Poverty, Inequality and Growth Revisited

Martin RavallionDevelopment Research Group

World Bank

Poverty and Inequality Analysis, Module 5, 2011

Outline

• Concepts: what do we mean by “poverty” and “inequality?• How does growth influence poverty and inequality?• How do poverty and inequality influence growth?• Conclusions and implications for policy

2

Concepts of “distribution”: “inequality” and “poverty”

3

4

What do we mean by “inequality”?

One can find different definitions in public debates and policy discussions

5

Relative vs. absolute inequality

• Relative inequality is about ratios; absolute inequality is about differences.– State A: two incomes $1,000 and $10,000 per year– State B: these rise to $2,000 and $20,000 – Ratio is unchanged but the “rich” can buy 10 times more

from the income gains in state B than can the “poor”• One is not right and the other wrong. Indeed, 40%

of participants in experiments view inequality in absolute terms.

• Whether one thinks about inequality as absolute or relative matters greatly to one’s views on the distribution of the gains from growth.

Further reading: Martin Ravallion, “Competing Concepts of Inequality in the Globalization Debate,” Brookings Trade Forum 2004, Washington DC: Brookings Institution, pp.1-38.

6

Two Gini indices of inequality

R e l a t i v e G i n i i n d e x

y

yyy

nG jir

221

A b s o l u t e G i n i i n d e x

jia yy

nG 22

1

The main focus here is on relative inequality.

7

What do we mean by “poverty”?

Again one can find two different definitions in public debates and policy discussions

Definition 1: Absolute poverty

• Poverty line aims to have a fixed real value over time and across countries/areas.

• At country level: all developing countries use poverty lines that aim to have the same real value at different dates and places. – Typically anchored to nutritional requirements for good health

and normal activities

• At the global level, the World Bank’s widely-used “$1-a-day” line is absolute in that it aims to have the same purchasing power in different countries and at different dates.

Definition 2: Relative poverty

The more common practice in most OECD countries and Eurostat has been to set the poverty line as a constant proportion—typically 40-60%—of the (date and country-specific) mean or median income:

)10( kkMZ ii

We can call this a strongly relative poverty line.

MeanPoverty line

The focus here is on absolute poverty.

Further reading: Martin Ravallion and Shaohua Chen, “Weakly Relative Poverty”, Policy Research Working Paper 4844, World Bank, 2009.

10

The two-way relationship:

Growth=>poverty reduction+

High current poverty=>low growth

=> slower subsequent progress against poverty

11

Growth=>Inequality and poverty

Does growth come with rising inequality?Does growth reduce poverty?

12

The Kuznets Hypothesis• The KH claims that relative inequality increases in the

early stages of growth in a developing country but begins to fall after some point,

• i.e., the relationship between inequality (on the vertical axis) and average income (horizontal) is predicted to trace out an inverted U.

Further reading: Sudhir Anand and Ravi Kanbur, “The Kuznets Process and the Inequality-Development Relationship,” Journal of Development Economics 1993, Vol. 40:25-52.

What is the evidence?

13

A simple quadratic relationship between Gini and GDI per capita, 1950-2000

2030

4050

6070

6 8 10 12ln(gdpppp)

GiniW + giniWY 95% CIFitted values

Pooled countries and dates; n=1,000

14

• A weak inverted U relationship (more than 1000 Ginis)• Huge variability in inequality; R2 only 0.08• The upward sloping part of the curve particularly hard to

discern• Turning point is quite unstable; here about $PPP 2,000

(level of Senegal or Zimbabwe)• Even this weak inverted U vanishes with country fixed

effects.

The most serious critique: With greater time series evidence, we find that very few developing countries have followed the prediction of KH.*

Further reading: * Bruno, Michael, Martin Ravallion and Lyn Squire (1998),

“Equity and Growth in Developing Countries: Old and New Perspectives

on the Policy Issues,” in Income Distribution and High-Quality Growth (edited by Vito Tanzi and Ke-young Chu), Cambridge, Mass:

MITPress.

15

Changes in relative inequality are uncorrelated with growth

1. Across 120 spells (between two surveys), virtually zero correlation between changes in relative inequality (the log Gini index) and economic growth (change in the log of the survey mean or PCE).* Figure=>

2. Mean income of the poorest 20% (say) has a regression coefficient of about one on the overall growth rate**

Further reading: * Martin Ravallion, “Inequality is Bad for the Poor,” in Inequality and Poverty Re-Examined, edited by John Micklewright and Steven Jenkins, Oxford: Oxford University Press, forthcoming.**David Dollar and Aart Kraay, ‘Growth is Good for the Poor,’ Journal of Economic Growth, 2002, 7(3): 195-225.

16

-.8

-.6

-.4

-.2

.0

.2

.4

.6

.8

-1.6 -1.2 -0.8 -0.4 0.0 0.4 0.8 1.2

Difference in log mean

Diff

eren

ce in

log

Gin

i inde

x

r = –0.13

17

-15

-10

-5

0

5

10

15

-0.2 -0.1 0.0 0.1 0.2

Annualized change in log mean

Ann

ualiz

ed c

hang

e in

abs

olut

e G

ini i

ndex

-10

-5

0

5

10

-0.2 -0.1 0.0 0.1 0.2

Annualized change in log mean

Ann

ualiz

ed c

hang

e in

rel

ativ

e G

ini i

ndex

Absolute GiniRelative Gini

Further reading: Martin Ravallion, “Competing Concepts of Inequality in the Globalization Debate,” Brookings Trade Forum 2004, Edited by Susan Collins and Carol Graham, Washington DC: Brookings Institution, pp.1-38.

However, not true for absolute inequalitySame data, but very different pictures

18

-1.0

-0.5

0.0

0.5

1.0

-0.3 -0.2 -0.1 0.0 0.1 0.2

Proportionate change in survey mean

Pro

por

tiona

te c

hang

e in

the

$1

/day

pov

erty

rate

Economic growth is typically “pro-poor”

Slope = -2

19

But the extent to which growth is pro-poor has varied enormously between countries

and over time

• A 1% rate of growth will bring anything from a modest drop in the poverty rate of 0.6% to a more dramatic 3.5% annual decline (95% CI).

• There have been plenty of cases of rising inequality during spells of growth. Indeed, inequality increases about half the time

Further reading: Martin Ravallion, “Growth, Inequality and Poverty: Looking Beyond Averages,” World Development, Vol. 29(11), November 2001, pp. 1803-1815.

-1.0

-0.5

0.0

0.5

1.0

-0.3 -0.2 -0.1 0.0 0.1 0.2

Proportionate change in survey mean

Pro

por

tiona

te c

hang

e in

the

$1

/day

pov

erty

rate

20

High inequality is an impediment to pro-poor growth

-20

-10

0

10

20

20 30 40 50 60 70Initial Gini index (%)

Gro

wth

ela

stic

ity o

f pov

erty

redu

ctio

n

Mean elasticity close to zero in high inequality countries

21

High inequality is an impediment to pro-poor growth

• Even when inequality is not changing, it matters to the rate of poverty reduction

• It is not the rate of growth that matters, but the distribution-corrected rate of growth

Rate of poverty reduction =[constant x (1 – inequality) ] x

growth rate

Further reading: Martin Ravallion, “Inequality is Bad for the Poor,” in Inequality and Poverty Re-Examined, edited by John Micklewright and Steven Jenkins, Oxford: Oxford University Press, forthcoming.

22

Rate of poverty reduction with a 2% rate of growth in per capita income and

a headcount index of 40%

• Low-inequality country (Gini=0.30): the headcount index will be halved in 11 years.

• High inequality country (Gini=0.60): it will then take 35 years to halve the initial poverty rate.

• Note: the argument works in reverse: high inequality protects the poor from negative macro shocks.

Is poverty itself the key aspect of initial distribution?

23

• In general, the growth elasticity of poverty reduction will depend on the initial distribution.

• Past work has focused on inequality as the relevant aspect of initial distribution.

• However, poverty may well be the more relevant proximate factor.

• By this view, inequality still matters, but only via its bearing on initial poverty.

Rate of poverty reduction =[constant x (1 – poverty) ] x

growth rate

Regressions for rate of poverty reduction

24

(1) (2) (3) (4) (5) (6)OLS IVE OLS IVE OLS IVE

Intercept 0.002(0.078)

0.008(0.267)

-0.012(-1.908)

-0.005(0.607)

-0.012**(-2.175)

-0.008(-1.365)

Initial poverty rate -0.004(-0.792)

0.008(0.267)

0 0 0 0

Growth rate (annualized change in log mean)

-2.674**(-6.660)

-3.564**(-4.325)

-2.615**(-6.608)

-3.323**(-4.560)

0 0

Growth rate interacted with initial poverty rate

2.780**(5.206)

3.492**(3.650)

2.621**(4.915)

3.101**(3.746)

0 0

(1-Poverty rate) times growth rate

0 0 0 0 -2.613**(-7.273)

-3.294**(-4.585)

N 91 86 91 86 91 86R2 0.537 0.439 0.535 0.458 0.535 0.466Homogeneity test 0.673 -0.215 0.037 -0.620 n.a. n.a.

Homogeneity tests passes => the relevant growth rate is the poverty-adjusted rate, as given by the growth rate times one minus the poverty rate.

ititiititi gHHg )()1()(

Rate of poverty reduction is proportional to the distribution-corrected rate of growth

25

-.4

-.3

-.2

-.1

.0

.1

.2

.3

-.10 -.05 .00 .05 .10

Poverty-adjusted rate of growth in survey mean

Annu

aliz

ed p

ropo

rtion

ate

rate

of p

over

ty re

duct

ion

ititiititi gHHg )()1()(

Poverty makes growth less pro-poor• The (absolute) growth elasticity of poverty reduction tends to

be lower in countries with a higher initial poverty rate. • Poorer countries tend to experience lower proportionate

effects on their poverty measures from any given rate of growth.

• At an initial poverty rate of 10% (about one standard deviation below the mean) the elasticity is about -3 while it falls to about -0.7 at a poverty rate of 80% (about one standard deviation above the mean).

26

27

High poverty and inequality=>Low growth

Does inequality impede growth?Does poverty impede growth?Is poverty self-perpetuating?

Theories of distribution-dependent growth

• Inequality restricts efficiency-enhancing cooperation amongst people, such that public goods needed for growth are underprovided or efficiency-enhancing policy reforms are blocked (Bardhan et al., 2000).

• Political-economy models of redistribution argue that high inequality leads democratic governments to implement distortionary redistributive policies, e.g., Alesina and Rodrik (1994).

28

Theories based on credit-market failures

• Market failure attributed to information asymmetries, notably that lenders are imperfectly informed about borrowers.

• Key analytic feature: a suitably nonlinear relationship between an individual’s initial wealth and her future wealth (the “recursion diagram”). – With diminishing marginal products of capital, the mean future wealth will be

a quasi-concave function of the distribution of current wealth– Thus higher current inequality implies lower future mean wealth at a given

value of current mean wealth, i.e., lower growth. – Examples: Galor and Zeira (1993), Benabou (1996), Aghion and Bolton (1997)

and Banerjee and Duflo (2003).

29

A neglected implication of such models: Poverty also impedes growth

Some theoretical models => Poverty itself can impede growth and (hence) poverty reduction

1. Lasting (adverse) productivity effects of poor nutrition, esp., in childhood (Dasgupta and Ray, 1986; Cunha and Heckman, 2007).

2. Lopez and Servén (2009) introduce a subsistence consumption requirement into the utility function in the model of Aghion et al. (1999) and show that higher poverty incidence (failure to meet the subsistence requirement) implies lower growth.

30

Implications of borrowing constraints

3. Banerjee and Duflo (2003) provide a simple but insightful growth model with a borrowing constraint. – Those with sufficient wealth will reach their unconstrained optimum,

equating the marginal product of capital with the interest rate. – But the “wealth poor,” for whom the borrowing constraint is binding, will

not be able to do so. – Banerjee and Duflo show that higher inequality in such an economy

implies lower growth. • However, their model also implies that higher current wealth

poverty for a given mean also implies lower growth. • In the Banerjee-Duflo model an unambiguously higher initial

headcount index of poverty holding the initial mean constant implies a lower growth rate.

31

Further reading: Martin Ravallion, “Why Don’t we See Poverty Convergence?” Policy Research Working Paper 4974, World Bank.

Past evidence on growth and inequality

• Empirical support from cross-country growth regressions for the view that a Gini index of inequality impedes growth at country level. – Alesina and Rodrik (1994), Persson and Tabellini (1994), Birdsall et al.,

(1995), Clarke (1995), Perotti (1996), Deininger and Squire (1998) and Knowles (2005)

• However, not all the evidence has been supportive; also see Li and Zou (1999), Barro (2000) and Forbes (2000).

• The main reason why these studies have been less supportive appears to be that they have allowed for country-level fixed effects. Limited power for FE models with these data.

32

Poverty and growth• Benchmark regression:

where

• The regression is consistent with a derivative of current mean with respect to lagged mean that is less than unity, but fades toward zero at sufficiently long gaps between survey rounds.

33

ititititi Hg ˆln017.0ln035.0234.0)()626.3()131.5()183.5(

/)/ln()( itititig

Further reading: Martin Ravallion, “Why Don’t we See Poverty Convergence?” Policy ResearchWorking Paper 4974, World Bank.

Alternative specifications

34

(1) (2) (3)Sample with two surveys

Full sample Consumption surveys only

NAS consumption per

capitaIntercept 0.024**

(5.183)0.300**(5.850)

0.151**(3.705)

Log initial mean -0.035**(-5.131)

-0.044**(-5.318)

-0.020**(-3.037)

Log initial headcount index

-0.017**(-3.626)

-0.025**(-4.845)

-0.011**(-2.711)

R2 0.147 0.201 0.128N 92 70 81

ititititi Hg lnln)(

Using countries with 3+ surveys

35

(4) (5) (6)Means from first

two surveys used as initial conditions

GMM estimator with IVs from earliest survey

rounds

As for (5) but using NAS consumption instead of survey

meansIntercept 0.235**

(4.569) 0.180**(2.772)

0.169**(3.517)

Log initial mean -0.029**(-3.264)

-0.020*(-1.994)

-0.014*(-2.017)

Log initial headcount index

-0.022**(-6.305)

-0.020**(-3.381)

-0.022**(-4.749)

N 77 64 58

Encompassing regressions

36

(1) (2) (3) (4)Survey Means

Consumption from NAS

Survey Means

Consumption from NAS

Intercept 0.442(0.795)

0.617(1.234)

0.26(1.279)

-0.275(-1.914)

Initial mean -0.058**(-5.961)

-0.035**(-3.657)

-0.060**(-6.912)

-0.030**(-3.764)

Initial poverty rate -0.027**(-5.482)

-0.017**(-3.033)

-0.027**(-5.750)

-0.014**(-3.024)

Initial Gini index -0.020(-0.400)

-0.081(-1.784)

0 0

Initial income share of middle three quintiles

-0.117(-1.477)

-0.167*(-2.167)

-0.091**(-3.985)

0

Initial share of population in Western middle class

-0.102*(-2.284)

-0.128**(-2.815)

-0.110**(-2.432)

-0.133**(-3.691)

Initial primary school enrolment rate

0.007(0.700)

0.003(0.271)

0 0

Initial life expectancy 0.117**(2.768)

0.154**(3.653)

0.129**(3.068)

0.139**(3.665)

Initial price of investment -0.014**(-2.650)

-0.016**(-3.140)

-0.014**(-2.698)

-0.017**(-3.434)

N 0.434 0.470 0.430 0.453R2 88 84 88 87

Aggregate results hide much of importance

• What aspect of “initial distribution” matters to subsequent growth and poverty reduction?

• Inequality or poverty? Poverty may well be the more important factor, but inequality still matters via poverty

• Inequality of results or inequality of opportunity? Inequality of opportunity may well be more relevant

• Inequality/poverty in what dimension (income vs. non-income factors)?

37

Conclusions

38

Growth=>Poverty reduction

• Growth typically comes with lower absolute poverty.• Although the same rate of growth can have very different

implications for progress against poverty.• High initial inequality (income and non-income dimensions)

and/or high initial poverty mean that growth is less effective against absolute poverty.

• The main way initial inequality matters is through poverty.• A high initial incidence of poverty entails a lower subsequent

rate of progress against poverty at a given growth rate (and poor countries tend to experience less steep increases in poverty during recessions).

39

Higher initial poverty=>Lower growth

• Consistently with theoretical models of economic growth incorporating borrowing constraints, there is an adverse effect on consumption growth of high initial poverty at a given mean.

• For many poor countries, the growth advantage of starting out with a low mean ("conditional convergence") is lost due to their high poverty rates.

• High current inequality is only a handicap if it entails a high incidence of poverty relative to mean consumption.

40

41

What does all this mean for development policy?

Traditional view: • “Rising inequality is inevitable to some degree.”

– Arthur Lewis: “Development must be inegalitarian because it does not start in every part of the economy at the same time.”

• “Policy makers should focus on promoting economic growth.” (Possibly with occasional attention to add-on compensatory redistributive policies.)

The research reviewed here casts doubt on this traditional view

• Inequality can be a constraint on growth and (hence) poverty reduction.

• But not all “inequalities” matter: The key aspect of initial distribution in poor countries is poverty itself.

• Policies that act directly to reduce poverty—including those that make markets work better for the poor—can be seen as important elements of longer-term poverty reduction through economic growth.

42