Transcript No real convergence

The role of large countries (China
and India in particular)
Milanovic, “Global inequality and its
implications”
Lecture 10
1. Large countries: an overview
Population in
2000 (in m)
Land area in
2000 (in m
sq.km)
China
1271
9.3
Undefined
India
1033
3.0
Union
United States
283
9.2
Federation
Indonesia
213
1.8
Unitary state
Brazil
172
8.5
Federation
Russia
144
16.9
Federation
Pakistan
141
0.8
Federation
Bangladesh
133
1.3
Unitary state
Nigeria
130
0.9
Federation
Japan
127
0.4
Unitary state
China
India
USA
Brazil
Idn
Period
78-01
80-00
77-01
85-01
83-01
Regions
27/30
14/25
50/50
26/26
26/26
92
100
99
100
Pop.
99
coverage
Richest Shanghai
Poorest
(22)
Guizhou
(1.6)
Mahar.
(3)
Bihar
(0.7)
Ratio rich
to poor
13.6
4.4
See also Table 4
Conn. S Paulo E. Kalim
(42)
(9)
(14)
W. Virg Maranhao E Nusa T
(1.1)
(1.3)
(20)
2.1
7.0
12.5
2. Concept 1 and Concept 2
inequalities in large countries
Three concepts of inequality
• Concept 1: unweighted inequality of regions (or
countries)
useful for study of
convergence (is growth faster in poorer regions?)
• Concept 2: population weighted inequality of
regions (countries); "feeling" of inequality,
particularly if there are regional cleavages. Also
proxy to...
• Concept 3: inequality between individuals in a
country (or world)
Example: population weighted divergence
•
•
•
•
2 rich and small regions, A and B
2 poor and populous regions, C and D
A and C grow fast, B and D slowly, then
no change (or small change) in Concept 1
inequality, no income convergence.
• no ρ between population size and growth
• But Concept 2 inequality goes up,
population weighted divergence (since C
and D become dissimilar)
Why it matters?
• Concept 1. An economic question. Will
there be convergence if L,K, goods move
relatively freely (compared to impediments
that exist between countries)
• Concept 2. A social question. What is the
"feeling" of inequality/exclusion
(particularly if there are ethnic/religious
cleavages). Threat to national cohesion.
The data we use
• Regional GDPs per capita
• Concept 1 & 2 inequality calculated across
nominal and real GDP per capita;
overestimate of inequality (some regional
redistribution; price levels higher in richer
regions)
• Also in PPPs
Yi , j , t ,95,$ 
Yi,j,t  1,78, d *(1  ri,j,t)
* PPP95
DDj (78,95)
Concept 1 Gini (unweighted inter-regional inequality)
(across nominal GDPs per capita)
0.400
China
0.300
Brazil
0.200
China: regional
convergence in the '80s
India
USA
0.100
India & Indon. regional
divergence throughout
US: regional convergence
since early 80's
0.000
19
77
19
78
19
79
19
80
19
81
19
82
19
83
19
84
19
85
19
86
19
87
19
88
19
89
19
90
19
91
19
92
19
93
19
94
19
95
19
96
19
97
19
98
19
99
20
00
20
01
Gini
Indonesia
Highest regional inequality
in China; lowest in the US
(despite having 50 units)
Years
Provincial mean-normalized incomes in 1980 and 2000
(mean is unweighhted all-China mean)
6
5
Shanghai
2000
4
3
Beijing
2
Zhejiang
Tianjin
Jiansu
1
0
0
1
2
3
1980
4
5
6
China: Concept 1 Gini inequality in nominal
and real terms
0.40
0.36
reals
No real convergence:
no systematic difference
in real growth rates btw.
the provinces
0.32
nominals
0.28
0.24
0.20
1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000
Between 1978 and 1990
prices rose faster in
poorer regions
India: Real and nominal divergence
0.3
0.2
nominals
reals
Nominal and real
inequality rise step in
step up to about 1991
Since then nominal
divergence stops while
real continues
0.2
0.1
Price catch-up of
poorer provinces (better
integrated domestic
market?)
0.1
0.0
1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000
China (1980-2000)
Red: fast growth (1σ above the mean)
Yellow: average
Light yellow: slow (1σ below the
mean)
North to South
Shandong
Jiangsu
Zhejiang
Fujian
Guangdong
India (1980-1999)
Maharashtra (Bombay)
Karnataka (Bangalore)
Tamil Nadu (Madras)
United States
New Hampshire
Massachusetts
Connecticut
Brazil
West to East
Amazonas
Para
Mato Grosso
Indonesia
Does not include oil and gas sectors.
West to East
West Nusa Tenggara
Jakarta/ Bali
Lampung
Irian Jaya
11
9
7
5
5
7
9
11
average growth rate of GDP per capita 1990-00
Chinese provincial growth 1978-90 and 1990-00
300
500
700
900
GDP per capita in 1978
1100
1300
800
1300
1800
2300
GDP per capita in 1990
2800
In 1990-2000, poorer regions growing slower than the average
Beijing, Shanghai and Tienjin not shown
kdensity gdpppp
China's rural and urban mean provincial
incomes in 2000
0
5000
10000
15000
GDP per capita in 95 PPP
rural
20000
25000
urban
Source: from Kanbur and Zhang; 26 provincial means for rural and 26 for urban.
Concept 2 Gini (population-weighted
inter-regional inequality)
0.400
1990's: Increasing
Concept 2 inequality in
the three Asian countries
Brazil
0.300
0.200
Indonesia
India
0.100
USA
0.000
19
77
19
78
19
79
19
80
19
81
19
82
19
83
19
84
19
85
19
86
19
87
19
88
19
89
19
90
19
91
19
92
19
93
19
94
19
95
19
96
19
97
19
98
19
99
20
00
20
01
Gini
China
Years
Highest inequality in
Brazil. If all people in
each state had the same
income, Gini would be
still more than 30. In the
United States less than
10!
What drives Concept 2
inequality?
• Different population growth rates by region
• Correlation between growth rates and
population size (do more populous states
grow faster
implications for the
productivity view of growth; poverty
reduction)
Impact of differential population and GDP per
capita growth on Concept 2 Gini
1980-90
1990-2000
Diff.
Population
effect
Diff. Growth
effect
Population
effect
Diff. Growth
effect
USA
0
+1.8
+0.1
-0.6
China
0
-2.9
+0.4
+2.6
India
0
+1.3
0
+2.5
Brazil
+0.1
-0.4
0
-3.0
Indon.
-0.8
+0.3
-0.1
+1.1
Results (for Concept 2 inequality)
• Differential population growth not
important
• Growth disequalizing in India throughout
• China: differential growth rates equalizing
in 1980-90, then disequalizing in 1990-2000
Importance of population-weighted
divergence
0.12
Pop Y in 1980 Growth
1980200
Mahar 95
1300
60
Betavalue
0.08
0.04
UP
170
680
15
Bihar
107
512
0
0.00
1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
-0.04
Years
India: β and 95% confidence interval
Economic and "political economy"
convergence
Real incomes
Nominal incomes
Concept 1
Are growth rates
negatively related
to intial real
income?
Are prices
moving the same
way in poor and
rich regions?
Concept 2
(Basically)
convergence
among the subset
of populous
regions
Are prices
moving the same
way in poor and
rich populous
regions?
Conclusions
• Asia: increasing regional inequality in the
1990's (India and China; not Indonesia)
• Concept 2 increases important for national
cohesion (India and China)
• Growth disequalizing; higher income level
equalizing; no evidence that nation-wide
openness positively related to Concept 2
inequality
• Populous states’ outcomes diverge in both
India and China
Complexity of the process
• In both China and India, a process directly
opposite to what we observe at global level
• In China & India: Concept 1 inequality
going down, Concept 2 inequality up
• World: Concept 1 inequality up, Concept 2
inequality down (and the latter solely due to
high average growth of China & India)
3. China and India Concept 3
inequalities
China: Inequality according to HS data
• Increase in Concept 3 between 1980 and 2000
about 14 Gini points (according to Ravallion and
Chen)
• Explained by rising differences between mean
provincial incomes (~8 Gini points),
• rising differences urban and rural areas (~2 Gini
points)
• rising differences within urban and rural areas
(another 3 Gini points)
Illustration of Concepts 2 and 3: China,
inequality according to HS data
0.5
0.4
Concept 3: from
Ravallion and Chen
Inequality within urban and
urban areas (10%)
0.3
Urban/rural
inequality (35%)
Concept 2: provinces and U/R
(from K-Zh)
0.2
Concept 2: provinces only (from
K-ZH, BM calcul)
Inequality between
provinces (55%)
0.1
0
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000
Decomposing total inequality in China
Gini = 44
0.50
0.40
Gini = 31
Other inequality
0.30
Rural/urban
inequality
0.20
0.10
Inequality
between
provinces
0.00
1980
2000
Based on Ravallion & Chen (2004), Kanbur & Zhang (2002), Milanovic (2004)
China and India compared
(Gini points)
China 2000
India 1997
Inequality between
provinces/states
Rural-urban inequality
24
22
13
7
Inequality within R/U
areas
Total inequality
7
9
44
38
3.1-1
1.8-1
Urban-rural ratio
From IndiaChina.xls file; China: based on HBS data; India based on state GDIs, italics: estimates
4. Role in global income
distribution
Shares of US, China and India in world
GDI (in $PPP terms)
Recall Concepts 2 calculation:
• In Gini terms:
1
n
n
( y  y ) pip


j
i
i
j
j i
• where Gi=individual country Gini, π=income share, yi =
country income, pi = population share, μ=overall mean
income, n = number of countries
• For each pair of countries depends on the mean-normalized
gap between their per capita incomes and population
shares
• As China’s GDI pc (in $PPP terms) is some 10
times less than the US’s, if China grows at 10%
per annum, US needs to grow only 1% to keep the
numerator the same.
• Then, only if world mean income grows, will the
China-US contribution to international ineqaulity
go down.
• Almost all of China’s contribution to reduced
Concept 2 inequality comes from its catching up
of other countrieds (not the United States); and (as
we shall see below) only 2/3 of it is due to growth.
Mean-normalized income distances between
China, India and the US
5.0
US-India
US-China
4.0
3.0
2.0
1.0
China-india
0.0
1965
1978
2000
Contributions (in Gini points) of differences in mean
incomes between Ch, In, US to Concept 2 inequality
12
10
8
China-US
6
4
US-India
2
India-China
0
1965
1978
2000
• About 20% of Concept 2 inequality explained by
the “triangle”
• US-China mean-normalized GDI per capita gap
decreased from 4.5 to 4 (btw. 1965 and 2000)
• Gini contribution of US-China decreased 6.3 to
4.2 points (over the same period)
• Between 1978 (reforms in China) and 2000, more
than 1/3 of the China decrease to Concept 2
inequality due to the population effect (↓ share of
world population; from 24% to 22%)
• Difference between China and India adds to global
inequality
China component in Concept 2 inequality
1978
2000
Change
Concept 2 inequality
59.4
53.4
-6.0
China component
20.9
16.1
-4.8
China economy
component (if pop. share
at 1978 level)
20.9
17.8
-3.1
China population
component (if GDI pc
relative to the world at
1978 level)
20.9
19.2
-1.7
Memo: Mean-normalized
distance to the US
4.25
4.0
Source: Jiang Zhiyong (2005)