Institutions and Development: The Role of Education and the Distribution of Income

1
Institutions and Development The Role of
Education and the Distribution of Income

• Cecilia García Peñalosa

2
Inequality across countries GDP pc relative to
world mean
INTRODUCTION

• University of California, Santa Cruz Atlas of
  Global Inequality

3
Inequality within countries Gini coefficients
around the world
INTRODUCTION

• Sources United Nations Development Programme
  Human Development Report 2003

4
Inequality within countries Gini coefficients
around the world
INTRODUCTION

• Sources United Nations Development Programme
  Human Development Report 2003

5
Inequality within countries Gini coefficients
around the world
INTRODUCTION

• Sources United Nations Development Programme
  Human Development Report 2003

6
Can inequalities within countries help us explain
inequalities across countries?
INTRODUCTION

• This requires asking the question of
• what is the relationship between
• inequality, education, and development

7
Plan of talk
INTRODUCTION

• Early theories of inequality, education and
  development
• Evidence I
• Inequality and democratization
• - Education, political participation and
  inequality
• Bourguignon and Verdier 2000
• - Inequality, revolution and franchise Acemoglu
  and Robinson 2000
• 4. Evidence II
• 5. Education, inequality and corruption
• Eicher, García Peñalosa and van Ypersele 2006

8
EARLY THEORIES

• Education, inequality and development
• Early theories

Imperfect capital markets and fixed cost of
education
The distribution of wealth determines who
studies
Distribution determines aggregate level of
education and hence income level
Output growth depends on aggregate level of
education
Greater inequality reduces education and
aggregate output
9
EARLY THEORIES

• Similar theories concerning physical capital
  accumulation
• Political economy arguments
• Galor and Zeira 1993, Alesina and Rodrik 1994,
  Persson and Tabellini 1994, Aghion and Bolton
  1997
• Surveyed in
• Aghion, Caroli and García-Peñalosa 1999
• Bertola, Foellmi and Zweimuller 2006

10
EVIDENCE I

• Evidence I
• Development and Institutions

• Effect of inequality on growth unclear
• Physical and human capital only play a moderate
  role in explaining cross-country differences in
  output levels
• Institutions have emerged as the main determinant
  of income differences across countries

11
EVIDENCE I
Ratio of richest 5 to poorest 5 countries 1990


Can institutions explain output
differences? Results log yi constant 5.14
Insi , R² 0.58

• Hall and Jones 1999

12
EVIDENCE I

• Evidence I
• Development and Institutions

• Effect of inequality on growth unclear
• Physical and human capital only play a moderate
  role in explaining cross-country differences in
  output levels
• Institutions have emerged as the main determinant
  of income differences across countries
• Evidence raises the question of whether the
  impact of education and inequality opperates
  through institutions

13
PARTICIPATION

• Education and Participation
• Bourguignon and Verdier 2000

• Two crucial assumptions
• - Electoral participation depends on
  individuals level of education
• - Technological externality stemming from
  average education level
• Income groups
• - Oligarchy educated make political decisions
• - Uneducated poor
• Redistribution
• - democratically chosen

14
PARTICIPATION
Higher human capital raises productivity and
hence income of oligarchy
Redistribution some poor can study
Tradeoff between increased productivity and
more redistribution
Middle class now participates in
elections increases redistribution
Oligarchy redistributes if inequality is not too
large
15
PARTICIPATION
Higher human capital raises productivity and
hence income of oligarchy
Redistribution middle class can study
Tradeoff between increased productivity and
more redistribution
Middle class now participates in
elections increases redistribution
Less inequality leads to increased
democratisation and output
16
FRANCHISE

• Inequality, Revolution and Franchise
• Acemoglu and Robinson 2000

• Two groups
• - Elite rich, has political power
• - Poor no political power
• The elite
• - sets a redistributive tax each period
• - can decide to extend franchise
• The poor
• - can undertake risky revolution
• - gain more from revolution when inequality is
  high

17
FRANCHISE
No commitment value
Redistribution ? yp
Avoids revolution if inequality is low
Costly high future redistribution
Extend franchise
Commitment
Avoids revolution even if inequality is high
18
EVIDENCE II

• Evidence II
• Development and Democratization

• Predictions
• - Unequal countries are more likely to extend
  franchise eg. Britain and France in mid 19th c.
• - Equal countries are more likely to
  redistribute eg Germany
• Good at explaining European historical evidence
  but not experience of l.d.c.s in second half of
  the 20th c.
• - Fast growth despite lack of democratization
  Taiwan, South Korea, China
• - Many formal democracies perform badly in terms
  of output and de facto institutions

19
EVIDENCE II

• Evidence II
• Development and Democratization

• Correlation education-democracy cross-country
  but not over time
• - Should we think about other institutions?
• - Constraints on the executive and corruption?
• Does education cause institutions or institutions
  cause education?
• - link between political institutions and
  education is hard to establish despite
  substantial macroevidence
• (Glaeser et al. 2004, Acemoglu et al. 2005)

20
CORRUPTION
Education, Corruption and the Distribution of
Income Eicher, García Peñalosa and van Ypersele
2006

• Key element more educated electorates are better
  at identifying corrupt politicians
• Substantial evidence (see Galston for a survey)
• More educated individuals
• - better identify the quality of institutions,
  politicians, and policies
• - have more consistent political views
• Less educated individuals
• - more likely to rely on character impressions
  rather than policy evaluation in their voting
  behavior
• Historical evidence for US states (Glaeser and
  Saks 2005)
• - more educated states have less corruption
• Implication education affects politicians
  strategies

21
CORRUPTION
Production, Education and Bequests

• Galor and Zeira 1993
• Education determines aggregate output
• No capital market and fixed cost of education ?
  only the rich can study
• OLG model parents leave financial bequests ?
  intertemporal bequest function linking the
  inheritance received and the bequest left
• There is a public good and income taxation
• Level of taxation shifts the bequest function

22


e
23


e
x0
24


e
25
Reductions in the tax rate


e
26
Inequality matters

x1 gt e

x1
e
x0
27
CORRUPTION
What does this mean?

• Reductions in the tax rate will increase the
  level of education
• If inequality is high, they will increase the
  level of education in the long-run
• If inequality is low, they will increase the
  level of education the next period ? implications
  for the behaviour of the politician

28
CORRUPTION
Politician

• Cost of public good c unkown to voters
• Politician sets the tax rate each period
• Honest c ? no rent
• Corrupt ? rent p
• Can pass constitutional reform
• - disclose c
• Lives for 2 periods
• There are elections at the end of 1

29
CORRUPTION
Politicians utility

• Utility
• Rents
• depend positively on level of education
• Prob. of reelection depends on education at
  t1 and on politicians action at t

30
CORRUPTION

Probability of reelection

• Key the educated can observe corrupt behaviour,
  the uneducated cannot

31
CORRUPTION
Higher human capital raises output and hence
rents increases corruption incentives
Effect of education on politicians expected
utility
Tradeoff between increased rent and lower
reelection probability
Education reduces probability of being reelected
if corrupt
Non-monotonic effect of education on corruption
32
CORRUPTION

• Need to consider two cases
• High initial inequality
• Bequests of the unskilled are low
• imposing doesnt allow them to study next
  period
• Low initial inequality
• Bequests of the unskilled are high
• imposing allows them to study next period

33
CORRUPTION
High inequality

1
34
CORRUPTION
Low inequality

• Imposing allows the unskilled to study next
  period
• This increases the monetary rent
• Hence U(H,C) is a viable strategy forgo rent at
  t in order to get higher one at t1

35
CORRUPTION
Low inequality

1
36
CORRUPTION
Implications

• High-education countries can escape from
  corruption
• Middle-education countries remain in a poverty
  trap with corrupt politicians
• Low-education countries can only escape the
  poverty trap if they are sufficiently equal

37
CORRUPTION
Implications 2

• Non-monotonic relationship between education and
  corruption/institutions at aggregate level
  despite micro-level links
• Path to development varies
• - Institution-led development more likely in
• unequal societies and for high human capital
• - Education-led development only takes place
  in
• highly equal economies
• Causality Do institutions increase education or
  does education lead to institutions?

38
CORRUPTION
Some examples

• Latin-America middle levels of education ? low
  income/corruption trap
• Africa vs East-Asia very low levels of education
  in the 60s
• In the 60s development economists
• - Africa was resource rich ok
• - Asia was seen as the problem

39
INEQUALITY
40
Measures of institutions
INSTITUTIONS

• Glaeser et al. 2004

41
CORRUPTION
High inequality low productivity

1