Global income inequality: current trends and future developmentspubdocs.worldbank.org/en/836191550771613249/BMilanovic... · Global income inequality: current trends and future developments

Global income inequality: current trends and future

developments

Branko Milanovic, Lead Economist Policy Research Talk

October 15, 2013

World Bank Development Research Group

Washington, DC

Branko Milanovic

A. National inequalities mostly increased

Branko Milanovic

Ginis in 1988 and twenty years later

1988 2008 Change

Average Gini 36.0 38.5 +2.5

Pop-weighted Gini

33.9 37.3 +3.4

GDP-weighted Gini

32.2 36.4 +4.2

Countries with higher Ginis (38)

33.7 38.5 +4.8

Countries with lower Ginis (20)

40.5 37.7 -2.7

From final-complete3.dta and key_variables_calcul2.do Branko Milanovic

Ginis in 1988 and 2008

From key_variables_calcul3.do

ARG

AUTBEL

BGD

BGR

BOLBRA

CAN

CHL

CHN-R

CHN-U

CIV

COL

CRI

CZE

DEU

DNK

DOMECU

EGYESP

FIN

FRAGBRGRC

HND

HUN

IDN-R

IDN-U

IND-R

IND-U

IRL

ISR

ITA

JOR

JPN

KOR

LKA

LVA

MEX

MRT

MYS

NGA

NLDPAK

PHL

POL

PRT

PRY

SLV

SVKSVN

THATUR

UGA

URY

USAVEN

20

30

40

50

60

Gin

i in

200

8

20 30 40 50 60Gini in 1988

Branko Milanovic

Ginis in 1988 and 2008 (population-weighted countries)

From key_variables_calcul3.do

Branko Milanovic

RUS

IND-U

MEX

BRA

NGA

IND-R

USA

CHN-U

CHN-R

20

30

40

50

60

Gin

i in

200

8

20 30 40 50 60Gini in 1988

Inequality and income, 1950-2012

Use fc:…\finance_and_development\figure2.do Use gdpppppreg4.dta

ChinaUnited States

Brazil

Russia

30

40

50

60

70

Gin

i coe

ffic

ien

t

1000 5000 10000 40000GDP per capita in PPP dollars

Issues raised by growing national inequalities

• Inequality as one of the causes of the global financial crisis.

• Real inequality greater than measured inequality because rich people refuse to participate in surveys and hide their income and assets (growing issue)

• Perception of inequality outstrips real increase because of globalization, role of social media and political (crony) capitalism (example of Egypt)

• The effect on the political system

Branko Milanovic

Difficult to explain developments using a single theory

Branko Milanovic

No downward portion plotted against time or income: example of China 1964-2007

Against income

twoway (scatter Giniall gdpppp if contcod=="CHN" , connect(l) ylabel(30(5)50)) (qfit Giniall

gdpppp if contcod=="CHN")

From global_new2.dta

Against time

twoway (scatter Giniall year if contcod=="CHN" & year>1960,

connect(l) ylabel(30(5)45)) (qfit Giniall year if contcod=="CHN" &

year>1960)

Based on giniall.dta

30

35

40

45

1960 1970 1980 1990 2000 2010year when the survey was conducted

Combined Gini coefficient Fitted values

30

35

40

45

50

0 2000 4000 6000 8000constant 2005 ppp, based on icp05

Gini from my allGini file Fitted values

No downward portion plotted against time or income: example of the USA, 1950-2010

twoway (scatter Giniall year if contcod=="USA", connect(l) ylabel(30(5)45)) (qfit

Giniall year if contcod=="USA" & year>1960, legend(off))

From allginis.dta.

twoway (scatter Giniall gdpppp if contcod=="USA" , connect(l) ylabel(30(5)50)) (qfit Giniall

gdpppp if contcod=="USA", legend(off))

Using gdppppreg4.dta

30

35

40

45

1940 1960 1980 2000 2020year when the survey was conducted

30

35

40

45

50

10000 20000 30000 40000 50000gdpppp 2005 icp; for years<1980, old growth rates

B. Between national inequalities increased (if unweighted) or

decreased (if population weighted—the role of China)

Branko Milanovic

Unweighted GDP per capita (in $PPP) in rich world and Sub-Saharan Africa

Africa

WENAO1

00

05

00

02

00

00

400

00

1940 1960 1980 2000 2020year

twoway (scatter aa year if contcod=="USA", yscale(log) ylabel(1000 5000 20000 40000) legend(off) text(2000 2000 "Africa") text(40000 2000 "WENAO")) (scatter bb year if contcod=="NGA", xline(1976, lpattern(dash)) xline(2000, lpattern(dash))) Using interyd.dta

Different countries and income classes in global income distribution in 2008

From calcu08.dta

USA

India

Brazil

China

Russia

1

10

2

0

30

4

0

50

6

0

70

8

0

90

1

00

p

erc

en

tile

of w

orl

d in

co

me

dis

trib

utio

n

1 20 40 60 80 100 country percentile

Branko Milanovic

C. Global inequality is the product of within- and between-county

inequalities How did it change in the last 25 years?

Branko Milanovic

Essentially, global inequality is determined by three forces

• What happens to within-country income distributions?

• Is there a convergence or not of country mean incomes?

• Are mean incomes of populous & large countries (China, India) growing faster or slower that the rich world?

Branko Milanovic

Global inequality 1950-2012: three

concepts

Branko Milanovic

Concept 2

Concept 1

Concept 3

.45

.5

5

.65

.7

5

Gin

i co

eff

icie

nt

1950 1960 1970 1980 1990 2000 2010 year

Divergence begins

Divergence ends

China moves in

International unweighted and population- weighted inequality, 1952-2010

Graph in interyd\dofiles\defines.do; using gdppppreg.dta

Concept 2

Concept 2 without China

Concept 1

.45

.5

.5

5

.6

.65

G

ini c

oef

fici

ent

in p

erce

nt

1950 1960 1970 1980 1990 2000 2010 year

India as new engine of

equalization

Branko Milanovic

Number of surveys

1988 1993 1998 2002 2005 2008

Africa 14 30 24 29 32 23

Asia 19 26 28 26 23 27

E.Europe 27 22 27 25 27 27

LAC 19 20 22 21 18 18

WENAO 23 23 21 21 22 23

World 102 121 122 122 122 118

Branko Milanovic

Population coverage

1988 1993 1998 2002 2005 2008

Africa 48 76 67 77 78 78

Asia 93 95 94 96 94 98

E.Europe 99 95 100 97 93 92

LAC 87 92 93 96 96 97

WENAO 92 95 97 99 99 97

World 87 92 92 94 93 94

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

GDI (US dollar) coverage

1988 1993 1998 2002 2005 2008

Africa 49 85 71 71 70 71

Asia 94 93 96 95 90 93

E. Europe 99 96 100 99 99 98

LAC 90 93 95 95 98 98

WENAO 99 96 96 100 100 97

World 96 95 96 98 97 95

Branko Milanovic

What does Gini of 70 mean?

twoway (scatter gini_disposable year if contcod=="SWE", c(l)) (scatter gini_disposable year if contcod=="USA“ , c (l)) (scatter gini_gross year if contcod=="BRA" & source=="SEDLAC", c(l) legend(off) text(0.30 2005 "Sweden") text(0.42 2004 "USA") text(0.63 2001 "Brazil")) (scatter gini_disposable year if contcod=="WRL", c(l) text (0.72 2005 "World")) Using data_voter_checked.dta to which I added the world from my global data

Sweden

USA

Brazil

World

.2.3

.4.5

.6.7

1970 1980 1990 2000 2010year

Branko Milanovic

Large countries and the world, from 1950-60s to today

Branko Milanovic …finance_nd_devt/figure2.do Using gdppppreg4.dta

ChinaUnited States

Brazil

Russia

World

20

30

40

50

60

70

Gin

i coe

ffic

ien

t

1000 5000 10000 40000GDP per capita in PPP dollars

How many people (ranked from the poorest to the richest) you need to get to each 1/5th of

global income?

75

13

6.2

4.1

1.7

0 10 20 30 40 50 60 70 80

1

2

3

4

5

From forpogge.xls Branko Milanovic

Concept 3 inequality with survey data rescaled by various means

summary08.xls

0.60

0.62

0.64

0.66

0.68

0.70

0.72

0.74

1988 1993 1998 2002 2005 2008

Concept 3 inequality with survey data rescaled to various means

consumption

gdpppp

HH surveys

Branko Milanovic

Discrepancy between NA and HS • Normalization by GDP per capita is wrong because GDP

includes non-welfare components like retained π, investments

• Normalization by NA consumption biases Gini down because we know that missed income is concentrated among the rich (not proportional across income distribution)*

• “True” allocation of missing income will probably result in a medium position btw. HH survey and cons. lines (because missed income is particularly high in countries like India)

*With Pareto distribution of non-compliance Ginis are not affected, but means are biased downward (Deaton). The between component of global inequality may be overestimated with low compliance among the rich (counter-intuitive result).

Branko Milanovic

D. How has the world changed between the fall of the Berlin Wall and

the Great Recession

Branko Milanovic

Real income growth at various percentiles of global income distribution, 1988-2008 (in 2005 PPPs)

0

10

20

30

40

50

60

70

80

0 10 20 30 40 50 60 70 80 90 100

x

Percentile of income distribution

From twenty_years\final\summary_data

X“US lower middle class”

X “China’s middle class”

Branko Milanovic

$PPP2

$PPP4 $PPP11

$PPP 110

Global growth at 5-year intervals, 1988-2008

-20

-10

0

10

20

30

40

50

60

70

80

5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 99 100

1988 to 1998

1988 to 2003

1988 to 2008

Branko Milanovic

Global Lorenz curves in 1988 and 2008

0

20

40

60

80

100

0 20 40 60 80 100

2008 1988 Branko Milanovic

Generalized Lorenz curve: second-order dominance of 2008 over 1988 (average cumulative income at any percentile greater in 2008 than in 1988)

Branko Milanovic

1

10

100

1000

10000

0 20 40 60 80 100

2008

1988

Ginis 1988-2008

Branko Milanovic

65

66

67

68

69

70

71

72

73

74

75

1988 1993 1998 2003 2008

Ginis, 1988-2008 with one-standard error band

twenty_years\Summary_data.dta

Conclusions

• At any percentile, income level is 2008 greater than in 1988 (first-order dominance)

• At any percentile, cumulative average income greater in 2008 than in 2008 (second-order dominance; implied by the 1st)

• No Lorenz dominance

• Gini slightly lower in 2008 than in 1988: 70.7 instead of 72.2

Branko Milanovic

0 0 1 1 1 1 1 2 2 2 3 3 4 5 4

1 3

5

10

25 27

0

5

10

15

20

25

305

10

15

20

25

30

35

40

45

50

55

60

65

70

75

80

85

90

95

99

10

0

Dis

trib

uti

on

(in

pe

rce

nt)

of

gain

ventile/percentile of global income distribution

Distribution of the global absolute gains in income, 1988-2008: more than ½ of the gains went to the top 5%

From summary_data.xls

Branko Milanovic

Quasi non-anonymous GIC: Average growth rate 1988-2008 for different percentiles of the 1988 global income distribution

Quasi non-anonymous growth between 1988 and 2008: real absolute per capita gains at different fractiles of 1988 distribution

140 217 236 443 466 799 1246 1163 2218

3168

7190

22891

0

5000

10000

15000

20000

25000

D1 D2 D3 D4 D5 D6 D7 D8 D9 V19 P95-99 P100

Ab

solu

te p

er

cap

ita

real

inco

me

gai

n b

etw

ee

n

19

88

an

d 2

00

8

Decile/fractile of 1988 global income distribution

Global income distributions in 1988 and 2008

twoway (kdensity logRRinc [w=pop] if logRRinc>2 & bin_year==2008 & keep==1 & mysample==1) (kdensity logRRinc [w=pop] if logRRinc>2 & bin_year==1988 & keep==1 & mysample==1, legend(off) xtitle(log of annual PPP real income) ytitle(density) text(0.95 2.5 "1988") text(0.85 3 "2008")) Or using adding_xlabel.do; always using final_complete7.dta

1988

20080

.2.4

.6.8

1

den

sity

300

100

0

300

0

600

0

100

00

300

00

500

00

100

00

0

log of annual PPP real income

Emerging global “middle class” between $3 and $15

Asian income distributions in 1988 and 2008

20081988

0.5

11

.5

den

sity

2 3 4 5log of annual PPP real income

twoway (kdensity logRRinc [w=pop] if logRRinc>2 & region_old==2 & bin_year==2008 & keep==1 & mysample==1) (kdensity logRRinc [w=pop] if logRRinc>2 & bin_year==1988 & keep==1 & region_old==2 & mysample==1, legend(off) xtitle(log of annual PPP real income) ytitle(density) text(0.35 2.5 "2008") text(0.3 2.2 "1988")) Using final_complete7.dta

The explosion of the Asian “global middle class”

Between-country inequality still by far the most dominant type but in future it might decline

Branko Milanovic

0.81 0.75 0.74 0.70 0.74 0.73

0.00

0.20

0.40

0.60

0.80

1.00

1988 1993 1998 2002 2005 2008

The share of the between component (using Theil 0)

Global inequality of opportunity

• Regressing (log) average incomes of 118 countries’ percentiles (11,800 data points) against country dummies “explains” 77% of variability of income percentiles

• Where you live is the most important determinant of your income; for 97% of people in the world: birth=citizenship.

• Citizenship rent.

Branko Milanovic

All equal Different (as

now)

All equal

Different (as

now)

Mean country incomes

Individual incomes within country

Global Ginis in Real World, Rawlsian World, Convergence World…and Shangri-La World (Theil 0; year 2008)

98

68 (all country Ginis=0)

30 (all mean incomes same; all country Ginis as now)

0

Branko Milanovic

E. The pattern of global inequality change differed from the pattern of

national inequality changes

Branko Milanovic

Shape of global growth vs. US growth

0

20

40

60

80

100

120

0 20 40 60 80 100

Re

al in

com

e g

row

th

Percentile of income distribution

summary_data.xls

World 1988-2008

US 1986-2008

Branko Milanovic

US pattern is not unusual: in most countries increasing gains for the rich

Philippines and Bangladesh Mexico and Colombia

50

100

150

200

250

300

co

mbin

ed r

ea

l_g

row

th 1

and

2

1 2 3 4 5 6 7 8 9 10decile

120

130

140

150

160

co

mbin

ed r

ea

l_g

row

th 1

and

2

1 2 3 4 5 6 7 8 9 10decile

MEX

COL

BGD

PHL

Branko Milanovic

Increasing gains for the rich with a widening urban-rural gap

Urban and rural China Urban and rural Indonesia

170

180

190

200

210

220

co

mbin

ed r

ea

l_g

row

th 1

and

2

1 2 3 4 5 6 7 8 9 10decile

200

250

300

350

400

450

co

mbin

ed r

ea

l_g

row

th 1

and

2

1 2 3 4 5 6 7 8 9 10decile

From key_variables_calcul2.do

Branko Milanovic

urban

rural

urban

rural

The contradiction of inequality changes during Globalization II

• Most countries displayed an upward sloping GIC (US, China, India urban, Indonesia…)

• Perception that the rich are doing better than anybody else (true)

• But growth rates of countries are uneven; those that grew the fastest were in the lower middle of global income distribution, and they were also most populous

• This led to the humped (more exactly, reclining S) shape of the global GIC and decreasing global inequality Branko Milanovic

Back to Mandeville…

• Can something that is bad nationally (increased inequality) be good globally (decreased inequality) ?

• Can national vices produce global virtue?

Branko Milanovic

Political implications

• Possible crowding out of national middle classes, and the creation of a global one

• But the middle class is presumably a force for stability when there is a political community. There is no political community at the global level. What does global middle class mean?

• Would global middle class create a global polity?

• Or, global plutocracy: in the longer-term, reversal to the pre World War I situation

Branko Milanovic

Or are we at the end of capitalism’s long “periodo especial”?

• Three challengers to global capitalism were beaten off in the 20th century: depression (by reinventing gov’t), war (by marshalling resources), Communism (through Welfare State)

• Neither of these threats is any longer present; so why can’t capitalism go back to what it once was?

• Was the 1930-1980 period capitalism’s long detour?

• Do we have to get used to permanently higher levels of inequality?

Branko Milanovic

Who gained and who lost?

Branko Milanovic

Income levels of Chinese urban and US median (fifth) decile, 1988-2008

100

03

00

01

00

00

200

00

decile

inc in

200

5 P

PP

US

D

1990 1995 2000 2005 2010benchmark year

twoway (scatter RRinc bin_year if group==6 & contcod=="CHN-U" & keep==1 & mysample==1, msize(vlarge) lwidth(thick) connect(l)) (scatter RRinc bin_year if group==5 & contcod=="USA" & keep==1 & mysample==1, msize(vlarge) lwidth(thick) udsng \final_complete2.dta

Branko Milanovic

Global percentile position of US median and Chinese urban middle decile

5254

58

62

66

9392

93 93 93

50

60

70

80

90

100

accord

ing to

RR

inc a

nd

mysa

mp

le=

=1

1990 1995 2000 2005 2010benchmark year

twoway (scatter percentile bin_year if group==6 & contcod=="CHN-U" & keep==1 & mysample==1, msize(vlarge) lwidth(thick) connect(l) mlabel(percentile)) (scatter percentile bin_year if group==5 & contcod=="USA" & keep==1 & mysample==1, msize(vlarge) lwidth(thick) connect(l) mlabel(percentile)), legend(off) ylabel(50(10)100)

Branko Milanovic

Who is now around the global median which grew the most?

• The gainers

• Chinese 8-9 rural deciles (2.5-2.7x)

• Chinese 3-4 urban deciles (2.7-2.8x)

• India median urban decile (1.6X)

• Brazil 3 decile (1.9x)

• Some who lost of stagnated

• Philippines 7-8 decile (gained 30%)

• Mexico 3rd decile (lost 25%)

Branko Milanovic

F. Global inequality over the long-run of history

Branko Milanovic

Global income inequality, 1820-2008 (Source: Bourguignon-Morrisson and Milanovic; 1990 PPPs )

Theil

Gini

02

04

06

08

01

00

1820 1860 1900 1940 1980 2020year

twoway (scatter Gini year, c(l) xlabel(1820(40)2020) ylabel(0(20)100) msize(vlarge) clwidth(thick)) (scatter Theil year, c(l) msize(large) legend(off) text(90 2010 "Theil") text(70 2010 "Gini"))

Branko Milanovic

A non-Marxist world

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

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

• Global division between countries more than between classes

Branko Milanovic

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)

Based on Bourguignon-Morrisson (2002), Maddison data, and Milanovic (2005) From thepast.xls

0

20

40

60

80

100

1870 2000

Th

eil

0 in

de

x (

me

an

lo

g d

evia

tio

n)

Class

Location

Location

Class Branko Milanovic

G. Gaps between countries today: Les jeux sont faits

when you are born?

Branko Milanovic

The importance of global vs. national locational inequalities

Unit (year) Number of units R2

World (08) 118 0.48

USA (08) 50 0.01

India (08) 35 0.20

EU27 (08) 27 0.28

EU15 (08) 15 0.09

Egypt (05) 27 0.12

Always calculated as average income of national/state percentile = fct (mean national/state income) Branko Milanovic

The XXI century trilema

A. Globalization of ideas, knowledge, Communication, awareness of others’ living standards

B. Increasing differences in mean incomes

among countries

C. No movement of people

If A and B, then no C. Migration is the outcome of current unequal globalization. If B and C, then no A. Unequal globe can exist if people do not know much about each other’s living conditions or costs of transport are too high. If A and C, then no B. Under globalization, people will not move if income differentials are small.

Branko Milanovic

Growing inter-country income differences and migration: Key seven borders today

Branko Milanovic

H. Should there be a global equality of opportunity or not?

Branko Milanovic

Is citizenship a rent?

• If most of our income is determined by citizenship, then there is little equality of opportunity globally and citizenship is a rent (unrelated to individual desert, effort)

• Key issue: Is global equality of opportunity something that we ought to be concerned or not?

• Does national self-determination dispenses with the need to worry about GEO?

Branko Milanovic

The logic of the argument

• Citizenship is a morally-arbitrary circumstance, independent of individual effort

• It can be regarded as a rent (shared by all members of a community)

• Are citizenship rents globally acceptable or not?

• Political philosophy arguments pro (social contract; statist theory; self-determination) and contra (cosmopolitan approach)

Branko Milanovic

Rawls from A Theory of Justice

• “Injustice is…simply inequalities that are not to the benefit of all…and in particular to the poor” (p. 54)

• But this is the rule enounced for a single nation-state? Will it be valid for the world as a whole?

• As we shall see: No, it won’t be!

Branko Milanovic

Rawls on (a) inequality between countries and (b) global inequality

• Neither of them matters

• Income divergence is irrelevant if countries have liberal institutions; it may be relevant for liberal vs. burdened societies

• Irrelevance rooted in two key assumptions: (i) political institutions of liberalism are what matters; (ii) acquisition of wealth immaterial for both individuals and countries

• Global inequality between individuals similarly irrelevant once the background conditions of justice exist in all societies

• But within-national inequalities matter because the difference principle applies within each people (note however that the DP may allow for high inequality)

Branko Milanovic

The Rawlsian world

• For Rawls, global optimum distribution of income is simply a sum of national optimal income distributions

• Why Rawlsian world will remain unequal?

Branko Milanovic

All equal Different (as

now)

All equal

Different (as

now)

Mean country incomes

Individual incomes within country

Global Ginis in Real World, Rawlsian World, Convergence World…and Shangri-La World (Theil 0; year 2008)

98

68 (all country Ginis=0)

30 (all mean incomes same; all country Ginis as now)

0

Branko Milanovic

Why pace Rawls global inequality matters?

• Because the world is becoming globalized and global inequality will come to matter more and more despite the absence of global government (analogy with national vs. village inequality)

• Because it is associated with migration which is fast becoming a prime political issue

• Because it raises the issue of global equality of opportunities

Branko Milanovic

5. Conclusion and emerging issues

Branko Milanovic

Conclusion and 21st century policy issues

• To reduce significantly global inequality (and poverty) and citizenship rent there are two ways:

• A slow and sustainable way: higher growth rate of poorer countries

• A fast and possibly politically tumultous way: increase migration

• Either poor countries will have to become richer or poor people will move to rich countries.

• Should migrants be taxed additionally to pay native population’s losers and those remaining in their countries of origin?

Branko Milanovic

A new view of development

• Development is increased income for poor people regardless of where they are, in their countries of birth or elsewhere

• Migration and LDC growth thus become the two equivalent instruments for development

Branko Milanovic

A key difficulty

• How to manage:

• (1) Possible contradiction between the rise of the emerging market economies and rich world’s middle class

• (2) Rising domestic inequalities

• (3) Migration, while recognizing its potential for global poverty alleviation

Branko Milanovic

How China climbed up global income distribution

Branko Milanovic

sixth urban decile

tenth urban decile

50

60

70

80

90

glo

ba

l pe

rcen

tile

1990 1995 2000 2005 2010benchmark year

From key_calcul3.do

Follow me on Twitter: @BrankoMilan Branko Milanovic