Global income inequality today

Branko Milanovic

Perugia, June 2009

Email: bmilanovic@worldbank.org

Based on the book Worlds Apart, 2005 and updates

BM note: this is an update of moscow2.ppt

1. Inequalities today

Three concepts of inequality definedConcept 1 inequality

Concept 2 inequality

Concept 3 (global) inequalty

Inequality, 1950-2006:The mother of all inequality disputes

0.4

0.5

0.6

0.7

1950

1952

1954

1956

1958

1960

1962

1964

1966

1968

1970

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

Gin

i co

effi

cien

t

Weighted international inequality without China

Weighted international inequalty (Concept 2)

Global inequality (Concept 3)

Unweighted international inequality (Concept 1)

Three concepts of inequality using the new 2005 PPPs (1950-2007)

.45

.55

.65

.75

1950 1960 1970 1980 1990 2000 2010year

global_gini2 Global Concept 1 inequality

1982 1990 2000

Graph in interyd\dofiles\defines.do

The impact of new PPPs

• Concept 2 inequality increases by almost 10 Gini points (a level shift)

• Somewhat steeper decline of Concept 2 inequality in the last decade (because India and China now appear poorer)

• About 5 Gini points increase in Concept 1 inequality (shift effect; no trend effect)

• About 5 Gini points increase in global inequality

(cont.)

• World poorer than thought, Asia in particular

• Inequality (in all formulations) greater

• Two engines of “global equalization”: China and India

Concept 1 and Concept 2 inequality

Concept 2

Concept 2 without China

Concept 1

.45

.5.5

5.6

.65

1940 1960 1980 2000 2020year

Graph in interyd\dofiles\defines.do

2. Inequality between world citizens today

Methodological issues

• GDI per capita or HS mean• Definitional difference (H&E, undisbursed profits)

and• Practical difference (under-surveying of the rich

and under-reporting of property Y)• Mixing of the two biases both poverty and

inequality down• Moreover, movements in NA and HS statistics

are different • If HS mean is it HSY or HSX?

Methodological issues (cont.)

• Even if HS welfare indicator is selected definitions of X,Y vary in time & btw. countries

• Issues: self-employed Y, home C, imputation of housing, treatment of publicly provided H&E, use of top coding, under-estimation of property incomes

• What PPP to use• Equivalence scales & intra-HH inequality

The difficulty stems from contradictory movements

• (1) Greater inequality within nations

• (2) Greater differences between countries’ mean incomes (think of US vs. Africa)

• (3) But catching up of large and poor countries (China and India)

• All of these forces determine what happens to GLOBAL INEQUALITY (but they affect it differently)

3. First calculations of global inequality from household survey

data alone

Population coverage

1988 1993 1998 2002

Africa 48 76 67 77

Asia 93 95 94 96

EEurope 99 95 100 97

LAC 87 92 93 96

WENAO 92 95 97 99

World 87 92 92 94

Non-triviality of the omitted countries (Maddison vs. WDI)

GDI (US dollar) coverage

1988 1993 1998 2002

Africa 49 85 71 71

Asia 94 93 96 95

EEurope 99 96 100 99

LAC 90 93 95 95

WENAO 99 96 96 100

World 96 95 96 98

Number of surveys (C-based)

1988 1993 1998 2002

Africa 14(11) 30(27) 24(24) 30(29)

Asia 19(10) 26(18) 28(20) 26(18)

EEurope 27(0) 22(0) 27(14) 26(16)

LAC 19(1) 20(4) 22(2) 21(1)

WENAO 23(0) 23(0) 21(3) 21(2)

World 102(22) 121(52) 122(63) 124(66)

1988 1993 1998 2002

International dollars ($PPP)

Gini index

68.4

(2.0)

70.0

(1.4)

69.4

(1.8)

70.8

(1.4)

Between compnt

61.8 62.6 62.1 63.9

US dollars

Gini index

77.8

(1.5)

80.4

(1.4)

79.6

(1.3)

81.0

(1.1)

Global inequality (with 2005 PPPs)(distribution of persons by $PPP or US$ income per capita)

4. Importance of the differences in countries’ mean incomes

Concept 1 inequality in historical perspective: Convergence/divergence during different

economic regimes

0

10

20

30

40

50

60

1820 1870 1890 1900 1913 1929 1938 1952 1960 1978 2000

Gini

Theil

FIRST GLOBALIZATION DEGLOBALIZATION

WAR DEVELOPMENTAL STATE

NEOLIBERAL

• In Gini terms:

• where Gi=individual country Gini, π=income share, yi = country income, pi = population share, μ=overall mean income, n = number of countries, L=overlap term

LppyypG j

n

ij

iij

n

i

n

i

iii

)1

1

How are Concepts 2 and 3 related?

Concept 2

A non-Marxist world

• Over the long run, decreasing importance of within-country inequalities despite the reversal in the last quarter century,

• Increasing importance of between-country inequalities (with some hopeful signs in the last five years, before the current crisis),

• Global division between countries more than between classes

Composition of global inequality changed: from being mostly due to “class” (within-national), today it is mostly due to “location” (where people live; between-national)

0

10

20

30

40

50

60

70

80

90

1870 2000

C lass

Location

Location

Class

1870

2000

Based on Bourguignon-Morrisson (2002) and Milanovic (2005)

A literary illustration: Elizabeth’s dilemma (from Pride and Prejudice)

Income in 1810 (£ pa)

Approx. position in 1810 income distribution

Mr. Darcy

10,000 Top 0.1%

Elizabeth’s family

3000/7~430 Top 1%

Elizabeth alone

50 Median

Gain 100 to 1

Income around Y2K (£ pc pa)

270,000

57,000

6,500

20 to 1

1810 position estimates based on Colquhoun 1801-3 data. Y2K data from LIS (UK1999), and for 0.1% from Piketty (Data-central).

Define four worlds:

• First World: The West and its offshoots• Take the poorest country of the First World

(e.g. Portugal)• Second world (the contenders): all those

less than 1/3 poorer than Portugal.• Third world: all those 1/3 and 2/3 of the

poorest rich country.• Fourth world: more than 2/3 below

Portugal.

Four Worlds in 1960

Four Worlds in 2003

Growth over 1980-2002 period as function of initial (1980) income

Population according to income of country where they live (2007): an empty middle

histogram gdpppp [w=popu] if year==2007 & gdpppp<50000 & Dcont==1, bin(20) percent ylabel(0(10)40) xtitle(GDP per capita in 2005 PPP)

010

20

30

40

Perc

ent

0 10000 20000 30000 40000 50000GDP per capita in 2005 PPP

USAW Europe, JapanBrazil, Russia, Mexico

China

India, Indon, Bgd

The key borders today

• First to fourth world: Greece vs. Macedonia and Albania; Spain vs. Morocco (25km), Malaysia vs. Indonesia (3km)

• First to third world: US vs. Mexico.

In 1960, the only key borders were Argentina and Uruguay (first) vs. Brazil, Paraguay and Bolivia (third world), and Australia (first) vs. Indonesia (fourth)

Year 2007 Year 1980

Approximate % of foreign workers in labor force

Ratio of real GDI per capita

Greece (Macedonian/Albanians)

7.5 4 to 1 2.1 to 1

Spain (Moroccans)

14.4 7.4 to 1 6.5 to 1

United States (Mexicans)

15.6* 3.6 to 1 2.6 to 1

Malaysia

(Indonesians)

>14.0 3.7 to 1 3.6 to 1

* BLS, News Release March 2009; data for 2008 inclusive of undocumented aliens.

5. Global inequality (cont.)

More than fifty-fifty world (new PPPs)

Cumulative % of world population

Cumulative % of PPP world income/consumption

In a single country (UK)

5 0.24

10 0.6 2.0

25 2.1

50 6.6 25.0

75 17.8

90 42 71.5

Top 10 58 28.5

Top 5 38.6 18.4

How big is a Gini of 70? (Year 2002, 2005PPPs)

Top Bottom Ratio

In PPP dollars

5 percent 38% 0.24% 165-1

10 percent 58% 0.6% 95-1

In current $

5 percent 45% 0.15% 300-1

10 percent 67.5% 0.45% 150-1

10 top countries 31,850 580 55-1

Different countries and income classes in global income distribution (year 2002; new PPPS)

Germany

USA

Russia Brazil

India

110

2030

4050

6070

8090

100

perc

entil

e of

wor

ld in

com

e di

strib

utio

n

1 10 20 30 40 50 60 70 80 90 100country percentile

Note…• Richest people in India barely intersect with

poorest people in Germany • Bottom 20% of Americans worse off than

equivalent people in Germany• But this is not true for Brazil and Russia: about

half of the population of Brazil better off than the very poorest percentile in Germany; for Russia, it is 4/5.

• Russian better-off than Brazilians except at the top (note convexity at the top in Brazil)

• Important later for rules re. global transfers

Distribution of percentile of global income distribution across five world regions (02 WYD)

020

40

60

80

100

glo

bal p

erc

entil

e if

whole

==

1

Africa Asia Latin America Eastern Europe West

. graph box inc_c if maxgroup==20, over(region); use world2002.dta

GermanyItaly

Russia

Serbia

Hungary

110

2030

4050

6070

8090

100

perc

entil

e of

wor

ld in

com

e di

strib

utio

n

1 5 10 15 20country ventile

GDP per capita and Gini

HUNHUNHUNHUNHUNHUNHUNHUNHUNHUNHUNHUNHUNHUNHUNHUNHUNHUNHUNHUN

ARMARMARMARMARMARMARMARMARMARMARMARMARMARMARMARMARMARMARMARM UKRUKRUKRUKRUKRUKRUKRUKRUKRUKRUKRUKRUKRUKRUKRUKRUKRUKRUKRUKRCZECZECZECZECZECZECZECZECZECZECZECZECZECZECZECZECZECZECZECZESVKSVKSVKSVKSVKSVKSVKSVKSVKSVKSVKSVKSVKSVKSVKSVKSVKSVKSVKSVKBGRBGRBGRBGRBGRBGRBGRBGRBGRBGRBGRBGRBGRBGRBGRBGRBGRBGRBGRBGRSRBSRBSRBSRBSRBSRBSRBSRBSRBSRBSRBSRBSRBSRBSRBSRBSRBSRBSRBSRBBLRBLRBLRBLRBLRBLRBLRBLRBLRBLRBLRBLRBLRBLRBLRBLRBLRBLRBLRBLR SVNSVNSVNSVNSVNSVNSVNSVNSVNSVNSVNSVNSVNSVNSVNSVNSVNSVNSVNSVNKGZKGZKGZKGZKGZKGZKGZKGZKGZKGZKGZKGZKGZKGZKGZKGZKGZKGZKGZKGZ HRVHRVHRVHRVHRVHRVHRVHRVHRVHRVHRVHRVHRVHRVHRVHRVHRVHRVHRVHRVLTULTULTULTULTULTULTULTULTULTULTULTULTULTULTULTULTULTULTULTUKAZKAZKAZKAZKAZKAZKAZKAZKAZKAZKAZKAZKAZKAZKAZKAZKAZKAZKAZKAZBIHBIHBIHBIHBIHBIHBIHBIHBIHBIHBIHBIHBIHBIHBIHBIHBIHBIHBIHBIHALBALBALBALBALBALBALBALBALBALBALBALBALBALBALBALBALBALBALBALB

TJKTJKTJKTJKTJKTJKTJKTJKTJKTJKTJKTJKTJKTJKTJKTJKTJKTJKTJKTJK ROMROMROMROMROMROMROMROMROMROMROMROMROMROMROMROMROMROMROMROMMDAMDAMDAMDAMDAMDAMDAMDAMDAMDAMDAMDAMDAMDAMDAMDAMDAMDAMDAMDAUZBUZBUZBUZBUZBUZBUZBUZBUZBUZBUZBUZBUZBUZBUZBUZBUZBUZBUZBUZB MNEMNEMNEMNEMNEMNEMNEMNEMNEMNEMNEMNEMNEMNEMNEMNEMNEMNEMNEMNE

POLPOLPOLPOLPOLPOLPOLPOLPOLPOLPOLPOLPOLPOLPOLPOLPOLPOLPOLPOLLVALVALVALVALVALVALVALVALVALVALVALVALVALVALVALVALVALVALVALVAMKDMKDMKDMKDMKDMKDMKDMKDMKDMKDMKDMKDMKDMKDMKDMKDMKDMKDMKDMKDRUS ESTRUS ESTESTESTESTESTESTESTRUSRUSRUSRUS ESTESTRUSRUSRUSRUS ESTRUS ESTRUS ESTESTRUS ESTRUSRUS ESTESTRUSRUSRUS ESTRUS ESTESTRUS

GEOGEOGEOGEOGEOGEOGEOGEOGEOGEOGEOGEOGEOGEOGEOGEOGEOGEOGEOGEO

20

30

40

50

60

gin

i

1000 2000 10000 40000gdp per capita, PPP (Constant 2005 international $)

Global inequality of opportunity

• How much of variability of income person’s global income can we explain with two circumstances only: person’s country of citizenship and income class of his/her parents?

• Both circumstances basically given at birth• With citizenship person receives several public

goods: income of country, its inequality level, and its intergenerational income mobility.

Base case Optimistic Pessimistic

Mean country income ($PPP; in logs)

0.99

(0)

0.99

(0)

0.99

(0)

Gini index -0.019

(0)

-0.019

(0)

-0.019

(0)

Parents’ national income class (from 1 to 20)

0.105

(0)

0.100

(0)

0.110

(0)

Adj. R2 0.81 0.80 0.83

Based on 116 countries, 20 ventiles for each, year 2002

The answer is: about 80 percent!The answer is: about 80 percent!(dependent variable: HH per capita income in $PPP )(dependent variable: HH per capita income in $PPP )

Base (optimistic, pessimistic) case about intergenerational income mobility in different parts of the world

6. Global financial crisis and global inequality

15 largest annual GDP per capita declines in the United States history

-0.3

-0.25

-0.2

-0.15

-0.1

-0.05

0

1946 1932 1908 1930 1914 1931 1893 1945 1894 1938 1921 1917 1896 1947 1904

Financial crisis 1893-96: 16%

Great Depresssion: 25%

End of war economy, 34%

Plutocratic and people’s recessions are not the same

thing

Plutocratic and people’s global growth rate in years of plutocratic recessions

-0.080

-0.060

-0.040

-0.020

0.000

0.020

0.040

0.060

0.080

1930 1931 1932 1954 1958 1960 1975 1982 1991

Plutocratic growth rate

People's growth rate

Shrinking GDPs People living in countries with negative growth rates (in milion)

(years of plutocratic recession in yellow)

0

500

1000

1500

2000

2500

Who is affected more: poor or rich countries?

The financial crisis 1890-95: relationship between initial GDP per capita and its

change during the crisis

AUS

AUT

BEL

CAN

CHEDEU

DNK

ESP

FIN FRA

GBR

ITA

NLD

NOR

NZL

PRT

SWEUSA

.8.9

11.1

ratio betw

een G

DPs

1000 2000 3000 4000 5000GDP per capitain 1890

The Great Depression 1928-33: relationship between initial GDP per capita and its

change during the crisis

AUS

AUT

BEL

CAN

CHE

DEU

DNK

ESP

FIN

FRA

GBRGRC IRL

ITA

NLD

NOR

NZL

PRT

SWE

USA.8.9

11.1

1.2

ratio betw

een G

DPs

2000 3000 4000 5000 6000 7000GDP per capitain 1929

World, by population, in the crisis 1930

twoway (scatter dlngdpppp laggdpppp if year==1930 [w=pop]) (lowess dlngdpppp laggdp> ppp if year==1930), yline(0) ytitle(growth rate) legend(off) xtitle(GDP per capita > in 1929) [from maddison_polity2.dta

-.2

-.1

0.1

grow

th r

ate

6 7 8 9GDP per capita in 1929

Unweighted correlation coefficient between growth rate and GDI per capita (in years of global recession)

-0.25

-0.20

-0.15

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

1930 1931 1932 1954 1958 1960 1975 1982 1991

Negative coeff: rich countries decline more.

World, by population, in the crisis 1991

-.1

-.05

0.0

5.1

grow

th r

ate

6 7 8 9 10GDP per capita in 1990

twoway (scatter dlngdpppp laggdpppp if year==1991 & dlngdpppp<0.1 & dlngdpppp>-0.1 [w=pop]) (lowess dlngdpppp laggdpppp if year==1991), yline(0) ytitle(growth rate) legend(off) xtitle(GDP per capita in 1990) [from maddison_polity2.dta]

But after 1973-75, the Third World took 25 years to recover

Average people-weighted growth rate in Third World (excl. China)

0.0000

0.0100

0.0200

0.0300

0.0400

0.0500

0.0600

Implications

• First negative effect will be mostly felt by rich people (in particular, those with high financial assets) and rich countries

• Rates of GDP will decline more in rich countries• Then, the crisis will spread wider• Difficult to predict next effects: prices of raw

materials, dependence on exports, amount of debt

• Possible replay of the period 1980-2000 when growth rate of Third World declined by more than 1 percentage point

6. Globalization and income inequality

Causal effect of globalization (openness) on global inequality

• Channel 1. Different effect on within-national income distributions (difference between poor and rich countries; HOS and revisions)

• Channel 2. Different effect on growth rates of poor and rich countries (the openness premium should be higher for poor countries)

• Channel 3. Different effect on populous and small countries

• Depends on history: are populous countries rich or poor at a given point in time?

• Assume globalization is good for for poor, populous countries, no effect on within-national distribution

• In the current constellation, India and China grow faster => global inequality ↓ (mean income convergence, lower global inequality)

• Decouple poor and populous; let China and India be rich

• No change in individual effects of gloablization; mean convergence continues but global inequality may now go ↑

• Conclusion. Even if effects are known and unchanged, the outcome may differ.

Conclusion: “The age of inequality”?

Inequalities between countries have increased

Population weighted inequality between countries went down thanks to fast growth in China and India (Caveat: R/U differences in China and India have global implications)

Inequality among people in the world is very high (Gini around 70) but its direction of change is not clear

Within-country inequalities have increased in many countries including in the largest (US, UK, China, India, Russia)

7. Does Global Inequality Matter?

• No one in “charge” of it; there is no global government

• No one can do much about it

• No global taxation authority

Does global inequality matter?

• NO, according to Ann Krueger (2002):

“Poor people are desperate enough to improve their material conditions in absolute terms rather than to march up the income distribution. Hence it seems far better to focus on impoverishment than on inequality.”

• YES, according to Kuznets (1954)“…reduction of physical misery associated with

low income and consumption levels…permit[s] an increase…of political tensions”

BECAUSE

“the political misery of the poor, the tension created by the observation of the much greater wealth of other communities…may have only increased.”

What may be the effects of global inequality?

• Globalization increases awareness of differences in living standards (aspiration level changes; empirical studies show it)

• Leads to migration

• Greater likelihood of conflict (Jennifer Government)

We need some rules for global transfers

• They should flow from a rich to a poor country. That is easy.

• But they have to satisfy the same rules as at the national level, i.e.

• transfers should be globally progressive, that is flow from a richer person to a poorer person.

In addition transfers have national income inequality implications

Progressive transfer at the global level and worsening national distributions (may not

be politically sustainable)

T B

Income

Income distribution in poor country

Income distribution in rich country

Thus transfers have to satisfy

• Progressivity 1: reduce mean income differences between rich and poor countries

• Global progressivity: tax payers should be richer than beneficiaries

• National progressivities: in rich country, tax payers should be relatively rich (reduce rich country inequality) and in poor country, beneficiaries should be relatively poor (reduce poor country inequality)

• Book “Worlds Apart: Measuring International and Global Inequality”, Princeton UP, 2005

• Email: bmilanovic@worldbank.org

• Website: http://econ.worldbank.org/projects/inequality

Extra slides

1988 1993 1998 2002

Between country Gini (PPP dollars)

61.8 62.6 62.1 63.9

Share of total inequality (in %) 90 89 89 90

Between country Gini (US dollars)

69.5 71.7 70.8 73.3

Share of of total inequality (in %) 89 90 89 91

Share of between-country inequality in total inequality (05 ICP)

The rich and the poor (equal total income)

57, 3316

29, 2531

3, 990

0

500

1000

1500

2000

2500

3000

3500

0 10 20 30 40 50 60

Rich people (in million)

Po

or

peo

ple

(in

mil

laio

n)

Top US percentile

Top US decile

Top global 1%

Percentage of global population living in countries with negative growth (in years of global plutocratic

recession)

0

10

20

30

40

50

60

70

1930 1931 1932 1954 1958 1960 1975 1982 1991

% o

f p

op

ula

tio

n li

vin

g in

co

un

trie

s w

ith

neg

ativ

e g

row

th