The Ef f ec t s o f Globa l i za t ion

o n I n e q u a l i t y :

A C r o s s - N a t i o n a l

A n a l y s i s

Hal le Inst i tute Occas ional Paper

Linda Beer, PhD & Terry Boswell, PhD

Department of Sociology

Emory University

The Claus M.

HALLEINSTITUTEF O R G L O B A L

L E A R N I N G

THE EFFECTS OF GLOBALIZATION ONINEQUALITY:

A CROSS-NATIONAL ANALYSIS

Linda Beer, PhDDepartment of Sociology

Emory University

and

Terry Boswell, PhD*Department of Sociology

Emory University

Running head: The Effects of Globalization on InequalityApprox. word count: 10,927

* Please direct all correspondence to: Linda Beer, Southeast AIDS Training andEducation Center, Emory University, 735 Gatewood Road, Atlanta, GA 30322.

Abstract

Why does globalization produce income inequality? Over the last twenty years we haveseen a steep rise in income inequality worldwide that reversed a prior equalizing trend.The rise in inequality appears closely linked to the rapid increase in the integration ofworld trade and in international investment that we call economic globalization. Weexamine the relationship between foreign investment dependence and income inequality,and compare it to theories that focus on world trade. The evidence provides strongsupport for the theory that nations highly dependent on foreign capital experience highand worsening income inequality. Exploitation also increases inequality while democracyand education reduce it. Other sources of inequality derived from modernization theoryor dual sector models do not remain significant in our models. Considering the dramaticincrease in foreign investment in recent years, the implications of this finding for povertyreduction and political stability are stark.

Introduction

The contemporary era is one of both accelerated economic globalization and rising

inequality. International markets for goods, services, and capital have become

increasingly integrated and, since the 1980s, this trend has shown a sharply upward curve

(Rodrick 1997; Brady and Wallace 2000). Economic inequality has increased during this

time period as well, whether measured between individuals, between nations, or within

nations (Berry et al 1991; Ram 1992; Korezeniwicz and Moran 1997). Milanovic (1999)

estimates that the world Gini index for the richest to the poorest income groups increased

one percent per year between 1988 and 1993, (from .63 to .66.), while the World

Development Report finds that GDP per capita in the richest 20 countries has grown to

37 times that of the poorest 20 nations, a gap that has doubled in the past 40 years

(2000/01).

There is an increasing awareness among both academic scholars and development

professionals that globalization puts certain populations at risk (Rodrick 1997; Birdsall

1999, UNCTAD 2000). This contrasts with a “Washington consensus” among global

elites that emphasized trade and investment liberalization as the panacea to development

problems in the 1980s and early 1990s. The key turning point was the impoverishment

left behind by the East Asian currency crisis of 1997 and subsequent meltdowns from

Russia to Brazil, which produced earnest calls, often by the same elites, to take heed of

the ways in which globalization has had unequal effects among the world’s population,

both within and between nations. Since the agenda-setting success of the “Battle of

Seattle,” ‘globalization’ has become the unifier of diverse grass-root social movements in

a string of large scale protests when the WTO, IMF, or other global policy makers try to

meet. Inequality is back on the global agenda, according to the World Bank’s World

Development Report (2000/01). However, there has been inadequate theoretical analysis

and up to date empirical studies that explain just how contemporary globalization affects

inequality and the well-being of individuals (Paus and Robinson 1997).

To examine the effects of globalization on inequality, we start with world-system

theory, which emphasizes the developmental consequences of global relations between

unequally powerful nations, in particular, relations of dependency. Since the 1970s, much

of this work is concerned with the effects of accumulated investment from transnational

corporations (TNCs) in the developing world, or periphery.1 Specifically, studies of

‘capital dependency’ or TNC ‘penetration’ contend that disproportionate control over

host economies by transnational corporations increases inequality by altering the

development patterns of these nations. Although the vast majority of foreign direct

investment (FDI) is located within developed nations, the impact of FDI on a developing

nations economy is much more significant (WIR 2000). Therefore, world-system

scholars focus on dependency in the form of accumulated stocks of foreign investment as

a share of the host nations GDP.

Our focus on dependency is in sharp contrast with most globalization studies.

While globalization has multiple economic, political, and cultural facets, when studying

1 See, for instance, Chase-Dunn 1975; Bornschier 1981; Rubinson 1976; Bornschier and Ballmer-Cao

1979; Dolan and Tomlin 1980; Evans and Timberlake 1980; Sullivan 1983; Bornschier and Chase-Dunn

1985;Dixon and Boswell 1996; Beer 1999.

inequality most have focused on the effects of international trade, neglecting the

significance of foreign ownership (Rodrick 1997; Ferreira and Litchfield 1998; Lachler

1998; WDR 2000/01). Similarly, with some important exceptions (i.e. Tsai 1995; Dixon

and Boswell 1996; Alderson and Nielsen 1999), most recent cross-national studies of

income inequality have moved away from examining global forces like foreign direct

investment (FDI), focusing instead, on economic and socio-cultural dualism (Williamson

1991; Nielsen and Alderson 1995) or technoecological heritage (Lenski and Nolan 1985;

Crenshaw and Ameen 1994). This is surprising for two reasons. First, dependency

arguments concerning the impact of FDI on inequality have received relatively robust

empirical support, warranting further examination.2 Second, foreign direct investment

has dramatically increased in importance over the past two decades, and is currently the

primary source of resource flows to developing nations (Froot 1993; Tsai 1995). In 1988,

FDI surpassed all other forms of lending as a source of foreign capital to developing

nations (WDR 1991). In 1982, the total value of global inward FDI stock stood at almost

6 billion (US$), by 1990 that figure had reached 1.7 trillion (US$), and by 1999 it had

reached 4.7 trillion (WIR2000). The ratio of world FDI stock to world GDP increased

from 5% in 1980 to 16% in 2000 (WIR 2000). Indeed, less developed countries are

encouraged to attract foreign investment as the primary route to economic growth and

well-being in the contemporary world-economy. Foreign investment is promoted by

development agencies such as the World Bank and the International Monetary Fund as an

efficient way to add to existing domestic pools of capital, technology and entrepreneurial

2 Some seem to conclude that the insignificance of core/periphery dummy variables invalidates any world-

talent (McMichael 1996; Rothgeb 1996; WIR 1991). Implicit in the logic of investment

liberalization is the idea that the free flow of unregulated capital is the best means to

national development (Ranney 1998). It is only recently that traditional development

agencies have begun to realize that FDI may disadvantage certain groups of individuals

and that it is incumbent upon policymakers to safeguard vulnerable populations

(UNCTAD 2000).

This study attempts to specify the conditions under which transnational corporate

(TNC) penetration and other global factors influence changes in domestic income

distribution in the hopes that this knowledge may allow us to develop more effective

policies to mitigate these inequalities. Due to a prior lack of high quality time-series

income inequality data, most cross-national studies of income distribution have employed

regression models with a cross-sectional design. It is clear that the lack of time-series

and longitudinal analyses covering many countries is a significant gap in the literature

concerning cross-national income inequality. Moreover, examining changes in inequality

over time has theoretical and empirical benefits. It enables an analysis of the impact of

contemporary globalization on inequality and an evaluation of the sometimes

contradictory findings of prior research. This paper presents an analysis of the

determinants of change in national income distribution using linear regression models

with a panel design. The data set contains inequality data for 65 nations at two points in

time, circa 1980 and 1995. Our central concern is the effects of economic globalization

on national income distribution, with an emphasis on transnational corporate penetration.

system approach, even though such dummies are the crudest possible measures (i.e. Muller 1988).

Theoretical Perspectives on Cross-National Income Inequality

There are three main world-system arguments concerning the global sources of domestic

income inequality, which follow a rough temporal pattern. The first focuses on

inequalities arising from the concentration of land ownership that generates severe

income inequality. Land concentration begins as a legacy of colonialism and continues

through corporate agriculture (Furtado 1970; Muller and Seligson 1987; Boswell and

Dixon 1990). Many cross-national studies have found a positive association between

land inequality and inequitable distribution of income (Simpson 1993; Crenshaw 1993;

Crenshaw and Ameen 1994). In the examination of the relationship between growth and

inequality, some researchers have found that land redistribution prior to the onset of

economic expansion is a crucial intervening variable (Bowman 1997; Deininger and

Squire 1997).

The second argument emphasizes the export structure of developing nations. Trade

between industrial and industrializing countries creates dependent patterns of unequal

exchange, leading to high levels of income inequality within the developing world (Baran

1957; Frank 1967; Galtung 1971). Export-oriented production for the world market

creates sector dualism in which the primarily foreign-owned export segment of the

developing economy monopolizes internal capital and repatriates profits, stagnating the

domestic sector. The empirical evidence indicates that large export sectors are positively

related to income inequality (Stack 1980; Prechel 1985). The more capital intensive

nature of export production results in higher returns to capitalists at one end and the

underemployment of the indigenous labor force at the other. Moreover, wages are

depressed by limited labor force mobility between sectors, due to lack of skill

transferability, low education levels, and various social and legal barriers (Amin 1976;

Prechel 1985).

Prechel (1985) argues that the productive capacities and structure of the export and

traditional economic sectors of developing nations are linked and not simply temporarily

disarticulated. The growth of the first depends on the stagnation of the latter.

Furthermore, creation of a high-wage high-profit oligopolistic capitalist sector not only

creates a minority of high-income employees, it further increases inequality by

encouraging urban migration and increased competition for unskilled jobs (Evans and

Timberlake 1980). Competition in crowded urban areas reduces wages by decreasing the

bargaining power of labor. These tendencies are exacerbated by the development

strategies most developing nations pursue. Prechel discusses the implications of these

strategies for increasing income inequality, arguing that export-oriented production is

usually foreign-owned or financed. This leaves these economies vulnerable to

fluctuations in the world market, which along with the factors already discussed,

contributes to maintenance of high levels of income inequality. The subordinate position

of peripheral governments vis-a-vis core capital and prominent transnational actors such

as the IMF, decrease their ability to implement autonomous social and economic policies

(McMichael 1996).

Other discussions of trade dependency focus on the character of a nation’s

participation in the global trading system. Commodity concentration refers to the degree

to which a nation’s export role is limited to the production of only a few commodities. In

comparison, nations with a more diversified array of exports have more options in

responding to fluctuations in the world economy, for example being able to better

weather downturns in the global commodity market. It is therefore not surprising that

commodity concentration has been found to have negative effects on physical quality of

life (Ragin and Bradshaw 1992).

While the size of the export sector has been found to increase inequality, other

dependency measures such as commodity concentration and debt service have not

(Weede and Tiefenbach 1981; Prechel 1985; Chan 1989). However, dependency

scholars stress the dynamic nature of global capitalism, and examine how the changing

character of capitalist exchange on a global scale is coupled with alterations in the nature

of the relations of dependency among the participants. Many of these researchers argue

that, for the past twenty to thirty years, cross-national capital transfers are more indicative

of dependency than are trade-based measures (Prechel 1985; Chan 1989).

The third wave of the literature, which is our main emphasis, focuses on foreign

investment as the primary means through which the modern capitalist world-system

creates and maintains intra- and international socioeconomic inequities. Numerous

empirical studies have confirmed a significant association between foreign corporate

penetration and inequality (Evans and Timberlake 1980; Kohli et al 1984; Bornschier and

Chase-Dunn 1985; Chan 1989; London and Robinson 1989; Crenshaw and Ameen 1994;

Dixon and Boswell 1996; Alderson and Nielsen and 1999; Beer 1999). Others have

found foreign penetration effects only in certain geographical regions (Rothgeb 1989;

Tsai 1995). Even those studies that fail to confirm this relationship generally report their

conclusions with reservation and do not dismiss foreign corporate penetration as a

potentially important determinant of income inequality (Weede and Tiefenbach 1981;

Crenshaw 1992).

Three main mechanisms are hypothesized to link capital dependency and social

inequality (Crenshaw and Ameen 1994). First, foreign investment in developing

countries generates large sectoral disparities. The sector dualism in this case is between

the foreign owned and domestic sectors. The former includes a disproportionate share of

the export sector in developing countries, but is not limited to it. Compared to the

domestic sector, higher capital intensity and lower utilization of indigenous labor

polarizes income distributions.

Second, transnational corporations operating in developing nations accrue a

disproportionate share of local sources of credit and repatriate a portion of profits rather

than reinvesting them in the local economy. Most importantly, compared to domestic

capital, they do not facilitate near as many links to local businesses and may even

displace small and medium business suppliers, professionals, and retailers who fuel the

entrepreneurial and professional middle class. The lack of linkages between sectors is the

prime difference between FDI in the periphery, and in the core where linkages are

common and foreign investment has a large multiplier effect on local business.

Finally, the governments of these nations, motivated by the need to attract and

retain highly mobile foreign investment, implement policies and strategies that decrease

the bargaining power of labor and inhibit vertical mobility by the lower classes, while

simultaneously enhancing the mobility and training of the managerial and TNC technical

elite. These policies and strategies include tax concessions, guarantees of profit

repatriation, and labor laws unfavorable to workers (London and Robinson 1989;

O’Hearn 1989). Engineers or MBAs with degrees from the US or EU expect pay and

living conditions similar to their core peers, but one major reason a TNC locates in the

periphery is for low cost labor. While globalization is designed to ever increase the

mobility of capitalists to seek higher returns across borders, workers face decidedly

nationalist laws that criminalize their like behavior.

More recent work along this line emphasizes the decreased autonomy of peripheral

national governments to the workings of the global economy and emerging transnational

actors. Scholars argue that nations that are highly dependent on foreign capital and

encumbered by enormous debt are subordinated to TNCs and multilateral development

agencies such as the IMF and the World Bank (McMichael 1996). These global actors

influence economic and social policies both directly (by tying loan restructuring to the

implementation of structural adjustment policies) and indirectly (by creating a

competitive environment among developing nations for foreign investment that depresses

wages and encourages lowered labor and environmental standards). DeMartino (1998)

asserts that the increased mobility of capital weakens the ability of developing nations to

tax capital and provide social insurance for workers. Liberalized investment rules

undermine national policies that encourage employment and wage enhancement such as

targeted job training, priority hiring and local purchasing requirements (Ranney 1998).

Alternative Theories of Cross-National Income Inequality

The “classic” modernization argument is based on the work of Simon Kuznets (1955,

1963, 1976) who found a curvilinear association between income inequality and

economic growth, and was among the first to develop a theoretical argument to explain

this finding. Modernization theorists argue that wealth concentrates in the hands of a few

entrepreneurs in the early stages of industrialization, as this is the most efficient use of

scarce capital. Increasing the rate of capital investment, both foreign and domestic,

depends on the development of modern economic segments of the economy. Inequality

eventually decreases, however, as modern values and technology diffuses through out the

economy.

Recent research seeks to explain inequality by reference to dualism between a

modern (industrial) and traditional (agricultural) sectors, which harkens back to

modernization theory (Paukert 1973; Cheney and Syrquin 1975; Ahluwalia 1976;

Kuznets 1976). It is not development per se, but dualism and diffusion processes that are

the keys to explaining income inequality (Nielsen and Alderson 1995)3. Dualism is also

found in dependency arguments (Prechel 1985), differing in the type of sectors and in the

predictions of readily increasing sector integration by diffusion alone due to structural

constraints that prevent the lessening of inequality. Nielsen (1994) argues that part of the

effect of development is explained by two transitional processes: sector dualism and

generalized sociocultural dualism. Sector dualism is primarily economic and is

associated with labor force shifts from the low productivity, low-wage traditional sector

to the high productivity, high-wage modern sector. He argues that the movement of labor

from one sector to another increases income inequality as an automatic numerical

consequence, regardless of the level of income inequality within the various sectors. This

notion of sector dualism is drawn from the work of Kuznets (1955) and Lecaillon et al

(1984). Both sectoral labor force shifts and the amount of inequality due to differences in

income between traditional and modern economic sectors of developing nations produce

income inequality. That is, inequality is a function of both the differences in average

incomes between sectors as well as the relative size of the sectors. Urbanization and

internal migration are related to these processes. Indicators such as percent of labor force

in agriculture have been found to be associated with income inequality at lower levels of

development (Crenshaw 1992, 1993; Simpson 1993; Nielsen and Alderson 1997). Direct

measures of sector dualism, the gini coefficients for the difference between agricultural

shares of the labor force and its share of the total income of society, are positively

associated with inequality (Nielsen 1994; Nielsen and Alderson 1995).

Generalized sociocultural dualism is associated with demographic transition. The

transition is the increased rate of growth in population among newly developing countries

(due to reductions in the death rate through the diffusion of medical technology) that is

not yet offset by a reduction in the birthrate. As a consequence, societies experiencing

the transition have a high natural rate of population growth that increases income

inequality through its impact on labor surplus, decreasing the relative bargaining power

of low-skilled workers. Nielsen (1994) argues that any variable associated with

3 Nielsen and Alderson (1995) are hard to classify as their rhetoric is quite critical of world-system theory

development that generates heterogeneity due to partial or selective diffusion will

contribute to inequality. Generalized sociocultural dualism entails all of the dimensions

of industrialism that spread unevenly and therefore affect the distribution of income,

including the diffusion of education and political democracy, which have traditionally be

explained by other scholars from a different perspective.

Another key process is access to education. Some theorists argue that education

allows for the attainment of credentials and skills necessary for employment in the

modern industrial sectors of the economy (Simpson 1990; Crenshaw 1992). This

argument is derived from the modernization perspective in that the relationship is

dependent on national economic growth and increasing internal sectoral complexity.

Educational institutions in the early stages of economic growth are assumed to be

concentrated in urban areas and primarily accessed by elites. As industrialization

continues, however, mandatory and open educational policies are instituted nationally,

allowing for educational attainment and hence increased employment opportunities for

the rest of the population. This occurs through a process of diffusion of institutional

forms and practices from urban to rural areas. It is also a result of the growth of effective

popular demand generated by improving economic conditions.

Some researchers have confirmed this inverted-U shaped relationship between

educational enrollments and income inequality (Simpson 1990; Crenshaw 1992). Others

have found negative effects on income inequality (Weede 1993; Nielsen 1994). Still

others agree that the spread of education is curvilinear related to income inequality, but

while their actual empirical findings confirm penetration effects previously found in similar models.

assert that the relationship is U-shaped (Crenshaw and Ameen 1994). These scholars

argue that along with the spread of educational credentials associated with development

come increased competition for those positions requiring credentials, which reduces wage

differentials between the educated and uneducated and decreasing income inequality

(Nielsen and Alderson 1997). At a certain point of institutionalization, however, a new

set of post-industrial social inequalities are established and the relationship becomes

increasingly positive (Crenshaw and Ameen 1994).4

Finally, development economists focus on the effects of trade openness. Openness

is a neoliberal concept based on the magnitude of a nation’s participation in world trade,

often measured as imports and exports as a percent of GDP. Openness is expected to

increase inequality in developed countries, due to a number of factors/processes: through

‘deindustrialization,’ the replacement of high paying unionized and protected industries,

especially those with low technology and, through competition with imports from

developing countries (Bluestone and Harrison 1988; Braun 1991). Rodrick adds that trade

and investment liberalization alters the relations of production in advanced industrial

economies by both reducing the demand for low-skill labor and increasing the elasticity

of that demand (1997). This affects the unskilled, semi-skilled, and middle management

4 Even when empirical findings do not differ, often divergent theoretical explanations are offered. For

example, Crenshaw (1992, 1993, 1994) and Simpson (1990, 1993) engage in a debate as to whether the

effects of such factors as the diffusion of education and the spread of political democracy are primarily

political or economic phenomena. This is in large part a theoretical, rather the empirical, debate.

by raising their non-wage burden, increasing job volatility and insecurity, in addition to

decreasing their bargaining power.

In developing countries, however, the logic of market liberalization suggests that

openness will cause inequality to fall as employment rises among low skill workers in

export processing zones. This effect may be mitigated where there is a large labor supply,

where employers concentrate hiring on women workers that had previously been outside

the labor market, and where labor laws are repressive. Several studies have argued that

globalization in the production of goods and services widens the divide between skilled

and unskilled workers by making each year of education worth far more in developing

nations (Ferreira and Litchfield 1998; Lachler 1998; WDR 2000/01).

We will consider a variety of other specific arguments below when explaining the

independent variables. But for now, let us turn to the variables and methods utilized in

this study.

Data and Methods

The Measurement of Income Inequality

Many researchers have recognized problems with the data quality and comparability of

cross-national measures of income inequality (e.g. Ahluwalia 1993[1974]; Muller

1993[1984] Hoover 1989). Fortunately, data collection procedures have improved

substantially in recent years and much work has been done in assessing their

comparability. Our primary source for the inequality data is a World Bank data set

constructed by Deininger and Squire (DS) (1996). In order to increase our sample size as

much as possible we supplement the DS data with two other high quality sources: the

ILO’s “Statistics on Poverty and Income Distribution” (ILO) (1996) and the World

Development Indicators 1999 (WDI) (1999). All three data sets were themselves

compiled from various sources, but with careful attention to issues of quality and

comparability.

Our primary source, the DS data set, has proven extremely useful in preliminary

analyses (Deininger and Squire 1996, 1997). As many of the theoretical explanations of

inequality contain a temporal element, lack of high quality time-series data has hampered

empirical testing of these hypotheses. Indeed, Deininger and Squire argue that the use of

inferior data calls into question the results of many studies, especially those examining

changes in inequality over time (1996: 570, 573). They impose stringent quality

standards in the construction of their high-quality data set. Each observation must meet

three requirements: 1) household or individual as the unit of observation, 2)

comprehensive coverage of the population, and 3) comprehensive coverage of income or

expenditure. The researchers argue that their data set improves upon those used

previously in three ways: it contains a larger number of high caliber observations,

includes a greater number of nations and provides a more reliable basis for time-series

analysis.

In terms of the issues discussed by Deininger and Squire mentioned above that

may affect comparison of measures of inequality we take the following steps. Where

possible we attempt to use observations where the income recipient unit is the household

rather than the individual, as household-based measures yield lower estimates of

inequality. Similarly, we selected measures where income is reported net of taxes where

possible, as these tend to generate more equally distributed estimates of fractile income

share. As previously mentioned, we supplement the DS data with ILO and WDI data sets.

The data contained in each frequently overlaps and is quite similar. In terms of

preference, we selected measures based on the income earning population first, then on

economically active population, followed by tax records. Based on the findings of

previous studies, however, we do not expect any systematic bias based on differences in

measurement in these areas (Deininger and Squire 1996; Alderson and Nielsen 1999). 5

Using this methodology, we constructed a 1980 to 1995 panel data set. The

population consists of all nations with a population of over one million for which data

was available. The result is a data set of 65 nations for which we have inequality data at

two points in time. The measurement year of the earlier estimate ranges from 1968 to

1986 (mean of 1979), and the range for the later estimate is from 1988 to 1995 (mean of

1991). The 65 cases average an 11.5 year lag between inequality measures, ranging from

3 years to 21 years6.

5 The majority of indicators in our data set are based on income, but in the interest of expanding coverage

we include 19 expenditure measures. To address the potential error involved, we follow the steps

suggested by Deininger and Squire (1996). We calculated the mean difference between income- and

expenditure-based quintile income shares and adjusted the expenditure-based data in the following way:

adding .0168 to the top quintile, subtracting .0008 from the upper middle, subtracting .0039 from the

middle, subtracting .0009 from the lower middle, and finally subtracting.0115 from the bottom quintile (7

cases in the 1995 data and 12 cases in the 1985 data).

6 The decision to include one case, Cote d’Ivoire, where change in inequality is only measured over three

years was made in the interest of retaining as many African nations in the data set as possible.

Because income data is frequently collected in quintile shares, the two most

commonly used measures in comparative studies of income inequality are the gini

coefficient, which looks at the disparity between equal and actual distribution of income

among quintile shares, and the concentration of income received by the top 20% of the

population. The majority of cross-national studies find no difference between models

based on top quintile income concentration and those based on the gini (although the gini

coefficient has been extensively critiqued on methodological and theoretical fronts (Chan

1989; Hoover 1989; Braun 1991; Muller 1993[1984]). It is, however, for theoretical

reasons that researchers should opt to use income concentration or the gini coefficient.

Muller (1993[1984]) argues that world-system theory points to the concentration

of income at the upper end of the distribution as the crucial indicator of income

inequality. As there is little variation in the bottom 20%, the difference between

concentration in the top percentiles and the gini score is due almost entirely to the

distribution in the middle. As Deininger and Squire (1996) note, when examining

changes in inequality the use of an aggregate measure such as the gini coefficient

obscures the character of shifts in income distribution because “…there is no unique

mapping between the changes in the index and the underlying income distribution…” (p.

567). That is, the gini coefficient for a particular nation may increase, but there is no way

of discerning whether the rise in inequality is due to redistribution from the bottom

quintile to the top or a result of a shift from the middle to the top. Also of theoretical

importance is that the use of upper proportional shares of income indirectly measures

asset inequality, another significant dimension of economic stratification (Boswell and

Dixon 1993). Considering these theoretical concerns, in the analyses that follow, we

concentrate on examining change in top quintile income share.

The dynamics of income inequality, 1980-1995

The data suggest that, overall, inequality as measured by top quintile income share has

increased within nations during the 1980s and early 1990s7. On average, the percent of

national income accruing to the top twenty percent of the population rose by 2.4%.

While this may seem rather small, it is important to consider that in real terms this

amount is quite substantial, especially in nations where great numbers of people live in

poverty. In El Salvador, for example, 2% of national incomes is equal to roughly

US$189,756,000.

In addition, the data indicate that the long held assumption that inequality changes

rather slowly over time does not hold for the more recent time period under study.

Nearly 30% of the nations in our data set exhibit top quintile changes of +/-10%, and

60% show changes of +/-5%. Of these nations, the majority experienced worsening

income inequality. For example, fourteen nations (22%) exhibited increases in top

quintile income share of over 10%, and twenty-four nations (37%) had increases of over

5%. An interesting issue is whether the generally held assumption that income inequality

is a relatively stable feature was based on flawed or incomparable data or, alternately,

whether this was true for earlier periods but no longer holds in the era of increasing

economic globalization.

7 Top quintile income share for all nations at both points in time are included in Appendix A.

The advanced industrial nations as a whole tend to have lower inequality than

other countries, especially the Scandinavian social democratic nations. There was not a

tremendous amount of change in inequality among developed countries, with the

exception of the “Anglo-American” liberal market nations: Australia (+5%), the United

States (+6%), and the United Kingdom (+8%). What is striking is that the latter all began

the decade with high inequality relative to other advanced industrial nations, defying any

ceiling effect on inequality provided by political institutions or global convergence

expected by world cultural theorists. Stallings (1995) suggests that all share similarities in

their “Anglo-American” variant of capitalism and many studies have documented rising

inequality in these nations in recent decades (Bluestone and Harrison 1988; Braun 1991;

Nielsen and Alderson 1995, 1997).

Measurement of the Independent Variables

The independent variables we include in our regression models are representative of the

theoretical perspectives that address cross-national income inequality discussed in the

literature review. This will allow the perspectives to compete freely with one another and

allow for a fuller specification of the empirical models. Where data is not available for

the target year for a substantial number of cases, we use averages over a specified time

period in order to retain as many nations as possible in the models.

World-System Indicators:

TNC Penetration (World Investment Report 1997):

Capital dependence is characterized by significant amounts of foreign control over the

national economy, represented by the accumulation of stock owned by transnational

corporations. The most common operationalization of investment dependence is

transnational corporate penetration (PEN), the ratio between inward foreign direct

investment (FDI) stock and GDP. This measure has been found to be significantly

associated with high levels of inequality in developing nations.8 The WIR provides

measures of inward FDI as a percentage of gross domestic product for multiple years.

On average, TNC penetration rose over the decade; nations averaged a 7% pen

score in 1980, whereas in 1990 that average had increased to 11%. The majority of

nations in the data set showed dramatic increases in foreign accumulation of stock

relative to the size of their economies, the average percent increase from 1980 to 1990

was 184%. Of the 73 nations for which we have pen data for both 1980 and 1990, all but

13 (18%) increased their pen scores, many quite substantially. Over half of the sample

(60.3%) increased pen by over 50%. The majority of nations actually increased their

dependence on foreign ownership by relatively large amounts, 42.5% increasing by more

than 100% and 20.6% increasing by more than 200%.

Exploitation (World Data 1995):

Following Boswell and Dixon (1993), class exploitation is measured as wages and

salaries as a percent of value added in manufacturing [(1 - WSPVA)/WSPVA].

Manufacturing is the only sector for which data is available, but is presumably correlated

to other sectors. This measure captures the cross-national disparity in bargaining power

between capital and labor in determining the returns to increases in productivity due to

factors such as differences in capital mobility, labor laws, social welfare, and so on. In

8 See Evans and Timberlake 1980; Kohli 1984; Bornschier and Chase-Dunn 1985; Chan 1989; London and

general, the Scandinavian social democratic countries have the lowest rates and the

highest are found in Latin America and among heavily indebted countries, including

those in Eastern Europe (even while nominally ‘communist’ (Boswell and Dixon 1993).

Although rarely used in cross-national studies of inequality, it has been found to be

positively associated with income concentration (Boswell and Dixon 1993)9.

Trade Dependency (Pen World Tables 5.6; World Development Indicators 1999;

UNCTAD various years):

We include three trade-related measures in the analyses that follow: trade dependency,

trade openness, and commodity concentration. Trade dependency is operationalized as

exports as a percentage of GDP, and is meant to capture the degree to which the national

economy is oriented to the global trading system. This measure has been found to have a

positive effect on inequality in some studies (Stack 1980; Prechel 1985), but others have

found insignificant effects (Crenshaw and Ameen 1994). We will also include a measure

of commodity concentration, which have been found to have negative effects on physical

quality of life (Ragin and Bradshaw 1992). Commodity concentration refers to the

degree to which a nation’s export role is limited to the production of only a few

commodities. In comparison, nations with a more diversified array of exports have more

options in responding to fluctuations in the world economy, for example, being able to

better weather downturns in the global commodity market. Our measure of commodity

concentration is the Gini-Hirschman index of concentration defined over 239 three-digit

standard international trade classification categories of exports as calculated by the

Robinson 1989; Crenshaw and Ameen 1994; Beer 1998.

United Nations Conference on Trade and Development (UNCTAD), and is available for

many years. Examining multiple trade measures in the models will help to specify the

potentially different effects of different forms of participation in the global trading

network. A recent study found that, while overall trade openness had a positive impact

on indicators of economic and social development, commodity concentration had

negative effects on development (Beer and Meyer 2000). This research illustrates the

importance of differentiating between magnitude of trade and the character of such

participation in the global economy when testing dependency hypotheses.

As mentioned in the literature review, some scholars argue that the harmful effects

of various forms of dependency are period dependent, and that trade dependency has

lessened in importance in recent decades (Prechel 1985; Chan 1989). Although based on

recent theoretical work we anticipate that investment dependence will exhibit the

strongest association with income inequality in the developing world for the past two

decades, we include other forms of dependence such as trade and debt dependency in

order to explore how, as Ragin and Bradshaw (1992) argue, inequality mirrors changes in

dependency over time.

Debt Dependency. (World Development Indicators 1999):

We use two measures of debt dependency in order to explore the effects of the 1980's

debt crisis on income distribution. The first is a simple measure of debt dependency,

which is measured as debt service as a percentage of GDP. Debt service is defined as the

series of payments of interest and principle required on a debt. Because this variable has

9 This measure is averaged over 1980-1986 for Nepal

been found to be insignificant by others (i.e. Chan 1989, Prechel 1985), we do not expect

this variable to have a significant influence on inequality. The second is a measure of

debt pressure, which is the ratio of the sum of principle repayments and interest payments

of external debt to export values, or debt service as a percent of total exports. Some have

argued that export-oriented strategies allow nations to escape debt crises. Stallings

(1995) notes that in the early-1980s South Korea had similar levels of debt/GNP as many

Latin American nations but had a much lower debt service/export ratio, accounting for

the lack of a debt crisis similar to that experienced in Latin America.

One limitation of foreign debt data is that it is rarely collected for advanced

industrial nations. As much of the emphasis of the international financial institutions

(such as the International Monetary Fund) that compile data on international debt is on

formulating policies aimed at providing avenues of relief for nations with severe debt

burdens, this lack of data for the relatively wealthy nations is unsurprising. These nations

face negligible amounts of debt relative to the size of their economies; in such cases debt

is not the crucial factor in economic planning or national politics, as it is in nations facing

debts many times the size of their economies. As our interest is in examining the ways in

which heavily indebted nations are affected by their debt burden, we take the step of

assigning a score of .0001 for the advanced industrial nations. Dropping these nations

from the equations looking at debt is not a desirable option, as this would severely

decrease our sample size. Furthermore, these nations are an important reference group

for more heavily indebted countries, simply omitting them from the equation would

effectively truncate the range of variance in international debt among nations. However,

caution should be used in interpreting the debt models. Therefore, we also run the

equations with the unrecoded debt data in order to ensure that the results are not due to

data manipulation.

Sectoral disarticulation (World Development Indicators 1998):

This measure is meant to capture the underabsorbtion of labor in the economic sectors of

the economy, it reflects the disproportionality of productivity across sectors. Some have

argued that disarticulation may be one of the mechanisms through which TNC pen affect

income distribution (Breedlove and Armour 1997). This measure is similar to the sector

dualism measure, but that measure only captures dualism in agriculture. It is constructed

by taking the sum of the differences between a sectors share of the labor force and that

sectors contribution to GDP, across all three major sectors of the economy: service,

industry and agriculture.

Change in the female labor force (World Development Indicators 1998):

Production for the world economy is increasingly young and female (Glasberg and Ward

1993; Fernandez-Kelly 1984). The motivations attributed to the TNCs that employ these

workers are that they are cheaper and easier to control than male workers. Therefore, it

may be that the entry of women into formal employment depresses wages both directly,

and through the threat they pose to male workers. We include a measure of change in the

percent of the labor force that is female to explore this issue. We measure the change in

female formal employment rather than the level of employment in order to differentiate

the relatively recent rise in female formal employment in the developing world from the

relatively high numbers of women workers in the advanced industrial nations.

Change in urban population. (World Development Indicators 1998):

This measure is meant to capture “overurbanization,” the rapid migration of rural workers

to the cities. This relocation to urban areas is assumed to be a result of the mechanization

and transnationalization of the rural sector in developing nations (Prechel 1985).

Overurbanization contributes to inequality by increasing the number of unemployed

workers and thereby decreasing wages. We operationalize this measure as change in the

percent of the population living in urban areas over a five year time period. We use this

instead of a measure of percent urban population in order to capture the effects of a rapid

increase in urban dwellers, rather than the slower trend of rising urbanization seen

throughout the world.

Change in service labor force (World Development Indicators 1998):

Evans and Timberlake (1980) argue that inequality is greater within the tertiary sector,

even though average incomes are higher in the tertiary as opposed to the agricultural

sector. This is because incomes are more polarized in the service sector, which contains

both professionals and low-skilled workers. Growth in the tertiary sector contributes to

inequality through its creation of a readily available reserve army of labor, which

depresses wages in other sectors through weakening of labor’s bargaining power.

Foreign investment increases the tertiary sector and change in tertiary sector has a

positive effect on income inequality (Evans and Timberlake 1980).

Alternative Indicators:

Level of Development. (Penn World Tables 5.6; World Development Indicators 1999):

We use real gross domestic product per capita as a measure of level of economic

development. Real dollars are those adjusted for differences in domestic prices using

purchasing power parities. In various equations we also include its square to capture the

hypothesized curvilinear effect of development on inequality found by some studies

(Weede and Tiefenbach 1981; Weede 1980, 1993). However, as mentioned previously,

many recent studies have called into question the robustness of the Kuznet’s effect.

Nielsen and Alderson (1995), for example, argue that development is a proxy for the

processes captured by the following four indicators:

Education (World Development Indicators 1997):

We operationalize this variable as secondary school enrollments and its inclusion in the

models is meant to capture levels of domestic human capital. Nielsen and Alderson

argue that high levels of secondary school enrollments indicate “skills deepening”

(Nielsen and Alderson 1995). The nature of this relationship, however, has not been fully

settled by the empirical evidence. Some have found education to be negatively related to

inequality (Weede 1993; Nielsen 1994; Nielsen and Alderson 1995), others have found

an inverted-U shaped pattern (Simpson 1990; Crenshaw 1992) and still others have found

a U-shaped association (Crenshaw and Ameen 1994). Therefore, this variable is included

in the models both singly and in quadratic form in order to explore this relationship

further.

Modern Sector dualism (World Development Indicators 1998):

As per Nielsen and Alderson (1995), this variable is operationalized as the absolute value

of percent of the labor force in agriculture minus agriculture as a percent of GDP, and is

expected to have a positive association with income inequality.

Percent of labor force in agriculture (World Development Indicators 1998):

While some have found a large agricultural sector to be related to high inequality

(Simpson 1993; Crenshaw 1993, 1992), they have not included a sector dualism measure.

Nielsen and Alderson (1995) argue that, when controlling for intersectoral differences in

inequality, a large agricultural sector (having a relatively more equal distribution of

income) should exhibit a negative association with the inequality indicators.

Natural rate of population increase (World Development Indicators 1999):

This measure is operationalized as the crude birth rate minus the crude death rate and is

intended to capture the effect of the demographic transition and generalized sociocultural

dualism on income distribution (Nielsen and Alderson 1995). It is expected to have a

positive effect on inequality (Ahluwalia 1976; Bollen and Jackman 1985; Simpson 1990;

Nielsen 1994).

Economic growth. (Penn World Tables 5.6):

We operationalize this measure as the percent change in real gross domestic product over

the time period of interest. The nature of the relationship between income inequality and

economic growth is far from settled in the literature, and is usually based on theories

regarding level of development. Central to the modernization/developmental perspective

is the notion that for an economy to develop in its early industrial phase, a certain level of

inequality is required (inequality/growth trade-off). Some argue, however, that there may

be a feedback effect where high levels of national income inequality put a drag on long-

term economic growth (Deininger and Squire 1997). It may also be the case that high

levels of growth allow for the maintenance of prior levels of inequality, as popular

pressure for redistribution may be muted if poorer segments of the population are better

off in absolute terms due to the expansion of the national economy.10

Trade openness (Pen World Tables 5.6; World Development Indicators 1999):

Trade openness is operationalized as imports and exports as a percent of GDP, and is

meant to measure the overall magnitude of a nations participation in international trade.

Openness is expected to increase inequality in developed countries, but may reduce it in

developing ones. As such, we consider interaction terms with GDP and GDP squared.

Political-Institutional Indicators:

Political Democracy (Jagger and Gurr, 1996):

The effect of national state structures on income inequality is a subject of contentious

debate in the literature. The results of studies examining the relationship between

democracy and inequality are inconclusive and contradictory (Hughes 1997). Many

studies find no causal relation between the two (Bollen and Jackman 1985; Chan 1989;

Weede 1989, 1993), while others find a negative association (Stack 1980; Nielsen 1994;

Muller 1988, 1995). Others conclude that democracy has a net positive effect on income

inequality (Simpson 1990; Crenshaw 1992). In addition, Bornschier and Chase-Dunn

(1985) argue that both position within the spacioeconomic hierarchy as well as

10 For the 1995 models, this measure is averaged over 1980-1988 for Tanzania, and over 1980-1986 for

Ethiopia and Nepal.

sociopolitical processes have an impact on levels of national economic inequality.

Clearly, this process needs further specification.

The democracy measure reflects the general openness of political institutions

within the polity (Jagger and Gurr, 1996). The measure is an additive index based on the

codings of three authority characteristics. The first is the competitiveness of

participation, or the extent to which non-elites are able to access institutional structures

for political expression. The second concerns executive recruitment competition,

referring to the extent to which executives are chosen through competitive election and, if

so, the degree of opportunity available to non-elites to attain executive office. Finally,

the democracy index adds executive constraints in an assessment of the operational (de

facto) independence of the chief executive.

Social Democracy (The Statesman’s Yearbook, various years; Huber and Stephens

2000):

An additional indicator of the effects of state socioeconomic and sociopolitical processes

will also be included in the models, a social democratic government dummy variable.

This variable addresses the argument that Muller (1993[1984]) and others have made: the

redistributive effects of leftist governments mitigate the relationship between income

inequality and other variables. Positive effects of social democratic parties on income

distribution have been found by some researchers (Hewitt 1977; Huber, Ragin and

Stephens 1993), but are not systematically included in most studies of regime type and

inequality. It may be that separating the effects of social democratic versus non-social

democratic government will explain the contradictory findings surrounding the effects of

democracy discussed above. Many studies have found that the cooperative public-private

structures that characterize social democratic regimes promote overall social welfare,

increasing national economic security and equality (Hicks 1988; Kenworthy 1995).

We construct the social democratic government indicator based on two sources.

The first is whether there was majority participation in parliament by a social democratic

party based on information obtained from The Statesman’s Yearbook (various years).

This information was compared with the categorization of social democratic or Christian

democratic state in Huber and Stephens (2000).

An Analysis of Change in Top Quintile Income Share, 1980-1995

Descriptive statistics and a correlation matrix of the variables used in the panel models

are shown in Tables 1 and 2. Several variables were logged to correct for skewness:

TNC penetration, real GDP per capita, exploitation, agricultural density, and average

exports as a percent of GDP. As mentioned earlier, income inequality increased in the

majority of nations in our sample. On average, top quintile income share rose in the 65

nations for which we have data at two points in time, from 45.51% to 46.80%. Also of

interest is the increase in average TNC penetration over the decade; nations averaged a

7% pen score in 1980, whereas in 1990 that average had increased to 11%. The majority

of nations in the data set showed dramatic increases in foreign accumulation of stock

relative to the size of their economies. The average percent increase from 1980 to 1990

was 184%. Of the 73 nations for which we have pen data for both 1980 and 1990, all but

13 (18%) increased their pen scores, many quite substantially. Over half of the sample

(60.3%) increased pen by over 50%. The majority of nations actually increased their

dependence on foreign ownership by relatively large amounts, with 42.5% increasing by

more than 100% and 20.6% increasing by more than 200%. The data indicate that the

acceleration of global investment is indeed one of the characteristic features of the

contemporary era.

[Table 1& 2 here]

Regression Analysis of Change in Top Quintile Income Share, 1980-1995

Table 3 presents a series of models predicting top quintile income share circa

1995 with a panel design. Equation A indicates that the lagged dependent variable, top

quintile income share in 1980 is a strong predictor of subsequent income inequality,

accounting for 68% percent of the variation. Nations with high levels of initial inequality

were more likely to have increased inequality in the later time period. With the inclusion

of logged TNC penetration in 1980 in equation B, the model reveals that, along with the

positive effect of the lagged dependent variable, increases in elite income concentration

are also associated with investment dependence. That is, pen is related to income

inequality in multiple ways. Structurally, high levels of investment dependence are

associated with high levels of income inequality, as seen in many other studies (Evans

and Timberlake 1980; Kohli 1984; Bornschier and Chase-Dunn 1985; Chan 1989;

London and Robinson 1989; Crenshaw and Ameen 1994; Beer 1999). Dynamically, high

levels of TNC penetration are related to larger increases in top quintile income share over

time. This is a significant finding that provides important support for the Dependency

hypothesis that states that relations of international economic dependence result in

increases in domestic income inequality for dependent nations. It is not simply that high

levels of pen and inequality are linked, these findings provide empirical evidence that the

globalization of investment that characterizes the contemporary world is related to

worsening income distribution in host countries.

[Table 3 about here]

Equation C indicates that TNC pen has an influence on change in income

inequality both in terms of its level and its increase. This model reveals that not only is

the magnitude of foreign economic ownership of the domestic economy associated with

higher increases in inequality, but that rising income concentration is also related to the

percent increase in investment dependence from 1980 to 1990. Over the decade, nations

that had high levels of foreign stock ownership and those that saw large increases in

foreign ownership experienced relatively greater intensification of income inequality

within their borders. This measure has not been empirically tested prior to this study and

further strengthens the support for Dependency arguments. For the first time we may say

that the quantitative cross-national evidence supports the contention that investment

dependence is associated with high income inequality, and that income concentration

among the elite rises as a nation’s dependence on foreign investment grows.

In order to more fully specify top quintile income share and to test the robustness

of the pen effect, other variables are added to the model in equations D through J.

Equation D includes the lagged dependent variable, logged pen, change in TNC pen, and

a measure of average secondary school enrollments (1980-1985). The education measure

has a significant and positive association with change in top quintile income share,

lending support to human capital arguments that assert that investments in post-primary

education are likely to lead to improvements in income distribution because skilled

workers are relatively more productive than unskilled workers. The effects of the other

variables in the model remain the same, and the significance of the pen effect is

increased. This is because controlling for level of education captures some of the

contextual effects of pen. The positive effect of investment dependence on increases in

income concentration are somewhat mediated by the level of secondary school

enrollments, for example in the advanced industrial nations. With the inclusion of the

education measure, the model then captures the effects of pen on income inequality

regardless of level of human capital development.

Equation E adds economic growth (1980-1990) to the model, which appears to

have a mild negative association with increases in top quintile income share. This

supports the assertion that improvements in economic productivity mitigate against

increases in inequality. However, with the inclusion of logged real GDP per capita in

equation F, the effect of growth weakens. With the inclusion of other variables in the

equations that follow, the relationship between growth and inequality becomes unstable

and never attains significance. Gains in economic growth should promote employment

and hence reduce inequality. However, the insignificance of this measure is due to the

inclusion of the other independent variables in the model, which are themselves

components of national economic growth. As was found in the cross-sectional models,

when we account for the factors of production --- productive labor (secondary school

enrollments) and domestic capital (logged real GDP per capita) --- we have captured the

impact of economic growth on national income distribution dynamics.

Therefore, we drop the growth measure from the model and retain education and

level of development (equation G). Logged real GDP per capita has a positive

association with increases in income inequality, which at first may seem counter-

intuitive. According to Modernization theory, wealthier nations should have lower

inequality as a result of the re-articulation hypothesized to occur as nations move along

the trajectory of economic development. However, many researchers have noted

growing inequality in advanced industrial nations in recent decades as they experience

the shift toward service-based economies (Bluestone and Harrison 1988; Braun 1991).

Nielsen and Alderson (1995, 1997) hypothesize that Kuznets’ “inverted-U” may in fact

take the form of a wave, a pattern that emerges when we combine the “inverted-U” with

the rise in inequality observed in nations at high level of economic development.

In equation H, we add a variable meant to capture national shifts in sectoral

employment, the relative change in employment in the non-agricultural sectors of the

economy between 1980 and 1985. This measure has the effect of decreasing income

concentration among the top quintile as a result of the equalizing impact of a more

productive labor force. All nations decreased the percent of economically active workers

in the agricultural sector over the time period, but those that increased proportional

employment in other sectors to a greater extent were less likely to experience worsening

income distribution. TNC penetration and change in pen remain significantly positively

associated with increases in income inequality.

Equation I adds a measure of exploitation in manufacturing to the model, which

has a positive relationship with increase in top quintile income share. This supports the

need to specify different forms of employment when examining cross-national income

distribution. While employment in non-agricultural sectors generally mitigates against

increases in income concentration, manufacturing employment in highly exploitative

firms and that which is a result of TNC investment, aggravates inequality.

Finally, in equation J we add average level of institutional democracy to the

model. This measure has a mildly dampening effect on increases in top quintile income

share, net of the effects of the other variables. While this measure has no effect on level

of inequality, as seen in other studies (Bollen and Jackman 1985; Chan 1989; Weede

1989, 1993) it seems that democracy has an equalizing impact on changes in inequality.

That is, when classes occupying the lower end of the income distribution gain access to

political institutions there is pressure to provide education and social services that may

reduce inequality. At least, it makes it more difficult to increase the income inequality

that already exists (although not impossible, witness Reagan and Thatcher). This is a

significant finding that may account for the empirical differences of various cross-

sectional studies of democracy and national income inequality.

The final panel model in Table 3 (equation J) explains almost 77% of the variance

in change in top quintile income share between 1980 and 1995. The level and the percent

increase in investment dependence, level of economic development, and exploitation in

manufacturing are associated with worsening income inequality. Improvements in cross-

national income distribution are related to levels of secondary school enrollments,

increases in the non-agricultural labor force, and level of political democracy. This

model will be discussed further in the conclusion.

[Table 4 here]

Table 4 adds other variables of interest to final panel model. These measures

capture various dimensions of particular domestic characteristics. The models reveal that

none of the additional variables attains the desired level of significance net of the other

measures in the equations, and in some cases certain relationships among the base

variables are changed due to multicollinearity. The national-level variables in equations

A through H indicate that none of these factors have a significant effect on change in top

quintile income share net of the variables included in the base model. This should not be

taken as an indication that these additional variables do not matter in and of themselves,

for example in certain nations or under specific circumstances. However, because

nothing adds significantly to our model, these findings show that our set of variables is

robust to alternative theories that seek to explain the dynamics on income inequality.

[Table 5 here]

In Table 6, we turn to an examination of the global-level variables that reflect the

nature of economic exchange with other nations in the world economic system via trade

and debt. As in Table 5, in each of the regression equations we include our base model

from Table 4, Equation J. These equations show that none of these other measures of

economic globalization significantly contribute to change in inequality in ways that are

not captured by the set of variables included in our base model. As we noted above, this

should not be seen as evidence that relations of trade and debt between nations have no

effect on national income distribution, but rather that the ‘story’ told by our model is

robust to alternate explanations of the dynamics of income inequality. The results

presented in Table 6 support the work of scholars who argue that the structure of

inequality in the contemporary era is increasingly affected by investment-based, as

opposed to trade-based, linkages (Stack 1980; Chen 1989; Ragin and Bradshaw 1992).

Discussion and Conclusion

As noted in the introduction, inequality is back on the agenda for many development

agencies. This is in part due to the failure of the past few decades to significantly reduce

global poverty in an era of increasing liberalization despite robust economic growth

(Milanovic 1999). The World Bank’s World Development Report notes that 2.8 billion

of the world’s 6 billion individuals live on less then $2 per day (WDR 2000/01).

Although the percent of individuals living in poverty has declined somewhat, the absolute

number of poor people has increased (Chen and Ravaillon 2000). Moreover, global

income inequality has rapidly grown in the past few decades (Berry et al 1991;

Korezeniwicz and Moran 1997). Development agency scholars have traditionally

stressed economic growth generated through integration with the world economy as the

primary route to improving the lives of those in developing nations (see, for example,

Dollar and Kray 2000). Now, however, there is an increasing emphasis on growth with

equity, as many recent studies have found that the benefits of economic growth to the

poor are highly dependent on the existing level of inequality within nations (WDR

2000/01; Weisbrot et al 2000; Wodon 1999). Chen and Ravaillon provide empirical

evidence indicating that inequality is a constraint on pro-poor growth (2000). The sense

that equitable distribution of income within nations is an important precursor to achieving

widespread and beneficial economic growth is becoming more widespread.

Recently, development scholars have begun to explore the ways in which

globalization puts nations at risk of increasing income inequality (Birdsall 1999). Rodrick

argues that the primary challenge for the world economy is “…ensuring that international

economic integration does not contribute to domestic social disintegration.” (1997p. 2).

As noted, a gap in most of these recent studies is that they focus on the effects of

international trade, neglecting the significance of foreign investment.11 Globalization

heightens the vulnerability of certain groups, not only in developing nations but also in

advanced market economies (UNCTAD 2000).

The analysis presented here illustrates the usefulness of panel models in

understanding the dynamics of income distribution. As Kohli et al (1985) points out,

while cross-sectional analyses may reveal long-term structural tendencies, panel analyses

are essential for uncovering causal forces affecting changes in income inequality. As the

global economy increases in both rapidity and volume, these changes may have greater

consequence. The panel models provide evidence of a strong link between foreign

capital dependence and income concentration among the elite. Moreover, this

relationship is revealed to have both structural and dynamic effects, both level of TNC

pen and its percent increase are positively associated with change in top quintile income

share. As discussed in the introduction, both pen and inequality increased quite

dramatically over the period, and this trend appears to be continuing.

11 One recent important exception is Brady and Wallace, who found that FDI has a negative impact on the

political and economic power of workers in U.S. states (2000).

The empirical models indicate that, over the 1980s and early 1990s, inequality

grew in nations that had the following characteristics: relatively high and increasing TNC

penetration, lower secondary school enrollments, relatively high GDP per capita, elevated

rates of exploitation in the manufacturing sector, lower employment growth in the non-

agricultural sectors overall, and also lower levels of institutional democracy.

However, other variables suggested by scholars working in the world-system

perspective are not supported by the models presented in this study. Net of the base

model variables, change in urban population, change in the female labor force, and

disarticulation are not related to significant increases in top quintile income share over

the 1980s and early 1990s. In addition, variables associated with the Modernization

approach of Nielsen and Alderson, sector dualism and population increase, do not receive

empirical support when the base variables are included in the model. The insignificance

of certain indicators, such as disarticulation and sector dualism, may be due to

multicollinearity problems as a result of their strong association with education and level

of economic development. However, these indicators certainly warrant further empirical

exploration.

The political-institutional indicators also deserve more attention from scholars of

cross-national income distribution. The social democracy measure does not significantly

predict change in top quintile income share net of the base model variables, but the

institutional democracy measure has a mildly equalizing effect over the time period. This

indicates that nations with open political institutions are less likely to increase income

inequality. This may explain the contradictory findings of much of the empirical research

on inequality and democracy. Although many find no clear relationship between a

nation’s level of political democracy and its level of income inequality (Bollen and

Jackman 1985; Chan 1989; Weede 1989, 1993), the evidence presented here suggests that

the measure of power given to citizens in nations with accessible political institutions

mitigates against increasing income concentration.

Turning to the global-level variables, this work supports scholars such as Stack

(1980), Chen (1989), Ragin and Bradshaw (1992), who argue that relations of

dependency among contemporary nation-states are increasingly structured by investment-

based linkages, as opposed to the earlier salience of trade-based ties. Whereas World-

system theories focused on issues of export structure and commodity or partner

concentration, the new wave of work in this tradition examines the increasing importance

of the character of investment relations in shaping the development potentials of

participants in the global economy. This study provides empirical support for these

notions, indicating that the greater a nation’s capital dependency, the more concentrated

income becomes among the elite members of the population.

In addition, some factors associated with alternative theories receive support.

Although not curvilinear, development has a positive relationship with income inequality

net of the other factors included in the model. While this does not fit a traditional

modernization view of the world, it tells us a great deal about the kind of unfortunate

world we live in today. Nielsen and Alderson (1997) have argued that the Kuznets

“inverted-u” actually takes the form of a wave if you consider the relatively recent

increase in inequality among the wealthy nations. The empirical evidence also supports

the importance of education for decreasing inequality; the education variable consistently

had a robust and highly significant negative association with change in top quintile

income share. Nations with high levels of secondary school enrollments appear to

experience fewer increases in income concentration, as human capital becomes more

widely dispersed. Nielsen and Alderson (1995) argue that this variable indicates a “skills

deepening,” where the accumulation of various productive skills generates income for

greater proportions of the population by providing them with the expertise to fill more

positions, thereby equalizing incomes. This suggests that nations with a desire to

decrease income inequality within their borders would do well to invest in programs that

increase secondary school enrollments. Expanding human capital, especially among

women, is the surest way to increase labor’s share of the productivity gains in value

added.

In countries that seek to reduce inequality without sacrificing economic

development, what does our research imply should be helpful, besides increasing

education? Expand democratization of institutions, which implies an equal voice for the

lower classes in elections, no sweet heart deals for corporations or ‘crony capitalism,’ and

a legal environment that allows free expression, union organizing, and so on. More

directly from the world-system approach, reducing exploitation in manufacturing and

elsewhere would increase their income share. However, this is tricky as capital reinvested

from the capitalist’s share (as opposed to consumed or sent abroad) increases growth and

development. As mentioned above, the most productive, although highly limited, route is

to expand education and skill training, especially among women and other more easily

exploited populations. Related productive steps include banning child labor and

sweatshops, guaranteeing occupational and environmental safety, improving public

health and welfare and so on, although the cost of these rises for poorer countries.

Increasing bargaining power through unionization and public policy would be the

broadest and most direct route, but is the most fraught with risk from capital flight. Long

term success under globalization will require international union organizing and changing

the rules of global exchange in the WTO, IMF, and other institutions.

Finally, TNC penetration and the acceleration of penetration have an impact on

income distribution beyond what one would expect based on prior levels of inequality

and various domestic factors. Can a country reduce TNC penetration, which increases

inequality, without reducing foreign investment, which increases growth? At first, this

may seem impossible as penetration is a factor of FDI. However, one of the main

problems of FDI in developing countries is the lack of linkages to domestic businesses.

Policies that expand linkages, including ones that make local capital more useful, would

multiply domestic growth (and a local middle class) alongside FDI, keeping the degree of

penetration in check. Even if penetration were to increase, the negative effects would be

muted, especially if matched with the policies listed above.

Globalization does indeed appear to put nations at risk of increasing inequality

(Birdsall 1999). Although some have dismissed the fears of those who have expressed

apprehension regarding the effects of accelerated globalization on social welfare as

“protectionist” or “inward-looking,” the study presented here supports the idea that undue

reliance on foreign investment may in fact benefit elite segments of the population over

others. The empirical evidence supports the developing view that globalization puts

certain populations at risk (Rodrick 1997). Moreover, the results of the study indicate

that this effect is not limited to only developing nations. The data suggest that those

advanced industrial nations with relatively high and increasing levels of foreign

investment, such as the United States and the United Kingdom also saw increased

inequality over the 1980s and early 1990s. That these very nations, along with the

majority of global financial institutions, vigorously espouse the continuing trend toward

liberalization of foreign investment policies and assert that fears of globalization are

unwarranted, is somewhat ironic. Global investment may certainly be likened to a rising

tide, and it may be a tide that lifts all boats, but the empirical evidence presented here

indicates that it surely lifts some boats higher than others. This study suggests that

dependence on foreign capital as a development strategy can be devastating for nations

concerned with equality. Net of other factors, foreign capital dependence benefits the

elite segments of the income-earning population over the poorer eighty percent.

Table 1 –Descriptive Statistics

N Minimum Maximum Mean S.D.top 20 (1980) 66 32.10 68.00 45.51 8.60top 20 (1995) 86 33.80 65.18 46.80 8.56TNC pen 74 0.00 0.53 0.07 0.08% ch pen 73 -0.99 29.50 1.84 4.43growth 76 -0.13 1.43 0.31 0.29rgdppc 83 322.00 15295.00 4416.49 3990.37av disart 68 -25.96 8.06 -2.61 4.14av sec dual 70 -1.47 61.31 22.86 17.49pop inc 84 -1.10 37.80 19.51 10.73av sec ed 78 3.00 105.00 50.35 30.76% ch non-ag lf 84 -0.18 51.24 9.44 8.16av dem 84 0.00 10.00 3.91 4.27% ch urb pop 85 -2.18 34.12 6.79 6.86av exploit 71 0.46 7.10 2.30 1.43av trade open 85 11.53 368.2 62.67 45.96av export/gdp 78 3.40 181.94 28.37 23.56av com con 75 0.08 0.95 0.36 0.21av debt/gnp 71 0.00 18.68 5.22 4.80av debt/exp 73 0.00 58.38 18.51 14.78% ch fem lf 82 -7.42 37.02 7.31 9.19% ch svc lf 83 -12.42 57.79 17.48 12.53soc dem 88 0.00 1.00 0.14 0.35

top 20 (1980): top quintile income share, circa 1980; Deininger and Squire 1996, ILO 1996, WIR 1999.top 20 (1995): top quintile income share, circa 1995; Deininger and Squire 1996, ILO 1996, WIR 1999.TNC pen: TNC penetration, 1980; WIR 1998.% ch pen: percent change in TNC penetration, 1980-1990; WIR 1998.growth: percent change in real GDP per capita, 1980-1990; Penn World Tables 5.6.rgdppc: real GDP per capita, 1980; Penn World Tables 5.6.av disart: average disarticulation, 1980-1985; WDI 1998.av sec dual: average sector dualism, 1980-1985; WDI 1998.pop inc: natural rate of population increase, 1980; WDI 1999.av sec ed: average secondary school enrollments, 1980-1985; WDI 1997.% ch non-ag lf: percent non-agricultural labor force increase, 1980-1985; WDI 1998.

av dem: average institutional democracy index score, 1980-1985;Jagger and Gurr, 1996.% ch urb pop: percent change in urban population, 1980-1985; WDI 1998.av exploit: average exploitation in manufacturing, 1980-1985; World Data 1995.av trade open: average trade openness, 1980-1985; Penn World Tables 5.6.av export/gdp: average exports as a percent of GDP, 1980-1985; WDI 1999.av com con: average commodity concentration index score, 1980-1985; UNCTAD various years.av debt/gnp: average debt service as a percent of GNP, 1980-1985: WDI 1999.av debt/exp: average debt service as a percent of total exports, 1980-1985: WDI 1999.% ch fem lf: percent change in female labor force, 1980-1990; Wistat 1994.% ch svc lf: percent change in labor force in services, 1980-1990; WDI 1998. soc dem: social democratic government, 1980; The Statesmen’s Yearbook, Huber and Stephens 2000.

Table 2 – Correlation Matrixtop 20(1980)

top 20(1995)

lpen(1980)

% ch pen(80-90)

growth(80-90)

lrgdppc(1980)

av disart(80-85)

av sec dual(80-85)

pop inc(1980)

av sec ed(80-85)

% ch non-ag lf(80-85)

av dem(80-85)

% ch u pop(80-85)

av exploit(80-85)

top 20(1980)

1.000(66)

.847**(65)

.091(58)

.148(57)

-.090(65)

-.391**(66)

.162(53)

.568**(55)

.700**(63)

-.571**(61)

-.154(63)

-.178(64)

.334**(64)

.386**(60)

top 20(1995)

1.000(86)

.231*(72)

-.004(71)

-.183(79)

-.329**(81)

.171(66)

.384**(68)

.603**(82)

-.563**(76)

-.220*(82)

-.180(82)

.229*(83)

.355**(69)

lpen(1980)

1.000(74)

-.413**(73)

.085(73)

.282**(73)

-.119(61)

-.250*(62)

-.064(71)

.194(68)

.295**(71)

.301**(72)

-.340**(72)

-.192(65)

% ch pen(80-90)

1.000(73)

-.150(72)

-.162(72)

.078(61)

.216*(62)

.135(70)

-.157(68)

-.175(70)

-.194(71)

.049(71)

.132(65)

growth(80-90)

1.000(81)

.048(81)

-.002(67)

-.052(69)

-.133(78)

.138(74)

.378**(78)

-.066(79)

.063(79)

-.063(71)

lrgdppc(1980)

1.000(83)

-.395**(67)

-.817**(69)

-.750**(80)

.851**(76)

.561**(80)

.644**(81)

-.700**(81)

-.474**(71)

av disart(80-85)

1.000(68)

.380**(67)

.331**(67)

-.394**(65)

-.164(68)

-.474**(66)

.238*(68)

.056(62)

av sec dual(80-85)

1.000(70)

.678**(69)

-.791**(67)

-.527**(70)

-.512**(68)

.614**(70)

.308**(63)

pop inc(1980)

1.000(84)

-.807**(77)

-.411**(83)

-.510**(81)

.481**(84)

.453**(69)

av sec ed(80-85)

1.000(77)

.494**(77)

.493**(77)

-.620**(78)

-.490**(67)

%ch nonag(80-85)

1.000(84)

.302**(81)

-.367**(84)

-.244*(69)

av dem(80-85)

1.000(84)

-.497**(82)

-.450**(69)

% ch u pop(80-85)

1.000(85)

.394**(70)

av exploit(80-85)

1.000(71)

* significant at the .05 level (1-tailed), ** significant at the .01 level (1-tailed)

Table 2 – Correlation Matrix (cont.)

trade op(80-85)

lavexp(80-85)

av com con(80-85)

debt/gnp(80-85)

debt/exp(80-85)

% ch fem lf(80-90)

% ch svc(80-90)

soc dem(1980)

top 20(1980)

.003(66)

-.032(60)

.338**(58)

.427**(56)

.543**(57)

.024(64)

.212(61)

-.320**(66)

top 20(1995)

-.051(83)

-.064(76)

.316**(73)

.386**(69)

.541**(71)

.040(81)

.150(81)

-.370**(86)

lpen(1980)

.374**(73)

.512**(68)

-.109(72)

.266*(64)

.081(65)

.332**(71)

-.112(71)

.041(74)

% ch pen(80-90)

-.007(72)

.057(68)

.457**(71)

.155(64)

.121(65)

-.110(70)

.024(70)

-.066(73)

growth(80-90)

.306**(80)

.148(75)

-.311**(74)

-.162(70)

-.183(71)

-.096(78)

-.048(76)

-.077(81)

lrgdppc(1980)

.236*(82)

.431**(76)

-.612**(74)

-.283**(70)

-.394**(72)

.460**(80)

-.194*(78)

.494**(83)

av disart(80-85)

-.196(67)

-.286**(67)

.358**(62)

.275*(62)

.361**(62)

-.273*(67)

-.120(68)

-.258*(68)

av sec dual(80-85)

-.247*(69)

-.307**(69)

.398**(63)

.261*(64)

.369**(64)

-.444**(69)

.224*(67)

-.386**(70)

pop inc(1980)

-.186*(81)

-.311**(77)

.626**(73)

.487**(70)

.615**(72)

-.201*(81)

.135(81)

-.569**(84)

av sec ed(80-85)

.189(77)

.373**(75)

-.609**(70)

-.300**(69)

-.455**(71)

.338**(76)

-.228*(75)

.486**(78)

%ch nonag(80-85)

.538**(81)

.491**(77)

-.364**(73)

-.069(70)

-.160(72)

.269**(81)

.056(81)

.186*(84)

av dem(80-85)

.040(82)

.237*(76)

-.364(74)

-.305**(71)

-.405**(73)

.465**(79)

-.055(79)

.479**(84)

% ch u pop(80-85)

-.179(82)

-.346**(78)

.407**(74)

.157(71)

.247*(73)

-.400**(82)

.239*(82)

-.313**(85)

av exploit(80-85)

-.224*(71)

-.368**(69)

.419**(65)

.327**(63)

.504**(64)

-.061(69)

-.045(68)

-.431**(71)

trade op(80-85)

1.000(85)

.822**(77)

-.061(74)

.262*(70)

-.182(72)

.094(80)

.001(80)

.103(85)

lavexp(80-85)

1.000(78)

-.060(70)

.324**(69)

-.081(71)

.170(76)

.066(75)

.250*(78)

av comcon(80-85

1.000(75)

.426**(66)

.416**(67)

-.261*(72)

.149(72)

-.353**(75)

debt/gnp(80-85)

1.000(71)

.780**(71)

-.039(69)

.037(68)

-.418**(71)

debt/exp(80-85)

1.000(73)

.025(71)

.060(70)

-.473**(73)

% ch femlf(80-90)

1.000(82)

-.080(79)

.103(82)

% ch svc(80-90)

1.000(83)

-.085(83)

soc dem(1980)

1.000(88)

* significant at the .05 level (1-tailed), ** significant at the .01 level (1-tailed)

Table 3 – Panel Model Predicting Change in Top Quintile Income Share (1980-1995)

(a) (b) (c) (d) (e) (f) (g) (h) (i) (j)constant 3.841***

(0.086)7.414*(4.512)

7.319*(4.423)

18.462***(6.205)

20.558***(6.308)

8.592(10.442)

5.329(10.154)

2.533(10.028)

-4.090(10.581)

-8.484(11.156)

top 20(1980)

0.931***(0.086)

0.913***(0.085)

0.913***(0.083)

0.775***(0.099

0.759***(0.098)

0.748***(0.105)

0.760***(0.098)

0.768***(0.095)

0.743***(0.101)

0.726***(0.101)

lpen(1980)

0.839*(0.520)

1.112**(0.553)

1.455***(0.552)

1.482***(0.547)

1.292**(0.557)

1.245**(0.559)

1.473***(0.558)

1.545***(0.549)

1.539***(0.547)

% ch pen(80-90)

0.753**(0.361)

0.776**(0.351)

0.716**(0.350)

0.668**(0.348)

0.712**(0.348)

0.678**(0.340)

0.587**(0.342)

0.478*(0.348)

av sec ed(80-85)

-0.066***(0.026)

-0.067***(0.026)

-0.117***(0.043)

-0.122***(0.043)

-0.115***(0.042)

-0.113***(0.041)

-0.114***(0.041)

growth(80-90)

-3.084*(2.084)

-2.636(2.139)

lrgdppc(1980)

1.778*(1.244)

2.003*(1.237)

2.581**(1.244)

3.424***(1.274)

4.292***(1.430)

non-ag lfinc (80-85)

-0.173**(0.082)

-0.182**(0.079)

-0.196***(0.080)

av exploit(80-85)

2.012**(1.197)

1.831*(1.226)

av dem(80-85)

-0.295*(0.201)

F 118.630*** 62.343*** 43.959*** 34.850*** 28.909*** 24.955*** 29.325*** 25.853*** 24.268*** 21.690***Adj. R2 .677 .687 .701 .719 .725 .731 .728 .741 .762 .768N 57 57 56 54 54 54 54 53 52 51

* p < .10, ** p < .05, *** p < .01 – significance levels are one-tailed.

Table 4 – Panel Models Predicting Change in Top Quintile Income Share with Table 5.4, Equation J Base Model

Variable Coefficient adjR2 N F(a) lrgdppc2

(1980).267a

(.978).763 51 18.863***

(b) soc dem(1980)

-1.663b

(2.332).765 51 19.111***

(c) ch urb pop(80-85)

-9.976(14.786)

.765 51 19.080***

(d) % ch fem lf(80-90)

-2.186(7.546)

.763 51 18.869***

(e) % ch svc lf(80-90)

-0.021c

(0.059).755 49 17.458***

(f) av disart(80-85)

.132c

(0.242).775 45 17.810***

(g) av sec dual(80-85)

-.042d

(0.102).769 46 17.667***

(h) pop inc(1980)

0.000(0.137)

.766 50 18.846***

a development insignificant; b exploitation insignificant; c democracy insignificant; d change in pen insignificant.* p < .10, ** p < .05, *** p < .01 – significance levels are one-tailed

Table 5 – Panel Models Predicting Change in Top Quintile Income Share with Table 5.4, Equation J Base Model, Globalization Models

Variable Coefficient adjR2 N F(a) trade op

(80-85)-0.009(0.018)

.764 51 18.971***

(b) trade op(80-85)

-0.007a

(0.005).766 51 17.382***

trade op2

(80-85)0.000

(0.000)(c) l av export

(80-85)-1.375b

(1.493).766 50 18.820***

(d) av com con(80-85)

0.024c

(5.022).756 50 17.901***

(e) av debt/gnp(80-85)

-0.016(0.177)

.781 47 19.199***

(f) av debt/exp(80-85)

0.074b

(0.060).772 48 18.634***

a exploitation insignificant; b exploitation and democracy insignificant; c change in pen and democracy insignificant.* p < .10, ** p < .05, *** p < .01 – significance levels are one-tailed.

Appendix A

COUNTRY T2080 year T2095 year TOP20 ch PEN80AAustralia 44.200 1981 46.400 1990 .05 .087Banglades 45.320 1981 37.900 1992 -.16 .004Belgium 36.100 1979 35.030 1992 -.03Brazil 61.600 1980 65.180 1989 .06 .069Bulgaria 32.930 1980 39.247 1992 .19Canada 37.930 1981 34.840 1991 -.08 .204Chile 51.400 1968 63.000 1989 .23 .032China 36.660 1980 41.650 1992 .14Colombia 58.760 1978 54.350 1991 -.08 .032Costa Rica 51.400 1981 50.700 1989 -.01 .139Cote d'Ivoire 47.430 1985 44.080 1988 -.07 .052Czechoslovakia 32.100 1980 35.570 1992 .11Denmark 37.210 1981 37.830 1992 .02 .063Dominican Rep. 47.800 1984 55.700 1989 .17 .036Egypt 43.200 1974 41.090 1991 -.05 .096El Salvador 53.200 1977 54.400 1995 .02 .043Ethiopia 41.300 1982 47.700 1995 .15 .027Finland 40.000 1980 33.800 1991 -.16 .011France 41.820 1979 40.100 1989 -.04 .034Germany 37.420 1981 37.100 1989 -.01 .045Greece 40.170 1981 41.180 1988 .03 .113Guatemala 53.900 1979 63.000 1989 .17 .089Honduras 59.500 1986 56.330 1992 -.05 .036Hong Kong 46.500 1980 49.370 1991 .06 .063Hungary 32.240 1982 38.680 1991 .20India 40.900 1977 41.100 1992 .00 .007Indonesia 42.270 1980 41.950 1990 -.01 .142Israel 39.600 1979 42.500 1992 .07 .033Italy 39.050 1980 37.430 1991 -.04 .020Jamaica 50.300 1975 45.120 1992 -.10 .187Japan 39.570 1980 38.200 1990 -.03 .003Jordan 51.000 1980 47.690 1991 -.06 .040Kenya 60.900 1982 61.857 1992 .02 .048South Korea 45.400 1980 42.240 1988 -.07 .018Malaysia 55.800 1979 53.730 1989 -.04 .248Mexico 54.500 1977 59.300 1989 .09 .042Morocco 46.150 1984 46.300 1991 .00 .010Nepal 59.200 1977 44.817 1995 -.24 .001Netherlands 35.670 1981 36.360 1991 .02 .113New Zealand 40.570 1980 44.730 1990 .10 .105Nigeria 44.200 1986 48.290 1992 .09 .026Norway 41.050 1979 41.550 1991 .01 .116Pakistan 41.270 1979 39.700 1991 -.04 .024Panama 52.420 1980 59.800 1989 .14 .108Peru 58.190 1981 50.397 1994 -.13 .043Philippines 53.300 1975 52.500 1988 -.02 .038Poland 34.564 1980 37.660 1992 .09 .001

COUNTRY T2080 year T2095 year TOP20 ch PEN80APortugal 42.500 1980 40.420 1991 -.05 .044Puerto Rico 52.600 1979 53.200 1989 .01Senegal 60.900 1970 58.637 1991 -.04 .051Sierra Leone 52.500 1968 63.417 1989 .21 .070Singapore 46.590 1980 46.590 1988 .00 .529Soviet Union 34.000 1980 36.600 1989 .08Spain 35.000 1980 35.280 1989 .01 .024Sri Lanka 36.820 1980 39.357 1990 .07 .057Sweden 39.450 1980 38.980 1992 -.01 .029Taiwan 37.040 1980 38.660 1992 .04 .058Tanzania 36.600 1976 45.440 1993 .24 .009Thailand 51.100 1981 58.500 1992 .14 .030Tunisia 50.000 1980 46.330 1990 -.07 .090UK 37.660 1980 40.840 1991 .08 .117USA 41.500 1980 44.100 1991 .06 .031Venezuela 48.200 1981 58.410 1990 .21 .027Yugoslavia 40.830 1978 39.035 1990 -.04 .002Zimbabwe 68.000 1969 62.357 1990 -.08

References

Ahluwalia, Montek S. 1993[1974]. “Income Inequality: Some Dimensions of theProblem.” Pp. 31-39 in Development and Underdevelopment: The Political Economy ofInequality, edited by Mitchell A. Seligson and John T. Passe-Smith. Boulder, CO:Lynne Rienner Publishers.

------- 1976. “Inequality, Poverty and Development.” Journal of DevelopmentEconomics: 307-342.

Alderson, Arthur and Francois Nielsen. 1999. “Income Inequality, Development, andDependence: A Reconsideration.” American Sociological Review 64: 606-631.

Amin, Samir. 1976. Unequal Development: An Essay on the Social Formations ofPeripheral Capitalism. New York, Monthly Review Press.

Baran, P. 1957. The Political Economy of Growth. New York: Monthly Review Press.

Beer, Linda. 1999. “Income Inequality and Transnational Corporate Penetration.”Journal of World-Systems Research. http://csf.colorado.edu/wsystems/jwsr.html 5(1):1-25.

Beer, Linda and Lisa B. Meyer. 2000. “The Global and the Local: A Cross-NationalAnalysis of Women’s Status, Economic Dependency and Child Survival.” unpublishedmanuscript.

Birdsall, Nancy. 1999. “Globalization and the Developing Countries: The InequalityRisk.” Overseas Development Council program on The New Global Economy andDeveloping Countries (http://www/odc.org/programs/birsall.html).

Bluestone, Barry and Bennett Harrison. 1988. The Deindustrialization of America. NewYork: Basic Books.

Bornschier, Volker. 1981. “Dependent Industrialization in the World Economy: SomeComments and Results Concerning a Recent Debate.” Journal of Conflict Resolution.25 (3): 371-400.

Bornschier, Volker and Christopher Chase-Dunn. 1985. Transnational Corporations andUnderdevelopment. New York: Praeger Publishers.

Bornschier, Volker and Ballmer-Cao, Thanh-Huyen. 1979. “Income Inequality: ACross-National Study of the Relationships Between MNC-Penetration, Dimensions ofthe Power Structure and Income Distribution.” American Sociological Review 44: 438-506.

Bollen, Kenneth A. and Robert W. Jackman. 1985. “Political Democracy and the SizeDistribution of Income.” American Sociological Review 46: 651-659.

Boswell, Terry and William J. Dixon. 1993. “Marx's Theory of Rebellion: A Cross-National Analysis of Class Exploitation, Economic Development, and Violent Revolt.”American Sociological Review 58, October: 681-702.

------- 1990. “Dependency and Rebellion: A Cross-National Analysis.” AmericanSociological Review 55:540-549.

Bowles, Samuel, David M. Gordon and Thomas E. Weisskopf. 1984. Beyond the WasteLand: A Democratic Alternative to Economic Decline. Garden City, NY: AnchorPress/Doubleday.

Bowman, Kirk S. 1997. “Should The Kuznets Effect be Relied On to Induce EqualizingGrowth: Evidence from Post-1950 Development?” World Development 25 (1): 127-143.

Brady, David and Michael Wallace. 2000. “Spacialization, Foreign Direct Investment,and Labor Outcomes in the American States, 1978-1996.” Social Forces 79 (1): 67-99.

Braun, Denny. 1991. The Rich Get Richer: The Rise of Income Inequality in the UnitedStates and the World. Chicago: Nelson-Hall Inc.

Breedlove, William L. and J. Michael Armour. 1997. “Dependency, Techno-EconomicHeritage, Disarticulation, and Social Development in Less Developed Nations.”Sociological Perspectives 40 (4)winter: 661-680.

Chan, Steve. 1989. “Income Inequality Among LDCs: A Comparative Analysis ofAlternative Perspectives.” International Studies Quarterly 33: 45-65.

Chase-Dunn, Christopher. 1975. “The Effects of International Economic Dependenceon Development and Inequality: A Cross-National Study.” American SociologicalReview 40 (6): 720-738.

-------. 1989. Global Formation: Structures of the World-Economy. Cambridge:Blackwell Publishers.

Chen Shaohua and Martin Ravaillon. 2000. “How Did the World’s Poorest Fare in the1990s?” World Bank Development Research Group paper.

Chenery, Hollis B. and Syrquin, Moshe. 1975. Patterns of Development 1950-1970.London: Oxford University Press.

Crenshaw, Edward. 1993. “Polity, Economy and Technoecology: AlternativeExplanations for Income Inequality.” Social Forces 71 (3) March: 807-816.

------- 1992. “Cross-National Determinants of Income Inequality: A Replication andExtension Using Ecological-Evolutionary Theory.” Social Forces 71: 339-363.

Crenshaw, Edward and Ansari Ameen. 1994. “The Distribution of Income AcrossNational Populations: Testing Multiple Paradigms.” Social Science Research 23 (1)March: 1-22.

Deininger, Klaus and Lyn Squire. 1997. “Economic Growth and Income Inequality:Reexamining the Links.” Finance and Development March: 38-41.

------- 1996. “A New Data Set Measuring Income Inequality.” The World BankEconomic Review 10 (3):565-591.

DeMartino, George. 1998. “Foreign Direct Investment.” Pp in Foreign Policy in Focusedited by Martha Honey and Tom Barry, May.

Dixon, William J. and Terry Boswell. 1996. “Dependency, Disarticulation andDenominator Effects: Another Look at Foreign Capital Penetration.” American Journalof Sociology 102 (2): 543-562.

Dolan, Michael B. and Brian W. Tomlin. 1980. “First World-Third World Linkages:External Relations and Economic Development.” International Organization 34 (1): 41-63.

Dollar, David and Aart Kray. 2000. “Growth Is Good for the Poor.” The World BankDevelopment Research Group. World Bank, March.

Evans, Peter B. and Michael Timberlake. 1980. “Dependence, Inequality, and theGrowth of the Tertiary: A Comparative Analysis of Less Developed Countries.”American Sociological Review 45, August: 531-552.

Fernandez-Kelly, Patricia. 1984. For We Are Sold, I and My People: Women andIndustry in Mexico’s Frontier. Albany: State University of New York Press.

Ferreira, Francisco H. G. and Julie A. Lichfield. 1998. “Calm After the Storms: IncomeDistribution in Chile, 1987-1994.” World Bank Poverty Reduction Group paper.

Frank, Andre Gunder. 1967. Capitalism and Development in Latin America. New York:Monthly Review Press.

Froot, Kenneth A. ed. 1993. Foreign Direct Investment. Chicago and London: TheUniversity of Chicago Press.

Furtado, Celso. 1970. Economic Development of Latin America: Historical Backgroundand Contemporary Problems. London: Cambridge University Press.

Galtung, Johan. 1971. “A Structural Theory of Imperialism.” Journal of Peace Research8: 81-117.

Glasberg, Davita Silfen and Kathryn B. Ward. 1993. “Foreign Debt and EconomicGrowth in the World System.” Social Science Quarterly 74 (4) December: 703-720.

Hewitt, Christopher. 1977. “The Effect of Political Democracy and Social Democracyon Equality in Industrial Societies: A Cross-National Comparison.” AmericanSociological Review 42: 450-464.

Hicks, Alexander. 1988. “Social Democratic Corporatism and Economic Growth.”Journal of Politics 50 (3) August:677-704.

Hoover, Greg A. 1989. “Intranational Inequality: A Cross-National Data Set.” SocialForces 67 (4) June: 1008-1026.

Huber, Evelyne, Charles Ragin and John Stephens. 1993. “Social Democracy,Christian Democracy, Constitutional Structure and the Welfare State.” AmericanJournal of Sociology 99: 711-749.

Huber, Evelyne and John D. Stephens. 2000. “Partisan Governance, Women’sEmployment, and the Social Democratic Welfare State.” American Sociological Review65: 323-342.

Hughes, Marion R. 1997. “Sample Selection Bias in Analyses of the PoliticalDemocracy and Income Inequality Relationship.” Social Forces 75 (3) March: 1101-1117.

International Labour Office. 1996. Statistics on Poverty and Income Distribution: AnILO Compendium of Data. Geneva: International Labour Office.

Jagger, Keith and Ted Robert Gurr. 1996. Polity III: Regime Type and PoliticalAuthority, 1800-1994. ICPSR data file 9263.

Kenworthy, Lane. 1995. In Search of National Economic Success. Thousand Oaks,CA: Sage.

Kohli, Atul and Michael F. Altfeld, Saideh Lotfian, Russell Mardon. 1984. “Inequalityin the Third World: An Assessment of Competing Explanations.” Comparative PoliticalStudies 17 (3) October: 283-318.

Korzeniewicz, Roberto Patricio and Timothy Patrick Moran. 1997. “World EconomicTrends in the Distribution of Income, 1965-1992”. American Journal of Sociology 102(4) January: 1000-1039.

Kuznets, Simon. 1976. Modern Economic Growth: Rate, Structure and Spread. NewHaven: Yale University Press.

------- 1963. “Quantitative Aspects of the Economic Growth of Nations. VIII: TheDistribution of Income by Size.” Economic Development and Cultural Change 11:1-80.

------- 1955. “Economic Growth and Income Inequality.” American Economic Review45: 1-28.

Lachler, Ulrich. 1998. “Education and Earnings Inequality in Mexico.” World BankMexico Country Department paper.

Lecaillon, J. and F. Paukert, C. Morrisson, D. Germidis. 1994. Income Distribution andEconomic Development: An Analytical Survey. Geneva: International Labor Office.

Lenski, Gerhard and Patrick Nolan. 1985. “Technoeconomic Heritage, Patterns ofDevelopment, and the Advantage of Backwardness.” Social Forces 64 (2) December:341-358.

London, Bruce and Tom Robinson. 1989. “The Effects of International Dependence onIncome Inequality and Political Violence.” American Sociological Review 54, April:305-308.

McMichael, Philip. 1996. Development and Social Change: A Global Perspective.Thousand Oaks, CA: Pine Forge Press.

Milanovic, Branko. 1999. “True World Income Distribution, 1988 and 1993.” PolicyResearch Working Paper 2244. Washington,D.C.: World Bank.

Muller, Edward N. 1995. “Economic Determinants of Democracy.” AmericanSociological Review 54: 966-982.

------- 1993[1984]. “Financial Dependence in the Capitalist World Economy and theDistribution of Income Within States.” Pp 267-293 in Development andUnderdevelopment: The Political Economy of Inequality, edited by Mitchell A. Seligsonand John T. Passe-Smith. Boulder, CO: Lynne Rienner Publishers.

------- 1988. “Democracy, Economic Development, and Income Inequality.” AmericanSociological Review 54: 868-871.

Muller, Edward and Seligson, Mitchell. 1987. “Inequality and Insurgency.” AmericanPolitical Science Review 81: 425-449.

Nielsen, Francois. 1994. “Income Inequality and Industrial Development: DualismRevisited.” American Sociological Review 59, October: 654-677.

Nielsen Francois and Arthur Alderson. 1997. “The Kuznets Curve and the Great U-Turn: Income Inequality in U.S. Counties, 1970 to 1990.” American SociologicalReview 62 (1) February: 12-33.

------- 1995. “Income Inequality, Development and Dualism: Results from anUnbalanced Cross-National Panel.” American Sociological Review 60: 674-701.

Nolan, Patrick D. 1992. “A Standardized Cross-National Comparison of Incomes.”Sociological Quarterly 33 (4) winter: 599-609.

Nolan, Patrick D. 1983. “Status in the World System, Income Inequality and EconomicGrowth.” American Journal of Sociology 89: 410-419.

O’Hearn, Denis. 1989. “The Irish Case of Dependency: An Exception to theExceptions?” American Sociological Review 54 (4): 578-596.

Paukert, Felix. 1973. “Income Distribution at Different Levels of Development: ASurvey of the Evidence.” International Labor Review 108: 97-125.

Paus, Eva D. and Michael D. Robinson. 1997. “The Implications of IncreasingEconomic Openness for Real Wages in Developing Countries, 1973-1990.” WorldDevelopment 25 (4): 537-547.

Prechel, Harland. 1985. “The Effects of Exports, Public Debt and Development onIncome Inequality.” International Labor Review 108: 97-125.

Ragin, Charles C. And York W. Bradshaw. 1992. “International EconomicDependence and Human Misery, 1938-1980: A Global Perspective.” SociologicalPerspectives 35 (2): 217-247.

Ram, Rati. 1992. “Intercountry Inequalities in Income and Basic-Needs Indicators: ARecent Perspective.” World Development 20 (6): 899-905.

Ranney, David C. 1998. “Investment Liberalization Agenda.” Pp in Foreign Policy inFocus edited by Martha Honey and Tom Barry, July.

Rodrick, Dani. 1997. Has Globalization Gone Too Far? Washington, D.C.: Institute forInternational Economics.

Rothgeb, John M. Jr.1996. Foreign Investment and Political Conflict in DevelopingCountries. London: Praeger.

------- 1993. “A Regional Analysis of the Relationship Between Foreign Investment andPolitical Conflict in Developing Countries.” Journal of Political and Military Sociology21: 219-240.

Rubinson, Richard. 1976. “The World Economy and the Distribution of Income WithinStates.” American Sociological Review 41: 638-659.

Simpson, Miles. 1993. “Political Power versus Ecological Evolutionary Forces: WhatAre the Proximal Sources of Income Distribution?” Social Forces 71 (3) March: 797-806.

------- 1990. “Political Rights and Income Inequality: A Cross-National Test.”American Sociological Review 55: 682-693.

Stack, Steven. 1980. “The Political Economy of Income Inequality: A ComparativeAnalysis.” Canadian Journal of Political Science 13: 273-286.

Stallings, Barbara (ed.). 1995. Global Change, Regional Response: The NewInternational Context of Development. Cambridge, MA: Cambridge University Press.

The Statesman's Yearbook. 1980. New York: St. Martins Press.

------- 1975. New York: St. Martins Press.

Sullivan, Gerard. 1983. “Uneven Development and National Income Inequality inThird World Countries: A Cross-National Study of the Effects of External EconomicDependence.” Sociological Perspectives 26 (2): 201-231.

Tsai, Pan-Long. 1995. “Foreign Direct Investment and Income Inequality: FurtherEvidence.” World Development 23 (3): 469-483.

UNCTAD. 2000. United Nations Conference on Trade and Development: Plan ofAction. New York: United Nations.

Weede, Erich. 1993. “The Impact of Democracy or Repressiveness on the Quality ofLife, Income Distribution and Economic Growth Rates.” International Sociology 8:177-195.

------- 1989. “Democracy and Inequality Reconsidered.” American SociologicalReview 54: 865-868.

------- 1980. “Beyond Misspecification in Sociological Analyses of Income Inequality.”American Sociological Review: 32 (1) March:177-178.

Weede Erich and Horst Tiefenbach. 1981. “Some Recent Explanations of IncomeInequality: An Evaluation and Critique.” International Studies Quarterly 25 (2) June:255-282.

Weisbrot, Mark and Dean Baker, Rober Naiman, Gila Neta. 2000. “Growth May BeGood for the Poor-- But are IMF and World Bank Policies Good for Growth?” TheWorld Bank Development Research Group. World Bank, August.

Williamson, Jeffrey G. 1991. Inequality, Poverty, and History. Cambridge: BasilBlackwell.

Wodon, Quentin T. 1999. “Growth, Poverty, and Inequality: A Regional Panel forBangladesh.” World Bank Poverty Reduction and Economic Management Sector Unitpaper.

World Data 1995 (cd rom). 1996. NewYork: World Bank.

The World in Figures: Editorial Information Compiled by The Economist. 1987.London: Hodden and Stoughton.

World Development Report 2000/2001: Attacking Poverty. 2001. New York: WorldBank/Oxford University Press.

World Development Report 1991: The Challenge of Development. 1991. New York:World Bank/Oxford University Press.

World Development Indicators (cd rom). 2000, 1999, 1998, 1997.

World Investment Report 2000: Cross-Border Mergers and Acquisitions andDevelopment. 2000. United Nations, New York and Geneva: United NationsConference on Trade and Development.

World Investment Report 1997: Transnational Corporations, Market Structure, andCompetition Policy. 1997. United Nations, NY and Geneva: United NationsConference on Trade and Development.

World Investment Report 1991: The Triad in Foreign Direct Investment. 1991. UnitedNations, NY and Geneva: United Nations Conference on Trade and Development.