1 the long-run evolution of the distribution of income a b atkinson, nuffield college, oxford dulbea...
Post on 21-Dec-2015
215 views
TRANSCRIPT
1
The Long-Run Evolution of the Distribution of Income
A B Atkinson, Nuffield College, Oxford
DULBEA 1957-2007 Conference
2
1. Methodological introduction
2. The evolution of the overall distribution
3. Top incomes over the long-run
4. Earnings: Episodes of change?
5. Wealth: A return of the rich?
6. Factor shares and the macro-economy
7. Conclusions
3
1. Methodological introduction
Describing income inequality over the 20th century
Relating the explanation of inequality to the mainstream of economics
• Full run of years extended back in time
• Graded approach to data quality
• Draw on variety of sources
• Earnings
• Wealth
• Relative shares• Labour economics
• Accumulation and growth
• Political economy
4
1. Methodological introduction
2. The evolution of the overall distribution
3. Top incomes over the long-run
4. Earnings: Episodes of change?
5. Wealth: A return of the rich?
6. Factor shares and the macro-economy
7. Conclusions
5
Figure 1 Three Nordic countries
17.5
20.0
22.5
25.0
27.5
30.0
32.5
1966 1970 1974 1978 1982 1986 1990 1994 1998 2002
Gin
i co
effi
cien
t %
SWE 1967SWE 1SWE 2SWE 3FINO 1 unequivNO 2NO 3NO 4
6
Figure 2 Income inequality in UK and US 1929-2005
25
30
35
40
45
50
55
1929 1933 1937 1941 1945 1949 1953 1957 1961 1965 1969 1973 1977 1981 1985 1989 1993 1997 2001 2005
Gin
i co
effi
cien
t %
US BEA synthetic estumate GROSS
US CPS GROSS
US CPS new GROSS
UK Blue Book synthetic estimate NET
UK Economic Trends Market
UK Economic Trends NET
US
UK market
UK net
7
Figure 3 Belgium, (West) Germany and Netherlands
20
25
30
35
40
45
50
1945 1949 1953 1957 1961 1965 1969 1973 1977 1981 1985 1989 1993 1997 2001
Gin
i co
effi
cien
t %
NL 1 grossNL 2 tax unitsNL 3DE DIW synthetic estimateDE surveyDE panel studyBE (LIS data)
8
Figure 4 France and Italy
20
25
30
35
40
45
50
1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004
Gin
i co
effi
cien
t %
IT 1
IT 2
IT 3
FR 1
FR 2
FR 3
9
Figure 5 Czech Republic (Czechoslovakia), Hungary and Poland
15
20
25
30
35
40
1958 1963 1968 1973 1978 1983 1988 1993 1998 2003
Gin
i co
effi
cien
t
CZ per capita income
CZ per capita income 2
HU per capita income
HU Tarki equiv incomes
PL per capita incomes WORKER households
PL per capita incomes
PL LIS equiv incomes
10
Conclusions:• Popular view of a U-turn in income inequality finds some foundation
in the evidence for 12 OECD countries, but it has to be qualified.
• Timing of the downward arm of the U differed across countries, and in some cases was based on 1 or 2 influential observations.
• The recent upturn is clear in the Nordic and Anglo-Saxon countries, but picture in Continental Europe more varied.
• In the case of the UK, the Netherlands, Italy and possibly the Nordic countries, the increase in the Gini looks more like an episode than a continuing upward trend.
• In the Eastern European countries, the increase was associated
with the transition to a market economy.
11
1. Methodological introduction
2. The evolution of the overall distribution
3. Top incomes over the long-run
4. Earnings: Episodes of change?
5. Wealth: A return of the rich?
6. Factor shares and the macro-economy
7. Conclusions
12
UK
US
CA
AUS
NZ
A B Atkinson, and T Piketty, editors, Top Incomes over the Twentieth Century, Oxford University Press, volume 1, 2007.
Figure 6 Share of top 1% in English-Speaking countries
0
2
4
6
8
10
12
14
16
18
20
1900 1905 1910 1915 1920 1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000
Per
cen
t o
f to
tal
inco
me
AUS CA NZ
US UK
13
NL
DEU
CH
FRA
Figure 7 Share of top 1% in Continental Europe
0
5
10
15
20
25
30
1900 1905 1910 1915 1920 1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000
Per
cen
t o
f to
tal
inco
me
GER FR NL
SWI SE
14
Conclusions:• The income tax data provide only a restricted view of the
distribution of income, but they allow the changes over time to be tracked more closely.
• There is a broad commonality over the first three-quarters of the twentieth century, although differences in timing that reflect national specificities.
• The data show a marked difference in recent decades between the Anglo-Saxon countries and Continental Europe, with Sweden closer to the former. Although this may be changing …
15
1. Methodological introduction
2. The evolution of the overall distribution
3. Top incomes over the long-run
4. Earnings: Episodes of change?
5. Wealth: A return of the rich?
6. Factor shares and the macro-economy
7. Conclusions
16
Figure 8 Long run development of earnings distribution in United States 1939-2005
75
100
125
150
175
200
225
250
275
300
1939 1944 1949 1954 1959 1964 1969 1974 1979 1984 1989 1994 1999 2004
Per
cen
tag
e m
edia
n
30
40
50
60
70
80
90
100
110
120
Per
cen
tag
e m
edia
n
Top decileLH axis
P95LH axis
Upper quartile/quintileLH axis
Goldin and Margo
Bottom decileRH axis
Lower quartileRH axis
17
Figure 9 Long run development of earnings dispersion in France 1919-2004
100
120
140
160
180
200
220
240
260
280
300
1919 1924 1929 1934 1939 1944 1949 1954 1959 1964 1969 1974 1979 1984 1989 1994 1999 2004
Per
cen
tag
e o
f m
edia
n
50
55
60
65
70
75
80
85
90
95
100
Per
cen
tag
e o
f m
edia
n
Top decile LH axis
P95LH axis
Bottom decileRH axis
18
Figure 10 Long run development of earnings dispersion in United Kingdom 1954-2006
50
75
100
125
150
175
200
225
250
275
1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006
Per
cen
tag
e o
f m
edia
n
30
40
50
60
70
80
90
100
110
120
Per
cen
tag
e o
f m
edia
n
P95 LH axis
Upper quartile LH axis
Top decile LH axis
Bottom decile RH axis
Lower quartile LH axis
19
Figure 11 Long run development of earnings dispersion in West Germany 1929-2002
50
75
100
125
150
175
200
225
250
1929 1933 1937 1941 1945 1949 1953 1957 1961 1965 1969 1973 1977 1981 1985 1989 1993 1997 2001
Per
cen
tag
e m
edia
n
50
60
70
80
90
100
110
120
130
Upper quartile LH axis
P95 LH axis
Top decile LH axis
Top decile LH axis
Upper quintile LH axis
Bottom decile RH axis
Lower quartile RH axis
Note: geographical boundaries have changed
Lower quintile LH axis
20
Figure 12 Fanning out in the upper part of the distribution
0
5
10
15
20
25
30
35
40
45
50
AL DEGSOEP
DE GLE HU IT PL PT CH SE UK US
Note: AL 1975-2006; DE GSOEP 1984-1998; DE GLE 1962-1995; HU 1994-2003; IT 1977-2004; PL 1992-2004; PT 1982-2002; CH 1996-2004; SE 1983-2004; UK 1977-2001; US 1973-2005
Per
cen
tag
e in
crea
se r
ekat
ive
to t
he
med
ian
Growth in P75
Growth in P90
Growth in P95
21
Conclusions:
• “Great compression” in 1930s/1940s
• Rise in dispersion in Golden Age
• Post-1968 reaction
• Recent rise in dispersion: fanning out at the top
• W not V or U
22
1. Methodological introduction
2. The evolution of the overall distribution
3. Top incomes over the long-run
4. Earnings: Episodes of change?
5. Wealth: A return of the rich?
6. Factor shares and the macro-economy
7. Conclusions
23
Top 1% of wealth distribution in six countries from Ohlsson, Roine and Waldenstrom (2006)
UK data adjusted for break in 1960
10
15
20
25
30
35
40
45
50
55
60
1900 1905 1910 1915 1920 1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000
Sh
are
of
top
1%
FRA UK
SWE
DK
US
NOR
24
UKFigure 14 Wealth concentration in the US 1946-2000
0.00
0.25
0.50
0.75
1.00
1.25
1.50
1.75
2.00
1946 1951 1956 1961 1966 1971 1976 1981 1986 1991 1996
Pro
po
rtio
n o
f a
du
lt in
div
idu
als
% f
or
ric
h
1/1
00
ths
of
% f
or
su
pe
r-ru
ch
40
45
50
55
60
65
70
75
80
%
Gini coefficient among the rich (right scale)
Super-Rich(left scale)
Rich (left scale)
25
Conclusions:• Over the first three-quarters of the twentieth century, the reduction
in wealth concentration followed a similar path to that of overall income inequality, and specifically the top income shares.
• The fall in concentration reflected the positive force of the acquisition of popular wealth by the bottom 99 per cent, and the negative incidence of progressive income and estate taxation.
• In recent decades in Anglo-Saxon countries there has been a rise in the proportion of rich, but that has taken place against a background of a general rise in the wealth-income ratio, so that the impact on wealth shares is less marked.
26
1. Methodological introduction
2. The evolution of the overall distribution
3. Top incomes over the long-run
4. Earnings: Episodes of change?
5. Wealth: A return of the rich?
6. Factor shares and the macro-economy
7. Conclusions
27
Share of Earnings in Total personal income
= Earnings /Total compensation
x Total compensation/National income
x National income /Total personal income
28
Figure 15 Wage shares in the UK 1908-2006
50
55
60
65
70
75
80
85
90
95
100
1908 1913 1918 1923 1928 1933 1938 1943 1948 1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 2003
Per
cen
tag
e
Wages / total personal income
Wages / total compensation
Total compensation / National income
Wages / National income
29
Conclusion:• The macro-economic theory of distribution has been little discussed
in recent years, but should be revived. At the same time, once cannot read across directly from factor shares to the personal distribution. The experience of the UK, US and France suggests that the relation is one of some complexity.
30
Summary of Conclusions:
• There has indeed been a U-turn in overall income inequality, which is evident in Anglo-Saxon countries and Finland and Sweden, but it is less evident in Continental Europe and not apparent in France.
• In a number of cases, the recent increase in inequality looks more like a limited episode rather than a continuing upward trend.
• Top income shares show a marked U over the twentieth century as a whole in Anglo-Saxon countries and Sweden, but there has not so far been an upturn in Continental Europe.
• For individual earnings, if one is seeking a single-letter summary of the changes in the earnings distribution over the period examined here, then a “W” seems more appropriate than a “U”.
31
• The 1930s and 1940s experienced a reduction in wage differentials; this was reversed in the 1950s and early 1960s: with the exception of Germany, this “Golden Age” saw a rise in earnings dispersion. The later 1960s, following the events of May 1968, in France and other countries governments and unions achieved a narrowing of the earnings distribution. The rise in dispersion in recent decades has to be seen in this context.
• Over the first three-quarters of the twentieth century, the reduction in wealth concentration followed a similar path to that of overall income inequality, and specifically the top income shares. The fall in concentration reflected the positive force of the acquisition of popular wealth by the bottom 99 per cent, and the negative incidence of progressive income and estate taxation.
• In recent decades in Anglo-Saxon countries there has been a rise in the proportion of rich, but that has taken place against a background of a general rise in the wealth-income ratio, so that the impact on wealth shares is less marked.
• The macro-economic theory of distribution should be revived. At the same time, once cannot read across directly from factor shares to the personal distribution. The experience of the UK, US and France suggests that the relation is one of some complexity.