GLOBALIZATION AND CONFLICT: WELFARE, DISTRIBUTION, AND POLITICAL UNREST

The consequences of globalization for the development of a more peaceful world remain highly controversial. This article seeks to clarify the impact that the globalization of the economy may have on civil war and political instabil ity. Liberals argue that countries heavily dependent on the global economy (whether measured by trade or investment) are likely to experience higher economic growth, greater affluence, more democracy, and increasingly peace ful conditions at home and abroad. In stark contrast, most dependency theo rists argue that high levels of trade and investment tend to generate greater economic inequality. Relative deprivation theory suggests that such inequal ity will increase the risk of political instability. From these two broad per spectives, a set of hypotheses is developed and tested on a global dataset for the period 1965-93. The consequences of an open economy prove to be quite complex. A high level of trade does generate more domestic peace; at the same time, direct foreign investment also creates conditions conducive to political instability. However, the consequences of trade are dependent on what is being exported. Exports of manufactured goods create high levels of welfare and equality, while exports of agricultural products promote poverty and inequality. Inequality emerges as but one of many factors which lead to political instability.


INTRODUCTION
'The balance between international openness and domestic stability is a pivotal issue if globalization is to remain a constructive force." (Va yrynen, 1997:78).
E ver since the 1700s, liberals have held that mutual economic dependence between countries will promote cooperation rather than conflict. Representatives of the Manchester School explained the peace that followed in the wake of the Napoleonic Wars as resulting from increased interstate trade. A similar liberal line of reasoning was heard again during and after World War IL The USA promoted the establishment of financial institutions like the World Bank and the International Monetary Fund, arguing that strong trade, low tariff barriers, and competition on equal terms would prevent new crises (Knutsen, 1997). Today both trade and foreign investments are growing rapidly, and countries from the former Communist bloc are joining free-trade organizations. There seems to have been a breakthrough for the ideas of the Manchester School, in terms of using economics as a means toward peace. In recent years several scholars have found empirical support for the thesis that pairs of nations with a high share of mutual trade in GDP are also less war-prone.3 The number of civil wars has risen sharply since World War II (Gleditsch, 1996: 293). Trade and foreign direct investment (FDI) have been increasing-as has income inequality. 4 Is this merely a coincidence? Is globalization a new kind of Social Darwinism-a prescription for the survival of the fittest? Those inclined to a radical theory-dependency theorists or globalists-agree that high economic dependence on external markets increases income inequality at home, and that such inequality in turn may lead to domestic conflict. By contrast, liberals argue that increased trade acts to boost economic development, thereby reducing the likelihood of domestic conflict.
J . See e.g. Oneal, Oneal, Maoz & Russett (1996), and Polachek (1994).For a more skeptical view, see Barbieri (1996). 4 · From 1960 to 1990 the incomes of the richest one-fifth in the whole world grew three times faster than those of the poorest one-fifth. As a result , their share of the poorest one-fifth of the world economy declined from 2.3% to 1. 4% in 19604% in (Mufson, 1997, as cited in Vayrynen, 1997: 79). GLOBALIZATION AND CITIZENSHIP 329 In this article we test hypotheses derived from the liberal and the dependency-theoretical perspectives, using data from 96 countries for the years 1965-93. Lack of data makes it difficult to go beyond that time-frame.
Both the liberal and the globalist view receive some support from our empirical analysis. Liberals are correct in focusing on the relationship between openness and welfare, while dependency theorists are correct in arguing that openness leads to income inequality. However, our analysis indicates that trade does not have precisely the same effect as FOL The effects of globalization may be more complex than often assumed. FDI boosts inequality and political instability, while trade creates favorable conditions for peace. However, our analysis shows that the type of trade also matters. Exports of agricultural products are associated with poverty and inequality, while exports of manufactured goods go together with welfare and equality. However, in terms of triggering internal conflict, inequality is also found to be dependent on several other factors.
We define globalization in terms of high foreign trade and foreign direct investment in relation to GDP. We do not include other typical features of globalism such as transnational financial relations, currency trade, movements of people, or transfer of technology. For such variables it is hard to find data for a long time period, and for all countries. FDI and trade are common indicators of the globalization process. We also exclude such other possible explanations for high income inequality such as culture, attitudes, religion, and history, and we ignore ethnicity as an explanation for civil war. It would have been interesting to see how income inequality is associated ethnic affiliation, as the violent conflicts in South Africa and in Rwanda might be explained in such terms. However, in order to pursue this line of reasoning, we would have needed much more disaggregated economic data.

THE LIBERAL SCHOOL
'Peace is the natural effect of trade'- Montesquieu (1759) Writing at the end of the 18th century, Immanuel Kant (1795Kant ( /1992) explained how mutual economic dependence promotes cooperation instead of conflict. In the 1800s, the Manchester School, with David Ricardo,Jeremy Bentham, and Richard Cobden among its more prominent names, also argued that mutual trade and contact provide a stimulus for cooperative solutions (Blainey, 1973). Polanyi (1944Polanyi ( /1975 revived the liberal argument, stating that 'Trade has become linked with peace'. However, he relied on realist views, bringing in the balance of power as a means to peace. Polanyi explained the economic crises between 1914 and 1945 by an exaggerated belief in the self-regulating market. Politics and economics are mutually embedded, Polanyi argued. In his view, separating economic life from society gives rise to national and protectionist counter-reactions-as was the case especially after the dissolution of the gold standard stopped the world economy from functioning (Polanyi, 1944(Polanyi, /1975. The victory of the politics of economic nationalism over internationalism made possible the Great Depression and the rise of extreme ideologies. It also contributed to the military expansionism which ultimately led to World War II-the first genuinely global war (Vayrynen, 1997: 9).
Today the concept of 'less state, more market ' is on the offensive once again. Fukuyama (1989) has proclaimed the final victory of the marketbased economy over the centrally-planned economy. Weede (1995), in an analysis of the relationship between free trade and interstate war, has argued that strong trade links between nations raise the level of economic development in rich and poor countries alike. A high level of economic development increases the chances for a democratic system of government, and the existence of a democracy in turn reduces the likelihood of war against other countries with democratic governments. Thus, trade between nations will reduce the likelihood of interstate war. A similar line of argument will be presented in the following to show that economic openness can reduce the likelihood of civil war as well. Dollar (1992) and Eusufzai (1996) are among the many economists who have found countries with open economies to have a higher level of welfare. Dollar examined 95 less-developed countries for the period 1976-85 and found that outward-oriented countries grow faster than more inwardoriented countries. Eusufzai found that open countries have a higher level of human development as measured by the Human Development Index GLOBALIZATION AND CITIZENSHIP 331 (HDI), a lower under-five mortality rate, and a higher proportion of population with access to safe water. Firebaugh & Beck ( 1994: 649) draw a similar conclusion: in poor countries, dependency and economic growth have a significant, positive effect on expected lifespan, child mortality, and calorie consumption per capita. They conclude their study by saying that:'Even in the most dependent LDCs, the masses tend to benefit from economic growth.' The World Bank (1996, ch. 2) argues along the same lines, pointing out that China and Vietnam have experienced considerable economic growth after replacing parts of their centrally planned economies with free trade and market liberalism. A similar development can be observed in the rich countries, where an open economy gives higher level of welfare (Rodrik, 1997). However, Rodrik also notes the dilemma created by the lower ability of developed states to tax capital and highly-educated people, and the increased need for social insurance in the globalized economy. Both capital and highly-educated people are attracted by countries with lower levels of taxation, and they easily move to such places-making it important for countries to reduce their taxes to become more attractive (Rodrik, 1997: 55-64;Martin & Schuman, 1998). 5 In the long term, this development may reduce the level of welfare because the public sector will lack money for redistribution. However, Rodrik finds that openness in terms of international trade and FDI is positively associated with all aspects of welfare budgets. The rapid growth of globalization from 1966 through 1990 did not result in any cutbacks in public social spending among industrialized countries (Rodrik, 1997: 51-53). Rodrik's findings indicate that there may be institutional barriers that protect people against the unsettling forces released by the transition to free trade. States redistribute the assets from trade to income transfers, protecting workers against risks (especially illness and unemployment) or supporting them in old age (Vayrynen, 1997: 76).

From Development to Democracy
Some four decades ago, Lipset (1959) proclaimed that a high level of prosperity increased the chances for a democratic system of government. Higher income and better education for the lower strata would, he held, lead to a more compromise-oriented view of politics. Rich countries also have more surpluses to distribute; this permits modernization through education, occupational mobility, free flow of information, and organizational experience. Taken together, these factors encourage adaptability and compromise, tolerance, and moderation. Increased access to material and thus political resources, together with greater institutional diversity, were seen to act as preconditions for stable democracy. These views have found support in several empirical studies (see Burkhart & Lewis-Beck, 1994;Huntington, 1991;Muller 1985Muller , 1995a. One interpretation of this regularity is suggested by Einstein's remark that 'an empty stomach makes a poor political advisor' (quoted in Londregan & Poole, 1996: 1). From this point of view, democracy would be simply a beneficial side-effect of a high level of social and economic development. However, Londregan & Poole (1996: 20) find, after correcting for many features of the political and historical context, that the democratizing effect of income remains a significant factor promoting the emergence of democratic political institutions.
From Democracy to Domestic Peace A democratic system of government is frequently associated with lower likelihood of civil war. Democracies tend to enjoy greater acceptance among the general population, so dissatisfaction is not frequently expressed in the form of serious challenges to the regime. Actors can channel their expressions within the democratic system, thereby reducing the likelihood of outright rebellion (Eckstein & Gurr, 1975;Flanagan & Fogelman, 1971;Rummel, 1995). Rupesinghe (1992) and others, noting the re-kindling of conflict in Eastern Europe after the Cold War, have suggested that a democratic regime may permit suppressed conflicts to break into the open, in contrast to the situation under an authoritarian regime. East European countries found themselves with a transitional regime and a fragile democracy and this-for a while anyway-led to higher levels of conflict than under the previous totalitarian regimes. Combining these two tendencies, we would expect to find the least amount of domestic armed conflict in GLOBALIZATION AND CITIZENSHIP

Figure 1: A Liberal Model
Open Economy High level of eco nomic development 333 Peace Democracy established democracies, but less armed conflict in strict autocracies than in emerging democracies. Several empirical studies have shown that semidemocratic countries run the greatest risk of civil violence (Ellingsen, 1996;Ellingsen & Gleditsch, 1997;Muller & Weede, 1990). Does that mean we can equate transitional regimes with semi-democracies? Hegre, Ellingsen, Gleditsch & Gates (1999) find that regime change can explain some of the civil wars in semi-democracies-but not all of them.

From Development to Domestic Peace
There seems to be broad agreement that a high level of economic development increases the likelihood of domestic peace. Flanagan & Fogelman (1971:14) studied 65 nations from 1800 to 1960, and concluded that there is less likelihood of civil war where the level of economic development is high. Jacobsen (1996) found no civil wars at all in the period 1945-85 in countries with a high level of economic development. The best explanation for this relationship is probably that rich countries have a higher overall level of welfare, and have a more highly educated population.

A Liberal Model
This discussion can be summed up in a simple model (Figure 1). The liberal school of thought holds that a high degree of openness in the economy will strengthen the level of economic development. A prosperous country has a greater likelihood of having a democratic form of government. The final link in the chain is that both democracy and a high level of economic development have a positive effect on domestic peace. On this basis, then, we would expect globalization to have a generally peacebuilding effect.

DEPENDENCY THEORY
The gap between a small, wealthy elite and the impoverished masses has grown to such astronomic proportions due to so-called development that many former "Third World" countries are in a state of endemic civil war. ' Schrijvers (1993: 23)

From Globalization to Inequality in the South
According to dependency theory, the penetration of foreign capital into peripheral economies leads to the exploitation of local human and natural resources, and to a transfer of profit back to the imperial centers. This process results in impoverishment, inequality, and injustice (Galtung, 1971). The production of raw materials in poor countries serves to prevent competence-building, and the economy remains export-oriented (Hveem, 1996: 240). Ties are created between the local power elite and foreign interests, in turn increasing income inequality in the poor countries (Boswell & Dixon, 1990;Muller & Seligson, 1989;Rubinson, 1976). The production of raw materials will keep inequality high and the level of welfare low (Bourgignon & Morrison, 1989;Wood, 1994).
In the 1950s and in the 1960s most dependency theory focused on trade that exploited the poor countries. In the 1970s and 1980s dependency theory seems to have changed focus, to foreign direct investment (FDI). In this study we want to include both trade and investment to see whether they have similar effects on conflict.
In a classic study , Bornschier & Chase-Dunn (1985) looked at the consequences of the policies of multinational corporations in the periphery. Studying 72 countries for the period 1950~ 77, they concluded that foreign capital increased inequality in poor countries. Salaries tend to be higher in multinational companies than in the country's own companies (Bornschier & Chase-Dunn, 1985: 120). In a sense, multinationals presuppose the existence of a high level of inequality in the host country. The price of a globallymarketed product is approximately the same everywhere in the world. In the North, most people may be able to afford a given product, whereas in the periphery only the elite can afford it. According to Bornschier & Chase-Dunn, had there been greater equality in the poor country, no one would have been able to afford the product in question. Thus, as far as the multinational firm is concerned, a certain degree of inequality is desirable, since the elite is both employer and consumer.
Dependency theory claims that FD I in less-developed countries reduces economic growth while inequality increases. Foreign investment is thought less likely to contribute to public revenue, less likely to encourage indigenous entrepreneurship, less likely to promote links to other industries in the domestic economy, and more likely to use inappropriate capital-intensive technology (Firebaugh, 1992: 106). Earlier work by dependency theorists went further than Firebaugh, arguing that foreign investment was likely to decrease growth (Bornschier & Chase-Dunn, 1985). This argument has not been laid entirely to rest after the recent debate with Firebaugh ( cf. Dixon & Boswell, 1996a,b ). Similar arguments are frequently repeated in the political debate about globalization (Martin & Schumann, 1998). 6 6 · These points were widely accepted among sociologists and political scientists until Firebaugh (1992) launched a spirited attack on the theory as well as the empirical analysis. Firebaugh set out to show that the dependency school had systematically misinterpreted the data. He demonstrated that dependency researchers use flow (new investment) and stock ( accumulated flows) of foreign investment in the same regression equation , finding the effects of flow to be positive but the effects of stock negative. Therefore, the negative effect of stock was merely a denominator effect of flow/ stock. Since the investment rate is measured as flow/stock, the greater the initial level of stock, the slower the investment rate. This denominator effect was routinely misinterpreted by dependency researchers as a negative effect of the penetration of foreign capital on economic growth. Firebaugh concludes that foreign capital has a somewhat smaller positive effect for a poor country than does domestic capital, but that this does not mean that foreign capital represents something negative. In a recent paper, de Soysa & Oneal (1998) argue that even this is a misinterpretation: if you compare the effect of the two kinds of investment dollar for dollar, foreign investment emerges as about three times as productive as domestic in terms of economic growth. Moreover, in a Grangercausality analysis of foreign and domestic investment, foreign investment is found to cause domestic investment, while the reverse relationship is not significant. In other words, when multinational corporations express confidence in a third-world economy, this encourages domestic investment as well. But increased domestic investment is not by itself enough to attract foreign investment.

Globalization and Inequality in the North
According to Kuznets (1955), inequality is relatively low in agricultural societies because most people are engaged in small farming and have fairly similar profits. Inequality will increase as a country industrializes and people leave the countryside for the towns and cities, looking for paid work. The result is often cramped living conditions, as well as poor access to food. In factories and firms a few top leaders will draw sky-high salaries, whereas most laborers will be working for low wages. After a while, however, as the workers become integrated into the new industrial culture, they will seek to achieve better conditions. The struggle for higher wages and better legal protection will increase, and gradually the inequality will be reduced. This development is referred to as the Kuznets U-curve. Dependency writers are critical of this line of reasoning, however. In their view, worldwide capitalism is premised on there being some who have more control and power over production than others-which necessarily implies inequality (Rubinson, 1977: 656).
Today reference is frequently made to the increased income inequality in the rich countries. Some scholars have sought an explanation in terms of technological advance, whereas others have cited high immigration figures. But a strong opinion is emerging that globalization is the best explanation for the increased inequalities, especially in North America and in Great Britain ( Frank & Cook, 1995;Wood, 1994). Imports of cheap textiles and electronic goods often out-compete Western products (Bhagwati & Kosters, 1994;Borjas & Ramey, 1994), producing an increasing number of unemployed. A new group of working poor' is emerging because of weakening of the labor movement and because of companies' efforts to compete with low-cost countries. Multinational companies threaten their Western workers that if costs are not kept down, factories may be moved to countries with lower wages. To take one example, between 1990 and 1994 the Swiss-Swedish firm Asea Brown Boveri (ABB) closed down 40,000 jobs in North America and in Europe and created 21,150 jobs in Eastern Europe-mostly in Poland. Average hourly wages in a Western country were almost 12 times higher than in Poland. In addition a Polish worker would out in 400 more hours per year than a German worker (Thurow, 1996: 168). To avert such massive job losses, workers in Western countries have had to moderate their demands for higher wages. The real median income of families has barely GLOBALIZATION AND CITIZENSHIP 337 increased since the early 1970s, and although the poverty rate has declined slightly, numerous children are still growing up in poverty. Employment has been growing strongly in the USA since the early 1980s, but it is less impressive if we take into account the population growth and the low salaries in many jobs (Mishel, Bernstein & Schmitt, 1997: 381-416). This has led Luttwak (1994)-certainly no leftwing radical-to ask if the USA is becoming a third world country, since 15 million people-6% of the population-live under conditions similar to those found in poor countries (1994: 118, 125). Skarstein (1998: 52) asserts that globalization leads to the emergence of working poor in the USA and unemployed poor in Western Europe. Several years ago Harrison & Bluestone (1988) suggested that the Kuznets curve was turning downward again-a point of view that is still prevalent.

From Inequality to Conflict
Will inequality give rise to conflict? As early as in 1835, de Tocqueville (1835de Tocqueville ( /1961 expressed such a view: "Almost all of the revolutions which have changed the aspect of nations have been made to consolidate or to destroy social inequality. Remove the secondary causes which have produced the great convulsions of the world, and you will almost always find the principle of inequality at the bottom:' As Lichbach (1989: 433) points out, many revolutions have been based on egalitarian ideas. The rhetoric in the American Revolution was that 'all men are created equal'; in the French Revolution, the partisans shouted 'liberty, equality, fraternity'; the motto of the Russian Revolution was 'peace, land, bread'; and a wartime slogan of the Chinese Communist Revolution was 'those who have much give much, those who have little give little' (ibid.).
Two explanatory models link income inequality and political violence: economic discontent theory (Gurr, 1970) and political opportunity theory (Tilly, 1978). In the strict version, discontent-oriented theories maintain that inequality is the basis of all rebellion, and that if economic inequality is high, violent political conflict will occur. Conversely, politically-oriented theories maintain that economic discontent is not central, and that political resources and opportunities determine the extent of violent political conflict within nations.

Figure 2: A Dependency Model
Dependen ce Income P olitical Inequali ty In sta bility, Armed Conflic t Most studies of the connection between inequality and armed conflict have found a positive relationship (Boswell & Dixon, 1990;Muller & Seligson, 1989;Timberlake & Williams, 1989). In a survey article Lichbach ( 1989) critically reviewed this linkage ands uggested that these studies might be spurious because they failed to include control variables like the level of economic development and political regime type. Scholars had focused on relative deprivation, at the expense of more important explanatory factors. In addition, there are many different ways in which to operationalize income inequality. Shock ( 1996) writes that violent political conflict has typically been studied within either an economic-discontent or a political-opportunity framework. His empirical study shows that political opportunity structures moderate the relationship between economic inequality and violent political conflict. A further problem is that inequality changes only gradually over time, whereas armed conflict may erupt suddenly at any point (Muller, 1988), making it almost impossible to say when inequality results in armed conflict and when it does not.

A Dependency Model
As with the liberal model, this discussion can be summed up in a simple model ( Figure 2). The main hypothesis of the dependency school is that a high degree of openness in the economy leads to a high degree of income inequality, which in turn increases the likelihood of armed conflict. GLOBALIZATION AND CITIZENSHIP

Data
Our indicators of globalization are foreign trade and investment. Trade is the total sum of exports and imports divided by the Gross Domestic Product (GDP), and this variable is taken from Penn World Tables (Summers & Heston, 1991). We have two indicators of investment: One covers the whole period from 1965-93, while the other covers a shorter period . The long-term variable comes from Penn World Tables; it measures investment as a share of GDP. The weakness of this variable is the inclusion of both foreign and domestic investment. On the assumption that most investment in poor countries is from foreign countries or firms, we constructed a new variable for investment multiplied by a dummy variable for poor countries.7 However, this variable remains quite weak because we cannot distinguish between foreign and domestic capital in poor countries. The other investment variable measures FDI from 1980-93; it comes from collected by the World Investment Report (1997)issued by United Nations. Since the two investment variables have different weaknesses, we will include both in our analysis to see whether they produce the same results.
We measure economic development by energy consumption per capita, using data from Small & Singer (1993) for the period 1945-85. We have updated the dataset to 1994 by using the UN Statistical Yearbook. The correlation between Small & Singer's data for energy consumption per capita and the UN data for 1980 was 0.985. To reduce the skewness of this indicator we have taken the natural logarithm. To see whether the Kuznets CT-Curve still holds, we have squared the economic development variable.
For regime type, we use the index for Institutionalized Democracy in Polity III dataset, generated by Jaggers & Gurr (1995). The Democracy Index ranges from 10 (most democratic) to O (most autocratic). To test 7 · The classification of high and low development follows that of Human Development Report (1993). A highly developed society has an annual energy consumption of 1,500 or more coal equivalents, while a less-developed society is one that consumes less than 1,500 coal equivalents.
whether semi-democracies are more war-prone than any other regimes, we have also included the square of the democracy variable.
Data on inequality were drawn mainly from Social Indicators of Development (World Bank, annual). Income inequality is measured by the concentration of income in the top 20% of the population. We assume that inequality is measured before taxes-a weakness, because the redistributive effect of taxes is not included. 8 Data on income inequality are often not comparable, because some countries measure the inequality per household and others per person ; some before taxes and others after. Some countries use the Gini index of inequality and others the share of the income earned by the richest 20%. Our main objective here was to achieve maximum comparability, so we generally used Social Indicators of Development, which had inequality data for more countries than other sources.9 Recently, Deininger & Squire (1996) have published another set of inequality data. The correlation between our dataset and that of Deininger & Squire was 0.849. 10 We use three variables which indicate domestic unrest. One iscivil war as recorded by the Correlates of War project (COW) (Small & Singer, 1994). Civil war is defined as an internal war which involves: (a) military action, (b) having the national government at the time as one of the parties to the conflict, (c) effective resistance (as measured by the ratio of fatalities of the weaker to the stronger forces) on both sides, and (d) at least 1 1 000 battle-deaths in a single year (Singer & Small, 1994 ). The second variable is political instability . This index, which is taken from Alesina & Perotti (1996), is constructed on the basis of a principal component analysis on the following variables: number of politically motivated assassinations, number of people killed in conjunction with phenomena of domestic mass violence (as a fraction of the total population), number of successful coups, 8 · Social Indicators of Development does not state this explicitly, but most research in this field uses income inequality before tax. 9 · We also used some data from Atkinson et al. (1995), Muller & Seligson (1987), Sawyer (1976), and World Development Report (1996).
10 · Deininger & Squire (1996) have divided their inequality data into acceptable cases and unacceptable cases. When computing the correlation we included only the acceptable cases. GLOBALIZATION AND CITIZENSHIP 341 number of attemptedbut unsuccessful coups, a dummy variable that takes the value of 1 in a democracy, 0.5 in a semi-democracy and O in a dictatorship. Our third conflict variable comes from the Conflict Data Project at Uppsala University (Wallensteen & Sollenberg 1 1998). Their threshold for armed conflict is 25 casualties in a single year. However, this dataset is available only for the period after 1989. Since the analyses from the other two conflict variables yielded very similar findings, the results from the Uppsala dataset will not be reported here.
To avoid autocorrelation ( dependency between the units) in the time series for the COW data we use the outbreak of civil war, rather than the incidence, as our dependent variable. To avoid time dependency for the political-instability variable we used the average number for all the relevant data for the period 1965-85 in one of the tables. The weakness of this variable is therefore the small number of cases ( only 96). We have excluded Singapore because of extreme outliers on the trade variable.11

ANALYSIS
We use Ordinary Least Squares Regression when the dependent variable is inequality or political instability because these two variables are continuous. Theory suggests that we should study globalization in the South independently of globalization process in the North, because of the different mechanisms. Therefore we have two samples: one with all the countries in the dataset (N =96) and one for poor countries only (N =75) .12 We used the same procedure in the logistic regression with a dichotomous dependent variable: outbreak of civil war. We do not present separate analyses for poor countries only, because the results are quite similar to those shown in the Tables below. 1 1. The results changed a great deal when we excluded this case, especially for the inequality regression. Trade was found to lead to inequality when Singapore was included. It became insignificant, and the coefficient for inequality became negative, when we excluded Singapore.
12 · Because of missing data, the analysis covers the more limited time-span 1980-93 for Colombia, Venezuela, Guyana , Equador, and Peru. Tables 1 and 2 include the whole period (1965-93) because we use the variable Investment in poor countries. In Table 3 we use the FDI variable, which covers the period 1980-93. All the independent variables in this analysis are continuous. The unit of analysis in this, and in most of the subsequent Tables, is the country-year.
We start with the social consequences of globalization. Table 1 shows the relationship between economic openness on the one hand and economic development and inequality on the other, with several control variables. The results confirm Vayrynen's view that globalization leads to higher level of economic development as well as inequality. We have one exception, however: investment in poor countries leads to lower economic welfare. In contrast to this finding, we can note that FDI leads to welfare (Tables 6 and A-2). This is the same result that Dollar (1992) obtained for poor countries and Rodrik (1997) for rich countries. The Kuznets U-curve is also confirmed; inequality is highest in countries that are neither rich nor poor. Many of these countries are found in South America. Stark (1997) and Paris (1997) explain this by referring to the continent's stage of liberalization. GLOBALIZATION AND CITIZENSHIP In Table 2 we look at the relationship between economic openness and domestic conflict. Trade seems to lead to political stability and peace, while investment in poor countries leads to political instability and civil war. Inequality is associated with political instability, but not with civil war.13 Table 2 confirms that democracies are less often subject to political instability. We do not find a U-shaped relationship here, with semi-democracies as the least stable.
However, there are some problems with Table 2. First, the variable Investment in poor countries is not the best indicator of foreign investment 13 · With inequality included, the share of units with missing data is 27% and the number of civil wars is reduced by 50%. In any case, the regression analysis shows that inequality is not significant for civil wars.  since we simply have assumed that all the investment in poor countries is from foreign actors. Second, the regression for political instability suffers from time dependency, with the Durbin-Watson coefficient as low as 0.2. 14 Table 3 includes FDI and we obtain the same result as in Table L Regardless of which of the two investment variables is used, foreign investment leads to inequality in all the analyses here (Tables 1, 3, and 5). This result confirms most studies in this field, from Galtung (1971) to Vayrynen (1997) and Martin & Schumann (1998). Compared to Table 1 we see that the coefficient for Trade has changed. In the first Table, trade leads to inequality, but in Table 3 it is associated with less inequality. For the other variables we find similar results. The Kuznets U-curve is still confirmed, and democracy is still associated with greater equality.
To circumvent the problem of time dependency, in Table 4 we use the average value for each variable in the period 1965-85. A weakness of this 14 · In Table 3 time dependency is not as much of a problem, since the Durbin-Watson coefficient is 1.8 (Kennedy, 1998) GLOBALIZATION AND CITIZENSHIP method is, of course, that the number of cases decreases drastically. Unlike Table 2, we confirm the U-shaped relationship between democracy and conflict. Political instability is at a maximum for a democracy level of 3 ( on a scale from Oto 10). Thus, the highest level of conflict is found on the authoritarian side of the mid-point on the scale. The inequality variable is not significant in Table 4, but the direction of the coefficients confirms that inequality leads to greater political instability. However, from looking at the standardized coefficients we see that inequality is less important than many other variables in accounting for conflict. Lichbach (1989) is correct in emphasizing that inequality is not significant when other relevant variables are controlled for. Inequality is highly significant (0.00) when no control variables are included.
Except for democracy squared, none of the variables are significant in Table 4. The sign of the coefficients indicates that welfare and trade promote peace, while FDI leads to political instability. This result is similar to that  Table A-L The pattern is the same: trade leads to equality ( with one exception) and peace, while FDI leads to inequality and civil unrest, but these results are not significant. The different results for trade and FDI are surprising, since it is generally taken for granted that the process of economic integration has the same consequences regardless of the type of economic integration. The literature generally assumes that FDI and trade have similar effects, but emphasizes the difference in the effects of exports of agricultural products, minerals, and manufactures. This result may explain why those who look at the effects of trade conclude that it will produce peace (Oneal & Russett, 1996), while dependency theorists focus on foreign investment to explain why globaliza- GLOBALIZATION AND CITIZENSHIP 347 tion leads to lower economic growth (Bornschier & Chase-Dunn, 1985) and political instability (Boswell & Dixon, 1989).
There is another difference between these two types of economic integration: most researchers who focus on interstate war use trade as an indicator, whereas those who focus on internal war use FDI and type of export. Dependency theory has been mainly concerned with external explanations for civil war; foreign investment and exports of particular commodities have been seen as suitable indicators. Exports have always been interesting for dependency theory because they are indicative of the country's position in the world system.
There is also a substantial explanation for the different consequences of trade and foreign investment . FDI disrupts traditional economic patterns directly by establishing industry in the host country. Boswell & Dixon (1989) explained the revolutions in Cuba 1959 and Nicaragua 1979 in this way. The establishment of industry led to urbanization, which made it easier to organize a rebellion. Frequently, there was a common interest between the students and workers who fought against Western ideas and industry.
Openness through trade does not have the same effect, however. The literature in this field generally refers to mutual dependency as leading to economic growth and peace between states. It does not refer to civil war-and it is not equally obvious why domestic peace should result from trade. Here we have to go beyond aggregate trade. Table 5 gives us a better understanding of why trade creates welfare, equality, and internal peace. The consequences of high trade depend heavily on what type of commodities the country exports. We see that exports of manufactured goods promote equality, while exports of food are associated with inequality.15 Of course, the relatively egalitarian nature of most industrial societies may be a result of other typical features, such as wealth or democracy. The analysis shows, however, that the type of export commodities accounts better for inequality than does democracy or economic development.

Type of Trade Matters
Galtung argued in his frequently-cited article from 1971 that some products create spin-off effects, while others do not. Exports of minerals require few experts and many unskilled workers. Countries exporting agricultural produ cts do not need to promote education, and the country will remain poor and with high inequality. If a country can manage to produce manufactured and other highly-processed goods or exports, inequality is reduced-as has happened in several East Asian countries. They needed highly skilled workers in order to industrialize, and these workers eventually became active in the local labor movement. The result has been a higher level of welfare, and a decrease in inequality ( Bourgignon & Morrison, 1994;Wood, 1994). Table 6 confirms this view. Exports of manufactured goods promote prosperity, whereas exports of agricultural product do not.
1s. For this dummy variable we used Handbook of International Trade and Development Statistics (1976,1992). A country has been defined as 'agricultural' if exports of agricultural products account for more than 50% of total exports. These country-years are coded l, and all other country-years 0. The same coding procedure was used for manufactured goods and for minerals. GLOBALIZATION AND CITIZENSHIP

Globalization, Welfare, and Peace-the Middle-Income Countries
Most political unrest takes place in poor, non-democratic countries which are also not industrialized. Today there are numerous armed conflicts in Africa, in the Middle East, in South Asia, as well as in the successor states to the Soviet Union and in the Balkans. In Africa, ten out of thirteen wars are within developing countries. 16 Snow (1996: 96) points out that most internal wars take place in parts of the world that are most distant from the global economy. Thus, one can hardly argue that armed conflict primarily affects countries that suffer from excessive globalization.
One explanation for this is precisely the liberal theory that globalization promotes prosperity, which in turn creates conditions favorable to peace. India and China are good examples of the positive consequences of the liberalization and globalization process. China is the largest recipient of FDI in the developing world, and over the past 15 years its exports have increased more than tenfold. At the same time the country has made major improvements in health and education, and poverty now afflicts a tenth instead of a third of the total population (Human Development Report, 1997: 87).A similar pattern is found in India, Bangladesh, Indonesia, and Vietnam (ibid.). Several authors (Human Development Report, 1997;Snow, 1996;Vayrynen, 1997) have emphasized that globalization presents an opportunity for poor countries rather than a problem. Botswana is one of the countries that have managed to benefit from the inflow of FOL Revenues from its mining industry are invested to build up human capital in order to make the country attractive to other kinds of investment (Wor1d Investment Report, 1998: 191). Unlike most African countries, Botswana is defined as a 'lower-middle economy' rather than as a 'low-income economy' (Wor1d Development Report, 1996: 188 -221).
However, there are several problems in the globalization process, among them increased inequality. Human Development Report (1997: 88-89) reports a falling share of income for the poorest 20% in Argentina, Chile, the 1 e. The countries are Burundi , Chad, Kenya, Liberia Mozambique, Rwanda, Sierra Leone, Somalia, Sudan, and Zaire (Snow, 1996: 97), based on data from Project Ploughshares (1994). Their list is similar to, but not identical with, the armed conflict data used here. Dominican Republic, Ecuador, Uruguay, and Mexico. In 16 of 18 countries in Eastern Europe and the former Communist bloc, income distribution has become more skewed and poverty has increased during liberalization. Human Development Report explains the high inequality in poor countries by the entry of large companies into their markets. These companies offer higher wages than those found in the informal sector. A high level of trade increases the need for educated people, and poor people lose their jobs. This has been the case for many Latin American countries. Booth ( 199 3) explains rebellions in Somoza's Nicaragua (1977-79), El Salvador (since 1979), and Guatemala (since 1978) by export, income inequality, rapid economic growth, and political regime type. He points out that Central America's rapid growth in agricultural exports after 1950 and industrialization after 1960 markedly reduced the relative and even the absolute living standards of many members of the working class, who then mobilized to demand redress of their grievances. Where the state responded by accommodation and with limited repression (Costa Rica and Honduras), opposition mobilization stagnated or subsided. Where the state failed to ameliorate the growing inequality and instead employed heavy repression (Nicaragua, El Salvador, and Guatemala), opposition mobilization and unity increased, leading to a broad, rebellious challenge to regime sovereignty (Booth, 1993: 325).
However, according to the Kuznets U -curve, inequalities should decrease. In South-East Asian countries, inequality did decrease with industrialization (Wood, 1994). In the longer run, Latin America may experience the same development. If this coincides with a higher level of welfare, the future would look more peaceful.

Globalization, Poverty, and Political Unrest-the Poorest Countries
This optimistic view may not be valid for all countries. As noted, Snow (1996) pointed out that civil wars generally take place in countries that are most distant from the global economy. The globalization process does not include the poorest countries to the same extent as middle-income countries. Many poorer countries have scarcely experienced any expansion in world trade. Although the developing countries' share of the world population grew during the period 1970-91, their share of world trade hardly changed. The Least-Developed Countries, with 100/o of the world's population, have only 0.3% of world trade~half their share two decades GLOBALIZATION AND CITIZENSHIP 351 ago (Human Development Report, 1997: 83-84). This is also the case for FOL Latin America, Africa, and Asia (except Japan, Beijing, and eight Chinese coastal provinces), with 70% of the population, receive less than 10% of all FOL For a third of the developing countries, the ratio of FDI to GDP has fallen over the past decade (Human Development Report, 1997: 84). We have shown that trade with primary products is associated with lower levels of economic welfare, inequality, and political unrest. World Investment Report (1998) also points out that FDI in Africa leads to the extraction of primary products like oil and diamonds. The primary sector accounts for the largest share of FDI in Africa, with around 40% of the total stock of FDI in the period 1989-96. During that same period, the importance of FDI in manufacturing increased slightly (from 29 to 30%); the share of FDI in services to total FD I dropped from 33% in 1990to 27% in 1996(World Investment Report, 1998. Countries with a high inflow of FDI to extract primary products, such as Nigeria and Liberia, probably export precisely these products.17 If the same mechanism works for exports of primary products as for the inflow of resource-seeking FD I, the future does not look bright for these countries. They will remain poor, and with high income inequality. Primary produ cts are vulnerable in the external market, and income is hard to predict because prices are so unstable. Natural resources like oil, mining, and agricultural products seem to represent a problem for many poor countries rather than a means to obtain higher economic welfare and peace. Sashes & Werner (1995) argue that endogenous technical change does not occur in resource-rich countries because they become too dependent on natural resources and fail to innovate. Innovation progresses more rapidly within manufacturing, as distinct from agriculture, because the former sector offers greater opportunity for 'learning by doing' (Arrow, 1962). Congo is one of the countries which face problems with low welfare and political unrest because of its dependency 17 · Their main export article is minerals (Gissinger, 1997, appendix). World Investment Report (1998: 187-188) classifies the majority of the front-runner FDI countries in Africa as resource-seeking investments. Tunisia is the exception; it is defined as an efficiency-seeking FDI. on mining revenues. Molokai & Binswanger (1999) argue that dependence on mining revenue gives the rulers control over the revenue stream in the Congo; this control eliminates the need for bargaining for tax revenues. Since there is no bargaining, the population cannot influence how the revenues are used. The lack of powerful interest groups which can check each other as well as the governing elite makes for a weak state, and this in turn leads to civil unrest. Collier & Hoeffler (1999: 12-13) find that the risk of conflict peaks at a share of primary exports of 28% of GDP. A country with this level of resource has a risk of conflict 4.2 times greater than one without conflict. Hauge (1998) found that countries dependent on one single product are more prone to experience civil war. Our analysis similarly finds that export of minerals is associated with political instability, but this does not hold true for exports of agricultural products. 18 Our analysis also confirms that exports of agricultural products do not promote welfare. Since many poor countries are heavily reliant on the export of primary products, their future does not look bright.
However, not everyone shares such a pessimistic view of the future for the poorest countries. World Investment Report is basically optimistic (1998: 191-192). Sachs & Sievers (1998: 41) recommend that poor countries should make themselves more attractive to FDI by producing manufactured goods. Several countries in Africa stand a chance of developing a textile and apparel industry capable of competing in the US market, if they could have quota free and duty-free status, as is considered under the planned US-Africa Growth and Opportunity Acts. However, many Africans and leftwing critics are more pessimistic. The European Union excludes46 of South African agricultural products in negotiations between South Africa and EU for a free-trade agreement (Kit, 1999). Kit also points out that 53 manufactured goods will enter the local market in South Africa free of tariffs within the first few years of the agreement coming into effect-and well before South African industries have had time to restructure and face up to the 18 · The reason for this is probably the high threshold on this variable. In order for a country to be termed 'agricultural', exports of agricultural products have to make up at least 50% of GDP. GLOBALIZATION AND CITIZENSHIP 353 impact of such powerful competition. He goes on to say that' ... free trade agreements for industrial development and diversification [in South Africa] are usually overlooked or underplayed'. Pessimists also point to the Multinational Agreement on Investment (MAI), which was to remove virtually all barriers to the free flow of investment. The most controversial element in this agreement is that host countries have to protect foreign investment. First-World corporations are clearly concerned about the safety of their investment in regions where war can reduce their economic productivity and revolution can result in nationalization of the entire investment. To protect profit, MAI includes several provisions to ensure that governments of Third-World countries take measures to prevent such interference (Staples, 1998). Foreign corporations receive the same level of compensation from the government as do domestic corporations. Staples argues that this will be at the expense of social programs. Thus, developments in European Union and the ongoing discussion about the MAI agreement point in a pessimistic direction for the poor countries. Although overall trade has a positive effect on equality, welfare, and peace, this does not apply to the commodities which these countries export (agricultural and mineral products). Their chief source of revenue is exports of agricultural products, and such revenue becomes even harder to obtain when rich countries take protectionist measures against imports from poor countries. Protecting foreign industry in poor country through the MAI agreement is another problem. FD I has a positive effect on the level of welfare and a negative effect on distribution and conflict. If poor countries want to achieve a higher level of equality and peace, then granting compensation to firms from the rich countries in the case of political interference may not be a top priority item.

CONCLUSION
This article has dealt with consequences of the globalization process. Foreign direct investment (FDI) seems to have more negative effects on distribution and political unrest than does trade. However, the effects of trade, and probably FDI, depend on the stru cture of the economy. Exports of agricultural products lead to lower economic welfare, inequality, and political unrest, while exports of manufactured goods lead to higher economic development, equality, and political stability. Countries that export primary products are often poor countries with weak states. Thus, the globalization process may be positive for rich countries, whereas it may be negative for poor, agricultural societies.
However, both FDI and high levels of trade lead to higher levels of economic welfare, which in turn lead to peace. Here, we agree with Vayrynen (1997: 85), who argues that' ... openness to the international market is associated with better human conditions, while those left outside the international economic mainstream tend to suffer from their exclusion'. From this perspective, the challenge is to involve more countries in the globalization process, not fewer.
Dependency theory does account for the high level of inequality in poor countries. But the solution is not to close a country to trade or foreign investment. Rather, countries need to change their economic structure in order to gain from globalization. They need to obtain capital to strengthen their manufacturing industry, and they need time to protect their industry from foreign competition. Agreements like MAI may be counterproductive if they make it harder to regulate the market.
Globalization is positive in many ways-trade and foreign investment is necessary in order to increase revenues for poor countries. However, globalization will also make it more difficult for many states to become stronger. Even strong states are challenged by globalization. Vayrynen (1997: 85) argues' ... the state must pursue policies which foster equal opportunities and in that way economic and social equality. There may be a clash of interests here. Globalization appeals to strong states with a large public sector that can afford education and a good social policy, states able to distribute the economic gains and to establish a democratic system. At the same time, the nation-state with its democratic institutions loses some of its power in relation to multinational companies and currency traders (Martin & Schumann, 1998). This is a major challenge to the many nation-states today which are finding themselves increasingly powerless in the globalized world economy.