Chapter 17-Income Redistribution

and Welfare Programs

Binzhen Wu
Tsinghua University
 ■ Income
Redistribution

Outline

Why intervene?
Welfare program and its negative effects
• Moral hazard costs
 Work disincentives; reducing saving

How to overcome the negative effects: alternative

options
Cash transfer vs. in-kind transfer programs

Measures and reasons for Income Inequality and

Poverty
2
 ■ Income
Redistribution

Rationales for Income Redistribution and

Welfare Programs
Fairness
Social welfare may be maximized by redistributing from
high-income individuals to low-income individuals.
 SWF=σ𝑖 𝑤𝑖 ∗ 𝑈𝑖

Sometimes redistribution can improve efficiency

• E.g. people value insurance for being poor (Self-interest); altruism of
richer person;
Other concerns:
• Economic growth; Social Stability; Social Mobility; Commodity
egalitarianism
• Behavior effect: income inequality and saving (Jin, Li, Wu, 2012)

3
 ■ Income
Redistribution

5
 ■ Income
Redistribution

8
 ■ Income
Redistribution

Welfare Program

How to help the poor?

• It is often difficult for the private market to do redistribution.
We need government
How does government solve this problem?
• Typically: taxation + well-known projects: cash transfers and
in-kind transfers

Questions:
• Help whom?
• Help on what?
• Help how much?
• What are the effects?
9
 ■ Income

投票 最多可选5项
Redistribution

Before we dive in, should we know why they are

poor and what they need?
Why the poor is poor?

A Their parents are poor

B
They face too much risks that drive
them into the poverty trap
C They have low ability
D They do not work/study hard enough
E
They are impatient, cannot save or
work hard for the future
Submit
10
 ■ Income
Redistribution

11
 ■ Income
Redistribution

WELFARE PROGRAM AND ITS

EFFECTS

12
 ■ Income
Redistribution

Welfare Programs
Welfare programs try to help people with low
earning capability.
They can be categorical or means-tested.
Help whom？
• Categorical welfare: Welfare programs restricted by some
demographic characteristic, such as single motherhood,
disability, and age.
• Means-tested welfare: only individuals with sufficiently
“poor” are eligible.
 How to measure “whether poor enough”?
• Income or asset levels
13
 ■ Income
Redistribution

Welfare Programs
Help on what?
A patchwork of dozens of different programs. They can
be cash or in-kind.
• Cash welfare: Welfare programs that provide cash
benefits to recipients.
o US: EITC (Earned Income Tax Credit); TANF (Temporary Assistance to
Needy Families); SSI (Supplemental Security Income);
o China: Subsistence allowance (Dibao, Tekun in rural)

• In-kind welfare: Welfare programs that deliver

goods, such as medical care or housing, to recipients.
o US: Food stamps; Medicaid; Public Housing; Energy aid; Nutritional
programs; Child Care
o China: Infrastructure, Wubao (Five guarantees), health insurance for
low-income families
14
 ■ Income
Redistribution

The structure of anti-poverty expenditure

Total (100 mil) % of gov’t exp % of GDP agri (%) relief(%) fin aid(%) health(%) housing(%)

15
 ■ Income
Redistribution

The structure of agricultural related anti-poverty expenditure

Total % of gov exp infra(%) production(%) other (%) social (%) loans (%) manage(%)

16
 ■ Income
Redistribution

• Cash versus in-kind assistance in the US

Federal Expenditures on Major Need-Tested Programs

Program Federal Expenditures ($)
Health 319.3
Cash aid 129.6
Food assistance 77.5
Housing and development 59.9
Education 58.2
Social Services 44.3
Energy assistance 10.3
Employment and training 8.6

13-17
17
 ■ Income
Redistribution

Problems of Welfare Programs

A Leaky Bucket
• Administrative costs
 2019, expenditure on anti-poverty is over 914 billion RMB. We have
16.6 million people in poverty in 2018. Therefore, about 5500 a
person, much higher than the poverty line (2300 RMB).
• Taxation on high-income individuals may affect their labor
supply and savings.
 Efficiency loss
• Moral Hazard cost: by insuring against being poor, the
programs create an incentive for individuals to change
behavior to qualify for the transfers.
 Crowding-out
 Efficiency loss
Redistribution often involves efficiency losses.
18
 ■ Income
Redistribution

Moral Hazard Costs

Means-tested based on income/earning 

reducing working hours
• e.g. The TANF gives work disincentives.
Means-tested based on assets  reducing
saving
Only single mother eligible for some
welfare programs:  being single
Has cash welfare played a constructive or a destructive
role in the lives of lower-income groups? 19
 ■ Income
Redistribution

Moral Hazard of Cash Transfer Programs –

Basic Structure
Two parameters:
• The benefit guarantee, G, is the cash welfare benefit for
individuals with no other income.
• The actual benefit is reduced as income increases.
 The benefit reduction rate, t, is the rate at which benefits are reduced
per dollar of other income earned.
• This is essentially a tax rate on earnings, working disincentive.

Actual benefits received: B=G-t*E=G-t*w*h

“Break-even level” of earning (eligibility limit): the
earning level makes actual benefits equal to zero
• B=0  E*=G/t , h*=G/tw

20
 ■ Income
Redistribution

The cash transfers: work disincentives

The government’s welfare program changes

individuals’ budget constraint, and hence their
behavior.
• Max U(L,C) s.t. C = wh+B
 L: Leisure; h=T-L; C: spending on all other goods
 B=G-twh if Gtwh; B=0 if G<twh
 Budget constraint after considering welfare benefits
C=G+(1-t)wh if G  twh
C=wh if G<twh
Graphically, draw budget constraint with and
without the program!
21
Moral Hazard Effects of a Means-Tested Transfer System 2
Figure 17-4

100% Benefit Reduction Rate:

All families with income below the poverty line (X) and many individuals with income
above the poverty line immediately stop earning income so they can get more leisure
and consumption (Y).
 ■ Income
Redistribution

The cash transfer programs: work disincentives

The responses depend on individuals’ indifference

curve.
• X: mechanically eligible for welfare
• Better off
• Y: originally not eligible for welfare, and D is better
than original choice
 Better off: give up some consumption, but get much more leisure
• Z: originally not eligible for welfare and D is no
better than original choice
• No change

23
Solving Moral Hazard by Lowering the Benefit Reduction
Rate
Figure 17-5

50% Benefit Reduction Rate:

The net impact of new rate on labor supply is ambiguous and depends on
relative sizes and preferences of worker groups.
 ■ Income
Redistribution

Iron Triangle

Unintended effects: Eliminating poverty is potentially

much more expensive than in the case where people
do not change behavior.
The iron triangle means that there is no way to change
either the benefit reduction rate or benefit guarantee to
simultaneously encourage work(have desirable
incentives), redistribute more income (be
compassionate), and lower costs.
• If the tax rate t is lowered, work could be discouraged for some
and costs could go up.
• If the guarantee G is lowered, work increases and costs fall, but
redistribution falls.
25
 ■ Income
Redistribution

Empirical Evidence for work disincentives

Do high tax rates really matter for work behavior

of welfare recipients?
• Moffitt (2002) concluded that AFDC led to a 10-50%
labor supply reduction among welfare recipients.
• When TANF was introduced and tax rates were
lowered, the proportion of welfare recipients who had
any earnings increased from 6.7% in 1990 to 28.1% in
1999.
 Other factors, like work requirements and an improving
economy, clearly matter, too.

26
 ■ Income
Redistribution

Moral hazard effects of the cash transfer

programs: Other incentives
Effect on Saving
• Some significant negative effects
• Hubbard, R. G., J. Skinner and S. Zeldes (1995)
Effect on Family Structure
• Empirical research find that this effect is, at most,
very small.
• Moffitt (1992), Hoynes (1997)

27
 ■ Income
Redistribution

“Welfare Lock”： Medicaid

Medicaid is the largest spending program (in-

kind transfer) for the poor in the US.
Two main problems:
• Crowd-out
• To the extent that families give up costly private insurance
for the free Medicaid coverage, Medicaid “crowds-out”
private coverage.
• Strong empirical evidence
• Medicaid Notch: an example of welfare lock

28
 ■ Income
Redistribution

“Welfare Lock”
• For many low-income mothers, health insurance is
tied to non-work, through Medicaid.
 In the US, until the mid-1980s, Medicaid is taken away
entirely if a person is ineligible for TANF/AFDC
• Working more than a small amount produces a large
penalty from lost Medicaid benefits.
 This structure creates implicit tax rates on earnings far in
excess of 100% for becoming ineligible for TANF/AFDC.
• The discontinuous loss of Medicaid is known as the
Medicaid “notch” or Medicaid “welfare lock”,
because it creates a strong incentive to stay on
welfare.
• Example: The dilemma of Dibao
29
 Women’s eligibility for health
“Welfare Lock” insurance is linked with their
receiving cash welfare
This creates a payments
$ of consumption “notch” in the
per year budget constraint.

$30,000 A E Slope = net wage = −7.5

26,280
24,280 B

F
Health insurance =
G = 12,140 $2,000
D

Slope = net wage = −15 Cash =

$12,140

C
0 381 2,000 Hours of leisure
per year
 ■ Income
Redistribution

Removing “Welfare Lock”: “Medicaid Notch”

• Unlinking cash and in-kind benefits decreases

(increase) the relative value of welfare
(working), thus leads to exits from cash
welfare.
• Empirical evidence is mixed
 Even the largest estimates suggest that such uncoupling
only very modestly decreases the number on welfare rolls.

31
 ■ Income
投票 最多可选1项
Redistribution

Which of the following is a moral hazard cost of

transferring money from high-income to low-
income groups?

A
The taxation of higher-income individuals
may cause them to work less hard.
B
The taxation of higher-income individuals
may cause them to save
C
Individuals may reduce their labor supply
to qualify for the means-tested transfers.
D
The administrative costs of the transfer are
huge
Submit
32
 ■ Income
Redistribution

ADDRESS THE MORAL HAZARD

ISSUE: ALTERNATIVE OPTIONS

33
 ■ Income
Redistribution

How to overcome the moral hazard

problem? – Alternative policies
Alterative policies to help the poor
1) Targeting the benefits to some characteristics that
are hard to change, easy to verify and related to low
ability  Ex-ante better targeting
• Moving to categorical welfare payments

2) Design a system to get individuals to reveal their

true ability Ex-post better targeting
• Using ordeal mechanisms
• Increasing outside options

34
 ■ Income
Redistribution

Alternative Policies I: Moving to Categorical

Welfare Payments
• Moral hazard occurs because the benefits are
not well targeted.
• For example, the government uses earning as a
target to measure ability, but people can change
their income.
• If we could target benefits to some characters
related to low ability, hard to change (to
qualify), and easy to verify, there would be no
moral hazard.
• What makes a good target?
• Blindness, age, disability?
35
 ■ Income
投票 最多可选1项
Redistribution

Which of the following potential target groups for

welfare benefits would best satisfy the three
characteristics of a good targeting mechanism?

A single mothers
B people who have a missing leg
C people earning less than $8 per hour
D people without health insurance
E people who are short in height (Mankiw and Weinzierl 2009)
F poorly-educated individuals
Submit
36
 ■ Income
Redistribution

Alternative policies II - Using ordeal

mechanisms
Ordeal mechanisms are features of welfare programs that make
them unattractive, leading to self-selection of only the most
needy recipients.
• Work requirements of TANF
• In-kind transfers rather than cash
• Remove welfare Lock (delete in-kind benefits attached to cash transfers)
• “Stigmatizing” welfare programs
• Waiting in lines at public welfare offices.
The paradox of ordeal mechanisms: apparently making the less
able worse off actually makes them better off (See
homework)
• targeting will be more efficient because the non-needy leave the
program and the government can make a fixed budget go to the needy.
37
 ■ Income
Redistribution

Alternative policies II - In-kind Transfer

In-kind transfer: direct provide goods rather

than cash.
• Food stamps; Medicaid; Public Housing;
Energy aid; Nutritional programs; Child Care.
Reasons for providing in-kind transfer:
• Ordeal mechanism-discourage those who are not
“truly needy” from applying
• Commodity egalitarianism
• Political attractiveness
38
 ■ Income
Redistribution

In-kind Transfer - example

People prefer cash transfer
in this case.

Giving the transfer in-kind

is binding.

39
 ■ Income
Redistribution

People are indifferent between

in-kind and cash transfer in
this case.

Giving the transfer in-kind is

not binding

40
 ■ Income
Redistribution

Trade-offs for in-kind transfers

Problems for in-kind transfers

• A person can never be made better off with an
in-kind transfer that is equal in cost to a cash
transfer.
 It is difficult to measure the value of in-kind transfer. Not
always valued at dollar-for-dollar (if resale is difficult).
• Substantial administrative costs.

41
 ■ Income

投票 最多可选1项
Redistribution

Which of these types of welfare has the lowest risk of

moral hazard?

A Categorical, cash welfare

B Categorical, in-kind welfare

C Means-tested, cash welfare

D Means-tested, in-kind welfare

提交
42
 ■ Income
Redistribution

Alternative Polices III: Improving Outside Options

Increase the outside options available to

individuals so that it is no longer attractive to be
on welfare.
How?
• Why individuals do not choose the outside options
before the intervention? What is the benefit and cost
of outside options?

43
 ■ Income
Redistribution

Alternative Polices III: Improving Outside

Options
• Increasing effective wages
o Labor Market Subsidies
 Increase employment, reduces welfare use.
o Training
 Modest declines in welfare use, earnings increases.
• No “good” jobs
o Encourage firms to provide jobs?
o Improve amenities to make working more easily for the
disabled?
• Child care
o Increases mother’s employment.
o Why china’s female labor force participation declined
substantially?

44
 Increasing Outside Options

Higher wage makes welfare a

$ of consumption less attractive option relative
per year
E
to work.
$35,000
Slope = wage = −17.50

A
25,000
Y2
17,500

Slope = wage = −12.50 Y1

12,170 B F D
G = 11,170

C
0 1,000 1,106 1,381 2,000 Hours of leisure
per year
1,027
 ■ Income
Redistribution

Improve outside options- Training

The traditional approach to increase outside

opportunities is training welfare recipients.
• Most recipients have low skills.
Empirical evidence suggests training programs
lead to modest declines in welfare receipts and
increase earnings. Yet they cannot induce
sizable reductions in the welfare rolls. Heckman
(1999).
Return to education in evening/adult schools in
China?
46
 ■ Income
Redistribution

Appliance: The Canadian Self-Sufficiency

Project
• Randomized evaluation of a work subsidy program.
• Offered large wage subsidies to a (random) group of
Canadians on welfare for more than one year.
• The subsidy increased employment by 43% in the
short run relative to control group.
• The rate of welfare enrollment fell by roughly the
same amount.
• After five years, the impact on employment welfare
use fell to zero.

47
 ■ Income
Redistribution

The aid debate: why the poor is poor

The Aid Debate: Jeffrey Sachs vs. William Easterly

• “The End of Poverty” vs “The White Man’s Burden”
 Clinical economics (planned-based, outside help) vs. learn by doing (market-based,
inside search)
Banerjee and Duflo, 2011, book: “Poor Economics: A Radical
Rethinking of the Way to Fight Global Poverty”
• health? food? education? small business? Saving or borrowing?
• Wrong information, expectation; high risks and too much responsibility;
bad environment and bad policy on them
Mullainathan and Shafir, 2013, book “Scarcity: Why Having Too
Little Means So Much”
• The struggle for insufficient resources—time, money, food,
companionship—force the brain to focus on alleviating pressing
shortages and thus reduce the mental bandwidth available to address
other needs, plan ahead, exert self-control, and solve problems.
48
 ■ Income
Redistribution

How about Children?

How to help poor kids?

• Little moral hazard from the child-side
• Programs targeting in child all have very high returns.

APPLICATION: Child Care, Preschool, and

Child Outcomes
How shifting across modes of care for children, from
home to child care to pre-school, affects both the
short- and long-run outcomes for these children?
• Head Start program (early intervention, such as free
breakfast)
• Pre-kindergarten programs
49
 ■ Income
Redistribution

Expenditure Incidence-who benefit from the

welfare program (general equilibrium analysis)
Expenditure Incidence: the impact of expenditure
policy on the distribution of real income.
Generally, any government program sets off a chain of
price changes, and the incidence is unclear.
• Eg. Housing, Food Stamps
• How EITC affect non-eligible low-income group?
• How food stamps affects supermarket?
Families’ benefits depend on their preferences

50
 ■ Income
Redistribution

The last but not the least

Who should be responsible for the welfare

program? Local government or federal
government?
1996 welfare reform in the US gave the states
much greater freedom to redesign their
welfare system.
• No evidence for race-to-bottom
In China, most welfare programs are the
responsibility of the local governments.
51
 ■ Income
Redistribution

APPLICATION: Evaluating the 1996 Welfare

Reform
In 1996, President Clinton signed into law the Personal
Responsibility and Work Opportunity Reconciliation Act
(PRWORA).
1. Cash welfare was changed from an entitlement to a block
grant.
2. States were allowed and encouraged to experiment with
alternative structures of cash welfare payments.
3. Time limits were imposed on welfare recipients.
4. Work requirements were imposed on welfare recipients.
5. New efforts to limit unwed motherhood were introduced.

52
 ■ Income
CHAPTER 17: Income Distribution and Welfare Programs
Redistribution

APPLICATION: Evaluating the 1996 Welfare Reform 2

Figure
17-9

Note the association between the timing of this law and the reduction in welfare
caseloads. After reaching a peak in 1994, welfare caseloads had declined by more than
three-fourths by 2017.

53
 ■ Income
CHAPTER 17: Income Distribution and Welfare Programs
Redistribution

Effects of the 1996 Welfare Reform

Many different aspects of the welfare laws have changed so that it

is hard to assign a causal impact to any one of the changes.
Overall conclusion of evaluations:
• About one-third of the decline in the welfare rolls can be
ascribed to reform, with the rest due to the improving economy.
• Kleven (2019): Welfare reform, not the EITC, responsible for single
mother employment increases
• Despite enormous reduction in use of welfare, single mothers as
a group have not seen a drop in their income or their
consumption.
• Leisure clearly declined.
• There has been no noticeable effect of welfare reform on fertility
rates. 54
 ■ Income
Redistribution

INQUALITY AND POVERTY:

MEASURES AND FACTS

55
 ■ Income
Redistribution

Income Inequality and Poverty - Measures

Absolute Deprivation
• Poverty rate
Relative inequality
• Income distribution: Lorenz curve
• Relative income inequality: Gini coefficient
• Income share of different income groups
• Income gap or ratio between different
groups

56
 ■ Income
Redistribution

Absolute Deprivation: Poverty Rate

Absolute deprivation: The amount of income the poor have
relative to some measure of “minimally acceptable” income.
The poverty line: government’s standard for measuring absolute
deprivation; a fixed level of real income which is considered
enough to provide a minimally adequate standard of living.
• The poverty line is 4000 yuan per person a year in 2020 (about $1.66 a
day (PPP), 2300 RMB in 2011) in China, about 35% of the average income
in poor rural areas
• UN standard: $1.9 a day after 2015 (1.25 before); $5.5 a day for upper-
middle income in 2017
Poverty rate: the share of people whose income is below poverty
line.
• About 5.5 million (0.6%) are now below official poverty line in 2019 (16.6
million in 2018, 1.7%). Number of people in Dibao program: 18.6 million
China has set the goal of eliminating extreme poverty by the end
of 2020. 57
 ■ Income
Redistribution

58
 ■ Income
Redistribution

Proportion of people living on less than $1 a day (%)

$1.9 a day

59
 ■ Income
CHAPTER 17: Income Distribution and Welfare Programs
Redistribution

Poverty Lines by Family Size (2018)

Table 17-3

Size of Family Unit Poverty Line

1 $12,140
2 16,460
3 20,780
4 25,100
5 29,420
6 33,740
7 38,060
8 42,380
For each additional person, add 4,320
Data from: U.S. Department of Health and Human Services (2018a).

Since their development in 1964, these amounts have

simply been updated for inflation.
60
 ■ Income
CHAPTER 17: Income Distribution and Welfare Programs
Redistribution

Poverty Rates over Time in the United States

Figure 17-3

Poverty rates declined rapidly in 1960s and early 1970s and then oscillated for all age
groups. Poverty rates began to decline again after 2010. in 2017, poverty rate 12.3%,
about 39.7 million in poverty.
61
 ■ Income
Redistribution

Absolute Deprivation: Poverty Rate

Criticisms about poverty rate.

• Difference in cost of living across regions
• Subsistence consumption bundle has
changed
• Health care is not considered in the initial poverty line
• “Income” definition is incomplete
• consists only of cash receipts; but non-cash (in-kind
transfers) have grown substantially
• Ignores taxes (refund), but payroll tax and Income tax
credits (EITC, Child Tax Credit) have grown substantially
But, politically difficult to change definition
62
 ■ Income
Redistribution

Income Inequality – Relative inequality

A Lorenz curve is a diagram showing the
cumulative percentage of national income
received by a certain percentage of households.
• Families are ranked according to income; the x-axis represents
percentile in the income distribution. The y-axis represents %
of total family income received by those at or below a given
percentile.
The Gini coefficient is defined as the area
between the Lorenz curve and the perfect
equality line, divided by the area underneath
the perfect equality line
• The Gini coefficient ranges from 0 (perfect equality)
to 1 (perfect inequality).
63
 ■ Income
Redistribution

64
 ■ Income
Redistribution

65
 ■ Income
Redistribution

66
 ■ Income
Redistribution

67
 ■ Income
Redistribution

Source: Piketty, Yang,

and Zucman (2016)

68
 ■ Income
Redistribution

Piketty, Yang, and Zucman (2019)

69
 ■ Income
Redistribution

70
 ■ Income
Redistribution

71
 ■ Income
Redistribution

72
 ■ Income
Redistribution

Income Inequality - Reasons

Income Inequality has increased over time in both
developing and developed countries.
Possible Explanations
• Driving force is the increase in earnings differentials
between different skills (the returns to education), which
can result from demand changes due to skill-biased
technology changes and/or globalization (in the US), or due
to labor market reforms (in China).
• Katz and Autor (1999, HLE), Acemoglu (2002, JLE), Acemoglu and
Autor (2010, HLE)
• Increase in return to capital, Piketty and Saez (2019)
• Reduction in welfare payment due to institutional changes:
Meng(2004, 2005)
• Regional divergence: Cai, Chen, and Zhou (2010), Xie and
Zhou (2015)
73
 ■ Income
Redistribution

Source: Xie and Zhou (2015)

74
 ■ Income
Redistribution

Source: Cai, Chen, and Zhou (2010)

75
 ■ Income
Redistribution

Income Inequality - Reasons

In China, other important factors:
• Incomplete market reforms and political reforms
creates a lot of loophole for corruption, monopoly,
unfair opportunities
• Welfare system and tax system are very weak in
income redistribution
Review Papers:
Acemoglu and Autor (2010, HLE)
Atkinson, Piketty, and Saez, (2011, JEL)

76
Taxes and Transfers can help reduce income inequality

Gini before tax

UN R/P UN R/P Word Gini after tax
CIA Gini and transfers (late
10% 20% bank Gini and transfers
2000s)
China mainland 21.6 12.2 37 46.9
Hongkong 17.8 9.7 53.7
Taiwan 33.8
India 8.6 5.6 33.6 33.6
Brazil 40.6 21.8 52.9 51.9
Russia 12.7 7.6 41.6 42
South Africa 33.1 17.9 63.4 62.5
Philippines 15.5 9.3 43.0 46.0
Mexico 21.6 12.8 48.1 48.3 49.4 47.6
Korea 7.8 4.7 31.3 30.2 34.4 31.5
U.S. 15.9 8.4 41.1 45 48.6 37.8
Japan 4.5 3.4 32.1 37.9 46.2 32.9
Germany 6.9 4.3 30.1 27 50.4 29.5
UK 13.8 7.2 32.6 32.4 45.6 34.5
Argentina 31.6 17.8 42.3 45.8
Chile 26.2 15.7 50.5 52.1 42.6 32.4
 ■ Income
Redistribution

78
 ■ Income
Redistribution

79
 ■ Income
Redistribution

岳希明，2020
80
 ■ Income
Redistribution

Recap

The moral hazard costs of welfare policy

• Working disincentive
• Reducing saving
Reducing moral hazard of welfare
• Moving to categorical welfare payments
• Using ordeal mechanisms
• Increasing outside options
• Remove welfare lock
Compare cash transfer and in-kind transfer programs
Measures for Income inequality and Poverty
• Lorenz curve, Gini coefficients, poverty rate

81
 ■ Income
Redistribution

Reading list

Banerjee and Duflo, 2011, book: Poor Economics: A Radical

Rethinking of the Way to Fight Global Poverty
Chetty, R., Hendren, N., Kline, P., Saez, E., & Turner, N. (2014). Is
the United States still a land of opportunity? Recent trends in
intergenerational mobility. The American Economic
Review, 104(5), 141-147.
Ye Jin, Hongbin Li, Binzhen Wu,“Income Inequality, Consumption,
and Social-Status Seeking,”Journal of Comparative Economics,
39, 191-204, 2011.
Acemoglu, D and D Autor, 2011, “Skills, tasks and technologies:
Implications for employment and
earnings” Handbook of Labor Economics.

82