“A Cross-national Study on the Effect of Transnational Corporate Penetration on Income Inequality”

“A Cross-national Study on the Effect

of Transnational Corporate Penetration

on Income Inequality”

ABSTRACT

INDEX

ABSTRACT ……….. 2

INDEX ……….. 3

CHAPTER 1: INTRODUCTION ………. 4

CHAPTER 2: THEORETICAL BACKGROUND ……… 6

CHAPTER 3: METHODOLOGY …..……..………. 9

CHAPTER 4: RESULTS ………. 13

CHAPTER 5: CONCLUSION ….…..……….. 18

BIBLIOGRAPHY ……….. 19

CHAPTER 1: INTRODUCTION 1.1 Introduction to Research

“Nations do best when they are prepared to be open to the world. This means open in their economies, eschewing protectionism, welcoming foreign investment, running flexible labour markets.” Tony Blair (2007: 27). This quote seems to be the current ruling opinion of the world leaders. The United Nations Conference on Trade and Development “assist developing countries in attracting Foreign Direct Investment and in building their productive capacities and international competitiveness” (World Investment Report 2006).Whether in Blair’s piece of advice to all nations slumbers bitter self interest is a point of perception. It seems academic research does not whole heartedly agree with the viewpoint “openness” brings nothing but sweet fruits. Especially when we look at foreign investment, which is largely an instrument of Trans National Corporations, we observe “TNC affiliates do not systematically spur economic progress within their host countries” (Herkenrath and Bornschier 2003). However more and more countries adhere to this call for openness and for at least some of these countries the future seems bright. China, for instance, has opened up its economy and has the biggest growing economy of the world. Whether all countries are able to benefit as China is doing from participating in the global economy of today remains to be seen. Foreign investment played a significant role in the rise of China’s economy. The relationship between foreign investment and economic growth has been the subject of many debates between researchers of different schools. But without interference in this debate I want to look at another trend. Together with the rise of globalization the world has grown more economically unequal “whether measured between individuals, between nations, or within nations” (Beer and Boswell 2002). So our key concern is whether there is a relationship between these two observations, more specifically, between foreign investment and inequality.

Not only on philosophical and ethical grounds is an aversion to inequality justified, but also on pure economic grounds. At a function level, too much inequality can distort economic growth (Ray 1998).

Figure1 1.2 Problem Statement and Research Questions

This research focuses on the relationship between transnational corporate penetration and income inequality. This relationship has been the subject of multiple studies and an update is long overdue, especially since the explosive growth of FDI flows in the 1990’s. The objective of this research is to replicate Beers study (1999) with newer data in order to see whether her models are still relevant and if her conclusions still hold.

Therefore our specific research question is addressed as follows:

Is the positive effect of transnational corporate penetration on income inequality, which was found by Beers research for the mid- 1980s, still supported by empirical evidence if her data is updated?

Due to improvements in data collection, it is possible to expand the dataset used by Beer. Another research question will be:

Is the positive effect of transnational corporate penetration on income inequality, which was found by Beers research for the mid- 1980s, still supported by empirical evidence when Beers dataset is expanded?

1.3 Document Structure

CHAPTER 2: THEORETICAL BACKGROUND

In the early stages of development economics Simon Kuznets noticed the empirical relationship between the income inequality and the level of development of countries. It seemed that developing countries face rising levels of inequality till a certain point, where after the levels of inequality begin to drop. This inverted U-hypothesis leads to the well-known Kuznets-curve. Now whether this curve really exists and/or how it should be explained is a controversial subject. As Ray (1998) explains “The literature on the inverted-U hypothesis resembles, to some extent, the search for the Holy Grail; that is, the search for some evidence that supports an implacable law of development”. But even if there is no evidence for an “implacable law of development”, there is evidence in the literature/studies discussed in this paper that does support an inverted-U relationship in cross-section studies (Ahluwalia 1976; Bornschier & Chase-Dunn 1985; Tsai 1995; Beer 1999; Barro 1999). How this empirical relationship should be explained is still subject of debate.

In the classical modernization and development theories economic development involves a shift of persons and resources from the agricultural/rural sector to the industrial/urban sector. Because of a higher wage in the industrial sector, people who shift sectors experience a rise in income. The initial aggregate effect for a country which develops is a rise in inequality. After this initial effect wages will tend to stabilize in the agricultural sector because of its diminishing size. Also more, relatively poor, people from the agricultural sector are able to work in the higher income industrial sector. So at later stages of development the relationship between the level of per capita product and inequality tends to be negative (Barro 1999). This is, in a nutshell, the “implacable law of development”. That there doesn’t seem to be any evidence supporting this law seems strange considering its popularity. According to Bornschier & Chase-Dunn (1985) it isn’t a law but just a piece of history of the Western core countries. So the inverted U curve is not a story of the evolution of inequality or the economic growth over time within countries, but rather an empirical regularity for a cross section of countries at a point in time (Li, Squire, and Zou 1998).

region a country belongs to. Wallerstein proposes four different categories; core, semi-periphery, periphery and external. The gains that are achieved by the division of labor, low wage unskilled work in the periphery and high wage skilled work in the core, are mostly claimed by the core region. In the figures below the reinterpretation of the inverted U curve by Bornschier & Chase-Dunn together with the preliminary data investigation of this research is being presented.

∆ Core, □ Semi-Periphery, ○ Periphery

Figure 2 Figure 3

This reinterpretation of the inverted U curve by Bornschier & Chase-Dunn (Figure 2) is based on an argument by Lenksi (1966) that inequality increases as we move from hunter/gatherer trough advanced agrarian societies but inequality decreases when industrial society appears due to changes in the power of the producing class. Bornschier & Chase-Dunn combined this argument with the world-system approach in order to give an explanation of the inverted U curve.

In the preliminary data analysis (Figure 3) we can see the data of this study supports the inverted U curve (for details see Appendix). The core countries are where we would expect them to be, considering the theory. The distinction between the semi-periphery and the semi-periphery is less clear. Ethiopia, Tanzania, Zambia, Uganda, Nigeria, Kenya, Ghana, Nepal, Mauritania and Senegal represent the 10 countries with the lowest GDP’s per capita. In these countries 50 to 90, with a shared average of 70, percent of the labor force works in the agricultural sector and thus they can be described as horticultural/ simple and advanced agrarian systems (Food and Agriculture Organization of the United Nations 2006).

Not only a country’s position within the world-system determines the level of inequality but also the amount of foreign direct investments has an effect.

towards foreign investment and predict “that penetration of peripheral national economies by large companies based in the most powerful states will distort economic growth and extract resources which might otherwise be used for national development” (Bornschier & Chase-Dunn 1985: xi). And the penetration of peripheral countries also has effect on income inequality.

“The dependency/world systems perspective posits three specific mechanisms that mediate the influence of foreign investment on social inequality. First, capital-intensive foreign investment causes gross sectoral disparities, the development of labor aristocracies, and the tendency to underabsorb labor. Second, multinationals monopolize local credit sources and at the same time repatriate capital to the home country rather than reinvest locally…. One consequence of such practices may be retarded growth and spread of income. Finally, in order to attract and retain foreign investment, governments tend to exercise stringent controls over labor, which in turn fosters social class rigidity due to poor vertical mobility.”

(Crenshaw & Ameen 1994: 6)

One of the internal manifestations of the modernization process is the sociopolitical process. To capture this process Bornschier & Chase-Dunn included in their model a variable system style. This variable ranged from “centrally planned system” to “private capitalist system”. The hypothesis, which was confirmed, was that centrally planned systems exhibit lower levels of inequality than private capitalist systems. After the collapse of communism, new variables had to be found to measure system style. Beer tried to capture the effect of sociopolitical processes in her models by including two dummy variables, one measuring democracy and the other social democratic representation.

CHAPTER 3: METHODOLOGY 3.1 Framework

To answer our research question we will employ the following framework. This framework builds on the theories described in the previous chapter.

* Only used in the analysis of the expanded sample

Figure 4

Just as in the studies of Bornschier & Chase-Dunn (1985) and Beer (1999) this research used a Ordinary Least Square linear regression model to investigate whether the conclusions of Beer’s research still hold in today’s world. In our analysis we will employ the models used by Beer with the same variables and we also will try to use the same data sources. Due to the improvement in data collection procedures and the availability of these datasets it was possible to expand the sample of countries. In the analysis of the expansion this research will confine itself to the significant variables only and will add a variable to measure the extent in which a country is agrarian.

3.2 Sample

Because this study will replicate Beer’s study the same sample of countries is being used. Strangely, a motivation for the 66 countries selected by Beer is missing in her study. Bornschier & Chase-Dunn (1985) used 72 countries, “which is the largest possible sample of countries with a population above one million”. If we keep this criterion, measured in the year 2000, the largest possible sample of countries for which data on income inequality is available will consist of 126 countries.

Ideally, one would like to have data measurements at one point in time in a cross-national analysis. However for the dependent variable, income inequality, this is not possible since measurements of income inequality data are scattered through time. Most researchers agree that the measurements can be compared to one another as income distribution is a relatively stable characteristic of a country (Chan 1989, Nielsen 1994). Bornschier & Chase-Dunn their data refer to different years between 1956 and 1973 (Ballmer-Cao & Scheidegger 1979). However, the data of most measurements are around 1968. Beer aims at 1985 and limits the measurements of the income inequality variable to a 14 year period, from 1979 till 1993. This research aims at 2000. Without the extreme dates of Rwanda 1983, Sierra Leone 1989 and Trinidad and Tobago 1992, the data covers 10 years, from 1993-2003. Since Beer’s study covered the period 1979-1993 this study neatly continues where Beer’s data left off (see figure 1). Her research adopted 1985 as base year and this research adopts 2000, which is also the median of all inequality measurements, as base year.

The data of all independent variables are measured in the year 2000. So all inequality measurements are being treated as if they were measured in 2000 (most of them also are).

3.3 Data

Income Inequality

All values of income inequality used in this study are taken from the Human Development Report 2006, United Nations Development Programme (original source World Bank, World Development Indicators 2006). Just as in the study of Beer (1999) the top 10% and the top 20% percentage shares of total income are used as a measurement for income inequality. This measurement is an improvement over the use of the Gini coefficient, as it is also an indirect measurement of asset inequality, and according to Beer (1999) it “is a better test of World-system/Dependency arguments”.

Although the dataset of the World Bank is highly esteemed, some comparability issues still remain. The inequality measurement of 78 of the 126 countries has been compiled using data referring to expenditure shares and the rest has been compiled using data referring to income shares. The data from Argentina and Uruguay refer to urban areas only. These issues are not insuperable but in the tradition of transparency have to be mentioned.

Transnational Corporate Penetration

The measurement of transnational corporate penetration is the same measurement used by Beer. This is the ratio between FDI inward stock and GDP. This measurement is more directly interpretable than the original measurement, used by Bornschier & Chase-Dunn: total stock of foreign direct investment/√total energy consumption* population (Dixon & Boswell 1996).

We employ two measurements of transnational corporate penetration that are theoretically the same; the first we calculate ourselves (PENln), the second we take directly from the World Investment Report 2006 (PENWBln). In our first analysis to answer the first research question we will calculate our data because we want to update Beer’s data without changing sources or methods. To answer the second research question we will add the World Investment Report 2006’s measurement, since we do expect this measurement to be more consequent and therefore more suitable for future research. Our calculated penetration measurement for Belgium is also used in the analysis with World Investment Report figures since the World Investment Report does not report on Belgium’s inward FDI stock as percentage of GDP.

The data for country’s FDI inward stock, of course in the base year 2000, are taken from the World Investment Report 2006. Population figures, to convert GDP per capita to GDP, are taken from the same source as GDP per capita, namely the Penn World Table.

Some cases have been removed to adhere to the normality assumption. In the first analysis the countries Hong Kong, Nepal and Burkina Faso have been removed because they have extreme values for the penetration variable. Hong Kong has an extremely high penetration degree because of its small population in relation to its large amount of foreign inward stock. Nepal and Burkina Faso have extreme low penetration degrees for the exact opposite reason. In the second analysis we also removed Niger, Japan, Haiti and Iran. The removal of these cases has little impact on the validity of our results, since they represent different countries from different regions with different positions in the world-system.

GDP

The data used for estimating country’s real Gross Domestic Product per capita are taken from the Penn World Table (Heston, Summers & Aten 2006). In version 6.2 the base year has been moved from 1996 to 2000. The measurement of real Gross Domestic Product per capita is logged. We also add the logged GDP squared to our models to adjust for the curvilinear association between inequality and development levels.

Core/Periphery

System Style

Beer used two measurements, one for democracy and one for social democracy. The variable which measures democracy is taken from the Polity IV data set, originally designed by Ted Rober Gurr (Marshall & Jaggers 2005). We used the same threshold point of an average democracy score of 6.66 as Beer has to devise our dummy variable. In case of missing values for the year 2000, another year was used. Hong Kong was not present in the above mentioned dataset. Although forces are at work to turn Hong Kong into a more democratic region, as till yet Hong Kong remains undemocratic (The Economist, 2007).

The measurement for social democracy is an indicator whether before the base year there was a majority of social democrats in parliament. All information is obtained from The Statesman’s Yearbook 2000. However in the report of the elections in The Statesman’s Yearbook, due to the different political systems among democratic nations, it was not always clear whether a majority in the House of Commons, or equivalents thereof, was reached. The socialist nature of the political parties involved was sometimes also hard to determine. The website of The Socialist International was helpful. Although there is research which found a positive effect of social democratic parties on income distribution (Hewitt 1977; Huber, Ragin & Stephens 1993), I doubt a majority of a social democrats in parliament before the base year is an adequate measurement of social democratic influence. One would rather prefer a measurement which includes a bit of history.

Agrarian system

CHAPTER 4: RESULTS

Table 1 and 3 present the correlation matrices for all the variables in the models (descriptives see Appendix). There is a strong and significant correlation between the independent variables GDP2000logged and LoggedGDP2. This correlation is to be expected since the Logged GDP 2 is added in the model to predict a curvilinear relationship. The correlation between Core and Top10/Top 20 is negatively significant and the correlation between Core and GDP positively significant, which is to be expected (see figure 3). The correlation between PENln and GDP (and therefore also the correlation between PENln and Core) is positively significant. So countries with higher GDP’s also have higher transnational corporate penetration levels. There also seems to be a correlation between DummyDemoc and GDP, DummyDemoc and Core. Although we must admit our data displays more significant correlations than we would hope to find, we are bounded by the model of Beer.

In table 2 we present the results of our linear regression model. The first equation clearly indicates support for the inverted U curve. The income share percentage shows a curvilinear relationship with logged GDP per capita. This is consistent with the notion that the middle income countries have the highest levels of income inequality. The second equation tells us that transnational corporate penetration is a significant variable in predicting income inequality. It has a positive relation with income inequality. This observation is in line with Beers results and “provides strong support for the World-System/Dependency contention that nations whose economies are highly penetrated by foreign corporations exhibit greater degrees of income inequality” (Beer 1999).

Bornschier & Chase-Dunn found that “the main effect of the CC dummy variable is significantly positive, showing that the curvilinear model provides a poor estimate for the core countries themselves” and that “a detailed analysis of the core countries reveals that there exists virtually no relationship at all with the level of GDP per capita”. This observation is not supported by the results. The CC dummy, Core, is negatively related to income inequality, suggesting that the curvilinear model provides a good estimate for core countries. The correlation matrix clearly shows a significant relationship between core countries and income inequality.

Penetration in core countries (CORE*PEN) is negatively related to income inequality, which is the same as Bornschier & Chase-Dunn found. However the interpretation that the net effect for core countries is 3.14-2.75=0.39 and the effect for peripheral countries is 3.14, does not hold because the interaction term is not significant. This outcome is the same as Beer’s research and I will discuss the implications in the conclusion.

Table 1

TOP10

GDPpc 2000

logged GDPpc2Logged PENokln Core Core pen Dummy Democ

TOP10 -GDPpc 2000logged -.36(**) -Logged GDPpc2 -.38(**) 1.00(**) -PENln -.06 .55(**) .56(**) -Core -.55(**) .71(**) .74(**) .41(**) -Corepen .47(**) -.60(**) -.63(**) -.14 -.85(**) -DummyDemoc .06 .57(**) .56(**) .32(*) .40(**) -.34(**) -SOCDEM _{.11 .05 .04 .07 .03 -.02 .16}

** Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed).

The results for the top20%, presented in table 3 and 4 are very similar to the top10% results. Table 3 TOP20 GDPpc 2000

logged GDPpc2Logged PENokln Core Core pen DummyDemoc

TOP20 -GDPpc 2000logged -.32(*) -LoggedGDPpc2 -.35(*) 1.00(**) -PENln -.01 .55(**) .56(**) -Core -.54(**) .71(**) .74(**) .41(**) -Corepen .47(**) -.60(**) -.62(**) -.14 -.85(**) -DummyDemoc .08 .57(**) .56(**) .33(**) .39(**) -.33(**) -SOCDEM .09 .05 .04 .07 .03 -.03 .15

** Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed).

Model number four is being used from the previous analysis as the data set is being expended. Also the percentage of the labor force that works in the agricultural sector is being added. All correlations of this variable with the rest of the variables are significant. So in explanatory power this variable probably does not add anything to the model. The strong correlation with GDP per capita does prove that the reinterpretation displayed in Figure 2 was right in assuming low levels of GDP per capita has a relationship with how many percent of the labor force works in the agricultural sector.

As we can see from the statistics on the expansion of the data set in the appendix most countries that have been added have lower GDPs per capita, lower penetration levels and are less likely to be democratic than the countries in Beers sample. Also only 7 percent of the added countries belong to the core instead of 23 percent for the countries in Beers dataset. The added countries do not however deviate in inequality measurements. Table 5 percentage earnings of TOP10 of total percentage earnings of TOP20 of total natural log GDPpc 2000 Logged GDPpc

squared natural log PEN natural log PENWB democratic yes/no percentage earnings of TOP10 of total -percentage earnings of TOP20 of total .995(**) -natural log GDP2000 -.312(**) -.305(**) -logged GDP squared -.325(**) -.318(**) .998(**) -natural log PEN -.059 -.055 .545(**) .561(**) -natural log PENWB .057 .055 .177 .179 .807(**) -democratic yes/no .016 .016 .580(**) .579(**) .417(**) .106 -Percentage laborforce in Agriculture .357(**) .345(**) -.850(**) -.839(**) -.487(**) -.197(*) -.516(**) ** Correlation is significant at the 0.01 level (2-tailed).

In Table 6 the self calculated measurement of transnational corporate penetration is presented and in Table 7 the measurement from the World Investment Report 2006.

Table 6 Table 7

Predictor Top 10% Top 20%

RGDPpClog 74.53*** 89.52*** (21.59) (23.44) RGDPpClog2 -10.54*** -12.51*** (2.85) (3.10) PENln 1.71** 2.04*** (0.71) (0.77) DEM 4.04*** 4.46*** (1.51) (1.66) Agri 0.14*** 0.17*** (0.05) (0.05) Constant -99.72** -112.48** (40.86) (44.33) F 8.98*** 9.37*** Adj. R2 0.25 0.26 N 118 119 Significance levels: *** - p < .01, ** - p < 0.05, * - p < 0.10

Predictor Top 10% Top 20%

RGDPpClog 59.32*** 72.06*** (20.96) (22.90) RGDPpClog2 -8.44*** -10.07*** (2.75) (3.01) PENWBln 2.26** 2.63*** (0.91) (1.00) DEM 5.20*** 5.82*** (1.51) (1.66) Agri 0.13*** 0.15*** (0.05) (0.05) Constant -74.14* -83.54* (40.34) (44.02) F 8.33*** 8.51*** Adj. R2 0.25 0.25 N 114 115 Significance levels: *** - p < 0.01, ** - p < 0.05, * - p < 0.10

CHAPTER 5: CONCLUSION

The empirical findings of this study still support the positive effect of transnational corporate penetration on income inequality, which was also found by previous studies. Even if the sample is expended from 65/66 countries to 114/115, and the transnational corporate penetration measurement from the World Bank is added, the effect remains. Although this conclusion might suggest a more critical stance towards foreign investment is justified, we have to remind ourselves that linear regression models can not ascertain causal mechanisms. Also we have to be careful about the direction of a causal mechanism. It could be that countries with higher inequality measurements are more prone to be penetrated by foreign investment. This notion was tested in the study of Beer and Boswell (2002) and controlling for reverse causation they still found a positive and significant relation.

The positive effect of transnational corporate penetration on income inequality might became stronger because throughout the core countries during the 80’s and 90’s neo-liberal regimes came to power. With these regime changes, income inequality has been on a rise and is still rising in core countries. With the fall of communism one would expect that in the periphery income inequality also has risen.

For the second time since the research of Chase-Dunn and Bornschier the core/penetration interaction term is insignificant. The world-system theory and the dependency school both argue that development processes are different in the core than in the periphery. Thus one would also expect transnational corporate penetration to have a different effect in the core than in the periphery. This is not the case. In both the core as the periphery higher transnational corporate penetration entails higher levels of income inequality.

Our result that democracy has a significant positive effect on income inequality is not expected. Previous studies have found both a negative as positive effect, so more detailed research is needed before we can conclude that democracy spurs higher levels of income inequality. The notion that democracy is associated with inequality as an institutional background that converts the effects of public sector size on inequality from positive to negative (Lee 2005) might provide interesting leads. Also one must not forget that there is a difference between liberal and illiberal democracies (Zakaria, 2003), formal and institutionalized democracies and these differences can influence the relationship between democracy and inequality.

BIBLIOGRAPHY

Ahluwalia, M.S. 1976. Inequality, Poverty and Development. Journal of Development Economics, 3: 307-342.

Ballmer-Cao, T.H. & Scheidegger, J. 1979. Compendium of data for world-system analysis. Bulletin of the Sociological Institute of the University of Zurich, March. Barro, R.J. 1999. Inequality and Growth in a Panel of Countries. Harvard

University.

Beer, L. 1999. Income Inequality and Transnational Corporate Penetration. Journal of World-Systems Research, VX, 1, spring 1999: 1-25.

Beer, L., & Boswell, T. 2002. The Resilience of Dependency Effects in Explaining Income Inequality in the Global Economy: A Cross-National Analysis, 1975–1995. Journal of World-Systems Research, VIII, 1, winter 2002: 30–59.

Blair, T. 2007. What I’ve learned. The Economist, 383(8531): 25-27.

Bollen, K.A. & Jackman, R.W. 1985. Political Democracy and the Size Distribution of Income. American Sociological Review, 13: 438–57.

Bornschier, V., & Chase-Dunn, C. 1985. Transnational Corporations and Underdevelopment. New York: Praeger Publishers.

Chan, S. 1989. Income Inequality Among LDCs: A Comparative Analysis of Alternative Perspectives. International Studies Quarterly, 33: 45-65.

Chase-Dunn C. & Kawano Y. & Brewer B. 2000. Appendix to Trade Globalization since 1795: waves of integration in the world-system. American Sociological Review, February 2000, Millennial Symposium. Available online at:

http://www.irows.ucr.edu/cd/appendices/asr00/asr00app.htm.

Crenshaw, E. & Ameen, A. 1994. The Distribution of Income across National Populations: Testing Multiple Paradigms. Social Science Research, 23: 1-22. Dixon, J.W. & Boswell, T. 1996. Dependency, Disarticulation, and Denominator Effects: Another Look at Foreign Capital Penetration. The American Journal of Sociology, 102, No. 2: 543-562.

Food and Agriculture Organization of the United Nations. 2006. FAOSTAT Online Statistical Service. Rome: FAO. Available online at: http://faostat.fao.org.

Herkenrath, M., & Bornschier, V. 2003. Transnational Corporations in World Development – Still the Same Harmful Effects in an Increasingly Globalized World Economy? Journal of World-Systems Research, IX, 1, winter 2003: 105–139. Heston, A. & Summers, R. & Aten, B. Penn World Table Version 6.2, Center for International Comparisons of Production, Income and Prices at the University of Pennsylvania, September 2006.

Hewitt, C. 1977. The Effect of Political Democracy and Social Democracy on Equality in Industrial Societies: A Cross-National Comparison. American Sociological Review, 42: 450-464.

Huber, E. & Ragin C. & Stevens J. 1993. Social Democracy, Christian Democracy, Constitutional Structure and the Welfare State. American Journal of Sociology, 99: 711-749.

Human Development Report, 2006. Beyond scarcity: Power, poverty and the global water crisis. United Nations Development Programme. New York: Palgrave

Macmillan.

Lee, C.S. 2005. Income Inequality, Democracy, and Public Sector Size. American Sociological Review, 70: 158–181.

Lenski, G.E. 1966. Power and Privilege: A Theory of Social Stratification. New York: McGraw-Hill.

Li, H. & Squire, L. & Zou, H.F. 1998. Explaining International and Intertemporal Variations in Income Inequality. The Economic Journal, January, 108: 26-43. Marshall, M.G. & Jaggers, K. 2005. Polity IV Project. Center for Global Policy. www.cidcm.umd.edu//polity.

Nielsen, F. 1994. Income Inequality and Industrial Development: Dualism Revisited. American Sociological Review, 59: 654-677.

Ray, D. 1998. Inequality and Development: Interconnections. In, Development Economics: 197-247. Princeton: Princeton University Press.

The Economist. 2007. Hong Kong: One Country, no democracy. 383(8535): 11. The Statesman's Yearbook. 2000. New York: St. Martins Press.

Tsai, P.L. 1995. Foreign Direct Investment and Income Inequality: Further Evidence. World Development, 23, 3: 469-483.

Wallerstein, I. 1974. The Modern World-System I: Capitalist Agriculture and the Origins of the European World-Economy in the Sixteenth Century. San Diego, CA: Academic Press Inc.

Weede, E. 1989. Democracy and Inequality Reconsidered. American Sociological Review, 54: 865–68.

World Investment Report, 2006. FDI from Developing and Transition Economies: Implications for Development. U.N. Conference on Trade and Development, NY and Geneva: United Nations Publication.

APPENDIX

Figure 3

Case Processing Summary

N

Total Cases 64

Excluded Cases(a) 0

Forecasted Cases ₀

Newly Created Cases 0

a Cases with a missing value in any variable are excluded from the analysis.

Model Summary and Parameter Estimates

Dependent Variable: percentage earnings of TOP20 of total

Equation Model Summary Parameter Estimates

R Square F df1 df2 Sig. Constant b1 b2

Quadratic .313 13.924 2 61 .000 -180.092 128.859 -17.815

The independent variable is natural log GDP2000.

Categorization of Core, Peripheral and Semi-peripheral countries from Chase-Dunn, Kawano & Brewer (2000).

Descriptives Table 1

Descriptive Statistics

N Minimum Maximum Mean Std. Deviation

percentage earnings of

TOP10 of total 61 21.3 48.3 32.013 7.6342

natural log GDP2000 _{61 2.860560 4.536110} 3.8087513

8 .488539338

logged GDP squared _{61 8.18 20.58 14.7413 3.69230}

natural log PEN 61 -5.02 -.08 -2.3323 1.05659

Descriptives Table 3

Descriptive Statistics

N Minimum Maximum Mean Std. Deviation

percentage earnings of

TOP20 of total 62 35.7 70.3 47.894 8.6178

natural log GDP2000 _{62 2.860560 4.536110} 3.8095896

5 .484563327

logged GDP squared _{62 8.18 20.58 14.7440 3.66197}

natural log PEN 62 -5.02 -.08 -2.3243 1.04979

Core vs. periphery _{62 0 1 .24 .432} core*penetration level interaction 62 -4.10 .00 -.3811 .79872 democratic yes/no 62 0 1 .68 .471 socialdemocratic government yes/no 62 0 1 .24 .432 Valid N (listwise) 62

Statistics on expansion data set Group Statistics

in analyse Beer or

not N Mean Std. Deviation Std. Error Mean

percentage earnings of TOP20 of total wel in LBeer 64 48.042 8.5279 1.0660 niet in LBeer 59 46.842 8.9845 1.1697 percentage earnings of TOP10 of total wel in LBeer 64 32.577 8.1291 1.0161 niet in LBeer 59 31.629 8.5932 1.1187 natural log GDP2000 wel in LBeer _{64 3.80909846 .490316431 .061289554} niet in LBeer 59 3.58578168 .450898400 .058701972

natural log PEN wel in LBeer 64 -2.3350 1.21016 .15127

niet in LBeer 59 -3.0282 1.12028 .14585

democratic yes/no wel in LBeer 64 .67 .473 .059

niet in LBeer 59 .44 .501 .065

Core vs. periphery wel in LBeer 64 .23 .427 .053

t-test for Equality of Means t df

Sig.

(2-tailed) DifferenceMean DifferenceStd. Error