ECINEQ Summer School Universitat de les Illes Balears "New Perspectives on the Study of Inequality,

1
ECINEQ Summer School - Universitat de les Illes
Balears"New Perspectives on the Study of
Inequality, Poverty and Redistribution Equity
and Development"

• François Bourguignon
• Chief Economist and Sr. Vice President
• Development Economics
• The World Bank
• Palma de Mallorca, Spain, 17th 21st July, 2006

2
Equity and Development Motivation

• Development defined as poverty eradication
• Change in poverty results arithmetically from
  changes in the mean income of population and
  changes in the distribution of relative incomes
• 'Development' is the result of the interaction
  between growth and distribution
• What do we know about this interaction? What kind
  of development strategy and development policy
  can be built on this interaction?

3
Equity and Development Motivation (contd)

• Preceding argument is really about the inequality
  of income or some other welfare metric and
  growth.
• Important economic literature about that
  relationship will be briefly reviewed in these
  lectures
• However, this literature misses an important
  point
• "It is not the relationship between inequality
  and growth that matters for development, but that
  between 'equity' and growth.
• Equity is defined here as equality of
  opportunities theme of WDR 2006.

4
Outline

• Introduction the poverty-inequality-growth
  triangle
• Inequality and growth a brief selective review
  of the literature
• Equity and development
• Theoretical considerations
• Searching for evidence
• Policy implications

5
1. The poverty-inequality-growth triangle a)
decomposition of change in poverty measures

• Poverty defined as welfare metric (y) lt poverty
  line ( z)
• Additively decomposable poverty measures
• where C( ) increasing and convex and f( )
  density of the distribution of y.
• Change in poverty
• Change in distribution can be
  decomposed into change in mean, , (growth or
  scale effect), and change in the distribution of
  relative welfare,

6
Decomposition of Change in Distribution and
Poverty into Growth and Distributional Effects

7
Simple Arithmetic of Poverty, Inequality and
Growth
Change in Poverty
F ( average income, distribution, growth, change
in distribution)
8
Growth Elasticity of Poverty
Note See Bourguignon (2004). ? the elasticity
of poverty with respect to income. T-statistics
are included in parentheses in the regression
equation.
9
Growth, distribution and changing poverty levels
in a sample of growth spells
   indicates increase in poverty  - 
indicates decline
Numbers indicate the count of growth spells in
each cell
10
b) The Poverty-Growth-Inequality Triangle
Absolute poverty and poverty reduction
"Development analysis and strategy"
Distribution and distributional changes
Aggregate income level and growth
11
2. Inequality and growth a brief selective
review of the literature
a) A simple general equilibrium representation of
the growth-distribution relationship
Individual income from endowments Xi.
Accumulation
Economic equilibrium
Distribution Growth
12
b) The particular case of the (linear)
capital-labor one-sector growth model (YAK)
Individual income
Accumulation
Economic equilibrium a labor share, t tax on
capital
Distribution of relative
Growth income
and wealth
Wealth distribution and growth are independent
but (progressive) income redistribution (t)
affects growth negatively (Kaldor effect)
13
c) Other particular cases with the (non-linear)
capital-labor one-sector growth model
c.1) Solow-Stiglitz (YAf(k), kK/N)

• Redistribution of wealth does not matter for
  growth but growth influences distribution through
  declining marginal product of capital f'(k)
• Inequality-growth relationship Inverted-U
  (Kuznets) curve (See survey by Bertola
  (2000), Stiglitz (1969), Bourguignon (198x)

Income ineq.
k
14
c) Other particular cases with the (non-linear)
capital-labor one-sector growth model
c.2) Political economy models (YAK) Majority
voting on redistribution parameter, t median
voter theorem applies and equilibrium tax rate,
t depends on distribution of wealth. Higher
inequality leads to higher tax rate and to slower
growth. See Romer (1975), Alesina-Rodrik(1994),
Persson-Tabellini(1994), Bertola (1993), Benabou
(1996) and surveys by Bertola (2000), Aghion,
Penalosa and Caroli (1999)

15
c) Other particular cases with the (non-linear)
capital-labor one-sector growth model
c.3) Capital Market Imperfections

• Distribution of wealth affects total investment
  and growth, which in turn affects the evolution
  of the distribution of wealth
• Capital market imperfection generally explained
  by credit rationing due to asymmetric
  information see Banerjee-Newman (93),
  Galor-Zeira (93), Aghion-Bolton (96), Piketty
  (96), and survey by Bertola (2000) and Aghion,
  Penalosa and Caroli (99)
• Progressive redistribution of wealth may
  accelerate growth in some cases

16
A special case of imperfect capital market
r,s
r(ki)
s S(ki, r(ki))
ki
k
k3
k2
k1
Case of no access to credit market below k and
accumulation behavior increasing function of both
initial wealth and rate of interest.
Redistribution from k3 to k2 accelerates growth
but does the opposite from k1 to k2.
17
d) Other particular cases of general equilibrium
model of growth and distribution
Labor market imperfection and sectoral shift

• Zi is a variable that indicates whether
  individual i has access to high wage labor market
  e.g. 'modern' sector. It depends on the general
  level of development and the observable
  characteristics of individual i.
• This model includes the Kuznets-Lewis model of
  development based on sectoral shifts (leading to
  a Kuznets curve). Also consistent with
  Banerjee-Newman (1993) and others..
• Inequality here does not affect growth, except
  if savings depend on labor incomes.

18
e) Other sources of interaction between
inequality and growth

• Other examples of efficiency-enhancing wealth
  redistribution mechanisms linked more indirectly
  to growth include economies of scale, imperfect
  insurance markets, local commons, conflicts and
  crime,

19
f) Conclusion on theoretical mechanisms linking
growth and distribution

• There are circumstances under which an exogenous
  progressive redistribution of wealth may enhance
  economic efficiency and accelerate growth
• Growth is not necessarily distribution neutral.
  Several arguments support the Kuznets curve
  hypothesis.
• Asymmetry between wealth and income
  redistribution
• Fully general model suggests many factors may
  simultaneously affect growth and distribution
  technology, international prices, trade
  protection,
• Combining political economy model imperfect
  capital market models leads to a pretty good
  interpretation of main redistribution regimes in
  western countries.

20
g) Empirical evidence on the relationship between
distribution and growth

• Empirical literature dominated by cross-country
  studies
• Income inequality as a function of development
  level Kuznets curve in the 1970s Growth is
  good for the poor (Dollar-Kraay 2002)
• Growth as a function of initial income inequality
  in the growth regression literature of the 90s
• Both streams of literature equally inconclusive -
  not surprisingly.
• Case studies on the distributional effects of
  growth (pro-poor growth) showing very much
  country specificity

21
Cross-country 1
22
Cross-country 2

• Dollar-Kraay  Growth is Good for the Poor 

See Dollar and Kraay (2002)
23
Cross-country 3
24
Cross-country 4
25
Cross-country evidence Conclusion

• Growth ? Distribution
• Unsurprising lack of evidence on Kuznets curve
  many factors affecting change in distribution,
  including distribution policies
• Does not mean that causes for international
  differences in income distribution cannot be
  identified explained
• Initial analysis by Kuznets probably based on
  rudimentary data (general problem of data
  adequation and comparability)
• Distribution ? Growth
• Unsurprising lack of evidence theory does not
  necessarily imply that more inequality in income
  distribution leads to slower growth
• Few exogenous changes in distributional
  characteristics
• Danger of generalizing from very partial evidence
  (S. Korea, Taiwan, Brazil, South Africa, ..)

26
Case Studies of the Distributional Effect of
Growth

• Identifying forces affecting distribution over a
  given period and trying to relate them to growth
• Limitations of time series analysis
• Decomposition of distributional changes (MIDD,
  2004, project) specificity of countries and
  growth episodes
• Effects of prices, occupation allocation and
  endowments on observed evolution of
  distribution
• Example of Brazil (1976-1996) slow growth
  causes more relative poverty, an evolution
  compensated by educational gains and drop in
  fertility
• Example of Taiwan (1980-1996) fast growth
  causes increasing earnings inequality, an
  evolution compensated by rising female labor
  force participation
• No equivalent analysis for the effect of
  inequality on growth

27
Conclusion on "Inequality and Growth"

• Relationship between growth and distribution of
  income is complex and likely to be highly
  country-specific
• No conclusive evidence on a systematic
  relationship between growth and income inequality
• Not surprising, given the complexity of the
  effects of growth on distribution and the
  misunderstanding of the theoretical arguments
  relating growth to inequality
• From the latter point of view it is the
  inequality in wealth and access to credit (or
  labor) market that may have an impact on growth
• This logically leads to extending the concept of
  income inequality to that of equity or equality
  of opportunities

28
3. Equity and Development

• This part of the lectures borrows greatly from
  World Bank's "World Development Report 2006,
  entitled Equity and Development lead authors
  Francisco Ferreira and Michael Walton
• This report was very well-received and is
  inspiring much of the current work and policy
  formulation taken place in the Bank and within
  the development community
• It relies on new concepts the exploration of
  which is not fully completed and on hypotheses
  still to be confirmed.
• It thus offers a wide field of new research in
  development

29
A) Theoretical considerations

• Definition equity as equality of opportunities
• Why equity matters for development the positive
  effect of equity, efficiency and growth
• Tradeoffs vs. Complementarity of equity and
  efficiency in equity-promoting policies
• Some conceptual pending problems
• Quantity vs. Distribution of opportunities
• Is the 'equity-efficiency' tradeoff really to be
  set aside?
• The concept of 'inequity trap'

30
a) Defining and observing equity

• Equity is a normative concept of which component
  principles include
• Equality of opportunities (equality of
  endowments, process fairness, merit-based
  rewards, ) (see Dworkin, Roemer, Sen, )
• Aversion to consumption deprivation
• Almost never the same as equality in one of the
  outcome dimensions, but may require changes in
  distribution in a number of them.

31
Opportunities
Endowments Wealth, land, culture, social
background,
Process labor market, investment, schooling,
voice
Outcomes income, consumption, health,
environment,
Individual traits and preferences taste for
efforts, acquired talent,
32
Relationship between concepts and measures of
inequality and inequity

• Inequality usually refers to outcomes (welfare,
  education, health,..) rather than opportunities
• A society with some (possibly) significant
  inequality may be equitable. However, a very
  unequal society is likely to be inequitable
  (Income Gini as marker)
• Equity leads to consider both the top and the
  bottom (poor) of the distribution of
  opportunities
• The difficulty of measuring inequity

33
Difficulties with the concept of equity as
equality opportunities

• The difficulty of measuring inequity
• Relevance of the distinction between 'endowments'
  and 'individual traits'. What if 'individual
  traits' are inherited from parents and comparable
  to other endowments?
• From a practical (policy) point of view, may be
  better to distinguish between 'observable
  endowments' and 'unobservable endowments' and
  individual traits

34
Example (1) Infant Mortality Rate and Mothers
Education
No Education
Secondary or Higher
Source Demographic and Health Surveys (DHS),
latest available data since 2000
35
Example (2) Opportunities in health are important
especially at an early stage in life
El Salvador study on Early Childhood Development
36
Example (3) Enrollment Rates in India by
income quintile of parents
37
Example (4) how access to formal jobs depends
on parental education in Brazil
Source Bourguignon, Ferreira and Menendez
(2005)
38
Example (5) Access to Credit Large firms are
more likely to have bank loans

• with Bank Loans

100
90
80
70
63.5
60
51.9
50
42.4
40
32.3
30
20
7.6
10
0
Micro Small Medium
Large Very Large
Source World Bank ICS data
39
b) Why equity matters for development

• General principle 1.
• Unequal access to processes that facilitate
  certain outcomes implies an inefficient
  allocation of resources some profitable
  investments are not undertaken whereas average or
  mediocre projects are implemented.
• Notes
• The race metaphor
• Generalizes the imperfect capital market
  argument. Applies to many other areas insurance,
  education, health, infrastructure, jobs in the
  formal sector, etc..
• Unequal access to markets like credit or
  insurance result from a combination of market
  imperfection and inequality in endowments. Other
  cases may include discrimination behavior (e.g.
  labor market)
• For a general statement of this principle see
  Bardhan, Bowles and Gintis (2000)

40
b) Why equity matters for development
(contd)

• General principle 2.
• Unequal protection of property rights and unequal
  political rights generate economic inefficiency,
  suppress incentives to investment, and prevent
  reform of bad governance.
• Illustration 'elite capture' or 'rent-seeking
  elites'.
• See Acemoglu (2006), Acemoglu, Johnson and
  Robinson (2005), Acemoglu and Robinson (2000),
  Besley and Coate (1998), Bourguignon and Verdier
  (2000), Krueger(1974, 2003), Mauro (1995),
  Persson and Tabellini (2000), Rajan-Zingales
  (2000),

41
Principle 2 Economic Impact of Élite Behavior
The rent-seeking and benevolent élites
Income per capita
Benevolent élite
Rent (per capita of the Elite)
GDP per capita
Intensity of rent-seeking
Rent-seeking élite
42
Principle 2 Economic Impact of Élite Behavior

• Majority voting with transparency could
  lead to the 'benevolent elite equilibrium.
• Standard argument about inefficiency and
  non-competitive structures (barriers to entry as
  inequity)
• Numerous other examples decision about public
  infrastructure at the local level, unequal access
  to judicial system, etc..
• Institutions and growth

43
b) Why equity matters for development (contd)

• General principle 3. Excessive inequity and
  weak institutions may be the motive of crime,
  violence, political instability and conflicts,
  all deterrent of economic growth
• Note Difference between role of inequity and
  inequality in conflicts Collier and Hoeffler
  (2004)
• Acemoglu-Robinson (2000), Alesina-Perotti (1996),
  Fajnzylber et al. (2002), Bourguignon (2000)

44
c) Tradeoffs vs. Complementarity of equity and
efficiency in equity-promoting policies

• Preceding argument apply in the long-run or in a
  comparative static sense.
• Equity promoting policies must be taken into
  account, which may have an efficiency cost
• Example equalize quality of education through
  additional public spending in education
• Cost of additional tax (or less spending in some
  other area)
• Benefit from more equity in the long-run

45
d) Some conceptual pending problems (i)
Quantity vs. Distribution of opportunities

• General Principle 1 refers to the distribution
  of existing opportunities in the population
  rather than their quantity in the economy
• Access to education
• Issue is not so much that more people educated
  may accelerate growth,
• But that making (limited) access unconditional on
  individual characteristics including family
  background- improves efficiency.

46
d) Some conceptual pending problems (ii) Is the
'equity-efficiency' tradeoff really to be set
aside?

• Important to keep in mind that the efficiency
  gain from equity refers to aggregate output not
  to individual welfare.
• Equity promoting policies could lead to Pareto
  superior outcomes only if credible redistribution
  commitments are available.

47
d) Some conceptual pending problems (iii) The
concept of inequity traps (??)

• Possibility to generalize the concept of low
  income trap to inequity traps?
• Low income trap yt lt z ? ytn lt z for all n
  gt 0
• Low inequity trap I(kt) gt s ? I(ktn) lt z
  and for all n gt 0

48
B) Equity and Development in search of empirical
evidence

• Limits of aggregate approach
• example of cross country analysis of inequality
  and growth
• Problem of measuring equity
• The Microeconomic approach

49
Example (1) Inefficiency of capital
distribution in MexicoDifferential Returns on
Own Capital
Source McKenzie and Woodruff, 2004
50
Example (2) Social Inequality Instilled
Inefficiency in IndiaDifferential Performance
when Caste is Made Salient
Source Hoff and Pandey, 2004
51
Example (3) Good economic institutions are
associated with prosperity Association between
economic prosperity and protection against risk
of expropriation.
52
Example (3b). Governance Growth Rate of GDP
Macro Evidence
CHN
BWA
KOR
SGP
THA
IND
IDN
MLT
HKG
CHL
CYP
TUN
SDN
MYS
UGA
BGD
EGY
PAK
BHR
DOM
TUR
KWT
LKA
ECU
IRN
MLI
CRI
GHA
ISL
SLV
BRA
GMB
PNG
COL
POL
DZA
MWI
GTM
SEN
MEX
ISR
SYR
HND
JAM
BRN
URY
GUY
KEN
PAN
CMR
BOL
JOR
ZAF
HUN
BGR
TGO
TTO
PRY
ARG
ZWE
PHL
NER
PER
COG
VEN
HTI
ROM
ZMB
NIC
SLE
ZAR
Stephen Knack Lessons of 90s, The World Bank
53
Example (4) Crime and Inequality as a proxy
for inequityIncome Distribution Robbery Rates
1970-1994 (5-year-averages)
Source Fajnzylber, Lederman, and Loayza (2002).
54
Other types of evidence

• Land reforms in West Bengal (Banerjee, Gertler
  and Ghatak, 2002)
• Micro-finance (Burgess, Pande, Wong, 2005,
  Banerjee and Duflo, 2004)
• Press freedom (Besley, Burgess,Prat, 2002),
• Women as policy makers (Chattopadhyay and Duflo,
  2004)
• Etc
• But evaluation is incomplete..!

55
Other types of evidence

• Country studies
• The difference between Asian tigers and Latin
  America
• The paradoxical case of China (and Russia)
• The case of African countries

56
C. Policy Implications

• An equity lens adds three new perspectives to
  development policy
• Good policies for poverty reduction may involve
  redistribution - of influence or government
  expenditures - away from dominant groups.
• Equity-efficiency trade-offs in such
  redistributions need to be assessed in light of
  the full long-term benefits of equity.
• The perception of a dichotomy between growth
  policies and policies for equity is misguided.
• (Country-specific context is key in policy
  choice.)

57
C. Implications for Policy (ctd)

• Equitable accumulation access to education,
  health, land, infrastructure,
• Market fairness labor market, credit, goods
• Full political participation
• Reduce absolute poverty (safety nets)
• Avoid excessive inequality that may transform
  into future inequity (taxes, public management,
  macro-economics)

58
Example Early childhood development
59
C. Implications for Policy (ctd)

• Political Economy dimensions of most policies.
  What should be done?
• Analysis and dissemination are key.
• Basic principle Make sure that all actors have
  perfect knowledge of the distributional impact of
  institutional reforms.
• Dissemination work
• Enhanced work by the intellectual community in
  underestanding the role of institutions and the
  way they may be modified
• Data collection for improving institutional
  indicators, monitoring reforms and comparing
  experiences of institutional change
• Scope for action by external agents?

60
Conclusion

• - The need for evidence and proper evaluation of
  policies
• - The international dimension of the
  equity-efficiency debate
• - Equity as unifying factor of the main pillars
  of development strategies "Investment climate"
   , "Empowerment"  and "governance"