Jonathan Kelley and M.D.R. Evans.
"Does Economic Development Reduce Tolerance for Inequality?
A Comparative Analysis of 30 Nations"
Chapter 4 in Charting the Globe: The International Social Survey Programme 1984-2009.
Edited by Max Haller, Roger Jowell and Tom Smith. London: Routledge (2009).
4
ECONOMIC DEVELOPMENT REDUCES
TOLERANCE FOR INEQUALITY:
A COMPARATIVE ANALYSIS OF 30 NATIONS1
Jonathan Kelley
International Survey Center & Department of Sociology, University of Nevada, Reno
[email protected]
M.D.R. Evans
Departments of Resource Economics and Sociology, University of Nevada, Reno
[email protected]
Abstract
Do conceptions of just rewards vary with economic development? To investigate this
question we use the 1999-2000 “Inequality-III” round of the International Social
Science Project together with other data in the World Inequality Study. There are 30
countries and 19 568 individual respondents in the full-time labor force. We measure
inequality by the Gini coefficient for the general public's report of the legitimate
earnings for their own occupation. OLS and multilevel analyses show patterns of
influences very similar to those found in earlier research, with one striking exception.
By far the most important influence, not previously documented across so many
countries, is the prosperity of the nation: people in poor nations are much more
accepting of inequality than are people in prosperous nations. If this cross-sectional
pattern reflects developmental trends, as is likely, then it seems that economic
development creates equalitarian attitudes. However, true egalitarianism is not held
as ideal in any country, and so is not an appropriate goal for public policy. Instead
the ideal level of inequality differs among countries. These ideals are a more
appropriate benchmark for policy. We suggest that these benchmarks, available here
for 150 nations, should be the starting point for future assessments of income
inequality.
Does economic development reshape our attitudes and values as well as our standard of
living, occupational distribution, and the like? Karl Marx, one of sociology's founders,
thought so: “Does it require deep intuition to comprehend that man’s ideas, views, and
conceptions, in one word, man’s consciousness, changes with every change in the condition
of his material existence ...” (Marx and Engels 1848 [1972]: 351). Some prior research is
consistent with this hypothesis. For example, multi-level analyses show that economic
development lifts subjective social class, net of individual characteristics (Evans and Kelley
2004a). But what about other aspects of inequality-related attitudes, values, and perceptions?
1
It is well known that very few people in any modern society yearn for strict equality of
earnings (Gijsberts 2002; Kelley and Evans 1993; Verwiebe and Wegener 2000), but just how
much inequality they consider ideal and why is a body of knowledge that is just beginning to
develop.
Observers of social inequality beginning with de Toqueville have long suspected that there
are substantial international differences in ranges of just incomes (Tocqueville 1839). ISSP
data from the Inequality III survey enable us to quantify and systematize these differences.
4.1 Theory and hypotheses
On the one hand, some social changes suggest that as we get richer, we develop a taste for
smaller inequalities. In this vein, questions we might ask include: Has economic development
delivered us into a brave new world of post-materialism (Inglehart 1997) where the declining
marginal utility of income means that near-egalitarianism comes at a relatively small price?
Does the large scale and complex interdependence of functions within modern organizations
(Durkheim 1902 [1933]) render less visible individual productivity and hence undermine the
legitimacy of large pay inequalities (Aristotle 1985[322BC])?
On the other hand, if economic development had immiserated the working class (Marx and
Engels 1845, 1848 [1972]; World Bank 2006), stimulating a desperate struggle for upward
mobility and for conspicuous consumption (Veblen 1967) to demonstrate status, then
development could whet appetites for pay inequalities, at least on the part of aspirants and
winners. The evidence shows that immiseration did not occur (Firebaugh and Beck 1994), but
people might think it did, in which case it could have real consequences. More realistically,
economic development might enhance tolerance of inequality, because the citizenry's
apparent concerns about inequality were really about the material and physical hardships of
the poor. The very large gains that economic development brings to life chances for people at
the bottom of the social ladder could lead the population at large to feel that their altruistic
goals have been achieved, so redistribution is little needed any more. Alternatively, as
Durkheim worried in darker moments, has the uprooting of the populace from the seeming
immutability of the social order in the countryside unleashed insatiable yearnings for more
income without a compelling moral order to restrain and channel them?
2
These conflicting theories imply very different hypotheses about the net effect of economic
development on the citizenry's pay-inequality ideals: they cry out to be tested on a large
sample of countries using multi-level analyses. Little prior research has tackled this, because
of the burden of making huge datasets truly comparable, but two papers have made a start
(Evans and Kelley 2007; Hadler 2005). Building on that foundation, we explore the effects of
economic development on pay-inequality ideals, and investigate the degree to which the
micro-processes whereby stratification and demographic characteristics influence these ideals
change with economic development.
Concerning the thesis that economic growth leads people to prefer small inequalities,
pioneering prior research finds support over short time spans within several countries (Jasso
2000). Cross-sectionally there is also support in a study modeling GDP effects on one item of
the ISSP's 3-item "relative inequality goals" scale, "Differences in income in (country) are too
large." (Hadler 2005). The only important counterexample is the dramatic increase in the
magnitude of legitimate inequality in post-Communist East-Central Europe (Kelley and
Zagorski 2005). We suspect that this reflects the interaction of institutional arrangements
with economic growth, and that the main effect of economic development is to shrink the just
range of pay inequality (Haller and Sharda 2005). Thus, the working hypothesis is:
H1: Economic development reduces the public's ideal magnitude of earnings
inequality
This paper tests H1 by examining the effect of GDP per capita on a measure of the ideal
magnitude of earnings inequality, net of individual characteristics and net of population size,
which also influences just earnings (Evans and Kelley 2007).
Prior research on a smaller range of countries finds that both demographic and class-related
aspects of social position influence ideals about the magnitude of earnings inequality, so we
include these as controls. Prior research finds few socio-economic differences in the
citizenry's ideal earnings for low status occupations, but
substantial class-related and
demographic differences in the pay that people see as legitimate for high status occupations
(Kelley and Evans 1993), and therefore for the ratio of high to low (Austen 2002; Gijsberts
2002). When applied to the earnings that people see as legitimate for themselves, we expect
our findings to parallel these; to parallel the actual financial returns that accrue to education,
occupational status, supervisory authority; and to reflect actual earnings. Thus:
3
H2: The more education one has, the higher one's just earnings will be.
H3: Higher occupational status, supervisory authority, business ownership, and
higher actual earnings will increase one's just earnings.
It is well known that economic development decreases returns to education. So, to the degree
that this reflects education's lesser contribution to productivity in developed countries, then
we would expect an interaction with the just returns to education falling with economic
development. The Durkheimian claim about interdependence and decreased visibility of
individual contributions leads to the same conclusion. The same should hold for occupational
status and perhaps for other class-related aspects of one's job. Thus:
H4: The just returns to education will fall with economic development.
H5: The just returns to occupation will fall with economic development.
4.3. Data
Surveys
Most of the surveys we analyze are from the 1999-2000 “Inequality-III” round of the
International Social Survey Programme (ISSP); the rest are from the International Survey of
Economic Attitudes and the International Social Science Survey/Australia. Comparisons with
the national census, where available, show the surveys to be representative (Evans and Kelley
2002: Appendix; Sikora 1997; Zentralarchiv fuer Empirische Sozialforschung 2002). We use
all surveys with the relevant data: 30 nations and 19 568 individual respondents.
We recoded each survey to a common international standard in the World Inequality Study
(Kelley, Evans, and Sikora 2005). Details are in (Evans and Kelley 2004b).
Sample selectivity
The sample of countries is necessarily an opportunistic one, rather than randomly drawn.
Fortunately, selectivity analysis shows that the sample is representative with respect to
population size, Anglo-Celtic heritage, and actual inequality, although more developed
nations are significantly more likely to be included (see Appendix Table 4.7). Since level of
development is included in our model, it will not induce selectivity bias.
4
4.4. Measurement
Just earnings
Building on a long tradition (e.g. Verba and Orren 1985; Sarapata 1963), our questions first
appeared in the International Social Science Survey/Australia, then in the ISSP's three
“Inequality” surveys of 1987/88, 1992/93, and 1999/2000, and subsequently elsewhere. They
begin:
Next, what do you think people in these jobs ought to be paid -- how much
do you think they should earn each year before taxes, regardless of what
they actually get...
Please write in how
much they ought to
earn each year
a. First, about how much do you think a skilled
worker in a factory ought to earn?. . . . . . . . $ ___________
dollars
and then continue with seven or eight other occupations and end with a new question (Kelley
et al. 1997):
i. And someone who works in YOUR usual occupation
-- how much ought they earn? . . . . . . . . . . . $ ___________
dollars
This last is the question analyzed here. Answers were in local currency. Country-by-country
analyses report the Gini inequality coefficient for these. For pooled analyses, currency units
matter, so we standardize by scoring each answer relative to the average full-time earnings of
unskilled workers in respondent's country (Kelley and Evans 1993: 85-86). We call these
"minimum incomes".
Other individual level variables
Measurement of other individual-level variables is conventional and is detailed in (Evans and
Kelley 2007).
Sample analyzed: Full-time workers
5
The analysis is confined to people employed full-time (30+ hours a week), earning at least
one quarter of the average income of full-time unskilled workers in their nation.
National characteristics
We use the Gini coefficient because it is the most widely used measure of income inequality,
and has been previously used in an international assessment of inequality-related attitudes
(Hadler 2005). In theory, it can range from 0 to 1.0 but in practice it ranges from around .20
for very equalitarian societies like Sweden and Norway to around .60 for the most
inegalitarian societies like Brazil.
Population size is from World Bank data (World Bank 2002) with a few additions and
amendments.
4.5 Description
To introduce the issues, let us examine the distributions for the US of (1) ideal full-time
earnings and, for comparison, (2) actual full-time earnings. Both distributions cover a broad
span (Table 4.1 and Figure 4.1).
Legitimate earnings in the US.
Earnings of about $20 000 a year (in 2006 dollars) for their occupation are nominated as just
by about 10% of Americans. More, 19%, think around $30 000 would be right, and another
22% think $40,000 would be right. About 17% think $50 000 would be right and still quite a
few, 13%, feel $60 000 would be right. A further 9% think something around $75 000 would
be right and 6% hold out for a full $100 000. At the top, 2% claim around $150 000. Overall,
the average income Americans find legitimate for their occupation comes to a handsome
$51 000.
6
Table 4.1. What Americans
think their occupation ought to
earn, and what they do earn.
[1]
Dollars
per year
Ought to
earn
(percent)
Actually
earn
(percent)
$150,000
2
0
$125,000
0
1
$100,000
6
2
$75,000
9
7
$60,000
13
10
$50,000
17
10
$40,000
22
12
$30,000
19
21
$20,000
10
24
$10,000
1
13
Total
100%
100%
Cases
586
1870
Mean
$51,226
$36,876
Gini
.29
.32
s.d. log
.52
.59
[1] Men and women in the labor force, working full-time,
adjusted upward to year 2006 income levels
(approximate). "Ought to earn" questions asked in only
one survey and "actually earn" in three.
[Table 4.1 near here]
Thus, if each American were paid exactly what he or she thinks just, there would be
substantial inequality, with a Gini coefficient of .29 (Figure 4.1, left panel). An alternative
inequality measure, the standard deviation of the distribution of log income, conveys the
same impression (Table 4.1, bottom row). Almost no-one would be at the bottom, most
would be near the middle, and a fair few toward the top. This is close to the moderately
equalitarian "Type C" society that many in Western nations believe characterizes their
societies (Evans, Kelley, and Kolosi 1992: Table 1). Equality may be attractive to
philosophers, but it is not at all close to what ordinary Americans think right.
7
USA: Actually earns
(mean= $37,000; Gini= .33)
$150,000
$150,000
$125,000
$125,000
$100,000
$100,000
$75,000
$75,000
$60,000
$60,000
Earnings
Ought to earn
USA: Own occupation ought to earn
(mean= $51,000; Gini= .29)
$50,000
$50,000
$40,000
$40,000
$30,000
$30,000
$20,000
$20,000
$10,000
$10,000
Percent
Percent
Figure 4.1 What you think your own occupation ought to earn. For comparison, what they actually do earn. Full-time
workers in the USA.
[Figure 4.1 near here]
Actual earnings follow a similar pattern, but shifted downwards (Table 1, second column).
Quite a few are at the very bottom. The mean is $37 000, only 72% of what Americans think
they ought to be paid, almost $15 000 less. Inequality is about the same (Figure 4.1, second
panel).
Absolute versus relative incomes
In the US in 2006, for example, unskilled workers earned about 30 000 US dollars. So we
treat an American engineer who says he should earn 45 000 dollars as wanting 1.5 minimum
incomes and an American executive who says she should earn 90 000 dollars as claiming 3
minimum incomes. The assumption is that an Australian who claims 43 000 Australian
dollars (1.5 times the minimum income in Australia) is making a claim, relative to the
Australian economy, equivalent to the engineer's claim on the American economy. We ignore
the fact that on the international market, the American can buy more than the Australian,
focusing instead on how each stands relative to his own countrymen. This focus on position
relative to one's own country is usual in the inequality literature and, indeed, built into the
very definition of the Gini and most other indices of inequality (Allison 1978; Gijsberts 2002;
Hadler 2005; Slomczynski and Wesolowski 2001).
8
An example of this shift in units is in Table 4.2. An American claiming $150 000 is asking
for five "minimum incomes" (as we call them): five times what a typical American unskilled
worker earns. An American claiming $60 000 is saying her occupation deserves two
minimum incomes, and an American asking for $30 000 is asking for one minimum income.
The Gini coefficient is the same regardless of which way income is measured.
Table 4.2. Further description of what Americans think they ought
to earn.[1]
Number of
minimum
incomes[2]
5
Dollars
per year
Percent
$150,000
2
Description
(box & whisker plot)[3]
4
$125,000
0
3.4
$100,000
6
2.6
$75,000
9
2.1
$60,000
13
1.7
$50,000
17
1.4
$40,000
22
Median=$41,000
1.0
$30,000
19
First quartile=$31,000
0.7
$20,000
10
0.3
$10,000
1
Total
100%
Mean
$51,000
Minimum income [2]
$29,000
Upper adjacent value= $108,000
Third quartile=$62,000
Lower adjacent value=$5000
Measures of inequality:
Standard deviation
$34,500
Std.dev. of log income
.52
Gini coefficient
.29
Inter-quartile range $31,000
% going to top quintile 36
[1] Men and women in the labor force, working 30 or more hours per week, with annual incomes at least a quarter of the
average full-time male unskilled wage. Adjusted upward to year 2006 income levels (approximate). 1972-1999. N=586.
[2] Defined as the average earnings of full-time, male unskilled workers.
[Table 4.2 near here]
We can summarize the US distribution compactly with a box-and-whisker plot (on the right
in Table 4.2). The median is around $40 000 to $50 000 (there is some uncertainty due to the
small sample). There is a good deal of dispersion with the bottom quarter ending around
$30 000 and the top quarter beginning around $60 000. So half of Americans think they
should earn somewhere between $30 000 and $60 000.
9
Absolute earnings. How do American's views compare with other nations in these absolute
terms (Figure 4.2, left panel)? Americans mostly think they ought to earn more than what
other folk think right for themselves. There are considerable uncertainties in these figures,
due to difficulties in converting currencies, but the broad pattern is clear. The average Briton,
Australian, or other English-speaker, as well as Western European, is content with less,
perhaps two-thirds or three-quarters what Americans think just. At the top, the differences are
even larger, with many Americans claiming more than citizens of other countries would
allow themselves.
00
50
US dollars
75
,0
,0
00
00
# minimum incomes
10
00
00
5
4
3
2
25
,0
1
in
ita
tra
sp
lia
ea
Eu
ki
ng
ro
pe
N
EC
N
EC
,J
ap
an
Fo
3d
rm
W
er
or
ly
ld
C
om
m
un
is
t
Au
s
SA
En
gl
is
h
Br
U
ai
n
Au
s
tr a
sp
ea
lia
Eu
ki
ng
ro
pe
N
EC
N
EC
,J
ap
an
Fo
3d
rm
er
W
ly
or
ld
C
om
m
un
is
t
En
g
lis
h
U
Br
it
SA
0
0
Source: World Inequality Study. Results are rough approximations. N=17,622 in 29 nations.
Source: World Inequality Study. N=17,622 in 29 nations.
Shows 25th percentile, median, 75th percentile, and adjacent values
Shows 25th percentile, median, 75th percentile, and adjacent values
Figure 4.2 Legitimate income in absolute terms (US dollars, left panel) and expressed relative to the income of
ordinary workers in respondents' own country (ratios, right panel). Full-time workers. 29 Nations, 1999/2000.
[Figure 4.2 near here]
In contrast, people in the Third World, as well as those in ex-Communist countries, think
they should earn much less. Even those in the top quartile think it proper to earn much less in
absolute terms than Americans in the bottom quartile. So a modest working-class American is
likely to think he ought to be better paid than a high status professional in India – unequal this
is, but it is what both sides think legitimate.
This has a lot to do with Americans actually being quite well-paid compared to others
elsewhere in the world (World Bank 2006).
10
Relative earnings. The picture is very different in relative terms: prosperous, high status
Americans think they should earn more than unskilled American workers, but not a lot more
(Figure 4.2, right panel). Prosperous Britons and Australians have much the same views,
compared to British and Australian workers. So do Western Europeans and Japanese – if
anything, they may claim even a shade more than Americans think proper.
But in the Third World, high status people claim much more than their unskilled countrymen.
So do top people in ex-Communist countries (consistent with previous research using
different methods, Kelley and Zagorski 2005).
It is in these relative terms that income inequality is usually conceived and measured by the
Gini coefficient and similar measures. And that is the focus of this chapter as well. But keep
in mind that the great inequality that people in developing nations think proper does not mean
that their elites believe their pay should be higher than pay in rich countries. Quite the
contrary: the moral claims their elites make are low compared to even quite ordinary workers
in developed nations, and high only in comparison with what ordinary workers claim for
themselves in their own nations.
Legitimate earnings in other nations
Ideal earnings span a wide spectrum, with large differences both among occupations and
among nations. For example, Figure 4.3 gives the distributions for 4 disparate countries – the
USA, Sweden, Brazil, and Russia. (The distributions for all the countries in our sample are
below in Table 4.3.) To facilitate comparisons across countries, we present them in the metric
of minimum incomes.
11
8.0
7.5
7.0
6.5
6.0
5.5
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
.5
Austria (Gini=.22)
Number of minimum incomes
Number of minimum incomes
Sweden (Gini=.15)
8.0
7.5
7.0
6.5
6.0
5.5
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
.5
Percent
Percent
8.0
7.5
7.0
6.5
6.0
5.5
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
.5
Russia (Gini=.45)
Number of minimum incomes
Number of minimum incomes
USA (Gini=.30)
8.0
7.5
7.0
6.5
6.0
5.5
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
.5
Percent
Percent
Figure 4.3. Earnings people think legitimate for themselves. Selected nations.
[Figure 4.3 near here]
The well known egalitarian preferences of tiny Sweden are highlighted by the strong
concentration of legitimate incomes between 1 and 1.5, together with the paucity of incomes
over 4 or under 1. Sweden's Gini, .15, is the lowest of the nations in this study. Legitimate
incomes are more diverse in the US, with more at the low end and many more at the higher
end; the result is a Gini twice Sweden's.
Legitimate incomes are even more unequally distributed in Brazil and Russia, with many
more at both the low and, especially, the high end. Indeed, Brazil and Russia, both large
nations, have the largest Ginis, .44 and .45, of any countries in the sample, three times the
Swedish figure.
12
Most of the countries in our sample have Ginis between .20 and .30, with only a few lower
and a few higher.
Table 4.3. Economic development and legitimate inequality. Persons working full-time. 19,568
[1]
respondents in 30 nations, circa 1999-2000.
GDP (USA
[2]
1995=1) Gini
Number of minimum incomes people believe their occupation ought to earn (percent)
.5
1
1.5
2
2.5
3
4
5
6
7
8+
Total
Cases Mean
USA
1.21
.30
9
29
30
13
9
7
0
1
0
0
1
100%
601
1.72
Norway
1.01
.16
1
60
26
8
2
3
0
0
0
0
0
100%
680
1.35
Canada
.98
.25
9
32
33
16
2
5
1
0
1
0
0
100%
581
1.50
Switzerland
.97
.21
0
5
34
28
19
9
2
1
1
0
0
100%
505
2.12
Japan
.91
.28
7
15
23
20
23
4
4
3
0
0
1
100%
461
2.03
Australia
.90
.23
2
36
35
13
8
5
1
0
0
0
0
100%
3226
1.54
Netherlands
.88
.22
1
22
33
25
10
7
3
0
0
0
0
100%
1147
1.78
Sweden
.87
.15
1
60
25
10
2
1
0
0
0
0
0
100%
610
1.35
Austria
.87
.22
1
25
40
12
12
6
1
2
1
0
0
100%
335
1.74
Germany-West
.84
.25
1
4
16
22
21
19
8
5
1
3
1
100%
354
2.67
Northern Ireland
.82
.22
0
9
21
32
17
10
7
3
0
0
0
100%
249
2.26
Britain
.82
.23
6
36
36
10
7
4
1
0
0
0
0
100%
278
1.48
France
.82
.29
1
23
19
25
14
6
6
2
1
0
1
100%
908
2.13
Germany-East
.76
.22
1
7
25
29
17
12
3
5
0
0
0
100%
214
2.20
New Zealand
.70
.25
8
42
34
9
1
5
1
0
0
0
0
100%
550
1.38
Spain
.66
.19
1
32
40
13
10
1
2
0
0
0
0
100%
468
1.58
Cyprus
.62
.23
0
6
40
26
17
2
4
4
0
1
1
100%
707
2.13
Israel
.62
.28
1
5
28
16
21
14
3
6
1
3
1
100%
514
2.59
Portugal
.58
.28
0
10
26
23
5
23
6
2
2
2
1
100%
587
2.43
Slovenia
.57
.23
0
20
36
21
14
4
3
1
0
0
0
100%
497
1.76
Czeck Republic
.50
.24
1
21
40
23
7
4
2
1
0
0
1
100%
875
1.85
Slovakia
.43
.26
4
28
24
20
0
16
5
0
3
0
0
100%
583
1.96
Hungary
.38
.30
0
5
15
22
20
8
16
9
2
0
5
100%
465
3.01
Chili
.36
.35
3
14
23
26
8
9
8
1
3
2
3
100%
555
2.44
Poland
.32
.35
2
34
13
20
15
5
5
3
1
1
2
100%
1118
2.20
Brazil
.28
.44
14
20
14
11
14
5
6
8
1
3
4
100%
633
2.56
Latvia
.27
.27
0
2
14
15
4
22
28
4
5
5
2
100%
585
3.42
Russia
.27
.45
9
21
15
8
15
8
12
3
1
2
7
100%
583
2.94
Bulgaria
.20
.31
7
32
29
16
8
4
2
1
0
0
1
100%
869
1.71
Philippines
.13
.40
10
19
16
19
12
6
9
3
1
1
4
100%
471
2.48
[1] GDP is at parity purchasing power in year of survey, expressed as an index with USA in 1995= 1.0. Surveys were in 1994 to 2002, with most in 1999 or 2000. Data are for persons in the
labor force, working 30 hours or more, with positive incomes.
[2] Gini coefficients in boldface type are significantly different from the USA at p<.05, one-tailed.
[Table 4.3 near here]
13
4.6 Economic development and inequality: Cross-national evidence
More prosperous nations prefer less inequality; Figure 4.4 shows the bivariate relationship.
The x-axis shows GNP per capita and the y-axis legitimate inequality, as measured by the
Gini coefficient of the just earnings distribution. There is a strong positive relationship,
statistically significant at p<.001. It suggests that poor societies with GNP around a quarter of
US levels (like the Philippines, Brazil, or Russia) would typically see great inequality as
legitimate, with Gini coefficients around .40 or .45. In contrast, prosperous societies at
European levels of GNP (like West Germany, Switzerland, or Norway) would typically have
little inequality in legitimate income, with Ginis around .20. That is about half as much
inequality as citizens of poor societies typically find proper.
.50
Ideal inequality: Gini coefficient
RUS
BZL
.40
PHL
POL
CHL
BUL
.30
HNG
LVA
FRA
ISR
PORT
SVK
CZE
SVN
.20
JPN
G-W
N_Z
USA
CAN
GB
AUS
AUTN_IR
G-E
NL
SPAIN
CYP
SWZ
NOR
SWE
.10
10% of US
50% of US
100% of US
GNP (Index: USA 1995=1)
Source: ISSP Inequality-III, 1999/2000
Figure 4.4. GDP and ideal earnings inequality.
Full-time workers.
[Figure 4.4 near here]
The US is a bit of an outlier, thinking more inequality proper than other nations at its high
level of economic development (as many commentators have suspected). In contrast, Sweden
and (more surprisingly) Spain think less inequality is proper than other nations at their level
of development. At the other end of the GNP continuum, both Brazil and Russia think even
more inequality is proper than do other nations at their modest levels of development.
14
Controlling for population size
Next, we assess whether this relationship is robust to a control for population size, which is
known to be an important influence on attitudes toward inequality (Evans and Kelley 2007).
The regression analysis in Table 4.4 shows that net of population size GNP still has a strong
effect reducing legitimate inequality (t= -8.05, p<.001).
The standardized regression coefficient is a very substantial -.74, emphasizing the key result
that the higher the level of development, the less the ideal inequality, ceteris paribus.
Table 4.4. Economic development and legitimate inequality as
measured by the Gini coefficient for the income people think
proper for their own occupation. Aggregate results for N=30
[1]
nations, circa 1999-2000.
StandardMetric
s.e.
t
ized
GDP per capita (index; USA 1995=1.0)
-.216
.027
-8.05
-.74
.00068
.00011
6.05
.56
Constant
.368
.018
20.46
.
R-squared
76%
Population of nation (millions)
[1] Nations are the unit of analysis. The dependent variable is the Gini coefficient for legitimate earnings. All coefficients
are statistically significant at p<.001, two-tailed.
[Table 4.4 near here]
In metric terms, the difference between a very poor society at just 10% or 15% of US GDP,
like the Philippines, and a very rich one like the US or Switzerland would typically be about
.19 on the Gini scale. So, for example, .37 for the poor society versus .19 for the rich one
assuming both had a population of 50 million – thus the rich society has about half as much
ideal inequality as the poor.
In sum, the countries that prefer the greatest inequality are large, poor countries such as
Brazil, the Philippines, and Russia. Those preferring the least inequality are small, developed
nations such as Norway and Sweden. Alternative analyses using a different measure of
inequality lead to the same conclusions (Appendix Table 4.8). This suggests that conflicts
over inequality will decrease over coming generations, assuming economic growth continues.
But it could increase in Europe should the European Union, with a very large population,
begin to emerge as a "country" in the eyes of its citizens (Evans and Kelley 2007).
15
4.7 What shapes views about legitimate earnings?
Model. By analogy to the conventional model for actual income, we propose that just
earnings are a function of actual earnings, education, occupation, supervision, selfemployment, and business ownership (Blau and Duncan 1967; Robinson and Kelley 1979;
Yun 2006). The unit of analysis is individuals. As is usual in the inequality literature, we
analyze the natural logarithm of legitimate income (Jasso 1978; Kelley and Evans 1993), thus
focusing on percentage changes. In practice, the model is robust across various specifications
of the dependent variable: For example, analyzing just earnings relative to the unskilled
wage in each country leads to the same conclusions.
A promising first hypothesis is that the socio-economic characteristics that matter in the
income determination process also influence just earnings. Figure 4.5 gives the total, direct,
and indirect effects of each variable from a series of OLS regressions.
Actual pay. Almost everyone thinks their occupation ought to earn more than it does – their
actual earnings are about three-quarters of their just earnings in the US, for example (see
Table 4.1 above). But in addition, those who earn a lot believe that they ought to earn even
more. The standardized effect, .41, is very large (Figure 4.5). Indeed, by far the largest
influence on how much people think their job ought to be paid is how much they themselves
actually earn.
However this is an upper bound estimate of the true effect, because it may well be that
people's views about their just earnings influence what they actually do earn rather than (as
we have assumed) the other way around. For example, a self-employed businessman may pay
himself only what he thinks is right; similarly, a consultant may charge only what she thinks
is proper, not whatever the market will bear; and, in market societies, people can decline jobs
that offer far less than they think is right for the job. If so, their pay is in part a consequence
of their views about legitimate earnings.
Education and training. Education also matters, with highly-educated people believing that
their jobs warrant more pay and poorly-educated people allowing that theirs warrant less.
Prior research has documented strong effects of education on actual earnings throughout the
world, from poor countries like Brazil (Haller and Saraiva 1992) to the very richest nations
(Ganzeboom, Treiman, and Ultee 1991). The results in Figure 4.5 suggest that this familiar
fact is fully in accord with people's normative views.
16
o2yourq
Education
edyrsi
occs2i
Occupational
status
superi
pettybi
Supervisor
captlsti
earnri
Solo self-employed
Beta
Beta
Beta
.08
0.224909.140.136048 0.083901
.
0.161018 0.092318
.09 .07
0.0961 0.038391
0.109438
0.09317 direct
More is legitimate,
.04 .06
0.087656 0.030642
More is 0.414963
legitimate, indirect
.
.09 .02
Less is. legitimate, direct
Less is legitimate, indirect
Business owner .03 .06
.41
Earnings
.00
.10
.20
.30
.40
.50
Importance: Standardized effect .
Figure 4.5. Structural influences on amount of
legitimate earnings.
[Figure 4.5 near here]
All in all, education is about half as important as actual earnings. This mostly comes about
indirectly: Well educated people get better paying jobs and earn larger incomes for that and
other reasons; these higher incomes then in turn legitimate their pay. That indirect effect
comes to .14 in standardized terms. The rest, .08 in standardized terms, comes about directly,
with better educated people feeling they are entitled to more pay than others with the same
occupational status, supervisory responsibilities, business ownership, and actual incomes.
Occupation. Occupational status is a smaller, but still important influence, with a total effect
of .16. Nearly half of this, .07, comes about indirectly, because people higher on the
occupational ladder tend to earn higher incomes and that leads them to feel that their jobs
should be highly paid, and the fatter “half”, .09, is a direct influence – even aside from how
much they actually earn, people working in high status occupations feel that their jobs merit
high pay.
Other aspects. The other features of a person’s work situation – Supervision, Solo selfemployment, and Business ownership are less influential (Figure 4.5), contrary to
expectations from Marx and Dahrendorf. We turn now to the key question: how is economic
development involved in all this?
17
How economic development influences legitimate inequality
One reason that people in poor countries find greater inequality to be legitimate may be that
the returns they think proper for education and occupation are greater, as we argued earlier.
To investigate that, we estimate education's total effects, both direct and indirect, from an
equation with a multi-level interaction term:
lnLegitIncome = b0 + b1Ed + b2GDP +b3GDP*Ed + e
where GDP*Ed is the multiplicative interaction of education and GDP. We must use a multilevel model in order to get correct standard errors and significance tests (DiPrete and Forristal
1994; Hox 1995). We estimated it using Stata 9's xtreg routine. The interaction effect is
clearly significant (p<.001, two-tailed).
In poor nations with a GDP around 10% of US levels, each year of education increases an
individual's just income by about 6.5% (Table 4.5, first row). This is education's total effect,
including those that come about indirectly through its influence on occupation, supervision,
ownership, and actual earnings. In richer nations around the US level of GDP, the effect is
under half of that, 2.9%. This is one of the reasons there is more legitimate inequality in poor
nations than in prosperous ones.
Table 4.5. GNP and legitimate earnings: Effects of education and
[1]
occupation are greater in developed nations. Total effects.
GDP of nation
Comparison
Variable
10% of USA
100% of USA
Education
1 additional year of education
6.5% more
2.9% more
Occupation
Higher professional (top)
versus farm laborer (bottom)
43% more
27% more
[1] Estimated from multi-level regressions.
[2] This is the average earnings of full-time unskilled male workers in respondent's nation, for example $29,000 in the USA.
[Table 4.5 near here]
Thus, education may be more productive, or at least scarcer, in poor nations. If education can
be viewed as a productive investment undertaken for economic reasons, (e.g. Becker 1975;
Mincer 1958), then the optimal return to investments in education should be close to its
effects on productivity. Objectively, in industrial societies a year of education increases
productivity by around 10% (Murphy and Welch 1994) and probably more in developing
18
nations. Abundant evidence (beginning with Aristotle in the Nicomachean Ethics) suggests
that ordinary people believe that rewards ought to reflect productivity. If so, a return of
roughly 10% should be viewed as just; and higher in poor societies. If all this is so, then the
6.5% we find legitimate in poor nations is, if anything, a bit on the low side, and the 2.9% we
find for prosperous nations decidedly low, perhaps because of the Durkheim's masking effect
of interdependence in large organizations.
Occupation. We estimate occupation's effects, both direct and indirect, from:
lnLegitIncome = b0 + b1Ed + b2GDP +b3GDP*Ed + b4Occ + b5Super + b6Solo +b7Bz +b8GDP*Occ + e
where GDP*Occ is the multiplicative interaction of occupational status and development.
GDP*Ed, as before, is the interaction of education and development. Both are statistically
significant in the multi-level analysis (p<.001, two-tailed).
On average, people in poor nations believe that high ranking professional occupations at the
top of the hierarchy should be paid 43% more than farm laborers at the bottom, ceteris
paribus (Table 4.5, second row). But people in prosperous nations believe 27% would be
appropriate, only two-thirds as much. This is another reason there is more legitimate
inequality in poor nations than in prosperous.
4.8 Implications
Credentialist and conflict theories of stratification
According to sociological functionalists, most economists, and almost all educators,
schooling confers skills that enable people to do their jobs better – that is, education enhances
productivity – and is rewarded for that reason. But a long “credentialist” tradition argues that
education has no intrinsic effect on productivity; hence that (factually) it is not rewarded for
that reason; and hence that (normatively) it provides no legitimate justification for inequality
(Bourdieu and Passeron 1977; Brown 2001; Collins 1979). Similar arguments harking back
to Marx and Dahrendorf suggest that conflict and coercion, rather than skills and
productivity, underlie the higher pay given to high status jobs, supervisors, and business
owners (Robinson and Kelley 1979).
By implication, there is no moral justification for their rewards either. If credentialism were
right, few people would find the rewards to education, occupation, supervision, and
19
ownership morally legitimate. Less educated workers, those in low status occupations, and
those who do not supervise would regard themselves as entitled to just as high pay as anyone
else. But, the results show that they do not, instead accepting that their pay ought to be
lower. In short, the stratification hierarchy is consensual, not imposed by coercion. Hence,
our results argue against credentialist and conflict theories.
Legitimate inequality
Many policy makers argue that the reduction of inequality should be as much a target of
policy as economic growth (e.g. Portes and Roberts 2005; Sen 1973; World Bank 2006). At
times, this has been a major theme not only of the left – inequality and growth are major
political issues in most Third World nations – but also of more conservative groups, such as
the World Bank (e.g. Velez, Barros, and Ferreira 2004). Our results provide strong new
evidence, reinforcing earlier arguments both theoretical and empirical (Kelley and Evans
1993; Welch 1999), that a great deal of inequality is morally legitimate in the eyes of
ordinary people. Ordinary people do not believe in the equal distribution of income, at least
not in any country for which we have good evidence. Far from it: they believe in a great deal
of inequality, inequality arising in good part from what they see as legitimate rewards to
education, occupational achievement, and job performance.
Brazil is a striking example. It has a very high level of inequality, but Brazilians think it
ought to have a very high level of inequality. In our data for full-time employees, the actual
distribution of income has a Gini coefficient of .40 but the legitimate distribution is, if
anything, even more unequal: .43. Outsiders might not like that, and many developmental
economists certainly do not (Velez, Barros, and Ferreira 2004), but it would take a very
authoritarian philosopher-king to wish to impose their personal views on an unwilling
citizenry.
In place of the naive assumption that all inequality is bad, we suggest that one should
compare how much income inequality there actually is with how much the nation's citizens
think there ought to be. These comparisons for 30 nations (based directly on our survey data)
and estimates for another 120 nations are in Table 4.6. The estimates use aggregate data to
impute individual level preferences and are admittedly uncertain.
20
We suggest that these benchmarks should be the starting point for future assessments of
income inequality.2
Table 4.6. Legitimate inequality in earnings for many nations: Gini coefficients. Entries
based on survey data are in bold face type; others are estimates as described in the
text. Also actual inequality in earnings where available, for comparison. Full-time
workers only.
Sweden
Norway
Cyprus (Greek)
Denmark
Kuwait
United Arab Emr
Spain
Netherlands
Hong Kong, China
Singapore
Switzerland
Belgium
Germany-East
Ireland
Finland
Puerto Rico
Northern Ireland
Austria
Australia
United Kingdom
Slovenia
Oman
Canada
Czech Republic
New Zealand
Taiwan
Mauritius
Germany-West
Trinidad
Gabon
Slovakia
Estonia
Italy
Botswana
Latvia
Greece
United States
Japan
Namibia
Uruguay
Portugal
Costa Rico
Hungary
Macedonia
Panama
Libya
Gambia
Bosnia & Herz
Israel
Guinea-Bissau
Lithuania
Jamaica
France
Gini:
Legi- Gini:
timate Actual
.15
.19
.16
.22
.17
.24
.18
.18
.18
.19
.24
.19
.26
.19
.20
.20
.25
.20
.21
.24
.21
.21
.21
.22
.27
.22
.19
.23
.29
.23
.27
.23
.28
.23
.24
.24
.24
.27
.24
.26
.25
.25
.25
.28
.25
.26
.26
.30
.26
.27
.27
.27
.36
.27
.27
.32
.27
.33
.27
.27
.28
.35
.28
.29
.29
.29
.29
.29
.29
.29
.29
.29
.30
.30
.30
.29
Croatia
Saudi Arabia
West Bank & Gaza
Yugoslavia
Lesotho
Bulgaria
Lebanon
Korea, Rep (South)
Paraguay
Jordan
Mongolia
Mauritania
Albania
Turkmenistan
Congo, Rep
El Salvador
Liberia
Argentina
Armenia
Papua New Guinea
Moldova
Central Afr Rep
Nicaragua
Tunisia
Dominican Republic
Kyrgyzstan
Belarus
Togo
Eritrea
Cuba
Honduras
Malaysia
Georgia
Lao PDR
Sierra Leone
Guatemala
Guinea
Bolivia
Tajikistan
Benin
Romania
South Africa
Azerbaijan
Chile
Venezuela
Haiti
Somalia
Ecuador
Burundi
Kazakhstan
Rwanda
Chad
Zimbabwe
Gini:
Legi- Gini:
timate Actual
.30
.30
.30
.30
.31
.31
.39
.31
.31
.31
.31
.31
.31
.31
.32
.32
.32
.32
.32
.32
.32
.32
.32
.32
.33
.33
.33
.33
.33
.33
.33
.33
.33
.33
.33
.34
.34
.34
.34
.34
.34
.34
.34
.34
.34
.44
.34
.34
.34
.35
.35
.35
.35
.35
.35
Iraq
Senegal
Poland
Syrian Arab Rep
Angola
Niger
Cambodia
Peru
Burkina Faso
Mali
Malawi
Colombia
Algeria
Sri Lanka
Cameroon
Cote D'ivoire
Madagascar
Ghana
Morocco
Thailand
Yemen
Mozambique
Uzbekistan
Turkey
Uganda
Mexico
Korea Dem (North)
Nepal
Sudan
Afghanistan
Iran
Ukraine
Kenya
Tanzania
Egypt
Myanmar
Congo, Dem Rep
Philippines
Ethiopia
Viet Nam
Pakistan
Bangladesh
Brazil
Nigeria
Indonesia
Russia
India
China
Gini:
Legi- Gini:
timate Actual
.35
.35
.35
.36
.35
.35
.36
.36
.36
.36
.36
.36
.36
.36
.36
.36
.36
.37
.37
.37
.37
.37
.37
.37
.37
.38
.38
.38
.38
.38
.38
.38
.38
.39
.39
.39
.40
.40
.41
.47
.41
.41
.42
.43
.43
.40
.43
.43
.45
.41
.48
.48
[1] In other studies, inequality is more commonly measured for family income rather than for individual earnings, and for all respondents rather than only for fulltime workers. Such figures are normally higher than the figures shown here. Full-time is defined as working 30 hours or more.
Source: Bold face entries are from the World Inequality Study; other entries are projected from those using coefficients from an OLS regression predicting
legitimate inequality in earnings on the basis of ln population size and GNP per capita.
21
[Table 4.6 near here]
4.9 Appendix:
Appendix Table 4.7. Sample selectivity: Probit analysis of selection
into the sample of nations.[1]
StandardMetric
s.e.
t
ized[2]
GDP per capita (index; USA 1995=1.0)
Population of nation (millions)
English speaking, English based law
Income inequality: Gini
Constant
Pseudo R-squared
.001
.001
4.87
.56
.00069
.00106
ns
.03
.057
.550
ns
.03
-1.567
1.807
ns
-.07
-1.342
.804
ns
--
35%
[1] Nations are the unit of analysis. N=155 nations with populations over 1 million, plus Cyprus.
[2] From the corresponding OLS analysis
ns -- Not statistically significant at p<.05, two-tailed.
[Table 4.7 near here]
Appendix Table 4.8. Sensitivity test using an alternative measure of
inequality: GDP and the standard deviation of the log of the income
people think proper for their own occupation. Aggregate results for
N=30 nations, circa 1999-2000.[1]
StandardMetric
s.e.
t
ized
GDP per capita (index; USA 1995=1.0)
Population of nation (millions)
-.368
.047
-7.78
-.71
.00128
.00020
6.49
.59
Constant
.636
.032
20.09
.
R-squared
76%
[1] Nations are the unit of analysis. The dependent variable is the Gini coefficient for legitimate earnings. All coefficients are
statistically significant at p<.001, two-tailed.
Source: World Inequality Survey.
[Table 4.8 near here]
22
References
Allison, Paul D. 1978. "Measures of Inequality." American Sociological Review 43:865-880.
Aristotle. 1985[322BC]. Nicomachean Ethics. Indianapolis: Hackett Publishing.
Austen, Siobhan. 2002. "An International Comparison of Attitudes to Inequality." International Journal of
Social Economics 29::218–37.
Becker, Gary S. 1975. Human Capital (2nd Ed.). New York: National Bureau of Economic Research.
Blau, Peter M. and Otis Dudley Duncan. 1967. The American Occupational Structure. New York: Free Press.
Bourdieu, Pierre and Jean-Claude Passeron. 1977. The Inheritors: French Students and Their Relation to
Culture. Chicago: University of Chicago Press.
Brown, D. K. 2001. "The Social Sources of Educational Credentialism: Status Cultures, Labor Markets, and
Organizations." Sociology of Education:19-34.
Collins, Randall. 1979. The Credential Society: A Historical Sociology of Education and Stratification. New
York: Academic Press.
DiPrete, Thomas A. and Jerry D. Forristal 1994. "Multi-Level Models: Methods and Substance." Annual Review
of Sociology, pp. 331-357.
Durkheim, Emile. 1902 [1933]. The Division of Labour in Society. [2nd Edition]. Translated by G. Simpson.
New York: Free Press.
Evans, M.D.R. and Jonathan Kelley. 2002. Australian Economy and Society 2001: Education, Work, and
Welfare. Sydney: Federation Press.
Evans, M.D.R. and Jonathan Kelley. 2004a. "Subjective Social Location: Data from 21 Nations." International
Journal of Public Opinion Research 16:3-38.
Evans, M.D.R. and Jonathan Kelley. 2004b. "Data, Measurement, and Methods." in Australian Economy and
Society 2002: Religion, Morality, and Public Policy. Sydney: Federation Press.
Evans, M.D.R. and Jonathan Kelley. 2007. "Population Size, Economic Development, and Attitudes Towards
Inequality: Evidence from 30 Nations." Population Review 46:1-23.
Evans, M.D.R., Jonathan Kelley, and Tamas Kolosi. 1992. "Images of Class: Public Perceptions in Hungary and
Australia." American Sociological Review 57:461-482.
Firebaugh, G. and F. D. Beck. 1994. "Does Economic-Growth Benefit the Masses – Growth, Dependence, and
Welfare in the Third-World." American Sociological Review 59:631-653.
Ganzeboom, Harry B. G., Donald J. Treiman, and Wout C. Ultee. 1991. "Comparative Intergenerational
Stratification Research: Three Generations and Beyond." Annual Review of Sociology 17:277-302.
Gijsberts, Merove. 2002. "The Legitimation of Income Inequality in State-Socialist and Market Societies." Acta
Sociologica 45:269-285.
Hadler, Markus. 2005. "Why Do People Accept Different Income Ratios?: A Multi-Level Comparison of Thirty
Countries." Acta Sociologica 48:131-154.
Haller, Archibald O. and Helcio U. Saraiva. 1992. "The Income Effects of Education in a Developing Nation:
Brazil – 1973 and 1982." Research in Social Stratification and Mobility 11:63-93.
Haller, Archibald O. and Bam Dev Sharda. 2005. "Theory of National Development and Societal Stratification."
Population Review 44:1-23.
Hox, J.J. 1995. Applied Multilevel Analysis. Amsterdam: TT-Publikaties.
Inglehart, Ronald. 1997. Modernization and Postmodernization. Princeton: Princeton University Press.
Jasso, G. 1999. "How Much Injustice Is There in the World? Two New Justice Indexes." American Sociological
Review 64:133-168.
Jasso, Guillermina. 1978. "On the Justice of Earnings: A New Specification of the Justice Evaluation Function."
American Journal of Sociology 83:1398-1419.
23
Jasso, Guillermina. 2000. "Trends in the Experience of Injustice: Justice Indexes About Earning in Six
Societies, 1991–1996." Social Justice Research 13:101–21.
Kelley, Jonathan and M. D. R. Evans. 1993. "The Legitimation of Inequality: Occupational Earnings in Nine
Nations." American Journal of Sociology 99:75-125.
Kelley, Jonathan, M.D.R. Evans, Petr Mateju, Knud Knudsen, Knut Skjak, Bogdan Cichomski, Pavel
Morawski, and Brina Malnar. 1997. ISSP Inequality-3 1999 Design Group: Preliminary Report.
International Survey Center (www.international-survey.org).
Kelley, Jonathan, M.D.R. Evans, and Joanna Sikora. 2005. "World Inequality Study, 1967-2004: Codebook and
Machine Readable Data File." International Survey Center (www.international-survey.org).
Kelley, Jonathan and Krzysztof Zagorski. 2005. "Economic Change and the Legitimation of Inequality: The
Transition from Socialism to the Free Market in Poland and Hungary, 1987-1994." Research in Social
Stratification and Mobility 22:321-366.
Marx, Karl and Frederich C. Engels. 1848 [1972]. "Manifesto of the Communist Party." Pp. 335-362 in The
Marx-Engels Reader, edited by R. C. Tucker. New York: Norton.
Marx, Karl and Friedrich Engels. 1845[1972]. "Alienation and Social Classes." Pp. 104-106 in The Marx-Engels
Reader, edited by R. C. Tucker. New York, NY: Norton.
Mincer, Jacob. 1958. "Investment in Human Capital and Personal Income Distribution." Journal of Political
Economy 66:281-302.
Murphy, Kevin M and Finis Welch. 1994. "Industrial Change and the Rising Importance of Skill." Pp. 101-132
in Uneven Tides: Rising Inequality in America, edited by P. Gottschalk. New York: Russell Sage
Foundation.
Portes, A. and B.R. Roberts. 2005. "The Free-Market City: Latin American Urbanization in the Years of the
Neoliberal Experiment." Studies in Comparative International Development 40:43-82.
Robinson, Robert V. and Jonathan Kelley. 1979. "Class as Conceived by Marx and Dahrendorf: Effects on
Income Inequality and Politics in the United States and Great Britain." American Sociological Review
44:38-58.
Sarapata, A. (1963). "Iustum pretium." Polish Sociological Bulletin 1:41–56.
Sen, Amartya. 1973. On Economic Inequality. Oxford: Clarendon.
Sikora, Joanna. 1997. "International Survey of Economic Attitudes in Australia, Finland and Poland:
Comparison with the Census." WwA: Worldwide Attitudes 1997.12.31:1-8. Available at www.internationalsurvey.org.
Slomczynski, Kazimierz M.
and Wlodzimierz Wesolowski. 2001. "Distributive Justice and Status
(in)Consistency: A Theoretical Debate and Empirical Evidence." Polish Sociological Review 3:299-312.
Tocqueville, Alexis de. 1839. Democracy in America. Translated by H. Reeve. New York: Adlard. Available on
Google Scholar.
Veblen, Thorsten. 1967. The Theory of the Leisure Class. London: Penguin.
Velez, C. E., R. P. d. Barros, and F. Ferreira. 2004. Inequality and Economic Development in Brazil: A World
Bank Country Study. Washington D.C.: World Bank.
Verba, Sidney and Gary R. Orren. 1985. Equality in America: The View from the Top. Cambridge, MA: Harvard
University Press.
Verwiebe, R. and Berndt Wegener. 2000. "Social Inequality and the Perceived Income Justice Gap." Social
Justice Research 13:123–49.
Welch, Finis. 1999. "In Defense of Inequality." American Economic Association Papers and Proceedings 89:117.
World Bank. 2002. "Population Size." in World Development Indicators, 2001: CD-Rom, edited by W. Bank.
Washington, DC: World Bank.
World Bank. 2006. World Development Report, 2006. New York: Oxford University Press.
24
Yun, Myeong-Su. 2006. "Earnings Inequality in USA, 1969-99: Comparing Inequality Using Earnings
Equations." Review of Income and Wealth 52:127-144.
Zentralarchiv fuer Empirische Sozialforschung. 2002. International Social Survey Programme 1999/2000:
Social Inequality-3, International Pooled File. Codebook and Machine-Readable Data File. Cologne,
Germany: Zentralarchiv fuer Empirische Sozialforschung.
1
This research was supported in part by Research Infrastructure and Equipment Facility (RIEF) grant
R19920093 from the Australian Research Council to the Melbourne Institute of Applied Economic and Social
Research, University of Melbourne. This paper draws heavily on our earlier paper (Evans and Kelley 2007)
which analyzed the same data but with a focus on population size. We thank Archibald O. Haller, Max Haller,
and Joanna Sikora for comments. Jonathan Kelley is director of the International Survey Center and Adjunct
Professor at the University of Nevada, Reno; he was previously Senior Fellow at the Australian National
University and Professorial Fellow at the University of Melbourne. email: [email protected].
M.D.R. Evans is Associate Professor in the departments of Sociology and Resource Economics at the University
of Nevada, Reno. email: [email protected].
2
They should not be the end point however. In addition to having the amount of inequality they think is right,
the inequality should arise from the sources they think right. Assessing that is a more complex task; Jasso has
made a beginning (Jasso 1999).
25