Lost in Globalization

Transcription

1 Lost in Globalization Economically Hurt or Culturally Threatened? Yotam M. Margalit Stanford University Prepared for Presentation at the Society for Comparative Research Yale University, May 2006 [DRAFT: Comments and corrections appreciated] Abstract A common assumption in the political economy literature holds that people s preferences on trade policy reflect the objective economic effects market openness has on them; those who gain from free trade support it, and those adversely affected by it oppose it. This paper studies the characteristics of the self-perceived losers from globalization and offers an alternative account that puts this assumption into question. I argue that many individuals view free trade as an inseparable component of a wider openness package, which also includes social and cultural dimensions. Peoples sense of loss from free trade therefore reflects, to a large extent, anxiety about non-economic cultural changes they associate with increased market openness. Using survey data from 44 countries, the paper offers an empirical test of the argument. The analysis also helps explain the oftenobserved anomaly of the support, by economic losers from free trade, of pro-trade yet socially conservative parties.

2 The representatives of the World Bank had better listen to us The World Bank and IMF and the transnationals must stop looting our natural resources, and stop privatizing basic services They must respect human life - Evo Morales, Bolivia s leading presidential candidate (November 2003) We need a president who will change the trade policy of this country. That s what I will do. And the folks who are 55 and 60 won t be losing their jobs to China and other countries around the world. - Richard Gephardt, Democratic Presidential Candidate Debate (October 2004). "It is wrong that our trade policies have caused 1 million good paying jobs to be shipped overseas because our companies can find cheaper labor and lower standards in another country." - John Edwards on the campaign trail (February 2004) What is the responsibility that American or European Markets have in keeping economies closed, in keeping a situation of hunger here in Brazil? - Fernando Enrique Cardoso, Brazil s President, (March 2000) "[Rich country s] approach is very simple. Should there be any new ideas in the social, economic or political field, they quickly come up with their own interpretations designed to benefit them. Thus we see them foisting on the world their moral values, political systems and economic globalization. Mahathir Mohammad, Malaysia s Prime Minister (April 2000) INTRODUCTION Economic globalization has increasingly become a high-profile political issue in many countries around the world. As the quotes above suggest, politicians identify the discontent of many voters with aspects of economic globalization and seek to seize upon this discontent to shore up political support. But, despite the growing salience of opposition to market openness as a political factor, there is very little that we actually know about the politics of the discontents. Who are those who view themselves as losing out from economic globalization 1? What aspects of it cause them to view themselves as such? How does the discontent translate into a political response? Existing literature on attitudes toward economic globalization - commonly taken to mean liberalization of markets, reduction of trade barriers, increased capital movement and foreign direct investments typically takes as a starting point the assumption that individuals are self-interested and that their support or opposition to increased market openness is a direct function of whether they stand to gain or to lose from it. Two central 1 Throughout the paper the terms economic globalization, economic openness, market openness and free trade policies are used interchangeably. 2

3 lines of argument are used in the political economic literature to explain attitudes towards market openness. The dominant approach holds that people s calculus is determined by their relative position in the labor market, i.e. whether the skill-set with which they are endowed stands to become more or less valuable in an open economy (e.g. Rogowski 1987, Irwin 1996, Scheve & Slaughter 2001). The more recent second approach holds that changes in individuals relative purchasing power shape their attitudes towards economic openness (Baker 2003, 2005). Importantly though, both lines of argument share the same underlying assumption, that attitudes towards economic openness ultimately reflect the objective economic conditions individuals are confronting. This paper questions the degree to which this assumption is useful when accounting for people s views on how trade openness affects them, especially of those who feel hard-done by it. I argue that in explaining people s sense of loss from economic globalization, the existing political economic theories offer limited insights. In contrast to the labor-market and consumption-based theories, this paper advances an alternative argument to explain people s belief that market openness is adversely affecting them. My argument holds that many individuals do not assess the economic effects of market openness independently of the non-economic effects they associate with the openness "package". The complex and all-encompassing nature of globalization means that, for many, attitudes toward trade are inseparably tied to views on other changes that come with openness, be they demographic, technological or cultural changes. Specifically, I contend that the selfperceived losers from economic globalization are, to a large degree, not those objectively hurt by it, but rather those who feel culturally threatened by the effects of openness to influences from other countries. Using recent cross-national survey data, the paper examines the degree to which the central arguments advanced in the literature on trade-policy preferences help account for who the self-perceived losers from economic globalization are. The analysis shows that both the factor-endowments model and the more recent consumption-based theory offer limited insight into the question at hand. In contrast, the cultural threat argument advanced in this paper is shown to carry significant empirical weight across all geographical regions: concerns about the effects of cultural openness play a large role in shaping people s sense of personal economic deterioration due to globalization. 3

4 In studying the characteristics and political views of the self-perceived losers of globalization (henceforth LOGs), this paper seeks to make three contributions 2. First, it offers an account of who the LOGs are in the different country settings. Second, it makes the case for a reconsideration of the assumptions underlying the study of trade policy preferences. Complementing the focus in the trade literature on changes in objective economic conditions relative position in the labor market, asset prices and purchasing power of consumers - the claim is made that individual perceptions of the economic effects of trade openness are largely driven by anxieties about the cultural changes associated with openness 3. The addition of this aspect to the analysis of trade preferences, the paper suggests, offers a more accurate account of the process by which people form attitudes. Lastly, the paper offers an insight into the variation in the political responses of the groups who are at the losing end of globalization. By arguing for the close relation in people s minds between the economic and the cultural effects of openness, the paper provides a framework for explaining the widely discussed anomaly whereby the losers of free trade often vote for parties that advocate pro-trade policies. It is important to emphasize that this paper, as well as the larger research project this paper is part of, is concerned with the self-perceived losers from economic globalization. Explaining someone s beliefs about the adverse effects free trade have had on her personally is not the same as explaining her support or opposition to free trade policies in general. As I later show, it is not uncommon for people to hold the view that free trade policies have benefited the country as a whole but adversely affected them personally, and vice versa. The analysis offered herein is therefore not an attempt to 2 Throughout the paper, the acronym LOGs refers to those who perceive themselves to be losing from economic globalization, not necessarily to those actually adversely affected by it. For the sake of convenience, I drop the self-perceived prefix when discussing the LOGs. 3 Important exceptions in the trade literature are the studies that use the ISSP dataset (O Rourke & Sinnott 2001, Rodrik & Mayda 2005), which find that measures of nationalism and patriotism are statistically significant predictors of attitudes towards trade. These studies, however, examine trade policy preferences rather than perceptions of personal economic gain or loss from free trade. The question used by both studies as the dependent variable is: Now we would like to ask a few questions about the relations between [respondent s country] and other countries. How much do you agree or disagree with the following statement: [respondent s country] should limit the import of foreign products in order to protect its national economy?. Note that this question is very different from the issue examined in this paper, which centers on individual s personal beliefs about how they and their families have been affected by economic openness. In another paper, Magaloni and Romero (2004) study cross-national attitudes variation in Latin American countries and find that partisan predispositions are significantly associated with individuals support or opposition to aspects of globalization. 4

5 confirm or disconfirm the validity of the existing trade-preference theories, which seek to explain variation in mass attitudes on trade policy. Rather, these theories are employed in order to examine the degree to which they help account for the specific phenomenon of discontent and sense of hurt caused by economic openness policies. The paper proceeds as follows. I start with a review of the literature on economic globalization and its losers. Section Two lays out my argument, as well as a set of alternative hypotheses, on the question who are the self-perceived losers of economic globalization. Section Three describes the data used to test these hypotheses and discusses measurement issues. Section Four describes the results of the empirical analysis. A discussion of the main argument of the paper then ensues: to what extent do views on cultural openness inform individuals' interpretation of the changing economic reality? Section Five situates the paper in the context of the larger project of the dissertation and concludes. I. GLOBALIZATION AND ITS LOSERS IN THE LITERATURE In recent years the debate over globalization and its effects has generated a vast literature. Debates on globalization-related topics range from whether it benefits the poor or not (e.g. Bhagwati 2001, Wolf 2004), through the role of international institutions in a globalizing economy (Rodrik 1997, Stiglitz 2000), to the extent to which globalization affects the role of the nation state as the key political unit (Ohmae 1990, Strange 1996). In this expansive literature, however, one cannot find any systematic account of globalization s discontents: who they are, why the view themselves as such, or how does the sense of losing out translate into a political response. Instead, the discussion concerning globalization s losers is often reduced to references to the anti-wto riots in Seattle and Genoa, or some mention of a potential backlash that could come about as the process of market integration expands. This section outlines the key insights the extant literature does offer on the characteristics of globalization s losers and on the political implications of their discontent. An important strand in the study of international trade seeks to explain the trade policies governments adopt over time. Initial arguments linked trade policies to larger theories of power and international conflict (e.g. Krasner 1976, Katzenstein 1985, Gilpin 5

6 1987,). More recently, directed by the view that government policies reflect the preferences of the electorate, attempts have been made to explain and predict individual attitudes on trade. This literature predominantly utilizes two international trade theories based on the notion of relative advantage and applies them to account for variation in individual preferences on trade openness. Each theory has well-defined implications regarding the distributive consequences of trade, and therefore offers clear hypotheses also on the question of who the losers from market openness are. If indeed the individuals that view themselves as losing out from market openness are also those that actually are worse off because of it, these theories could offer valuable insights regarding the composition and characteristics of the self-perceived losers. The most commonly used model in the trade preferences literature is the Heckscher-Ohlin (H-O) factors-endowment model. In most of its recent applications, countries are categorized as either skill-abundant or skill-scarce based on their supply of skilled workers relative to other countries 4. Openness of the market to international trade, according to the H-O model, brings a country to export goods whose production is intensive in its abundant factor while importing goods that are intensive in its relatively scarce factor. This logic therefore suggests that the relative losers from market openness are the unskilled in the skill-abundant countries and the skilled labor in the skill-scarce countries. Indeed, a wide range of empirical work has found evidence supporting aspects of the theory, through examination of historical accounts of political alliances and voting records (Rogowski 1987, 1989; Beaulieu 1996, Hiscox 2002), as well in analyses in the individual level through survey responses (Scheve and Slaughter 2001, 2004; O Rourke and Sinnott 2002; Rodrik and Mayda 2005) 5. However, arguments derived from the H-O model have struggled to account for certain empirical regularities, most notably the fact that cross-national survey work has 4 The original version of the H-O model focuses on capital and labor as the factors of production. Later modifications to the model include land and a division of the labor category to skilled and unskilled labor. 5 Some have argued that the H-O logic holds in the long-run analysis, but due to the immobility of certain production factors, the key cleavage along which support and opposition to trade lies in the short-run is the industry of employment. This specific-factors approach, often referred to as Ricardo-Viner (R-V) model, suggests that the losers of economic openness are first and foremost those working in the import-sensitive industries (Magee 1978, Irwin 1998). This approach has shown to be of little explanatory use when examined as a predictor of mass attitudes on free trade (Scheve & Slaughter 2001, Mayda & Rodrick 2005). 6

7 shown little evidence of a negative relationship between skill-level and pro-trade attitudes as a function of the country s relative skill-abundance. In fact, higher-skilled individuals are consistently found to be more in favor of free trade than their less-skilled counterparts almost everywhere, regardless of the relative skill-abundance in their respective economies. A more recent theoretical approach attempts to address this unexplained regularity, as well as others, by arguing that attitudes on trade reflect individuals interests as consumers (Baker, 2003, 2005). Consistent with the H-O logic described above, prices of a country s exportable goods rise with freer trade, while prices of the importable goods decrease. The purchasing power, or real wage, of heavy consumers of importable goods thus increases with free trade, while heavy consumers of exportable goods experience a relative loss. Since the poor have to expend a greater share of their income on necessities that are unskilled-intensive (e.g. food, clothes) and the wealthy have the highest rate of saving, the middle-class are, in relative terms, the key consumers of skill-intensive goods. The consumer-based theory of trade preferences thus holds that the relative losers of free trade are the heavy consumers of exportables, whom are the poor in skill scarce countries and the middle-class in the skill-abundant ones. The economic of happiness literature, which examines the relation between individuals economic standing and their subjective well being, offers insights on the sources of beliefs about economic loss (e.g. Easterlin 1995, 2001; Alesina et al 2002). This literature finds that the economic determinants of happiness are similar in countries that differ significantly in their level of economic development: income, age, or unemployment are strongly associated with reported levels of well-being in all countries. Graham and Pettinato (2001, 2002) analyze data from Russia and Peru, two countries that have experienced difficult market liberalization transitions, and find that relative income differences and volatility in income flows are two of the most important sources for people s economic grievances. Their analysis suggests that those in the middle or lower middle of the income distribution are more likely to be dissatisfied than are the very poorest groups. Another body of work this paper ties into is the literature on economic voting, which addresses the question of the extent to which people s vote is driven by an 7

8 economic rationale. Whereas some scholars claim to find an economic vote function (Fiorina 1981; Kramer 1983, Lewis-Beck 1988), cross-national analyses have failed to find a strong link between electoral and economic outcomes (e.g. Cheibub and Przeworski 1999). This debate has sparked research on the way people process and integrate economic data into their voting decisions, with significant work dedicated to exploring the rationality of individuals economic judgments (e.g. Rabin 1998, Caplan 2001). Well-documented deviations in individual behavior from rational expected utility calculus were a basis for constructivist accounts that focused on norms, ideas and ideology as important sources shaping economic beliefs and interests (e.g. Katzenstein 1996, Blyth 2002, Herrera 2005). The argument advanced in this paper speaks directly to such constructivist accounts, however two important distinctions need to be emphasized. First, my argument does not question the fact that objective economic conditions matter greatly in shaping people s beliefs about economic issues. This paper, in no way, is advancing the claim that all beliefs about the economy are imagined or socially-constructed. Second, unlike constructivist accounts, the argument advanced here does not suggest that context or shared understandings are what shape perceptions of economic self-interest. Rather, my contention is that due to insufficient information or lacking the cognitive skills needed to disentangle the different effects of globalization, individuals that are aggrieved by certain non-economic aspects of openness (mis)perceive themselves to be worse off from its economic aspects too. In simpler terms, perceived loss from cultural openness is translated to a perception of loss also from economic openness. It is to the task of making this point that I now turn. II. HYPOTHESES Who are the self-perceived losers? There is wide agreement that the benefits of globalization are unequally distributed. Clearly, some countries and some groups of individuals have gained tremendously from the increased integration of previously distinct economies and societies. Others, however, have been adversely affected by the changing economic environment, whether in absolute or in relative terms. Economic globalization has made many insecure about their 8

9 occupation prospects; some have seen their real income shrink, and yet others have grown to feel increasingly inadequate in face of the new skill demands of the globalizing market. Indeed, the difficult processes different countries are undergoing of adaptation to the changing market environment have generated many potential sources for discontent and anxiety. But of those potential sources of discontent, which ones actually matter, and to what extent do they vary across countries with different economic conditions? Figure 1 gives a rough indication of the variation in the size of the self-perceived losing group in different countries. It is evident that the size of the group varies significantly across countries, with as low as 5.6% in the UK, and rising up to 28% in Argentina. In Kenya (not shown the chart) this figure stands at 65%. What accounts for this variation? An intuitive explanation might be that the sizes of the losing group reflect the varying types of shocks different economies have undergone as a result of integration into world markets. Less developed countries that face greater barriers in adapting to a competitive global economy are, one might expect, likely to have a larger sector identifying itself as worse off from market openness. On the other hand, concerns unique to higher income states such as immigration, job loss from outsourcing of production to cheaper markets, and lowering of labor standards, might generate even wider levels of discontent in developed countries. % LOGS in Country % of total population ARG BRZ BUL EGY IND ITA LEB MEX PHI POL SAF TAN UK USA Figure 1. Categorization of LOGs is based on response to the question: Now thinking about you and your family do you think the growing trade and business ties between our country and other countries are very good, somewhat good, somewhat bad or very bad for you and your family? LOGs are those that answered the question with a somewhat bad or very bad response. 9

10 The data however does not support either argument. The correlation between a county s trade openness and the size of the LOG group is low (-0.04) 6. So is the correlation between the country s economic development and the size of its LOG group, (-0.09). In other words, the economic indicators of development and openness do not stand in a direct relation with the extent of discontent towards free trade. If so, what does account for variation in levels of discontent, and to what extent do the self-perceived losers in the different countries share similar characteristics? This section lays out a set of hypotheses to address this question. Starting with hypotheses derived from the major arguments in the literature on trade preferences, I then lay out my cultural protectionism theory of attitudes on economic openness. Lastly, two alternative explanations are offered. The relative strength of the theories is then examined in the empirical section Existing IPE Models 1. Relative Position in the Labor Market Hypotheses The dominant Heckscher-Ohlin theory, which closely follows the logic of the Stolper- Samulason theorem (1941), holds that the key determinant of individual attitudes on trade is the employable assets they have, i.e. the skill type they possess: those endowed with a relatively abundant skill will benefit from opening the economy while owners of a scarce skill will experience a decline in their relative standing in the labor market. If this logic indeed captures the way people form their views on how economic openness affects them, then LOGs could be predicted by their skill-type: H1.1. LOGs are those with factor-endowments that are in relative scarceness in their own country. In economies with relative scarcity of unskilled (skilled) labor, LOGs will predominantly be the unskilled (skilled). A second hypothesis relates to the core argument of both the factor-endowment (H-O) and the factor-specific (R-V) models discussed earlier. The argument is that people s attitudes on free trade are first and foremost a function of their perceptions of how it affects the labor market. The H-O model suggests that the employable skills of the individuals will determine their views on the impact of economic openness on their occupation prospects; R-V holds that the industry of employment, not the employable 6 Trade openness is calculated as the sum of a country s exports and imports as percentage of its GDP. 10

11 skills, will have the main influence on individuals views. Either way, the logic is that attitudes on market openness are driven by a self-interested calculation of whether one s prospects in the labor market are improving or worsening as a result of free trade. The implication of this logic is therefore: H1.2. LOGs are those with a strong sense of deterioration in occupation prospects and increased job insecurity. 2. The Consumption Based Hypothesis An alternative approach to account for attitudes on trade openness shifts the emphasis from individuals role as participants in the labor market to their role as consumers of goods. In brief, the logic goes as follows: Since protectionist trade policy raises the prices of tradable goods, the general consumer public is, in effect, made to pay for the economic rents with which the closed economy provides the protected sectors. Economic globalization drives down the prices of imports while raising the prices of exportable goods. This change in prices directly affects the purchasing power of consumers: it increases for heavy consumers of importable goods and decreases for consumers of exportable goods. The relative losers are therefore consumers of the exportable goods. Since the main consumers of unskilled-intensive goods, in relative terms, are the poor, while the middle-class are the heaviest consumers of skill-intensive goods (as a share of their incomes), the predictions of the consumption-based hypothesis can be summarized as follows: H2. The LOGs are consumers whose purchasing power has diminished. These consumers are disproportionately the poor in skill-scarce countries and the middle class in skillabundant countries. In substantive terms, the expectation is that perceptions of loss from free trade will be in negative relation with income, but that the interaction of income and skill-abundance in a country will be positively correlated with a negative view of free trade. Cultural Protectionism Theory The effects of globalization are dramatic and all encompassing. Cheap imports, enormous variety of products on the shelves, internet, off-shoring of production to markets abroad, growing presence of multi-national corporations, waves of immigration and emigration, 11

12 satellite dishes beaming broadcasts from around the world, unprecedented levels of international travel, etc. are the tip of an iceberg. To what extent are individuals able to disentangle the effects of this complex package of changes and processes that goes under the name 'globalization and assess a specific component of it - trade openness - independently of all its other components? The answer I propose to this question is that many people are not able to do this. I argue that because the effects of increased openness between countries are so intertwined, dramatic and complex, existing theories of individual attitudes to trade rely on an improbable assumption when they posit that people s views on trade directly reflect their calculations of personal economic gain and loss. I contend that, by divorcing the noneconomic aspects of openness from the study of attitudes toward trade, these theories miss an important part of the story. My argument is that the changes brought about by economic globalization are viewed by many as synonymous with other changes such secularization or Americanization. To be sure, many do not view the other changes as negative and associate them rather with development and progress. For others, however, the changes they associate with free trade are a source of uncertainty and anxiety, which translate into a negative view of the openness package' as a whole. Opposition to economic openness thus often reflects concerns about non-economic changes such as a perceived weakening of traditional family structures, secularization, emigration of the young, women in the workplace, etc. Such apprehension about the effects of these wider processes of openness, I conjecture, develops into beliefs about the detrimental effects of economic openness and leads many to believe that free markets adversely affect them 7. This argument suggests that much of the trade-preferences literature makes problematic assumptions about the level of economic knowledge and sophistication among the general population. For example, the often-used working assumption according to which individuals know with certainty the effects of trade policies on 7 The empirical section of the paper offers some data supporting the plausibility of this argument, although to detail the mechanism of this process convincingly is beyond the scope of this paper. In my dissertation I seek to show the causal link between fear of the effects of cultural openness on the one hand and beliefs about economic loss from trade openness on the other. To this end, I am currently working on an experiment that tests for the degree to which exposure to cultural concerns associated with openness affects individuals views on how trade openness affects them. Any suggestions for additional ways to test this presumed link would be greatly appreciated. 12

13 individual incomes and assets holdings (Scheve & Slaughter, 2001), or statements like voters may see trade liberalization as a remedy to high prices (Baker 2005), may offer reassuring tractability in the modeling of individuals preferences. However, they seem descriptively implausible in light of the extensively documented literature about people's knowledge about and judgments on economic issues (Blendon et al 1997, Dahl and Random 1999, Caplan 2001) 8. My expectation is that economic variables account for only a limited part of the variation in individual attitudes with regard to how economic openness affects them. Instead, I conjecture that people s concern with the social and cultural changes associated with foreign influences are the source of much of the apprehension from the perceived effects of free trade. The camp of self-perceived losers from economic openness, I argue, includes many who in objective terms have not lost economically from free trade (some may have even gained) but who associate economic openness with other processes which they feel affect them adversely. The implication of this argument in the political context is straightforward. The scholarly literature on the political implications of free trade has until now assumed a clear-cut division between those who benefit economically from free trade and those who do not (Rogowski 1987, Frieden 1991, Hiscox 2002). The theory presented here, in contrast, offers a framework to explain how free trade can generate political cleavage that is predicated on cultural elements. Approaches to Modeling the Cleavages Created by Market Openness Economic Impact of Openness Economic Impact of Openness Negative ~ Negative Negative ~ Negative Social-cultural Impact of Openness Negative ~Negative Figure 2 Existing PE Models Cultural Protectionism Approach 8 No doubt, the authors utilizing such assumptions are well aware of these assumptions limited descriptive accuracy. The issue then is ultimately the degree to which these assumptions serve as the tools in accounting for individual preferences. 13

14 The shadowed boxes in Figure 2 represent the potential pool of self-perceived economic losers. Note that in the figure on the left, the shadowed box includes only those adversely affected by the economic changes that come with openness. The figure on the right presents the division implied by the theory of this paper. Namely, the losers camp includes a larger potential pool of disaffected individuals: not only those objectively worse off because of trade openness but also those who feel culturally threatened by the changes associated with openness and who thus come to believe that trade openness as such is bad for them. Of course, the theory does not predict that all those culturally threatened by openness necessarily believe that freer trade is bad for them, but rather that there is a significant group to which this argument does apply. Relying on Prospect Theory s notion that losses loom larger than gains (Kahneman and Tversky 1979, 1984), my argument suggests that individuals whom openness has made better off in one dimension (economic or cultural) and worse off in the other, will tend to view themselves as overall losing out from the openness process 9. We should not expect therefore to encounter many instances where individuals who benefit from changes in the cultural dimension but lose in the economic dimension describing themselves as better off from free trade 10. The theory proposed here emphasizes that low levels of information, or lack of necessary analytical tools, are keys to explaining people s tendency to assess the different effects of globalization as a single bundle rather than to analyze them separately. Greater understanding of economic issues and higher cognitive skills are thus likely to lessen this tendency. I therefore expect that - holding all else equal - individuals with lower levels of education are more likely to associate cultural anxieties about openness with a sense of loss from economic openness, than individuals with similar levels of cultural anxieties 9 Obviously, in extreme cases where the gains in one dimension are so tremendous they may well trump small losses in the other dimension. But the theory holds that in cases where gain and loss are in the same range, losses will indeed loom larger. 10 A question this argument raises is whether the process described above operates also in the other direction, namely does economic loss translate into beliefs also about cultural loss? My sense is that this process is happening too, in significant levels, however demonstrating it empirically is difficult since there are no clear measures of objective cultural threat or loss. It is therefore very difficult to show that one s negative perception of the cultural effects of openness is unmerited and caused by something else such as economic loss. Panel data that tracks over time the economic performance of individuals as well as their changing beliefs about social and cultural issues would be an effective way of testing for the existence of such a relationship. It is something I hope to address in greater depth in the dissertation. 14

15 but with higher levels of education. The theory therefore makes the following two predictions: H3. Individuals' sense of losing out from trade openness is often driven not by an assessment of its objective economic effects, but rather by feelings of being threatened by cultural openness to foreign influences. H3.2 The tendency to form attitudes on the effects of free trade that reflect attitudes towards cultural openness is a function of one s available analytical tools. Ceteris paribus, lower levels of education will tend to correlate with the bundling of economic and cultural effects of openness. Alternative Explanations In addition to the extant trade-preferences theories in the literature and the one proposed in this paper, I examine two alternative arguments that seem plausible in explaining subjective assessments of economic loss. The first approach centers on individuals economic standing - absolute as well as relative in accounting beliefs about trade. The second approach assumed that a sense of being adversely affected by economic openness reflects a wider, and unrelated, dissatisfaction in life. 4. Personal Economic Standing Hypotheses An argument advanced in the economics of happiness literature holds that those perceiving themselves to be worse off economically are not those with the lowest earnings but rather those most discontent with their relative economic position, whether as a result of increasing income inequalities in the country or a prevailing perception of social immobility (e.g. Graham & Pettinato 2002). Subjective reports of economic dissatisfaction may thus correspond well with beliefs about the effects of economic openness. A plausible alternative could be that the sense of losing out from economic globalization is a straightforward outcome of financial struggle, regardless of whether it is caused by trade openness or not. If one s economic situation is bleak, goes this argument, any process associated with the country s economy may be considered to have had an adverse effect on the individual s personal standing. The predictions of these two arguments can be summarized as follows: 15

16 H4.1 Individuals that view market openness as bad for them are those most discontent with their relative economic standing. H4.2 Individuals that view market openness as bad for them are people that are economically struggling. In other words, LOGs are the poor. 5. General Dissatisfaction with Life Hypotheses The classification of LOGs used in the empirical section of this paper focuses on individual s own assessment of how increased market openness has affected them; it does not rely on an objective measure of loss by the individual. A possible conjecture might be that the self-perceived losers from globalization are simply the kvetchers, i.e. those that are generally discontent with life and that respond to a survey saying they are worse off because of free trade, just as they would about anything else being considered in the survey. In examining this argument, I distinguish between two types of discontents: First, those who have a very negative assessment of their current situation in life; second, those who feel that their personal situation has severely deteriorated over the past several years. The logic underlying these hypotheses is that for both types of discontents increased market openness is guilty by association, i.e. it is viewed as detrimental regardless of whether it actually had a negative effect on the respondent or not. H5.1 The LOGs are those most discontent with their current situation in life. H5.2 The LOGs are those who view their lives as having significantly worsened over recent years. In sum, the hypotheses outlined above can be described as falling into three types of explanations: (1) Hypotheses conforming to the main trade-theory models in the literature (H1 & H2); (2) Hypotheses based on the notion that dissatisfaction (with life in general or with one s personal economic standing) translates into a negative view of any precipitator of change, in this case free trade (H4 & H5) (3) The paper s main argument, namely that concern and anxiety about cultural changes associated with openness generate beliefs about the detrimental effects of trade openness (H3). In the following sections the hypotheses outlined above are tested empirically. I begin by describing the cross-national survey data utilized in the empirical analysis, and then move on to discuss the variables used to operationalize each of the hypotheses. 16

17 III. THE DATA & EMPIRICAL APPROACH The analysis in the paper is based on data from the Pew Global Attitudes Project. The data is comprised of 44 national surveys carried out in five continents during the year The surveys were all conducted in the native language by leading local polling firms, interviewing a representative sample of the population. The sample size ranges from 500 (e.g. Slovak Republic) to 3000 (China). Overall, the dataset consists of 38,263 survey responses 12. In almost all countries the exact same set of questions was asked, though in several cases certain questions had to be omitted (this applies mostly to China and Egypt). This dataset is uniquely suited for the purposes of this research since the survey includes questions about the perceived effects of free trade on the respondent herself, rather than questions about the respondent s general attitude on free trade policies, as is the case with other cross-national surveys. The dependent variable used in the analysis is based on a respondent s answer to the question Now thinking about you and your family do you think the growing trade and business ties between our country and other countries are very good, somewhat good, somewhat bad or very bad for you and your family? 13 The exact same question was asked also about the effects of trade and business ties on the respondents country, rather than on their family. The answers to the two questions indicate that respondents were clear-minded about the difference in the questions emphases, as responses often diverged: In fact, a majority (59.5%) of those who answer that growing trade and business ties were bad for them and their family believe that free trade is good for the country as a whole. The opposite figure is much 11 See appendix for complete list of countries. 12 Approximately 91% of the respondents answered the question on the effects of trade openness on themselves and their family. This reduced the total number of survey responses to 34,725. To avoid losing additional observations, missing observations were imputed for income (8.6% of observations), age and education (0.7% each). Imputation was done for each country separately, using Amelia (King et al, 2001). Results were unchanged for all variables when running the regressions with and without the imputed observations. 13 For example, the dependent variable used in Scheve and Slaughter s (2004) study on individual trade preferences is based on the following question in the NES Surveys of 1992 and 1996: Some people have suggested placing new limits on foreign imports in order to protect American jobs. Others say that such limits would raise consumer prices and hurt American exports. Do you favor or oppose placing new limits on imports, or haven t you though much about this? Clearly, such a question taps into a different issue from the one examined in this paper, namely individuals sense of how trade openness affects their family and them personally. 17

18 lower: only 6.4% of those who believe that they themselves benefit from growing trade openness think that it negatively affects their country. To test Hypothesis 1, namely the Heckscher-Ohlin factor-endowment argument, I use education level as a proxy for respondent s skill-type. This is a common assumption in much of the trade literature 14. The categorization is based on the highest level of education attained, and the categories are: 1. No formal education; 2. Completed primary education; 3. Some secondary education; 4. Completed high school; 5. Some university or college education; 6. Completed university or college degree. The factor-endowment argument holds that a survey respondent that completed his high school education in a skill-scarce country, say Bolivia, has a very different standing in the local labor market than a high-school-graduate in a skill-abundant country, say Germany. To account for the effect of skill as a function of the relative standing in the market I also include an interaction term of individual skill and the country s skill endowment. Measurement of a country s skill-endowment is problematic due to lack of comparable cross-national data. Even the most commonly used datasets measuring crossnational formal education attainment are partial and rely on extensive extrapolations (Barro and Lee, 2003). I therefore use the natural log of a country s per capita GDP as a proxy for a country s skill-abundance. The interaction term of skill and skill-abundance is represented with the Education X GDP per capita variable. As noted earlier, the arguments inspired by the H-O model focus on risk exposure in the labor market as the dominant source of attitudes on free trade. However, operationalizing the level of risk exposure in the labor market is problematic for several reasons. First, measuring accurately what industries in a given country have gained or lost is a difficult task that produces a lot of measurement noise. This in turn renders the exercise of identifying the workers in the losing industries highly problematic 15. Second, the occupation classifications in the surveys used in the paper are not specific enough to be able to get at an accurate measure of the industry effect. In order to get around these 14 For recent examples see Scheve & Slaughter 2002, Rodrik & Mayda Use of rough measures of growing or declining industries would not capture the variation in individuals changing standing in the labor market. For example, the significant variation in success of firms within an industry, or changes in production methods most notably in agriculture - that affect individuals job security would not be picked up by the categorization of the industry. 18

19 problems, I use a more direct measure of respondents assessments of changes in the job market. In addition, I include employment status data in the regression as another control. The variable Job availability is coded using the following question: Has each of the following gotten better or worse over the last five years in our country: The availability of good-paying jobs? (0 = Worse, 1= Not Changed 2= Better) 16. The employment status dummy variables are: (i) Self-employed; (ii) Part-time (iii) Unemployed, (iv) Pensioner; (v) Not employed (e.g. housewives, students, disabled). To test my argument that the sense of economic loss from free trade is derived from concern with or anxiety about non-economic (social-cultural) changes associated with openness, I code the Foreign Cultural Influence variable using the response to the survey question: Do you think that having the opportunity to watch movies and TV and listen to music from different parts of the world is very good, somewhat good, somewhat bad or very bad for you and your family? (Coded on a scale from 1-4, with 1 being very good and 4 very bad ) The two pocket-book hypotheses (H4.1 & H4.2) advance the logic that financial struggle leads respondents to view economic openness as having had an adverse effect on them personally, regardless of whether it actually had or not. To test these hypotheses, I include two variables in the analysis: the relative household income of respondents, and their subjective assessment of their economic standing. Income is measured by the question Here is a list of incomes. Which of these does your household fall into counting all wages, salaries, pensions and other incomes that come in? Just give the letter of the group your household falls into, before taxes and other deductions. Since the focus here is with one s relative income compared to that of others in the country, respondents in each country were divided into five income quintiles (5 denoting the highest earners). The second income-related variable is Economic Satisfaction. Measurement of the variable is based on the response to the question: How satisfied are 16 This question is problematic in that it is double barreled; in theory, one can interpret the question to be asking about job availability in general, while another might understand the emphasis of the question to be on good paying jobs. This leads to some difficulty in interpreting the meaning of the response. Nonetheless, the correlation of responses to the question with those to related questions on employment status and on satisfaction in the workplace suggests that this proxy does capture an important variation in individuals perceived prospects in the labor-market. 19

20 you with your household income? (1. Very dissatisfied; 2.Somewhat dissatisfied; 3. Somewhat satisfied; 4 Very satisfied) The final two hypotheses link general discontent in life to views on economic globalization. H5.1 emphasizes respondents view of their current state, and H5.2 focuses on respondents sense of deterioration, i.e. of a negative change to their lives. The following measures are used to test these two arguments: Current State is coded on a scale of 1 to 10, using the following question: Here is a ladder representing the ladder of life. Let's suppose the top of the ladder represents the best possible life for you; and the bottom, the worst possible life for you. On which step of the ladder do you feel you personally stand at the present time? (1-10) 17. Five Year Change is based on the question asking the respondent to rate her personal situation five years ago on the same ten-scale ladder of life : On which step would you say you stood five years ago? The coding of Five Year Change is then obtained by subtracting the ladder today rating from the response to the question about five years ago: Five Year Change= [(ladder five years ago) (ladder today)]. This means that the higher the score of Five Year Change, the bigger the perceived deterioration in one s life. Additional independent variables. In addition to the variables outlined above, the literature gives reason to include several additional control variables in the analysis. Age of the respondent is included since older people may find it more difficult to adjust to new skill requirements in the labor market (Tucker, Pacek and Berinskiy 2002). Studies of attitudes to free trade have repeatedly found gender gaps in respondents preferred level of protectionism. Possibly due to child-birth and child rearing considerations that affect their prospects in the labor market, women tend to be more apprehensive about free market economic policies (Burgoon and Hiscox 2003). A dummy variable for Female is therefore included. Measure of Religiosity is also controlled for, though omitted in the final models due to a large number of missing observations. I therefore conducted separate tests of the model for countries where data on religiosity was available. Results are reported in the appendix. Lastly, a variable measuring Nationalist Sentiment is 17 Emphasis appears in the original survey question. 20