Cointegration Analyses, Oktober
2007
Nr. 67: Hess, Sebastian; Cramon-Taubadel, Stephan von: Assessing General and Partial Equilibrium
Simulations of Doha Round Outcomes using Meta-Analysis, August 2007
Nr. 66: Eckel, Carsten: International Trade and Retailing: Diversity versus Accessibility and the
Creation of “Retail Deserts”, August 2007
Nr. 65: Stoschek, Barbara: The Political Economy of Enviromental Regulations and Industry
Compensation, Juni 2007
Nr. 64: Martinez-Zarzoso, Inmaculada; Nowak-Lehmann D., Felicitas; Vollmer, Sebastian: The Log
of Gravity Revisited, Juni 2007
Nr. 63: Gundel, Sebastian: Declining Export Prices due to Increased Competition from NIC –
Evidence from Germany and the CEEC, April 2007
Nr. 62: Wilckens, Sebastian: Should WTO Dispute Settlement Be Subsidized?, April 2007
Nr. 61: Schöller, Deborah: Service Offshoring: A Challenge for Employment? Evidence from
Germany, April 2007
Nr. 60: Janeba, Eckhard: Exports, Unemployment and the Welfare State, März 2007
Nr. 59: Lambsdoff, Johann Graf; Nell, Mathias: Fighting Corruption with Asymmetric Penalties and
Leniency, Februar 2007
Nr. 58: Köller, Mareike: Unterschiedliche Direktinvestitionen in Irland – Eine theoriegestützte
Analyse, August 2006
Nr. 57: Entorf, Horst; Lauk, Martina: Peer Effects, Social Multipliers and Migrants at School: An
International Comparison, März 2007 (revidierte Fassung von Juli 2006)
Nr. 56: Görlich, Dennis; Trebesch, Christoph: Mass Migration and Seasonality Evidence on
Moldova‟s Labour Exodus, Mai 2006
Nr. 55: Brandmeier, Michael: Reasons for Real Appreciation in Central Europe, Mai 2006
Nr. 54: Martínez-Zarzoso, Inmaculada; Nowak-Lehmann D., Felicitas: Is Distance a Good Proxy
for Transport Costs? The Case of Competing Transport Modes, Mai 2006
Nr. 53: Ahrens, Joachim; Ohr, Renate; Zeddies, Götz: Enhanced Cooperation in an Enlarged EU,
April 2006
Nr. 52: Stöwhase, Sven: Discrete Investment and Tax Competition when Firms shift Profits, April
2006
Nr. 51: Pelzer, Gesa: Darstellung der Beschäftigungseffekte von Exporten anhand einer Input-
Output-Analyse, April 2006
Nr. 50: Elschner, Christina; Schwager, Robert: A Simulation Method to Measure the Tax Burden on
Highly Skilled Manpower, März 2006
Nr. 49: Gaertner, Wulf; Xu, Yongsheng: A New Measure of the Standard of Living Based on
Functionings, Oktober 2005
Nr. 48: Rincke, Johannes; Schwager, Robert: Skills, Social Mobility, and the Support for the
Welfare State, September 2005
Nr. 47: Bose, Niloy; Neumann, Rebecca: Explaining the Trend and the Diversity in the Evolution of
the Stock Market, Juli 2005
Nr. 46: Kleinert, Jörn; Toubal, Farid: Gravity for FDI, Juni 2005
Nr. 45: Eckel, Carsten: International Trade, Flexible Manufacturing and Outsourcing, Mai 2005
Nr. 44: Hafner, Kurt A.: International Patent Pattern and Technology Diffusion, Mai 2005
Nr. 43: Nowak-Lehmann D., Felicitas; Herzer, Dierk; Martínez-Zarzoso, Inmaculada; Vollmer,
Sebastian: Turkey and the Ankara Treaty of 1963: What can Trade Integration Do for
Turkish Exports, Mai 2005
Nr. 42: Südekum, Jens: Does the Home Market Effect Arise in a Three-Country Model?, April 2005
Nr. 41: Carlberg, Michael: International Monetary Policy Coordination, April 2005
Nr. 40: Herzog, Bodo: Why do bigger countries have more problems with the Stability and Growth
Pact?, April 2005
Nr. 39: Marouani, Mohamed A.: The Impact of the Mulitfiber Agreement Phaseout on
Unemployment in Tunisia: a Prospective Dynamic Analysis, Januar 2005
Nr. 38: Bauer, Philipp; Riphahn, Regina T.: Heterogeneity in the Intergenerational Transmission of
Educational Attainment: Evidence from Switzerland on Natives and Second Generation
Immigrants, Januar 2005
Nr. 37: Büttner, Thiess: The Incentive Effect of Fiscal Equalization Transfers on Tax Policy, Januar
2005
Nr. 36: Feuerstein, Switgard; Grimm, Oliver: On the Credibility of Currency Boards, Oktober 2004
Nr. 35: Michaelis, Jochen; Minich, Heike: Inflationsdifferenzen im Euroraum – eine
Bestandsaufnahme, Oktober 2004
Nr. 34: Neary, J. Peter: Cross-Border Mergers as Instruments of Comparative Advantage, Juli 2004
Nr. 33: Bjorvatn, Kjetil; Cappelen, Alexander W.: Globalisation, inequality and redistribution, Juli
2004
Nr. 32: Stremmel, Dennis: Geistige Eigentumsrechte im Welthandel: Stellt das TRIPs-Abkommen
ein Protektionsinstrument der Industrieländer dar?, Juli 2004
Nr. 31: Hafner, Kurt: Industrial Agglomeration and Economic Development, Juni 2004
Nr. 30: Martinez-Zarzoso, Inmaculada; Nowak-Lehmann D., Felicitas: MERCOSUR-European
Union Trade: How Important is EU Trade Liberalisation for MERCOSUR‟s Exports?, Juni
2004
Nr. 29: Birk, Angela; Michaelis, Jochen: Employment- and Growth Effects of Tax Reforms, Juni
2004
Nr. 28: Broll, Udo; Hansen, Sabine: Labour Demand and Exchange Rate Volatility, Juni 2004
Nr. 27: Bofinger, Peter; Mayer, Eric: Monetary and Fiscal Policy Interaction in the Euro Area with
different assumptions on the Phillips curve, Juni 2004
Nr. 26: Torlak, Elvisa: Foreign Direct Investment, Technology Transfer and Productivity Growth in
Transition Countries, Juni 2004
Nr. 25: Lorz, Oliver; Willmann, Gerald: On the Endogenous Allocation of Decision Powers in
Federal Structures, Juni 2004
Nr. 24: Felbermayr, Gabriel J.: Specialization on a Technologically Stagnant Sector Need Not Be
Bad for Growth, Juni 2004
Nr. 23: Carlberg, Michael: Monetary and Fiscal Policy Interactions in the Euro Area, Juni 2004
Nr. 22: Stähler, Frank: Market Entry and Foreign Direct Investment, Januar 2004
Nr. 21: Bester, Helmut; Konrad, Kai A.: Easy Targets and the Timing of Conflict, Dezember 2003
Nr. 20: Eckel, Carsten: Does globalization lead to specialization, November 2003
Nr. 19: Ohr, Renate; Schmidt, André: Der Stabilitäts- und Wachstumspakt im Zielkonflikt zwischen
fiskalischer Flexibilität und Glaubwürdigkeit: Ein Reform-ansatz unter Berücksichtigung
konstitutionen- und institutionenökonomischer Aspekte, August 2003
Nr. 18: Ruehmann, Peter: Der deutsche Arbeitsmarkt: Fehlentwicklungen, Ursachen und
Reformansätze, August 2003
Nr. 17: Suedekum, Jens: Subsidizing Education in the Economic Periphery: Another Pitfall of
Regional Policies?, Januar 2003
Nr. 16: Graf Lambsdorff, Johann; Schinke, Michael: Non-Benevolent Central Banks, Dezember
2002
Nr. 15: Ziltener, Patrick: Wirtschaftliche Effekte des EU-Binnenmarktprogramms, November 2002
Nr. 14: Haufler, Andreas; Wooton, Ian: Regional Tax Coordination and Foreign Direct Investment,
November 2001
Nr. 13: Schmidt, André: Non-Competition Factors in the European Competition Policy: The
Necessity of Institutional Reforms, August 2001
Nr. 12: Lewis, Mervyn K.: Risk Management in Public Private Partnerships, Juni 2001
Nr. 11: Haaland, Jan I.; Wooton, Ian: Multinational Firms: Easy Come, Easy Go?, Mai 2001
Nr. 10: Wilkens, Ingrid: Flexibilisierung der Arbeit in den Niederlanden: Die Entwicklung
atypischer Beschäftigung unter Berücksichtigung der Frauenerwerbstätigkeit, Januar 2001
Nr. 9: Graf Lambsdorff, Johann: How Corruption in Government Affects Public Welfare – A
Review of Theories, Januar 2001
Nr. 8: Angermüller, Niels-Olaf: Währungskrisenmodelle aus neuerer Sicht, Oktober 2000
Nr. 7: Nowak-Lehmann, Felicitas: Was there Endogenous Growth in Chile (1960-1998)? A Test of
the AK model, Oktober 2000
Nr. 6: Lunn, John; Steen, Todd P.: The Heterogeneity of Self-Employment: The Example of Asians
in the United States, Juli 2000
Nr. 5: Güßefeldt, Jörg; Streit, Clemens: Disparitäten regionalwirtschaftlicher Entwicklung in der
EU, Mai 2000
Nr. 4: Haufler, Andreas: Corporate Taxation, Profit Shifting, and the Efficiency of Public Input
Provision, 1999
Nr. 3: Rühmann, Peter: European Monetary Union and National Labour Markets,
September 1999
Nr. 2: Jarchow, Hans-Joachim: Eine offene Volkswirtschaft unter Berücksichtigung des
Aktienmarktes, 1999
Nr. 1: Padoa-Schioppa, Tommaso: Reflections on the Globalization and the Europeanization of the
Economy, Juni 1999
Alle bisher erschienenen Diskussionspapiere zum Download finden Sie im Internet unter:
http://www.uni-goettingen.de/de/60920.html.