The Electoral Consequences of Offshoring: How the Globalization of Production Shapes Party Preferences

Abstract

How does offshoring affect individual party preferences in multiparty systems? We argue that exposure to offshoring influences individual preferences for those political parties with clear policy positions on issues relevant for individuals with offshorable jobs (left, liberal, and center-right parties) but does not affect voting decisions for parties concentrating on other issues (green or populist right parties). Examining individual-level data from five waves of the European Social Survey for 18 advanced democracies, we find that these effects vary by skill level and exposure. Offshoring increases preferences for liberal and center-right parties that advocate economic openness among the highly skilled. In contrast, low-skilled individuals exposed to offshoring are more likely to prefer leftist political parties that champion social protection and redistribution. Offshoring does not affect the propensity to vote for green and populist right parties.

Keywords

offshoring party preferences multiparty democracies globalization risk voting behavior elections

Introduction

In the wake of the recent successes of populist parties and candidates in the Western world, a prominent narrative has emerged that these achievements reflect the deep dissatisfaction of the losers of globalization. Who these losers are, how their dissatisfaction translates into political decisions, and how established parties are affected by globalization remains opaque, however. This article contributes to unpacking the mechanisms linking globalization and voting behavior. It focuses on one specific aspect of globalization that has accelerated rapidly over the last decades: offshoring—the migration of employment from one country to other countries (Blinder, 2009).

The enormous technological advances of recent decades have increasingly enabled firms to not just trade internationally, but also move production activities abroad. This phenomenon has confronted domestic workers not only with competition from foreign firms but also from within their own firms. Importantly, offshoring not only affects low-skilled employees of manufacturing firms, which build factories in countries with low labor and production costs, but is a phenomenon that nowadays affects many service-sector employees as well (Head, Mayer, & Ries, 2009; Jensen & Kletzer, 2010). Call center assistance, accounting services, or information technology (IT) support are increasingly provided by individuals located in foreign countries. As a result, many white-collar workers who have traditionally been sheltered from international competition have suddenly become exposed to global competition—a trend that is likely to intensify in the future (Blinder, 2006; Crinò, 2009). The number of workers directly affected by offshoring has indeed grown considerably in recent years, and offshoring has become a contentious and highly politicized issue in the public debate about globalization (Mankiw & Swagel, 2006; Owen, 2017). Some observers query whether offshoring is the “next industrial revolution” (Blinder, 2006, p. 113) threatening “virtually the entire employed middle class” (Luttwak, 1995, p. 7) or suggest that it poses “a greater threat than terrorism” (Roberts, 2014, p. 1), although others are more cautious, suggesting that offshoring “is not the tsunami that many claim” (Drezner, 2004, p. 29).

Surprisingly, we know comparatively little about the political consequences of this development, especially in the non-U.S. context. Several studies show that offshoring affects individuals’ policy preferences (Chase, 2008; Owen & Johnston, 2017; Walter, 2017). How exactly these preferences are translated into politically meaningful actions such as the vote is less clear, however. Existing work relies solely on single-country studies such as Switzerland (Walter, 2010) and the United States (Jensen, Quinn, & Weymouth, 2017; Margalit, 2011; Mughan & Lacy, 2002). Although insightful, these cases are not comparable with the bulk of developed democracies: Switzerland is a consensus democracy, where all large parties are always part of the government. And with its presidential political system, the dominance of two parties, and the very polarized political landscape, electoral politics in the United States exhibits very different dynamics than those in multiparty systems. Moreover, existing research either focuses on the vote for one specific party family or concentrates on the vote for the incumbent party.

In most modern democracies, voters have a choice between a broad range of political parties who pursue very different policy agendas in response to globalization (Burgoon, 2012; Garrett, 1998; Haupt, 2010; Kriesi et al., 2008; Swank, 2002). To understand how globalization affects partisan politics and national policymaking in multiparty democracies, however, it is important to understand how the objective individual-level risks and opportunities translate into voting behavior. Examining whether offshoring is a salient issue for voters’ electoral decisions at all, for which political parties offshoring is likely to matter most, and how it affects the electoral success of populist parties from a comparative perspective therefore improves our understanding of the link between globalization and partisan politics.

This article provides such an analysis. Building on the insight that the effects of offshoring vary significantly among citizens, we argue that offshoring represents a relevant issue for some political parties, whereas its salience for other political parties is low. Because highly skilled individuals tend to benefit from the opportunities of offshoring, they are more likely to support parties that advocate economic openness and international competition, especially liberal and center-right parties. In contrast, low-skilled individuals with easily offshorable jobs are threatened by the globalization of production, and are therefore expected to vote for parties that promise protection and compensation. Offshoring is hence important for those political party families with clear policy positions relevant for individuals exposed to offshoring, that is, left, liberal, and center-right parties. In contrast, the risks associated with offshoring are a much less salient issue for political parties who concentrate more on cultural and ideational issues, such as postmaterial issues in the case of green parties. Moreover, contrary to the widely held belief that globalization losers flock to populist parties across the board, we argue that this depends on the type of a voter’s globalization exposure: Offshoring exposure does not strongly affect the vote for populist right parties, that tend to focus predominantly on another dimension of globalization, especially immigration.

Empirically, this article utilizes cross-national survey data from 18 advanced West European countries over the period from 2002 to 2010 to examine how offshoring affects individual preferences for partisan policy positions and party families. Our results show that exposure to offshoring-induced risks and opportunities has significant effects on electoral behavior: Low-skilled individuals working in offshorable occupations are more likely to vote for compensatory policies put forward by left parties than low-skilled individuals working in sheltered occupations. In contrast, individuals in offshorable occupations are more likely to vote for parties advocating economic competition and openness, especially liberal and center-right parties, if they are highly skilled. At the same time, offshoring risks and opportunities play a minor role in the electoral support of populist right and green parties. Our findings thus support the notion that the effects of globalization on partisan politics are heterogeneous, affecting some political parties and party families more strongly than others. This implies that some political parties are more pressured than others in trying to reconcile their constituents’ policy demands with the demands of special interests and global competitive pressures emanating from general trends of globalization.

Offshoring and the Vote

How does offshoring affect partisan politics? We focus on electoral politics as a particularly salient arena of party competition and examine how offshoring affects individuals’ voting behavior. Building on the insight that the heterogeneous individual-level effects of offshoring create both winners and losers, we discuss how these distributional consequences influence policy preferences and, in turn, voting behavior. In a final step, we discuss how the effect of offshoring on vote choice differs among party families.

Risks and Opportunities of Offshoring

The jobs most at risk from offshoring in developed countries are routine jobs that can easily be provided from anywhere in the world. But even nonroutine jobs in the service sector that do not require face-to-face interactions are nowadays more and more likely to be moved abroad (Acemoglu & Autor, 2011). As the scope of offshoring has grown, public commentary has increasingly focused on the downside risks of offshoring. In line with this concern, existing scientific studies tend to assume that offshoring has a uniform negative effect on all workers in offshorable occupations (Mansfield & Mutz, 2013; Margalit, 2011; Owen, 2017; Scheve & Slaughter, 2004).

It seems intuitive that offshoring poses a substantial threat to workers whose job tasks can theoretically be performed abroad. Yet, offshoring not only carries risks but also brings opportunities (Feenstra & Hanson, 1996). Some jobs are offshored, meaning that their tasks are now performed abroad and that these jobs are lost to domestic workers. At the same time, other jobs are onshored, meaning that domestic workers perform tasks for use abroad. Although the phenomenon of offshoring is often thought of in terms of the migration of jobs from rich to poor countries, individuals in rich countries thus often also provide services for firms located in other rich countries or even poor countries. For example, some firms with headquarters in less developed economies have built up research centers in advanced economies and hired local engineers.

This suggests that offshoring has considerable distributive effects that vary across different groups of workers. A considerable number of people working in jobs most likely to be offshored face increasing difficulties of finding a new job in the same occupation, and their wages are likely to be depressed the more widespread offshoring becomes (Feenstra & Hanson, 1999; Hummels, Jørgensen, Munch, & Xiang, 2014). Typically, these are low-skilled workers, who perform routine tasks, which can be more cheaply provided from abroad (Owen & Johnston, 2017). Low-skilled workers in offshorable occupations may even experience downward pressure on their wages when their jobs are not actually offshored, because offshoring increases the supply of workers in their occupation (Grossman & Rossi-Hansberg, 2008). Not surprisingly, workers exposed to the negative risks of offshoring have been found to report higher levels of labor-market risk (Scheve & Slaughter, 2004; Walter, 2017).¹

But offshoring (in this case, better understood as “onshoring”) also offers opportunities for those workers who sell their services to foreign customers. These are usually high-skilled workers who perform nonroutine tasks.² As technological change and deregulation increase opportunities for offshoring, individuals providing such services can sell their skills to a wider set of customers worldwide. Research also shows that firms using high-skilled labor in tradable goods and services industries create new jobs and pay higher wages (Bernard, Jensen, & Schott, 2006). This improves job security and wages for individuals who possess skills that are competitive internationally. The benefits of offshoring thus predominantly accrue to well-educated individuals. Not surprisingly, offshoring has been found to increase the wages of high-skilled individuals (Hummels et al., 2014) and to be associated with higher levels of labor-market security among highly skilled individuals in offshorable occupations (Walter, 2010, 2017).

This discussion should not obscure an important point: Despite the accelerating offshoring trend, a majority of workers remain unaffected by this new form of global competition (Dancygier & Walter, 2015). In fact, many jobs simply cannot be offshored, because the services they provide require them to be on-site (Blinder, 2009). Even though individuals working in these occupations may be exposed to offshoring indirectly as consumers—for example, when calling a call center located in a foreign country—they are barely affected as labor-market participants.³ Workers employed in nonoffshorable occupations are therefore much more sheltered from the globalization of production than workers in occupations that provide more impersonal services or general manufactured goods, and this applies to both high-skilled and low-skilled workers in sheltered occupations.

Overall, this suggests that the effects of offshoring vary by an individual’s skill level (Owen & Johnston, 2017; Walter, 2010, 2017). Offshoring creates the highest labor-market risks for low-skilled individuals working in offshorable occupations (e.g., assembly-line workers). Equally low-skilled individuals working in sheltered occupations (e.g., cleaning personnel) are better off than their counterparts in offshorable occupations, although they continue to experience higher labor-market risks than high-skilled workers in sheltered occupations (e.g., doctors or teachers). Finally, highly skilled individuals in offshorable positions (e.g., engineers or consultants) are the main beneficiaries of the globalization of production. Labor-market risks are thus much more unequally distributed among workers exposed to offshoring than among workers in sheltered occupations.⁴

Offshoring and Party Preferences

How do these individual consequences of offshoring translate into voting behavior? Much research has shown that individuals support or oppose policies based on the material consequences of these policies (e.g., Rehm, 2009; Scheve & Slaughter, 2001; Wren & Rehm, 2013). With regard to offshoring, this suggests that low-skilled individuals with offshorable jobs should have a strong preference for protection from offshoring or, more indirectly, protection from these risks through a generous welfare state. In contrast, highly skilled individuals, who benefit from offshoring, have a lower need for a state-funded social safety net and are also among the main contributors to the financing of the welfare state. Individuals sheltered from offshoring should have more moderate policy preferences than their more exposed counterparts, with low-skilled individuals demanding more protection than high-skilled individuals.⁵ Existing studies on the effects of offshoring on policy preferences support these conjectures. For example, individuals working in routine jobs are significantly more supportive of protectionism when they work in offshorable occupations than those with nonroutine jobs (Owen & Johnston, 2017). Individuals in jobs most likely to be onshored are also much more skeptical of income redistribution than low-skilled individuals in offshorable jobs (Walter, 2010, 2017; Wren & Rehm, 2013).⁶

However, policy preferences can only have an actual impact on the policymaking process if they are effectively brought into the political arena. In democratic countries, the most straightforward instrument for individuals is their vote for a political party that champions the preferred policy in the political process.

Political parties differ with regard to both the policies they advocate and the saliency they put on different policy areas. This is particularly true for political parties in multiparty systems, where parties occupy a large range of positions in the political space usually demarcated by the traditional left-right (or economic) dimension and a cultural dimension (Benoit & Laver, 2006; Kitschelt, 1994; Marks, Hooghe, Nelson, & Edwards, 2006). Offshoring has clearly identifiable distributive effects, which predominantly affect preferences for policies located on the economic dimension—social and labor-market policies, protectionist and market-liberalizing policies, and fiscal policies. Exposure to offshoring should therefore primarily affect individuals’ party preferences for political parties with a distinct and salient position on these specific policies and the economic left-right dimension more generally.⁷

Parties located at the leftist end of the economic dimension are particularly attractive to offshoring losers—low-skilled individuals exposed to offshoring—because these parties typically pursue policies that strengthen the welfare state, redistribute income from the rich to the poor, and other policies that protect vulnerable workers from labor-market risks (Allan & Scruggs, 2004; Schmidt, 2010). Hence, we expect low-skilled individuals in offshorable occupations to exhibit a higher propensity to vote for left parties that advocate generous welfare policies. In contrast, the economic and fiscal policies pursued by left parties tend to conflict with the material interest of individuals benefitting from offshoring. Higher taxes and income redistribution are not only directly paid for by the high-income earners in internationally competitive jobs (Wren & Rehm, 2013), but these policies may also hamper international competitiveness and hence reduce the economic prospects of these individuals (Alesina & Perotti, 1997). As a result, the winners of the offshoring trend are less likely to vote for left parties. Notably, we expect significant differences not only in the voting behavior of low- and high-skilled workers but also within each of these groups. Although low-skilled voters are more likely to vote for left parties than high-skilled workers more generally, this difference should be more pronounced among those working in exposed occupations, because the need for protection is particularly high among low-skilled exposed workers, and the likelihood of paying into a redistributive scheme is particularly high among high-skilled exposed workers. Moreover, while left parties tend to push compensating the losers of the globalization of production, they also usually embrace the globalization of labor. Among left parties, we should thus see a clear difference between the voting preferences of the losers from offshoring (low-skilled individuals in offshorable jobs) and the losers from immigration who typically work in sheltered occupations (Dancygier & Walter, 2015). Low-skilled (high-skilled) individuals in offshorable jobs should therefore be significantly more (less) likely to vote for parties advocating leftist policies than low-skilled (high-skilled) individuals whose jobs cannot be offshored.

In contrast, parties located at the rightist end of the state-market dimension should be particularly attractive for high-skilled individuals in offshorable occupations and least attractive for low-skilled workers exposed to offshoring risks. As voters with high incomes and low labor-market risks, offshoring beneficiaries are not only net payers into the welfare system but also benefit from market-oriented policies, low levels of government spending, and lower taxes. As a result, they are more likely to vote for parties that pursue market-liberal policies, especially liberal and center-right parties. Likewise, these policies run directly counter to the policy preferences of offshoring losers, who should therefore be least supportive of these parties. Although the parties in these categories differ with regard to their position on other policy dimensions (Benoit & Laver, 2006; Schmidt, 2010), they share common grounds with regard to economic and social policies. Liberal parties are skeptical vis-à-vis state intervention in the economy, actively advocate free market policies, including a further opening of the economy, and promote lower levels of taxation and a less generous provision of social rights (Allan & Scruggs, 2004; Benoit & Laver, 2006; Zohlnhöfer, Obinger, & Wolf, 2008). Center-right parties tend to be located somewhat more to the center of the left-right dimension but also promote free market policies, although some, especially Christian democratic parties, favor embedding these policies in a resilient welfare-state system (van Kersbergen, 1995). Both liberal and center-right parties therefore carry a strong appeal to high-skilled individuals in offshorable occupations, whereas they are least attractive for low-skilled voters, especially those in highly offshorable occupations. Again, we expect the effect of skills on voting behavior to be weaker among sheltered individuals than among exposed individuals.

In contrast to party families who clearly position themselves on the state-market dimension and for whom social and economic policies are particularly salient, we do not expect offshoring to be an important issue for political parties for whom the cultural dimension of party competition carries greater importance or whose economic policies do not specifically benefit either the winners or the losers of the globalization of production. The two most important party families in this regard are populist right parties and green parties.

Even though previous studies and the popular press have argued that right-wing populist parties are particularly appealing to modernization and globalization losers (Betz, 1993; Kriesi et al., 2008; for a review, see Bornschier, 2018), the effect of offshorability on the propensity to vote for these parties is theoretically ambiguous. For one, these parties focus mainly on limiting immigration. This is an important topic on the cultural dimension making offshorability-related risks less of a salient issue. Immigration is of course not just a cultural issue but one with strong economic consequences as well, creating risks especially for low-skilled workers who are most likely to compete with immigrant labor. Low-skilled workers are therefore most opposed to immigration (Mayda, 2006), irrespective of whether they work in offshorable occupations or not (Dancygier & Walter, 2015). In terms of more classic economic policies, these parties tend to deemphasize the importance of this dimension and to deliberately blur their positions (Rovny, 2013). Interestingly, although they often advocate “welfare chauvinist” policies that limit social protection to nationals only, the overall economic policy position of radical and populist right parties tends to be on the market-liberal side of the political spectrum (Kitschelt, 2007), with some of these parties explicitly supporting free trade (de Lange, 2007).

For all these reasons, we contend that populist right parties are not particularly attractive to offshoring losers but appeal to low-skilled workers across the board. These are either threatened by offshoring and free trade when they work in economically exposed occupations or by labor-market competition through low-skilled immigrants when they work in sheltered occupations (Burgoon, 2012; Dancygier & Walter, 2015). This implies that low-skilled individuals should in general be more likely to vote for populist right parties but that there should be no reinforcing effect of offshoring exposure. High-skilled voters are expected to be much less likely to vote for the populist right across the board for both material reasons, because they tend to benefit from cheap labor, and immaterial reasons, because higher levels of education are associated with lower levels of xenophobia (Hainmueller & Hiscox, 2006).

Green parties, in contrast, share an emphasis on environmental protection and other postmaterialistic issues, and are both less homogeneous concerning questions about welfare-state expansion and free market policies (Benoit & Laver, 2006). With regard to globalization, the issue that primarily defines the Greens’ position is “the cultural aspect of globalization processes” (Dolezal, 2010, p. 548). As “the” postmaterialist party family, we therefore expect the material interests of individuals affected by offshoring to play a negligible role in explaining party preferences for green parties. Rather, we expect all high-skilled individuals to be more likely to vote for these parties than low-skilled individuals.

Table 1 summarizes our expectations about the effect of offshoring on individual party preferences. As discussed, this effect should be strongest for political parties with distinct and polar policy positions on the social-economic dimension of party competition; that is, parties with clear positions regarding welfare and market-liberal policies. In addition, because the salience political parties attach to economic issues varies, the effect of offshoring on voting behavior should vary among party families: Offshoring exposure should matter more for left, liberal, and center-right parties, who serve as natural agents for the losers and winners of off- and onshoring. Among high-skilled individuals, job offshorability should increase the propensity to vote for parties located at the market-liberal end of the left-right divide in partisan politics, whereas this effect should be reversed among low-skilled individuals. These expectations also suggest that the difference in voting propensity should be significantly larger between low- and high-skilled individuals in offshorable occupations than between low- and high-skilled individuals in sheltered occupations. At the same time, we expect that offshoring should not be associated with individuals’ propensities to vote for political parties that politicize more on the cultural dimension of party competition, especially populist right and green parties.

Table 1.

Expected Effect of Offshorability on Party Preferences.

	Low-skilled individuals	High-skilled individuals
Left-right position	−	+
Welfare-state policies	+	−
Market-liberal policies	−	+
Left parties	+	−
Liberal parties	−	+
Center-right parties	−	+
Populist right parties	0	0
Green parties	0	0

Research Design

We use survey data from five consecutive waves of the European Social Survey (ESS) conducted between 2002 and 2010 in 18 Western European countries to test the conditional effect of exposure to job offshorability on partisan preferences.⁸ This set of countries is especially useful because it represents developed capitalist democracies with established multiparty systems allowing us to test our argument about differentiated partisan effects. We focus on working-age respondents, because globalization-induced labor-market risks should be most important for this section of the population.⁹

Dependent Variables: Preference for Policy Position and Party Family

To examine how offshoring affects voting behavior, we proceed in two steps: We first concentrate on the propensity to vote for parties advocating specific policies, focusing on party’s overall left-right position on the economic dimension, positions regarding welfare state, and positioning on market-liberal policies. In a second step, we examine how exposure to offshoring influences individuals’ likelihood to vote for a specific party family, which continue to be the most relevant element in party competition and have converged ideologically across Europe in the last few years (Camia & Caramani, 2012).¹⁰

Our first set of dependent variables focuses on political parties’ economic policy positions, measured with data collected by the Comparative Manifesto Project (CMP; Volkens, Lehmann, Merz, Regel, & Werner, 2013). The CMP codes the direction and quantity of policy statements from electoral programs of all parties participating in a national election in a given year. Parties’ positions on the economic dimension are captured with the CMP score for the party’s overall orientation to the left or right, with higher values indicating a more rightist position. Party positions on welfare-state policies are measured with the respective CMP indicator, and contain partisan positions regarding social justice (statements about social equality or the need for a fair distribution of resources) and welfare-state expansion (mentions of the need to maintain or expand social security schemes). Finally, we measure party positions regarding market-liberal policies with the respective CMP indicator, which builds on statements about free enterprise capitalism (superiority of the individual enterprise over the state or favorable mentions to protect property rights) and economic orthodoxy (reduction of budget deficits or retrenchment in crises). Higher values indicate a stronger approval of the respective policies.

To operationalize the second dependent variable, we classify national parties into cross-nationally comparable party families based on two data sources: the dataset about the composition of governments in Organization for Economic Cooperation and Development (OECD) countries by Schmidt (2012) and CMP (Volkens et al., 2013). We focus on the five party families that have been most common in Europe, have converged regarding their policy positions, and have increasingly homogeneous voting distributions: left, liberal, center-right, populist right, and green parties.¹¹ To classify individual parties into these party families, we proceed as follows: First, we identify the party family separately on the basis of each database. We then merge these classifications in accordance with the following rules: If both databases report the same party family for a single party, we classify the latter accordingly. If one database codes a party as a member of one of the five party families and the other database codes it as a residual party or provides no information, we classify the party in line with the information-providing database. If both databases provide no information, we code the party family as missing. If the two databases disagree about the specific party family, we gather more data (e.g., the membership of a party in a political group in the European Parliament) to classify this party accurately.¹²

We match the information about each party’s policy positions and family to ESS respondents based on which national party they voted for in the last national election. For robustness, we additionally use information about respondents’ current closeness to a political party. Each individual is thus assigned his or her preferred party’s policy position and classified as voting for one of the five party families. Whereas party policy positions are continuous measures, we create five dummy variables recording whether a respondent voted for or feels close to each party family.

Independent Variables: Exposure to Offshoring and Skill Level

Our argument suggests that offshoring affects individual party preferences, but that this effect differs between low-skilled and high-skilled individuals. These considerations suggest three independent variables: exposure to offshoring, skill level, and an interaction term to address the conditional effect.

Exposure to offshoring

Jobs differ with regard to the degree to which they can be offshored. To measure respondents’ occupational offshorability, we match the information about respondents’ occupation contained in the ESS survey with information from an offshorability index developed by Blinder (2009). This index measures whether the service the job provides can theoretically be delivered over long distances with little or no degradation in quality, for more than 800 occupational categories.¹³ It allows us to assess individual exposure to offshoring on an occupational basis. Because it measures the potential for offshoring, it connects closely with our theoretical argument about offshoring risk. Moreover, in a validation study, Smith and Rivkin (2008) found Blinder’s classification to be highly correlated with a more intuitive coding of offshorability by business school students.¹⁴

Although Blinder distinguishes between different categories of offshorability, we use a dummy variable that distinguishes only between jobs that are potentially offshorable and jobs that cannot be offshored, because this distinction is less sensitive to country-specific contexts. We code all jobs that do not require workers to be at a specific work location in their country as potentially offshorable, taking the value of 1. This includes both workers who have to be physically close to their work unit and workers whose entire work unit may be moved to another country, or whose domestic presence is not required for other reasons. All other occupations, including those not listed by Blinder, are coded as not offshorable.¹⁵

Skill level

We operationalize an individual’s skill level based on his or her educational background, measured as total number of years a respondent has been in full-time education. Of course, individuals can also dispose of skills acquired through on-the-job training, and individuals with low levels of education can also deliver high-quality work, but empirical research has shown that higher educational achievement is positively related to higher occupational skills and higher levels of productivity (Jones, 2001; Spitz-Oener, 2006). Education years therefore serve as a proxy for individual skill levels.¹⁶ As a robustness check, we additionally use information on the highest level of education a respondent has achieved. The answers are standardized into the International Standard Classification of Education (ISCED) classification of education levels.¹⁷

Interaction between offshorability and skill level

Our argument suggests that the effect of offshorability on individual voting behavior depends on voters’ levels of education. Figure 1 shows the distribution of education years for respondents in nonoffshorable and offshorable occupations. It reveals that education is similarly distributed in both groups. This underscores our argument that both high- and low-skilled individuals can be exposed to offshoring.

Figure 1.

Distribution of education years by offshorability.

To capture the expected conditional effect of exposure to offshoring and an individual’s skill level on partisan preferences, we use an interaction term. Our argument makes clear predictions about the nature of this interaction term. As offshoring creates more labor-market risks for low-skilled individuals, this group of voters should be particularly likely to vote for parties that champion leftist and pro-welfare policies, and less likely to vote for those with a market-liberal policy profile. In contrast, highly skilled, exposed individuals should prefer parties on the right of the economic policy dimension, who favor market-friendly policies. This suggests a positive and statistically significant interaction term for the analyses of voting for the left-right policy dimension, parties’ stances on the market economy, liberal and center-right parties. In contrast, we expect a negative and statistically significant interaction term for welfare policy positions and voting for left parties. Finally, we do not expect a statistically significant interaction term for populist right or green parties.¹⁸

Control Variables

We consider a number of variables that control for alternative explanations of individual voting behavior. Following our theoretical argument, the selection of observable confounders bears on a risk-based model of voting behavior (Hellwig, 2008; Mughan, Bean, & McAllister, 2003; Mughan & Lacy, 2002). We include respondent’s income, gender, age, whether he or she is unemployed, whether he or she lives in an urban area, and cultural attitudes toward immigration in our preferred specification. Respondents’ income is measured by a self-classification into one of 12 income classes. To provide cross-national comparability, we recode this variable, so that it represents the deviation of the respondent’s income class from the country-specific median income class. We include age in years and a squared age term. Cultural attitudes toward immigration are measured on an 11-point scale, where higher values indicate that “cultural life is undermined” by people from other countries.¹⁹ The remaining variables are coded as dummy variables. Furthermore, we estimate an enhanced specification that controls for additional variables related to labor-market risk, such as outsider status (Emmenegger, Häusermann, Palier, & Seelaib-Kaiser, 2012; Rueda, 2005), skill specificity (Iversen & Soskice, 2001),²⁰ or routinization (Acemoglu & Autor, 2011). In additional robustness checks, we also control for self-employment, labor union membership, political interest, church attendance, economic attitudes toward immigration, ideology using self-placement on a 10-point left-right scale, employment in the public sector, sectoral exposure to international trade,²¹ and a proxy for risk aversion.²²

On the macro-level, we control for different national contexts in which respondents take their voting decision. The unemployment rate is a proxy for the state of the economy. The stock of foreign direct investment (FDI) captures the country’s overall exposure to the globalization of production. Both variables proxy the general level of labor-market risks (Arzheimer, 2009; Hellwig & Samuels, 2007; Kayser, 2007). In addition, we include the effective number of electoral parties to account for the fact that vote shares vary with the number of parties competing in an election (Bormann & Golder, 2013). Results are also robust to controlling for the country’s trade openness and the level of social expenditure.

Method

We perform our analyses on a dataset containing approximately 53,000 respondents in 18 countries at 5 points in time. Our preferred model specification is a multilevel model, where individuals (level 1) are nested within countries (level 2). This model allows us to account for the fact that respondents from the same country share a common context and are, thus, not necessarily independent from each other (Rabe-Hesketh & Skrondal, 2008; Steenbergen & Jones, 2002). To control for temporal variation, we include dummies for survey waves in all model specifications.²³ To analyze party position preferences, we rely on fixed effects ordinary least squares (OLS) specifications. Concerning preferences for party families, we employ random effects probit specifications. The disadvantage in modeling party preferences separately for all party families is that it does not allow to model simultaneous choice. As a robustness check, we therefore also use a multinomial logit model with country dummies (Long & Freese, 2006).

Empirical Findings

Does offshoring affect partisan politics through individuals’ electoral preferences? As we will show in detail below, our analyses of the effects of offshoring on voters’ preferences for partisan policy positions and party families indicate that job offshorability is indeed associated with variation in the voting behavior of individuals, and that this effect is conditional on skill levels. As predicted by our argument, voters take their offshoring-related material interests into account when making electoral choices: Offshoring losers vote for different political parties than offshoring winners. Importantly, this is only the case for party families that strongly advocate economic and social policies targeted toward compensating the losers or benefitting the winners of offshoring: left, liberal, and center-right parties. In contrast, voting for populist right and green parties does not differ among individuals in offshorable and sheltered occupations. Offshoring thus does not affect all parties equally.

Offshoring and Preferences for Partisan Policy Positions

How does exposure to offshoring affect individuals’ preference for specific party positions? Table 2 presents regression results for our analysis of individual preferences for parties’ general left-right position, their position regarding welfare-state policies, and their position regarding market-liberal policies.

Table 2.

Determinants of Individual Preferences for Parties’ Policy Positions.

	Overall left-right scale	Welfare-state policies	Market-liberal policies
Education years	−0.104***	−0.005	−0.029***
Education years	(0.03)	(0.01)	(0.01)
Offshorability	−1.608***	0.428**	−0.273**
Offshorability	(0.57)	(0.19)	(0.12)
Education × Offshorability	0.181***	−0.045***	0.035***
Education × Offshorability	(0.04)	(0.01)	(0.01)
Income	0.509***	−0.159***	0.085***
Income	(0.04)	(0.01)	(0.01)
Female	−1.710***	0.401***	−0.310***
Female	(0.15)	(0.05)	(0.03)
Age in years	−0.235***	0.074***	−0.039***
Age in years	(0.04)	(0.01)	(0.01)
Age squared	0.002***	−0.001***	0.000***
Age squared	(0.00)	(0.00)	(0.00)
Unemployed	−1.509***	0.436***	−0.273***
Unemployed	(0.33)	(0.11)	(0.07)
Urban resident	−1.430***	0.344***	−0.202***
Urban resident	(0.16)	(0.06)	(0.04)
Antiimmigration (culture)	1.308***	−0.262***	0.204***
Antiimmigration (culture)	(0.03)	(0.01)	(0.01)
Unemployment rate	−0.043	−0.359***	0.111***
Unemployment rate	(0.06)	(0.02)	(0.01)
FDI stock	0.070***	−0.013***	0.004***
FDI stock	(0.00)	(0.00)	(0.00)
Effective number of parties	0.477**	−1.717***	0.865***
Effective number of parties	(0.20)	(0.07)	(0.04)
No. of respondents	46,075	46,075	46,075
No. of countries	18	18	18
ICC (rho)	.237	.514	.211
Panel SD (sigma)	8.868	5.528	1.777
R² (overall)	.067	.001	.017
Prob > χ²	.000	.000	.000

Multilevel OLS estimates, standard errors in parentheses. FDI = foreign direct investment; OLS = ordinary least squares; ICC = intraclass correlation.

Level of statistical significance: *p ≤ .10. **p ≤ .05. ***p ≤ .01.

As predicted by our argument, low-skilled individuals working in offshorable occupations are less likely to vote for parties to the right of the political spectrum than low-skilled individuals in sheltered occupations (column 1). The positive and statistically significant interaction term between education years and offshorability, as illustrated in Figure 2A, indicates, however, that this relationship changes with higher levels of education. Among individuals who have received at least 11 years of schooling, those in offshorable jobs are more inclined to vote for more rightist parties at a statistically significant level, and this effect further increases the more education a voter has received.

Figure 2.

Conditional effect of offshorability on voting for policy positions.

Turning to more specific partisan positions on welfare-state expansion and market-liberal policies, we find the same pattern. Exposure to offshoring increases individuals’ probability to vote for parties advocating a strong welfare state when they are poorly educated, but decreases this probability when they are highly educated (column 2). In contrast, among the high-skilled, those working in offshorable occupations are significantly more likely to vote for parties with market-liberal policy positions than high-skilled workers sheltered from global competition, whereas exposure to offshoring significantly reduces this likelihood among the low skilled (column 3). In both cases, education reverses the relationship between offshorability and preferences for welfare-state and market-liberal policies, respectively (Figures 2B and 2C).

Interestingly, the policy preferences of individuals in sheltered occupations appear to be at odds with the conventional wisdom of traditional partisan models. Among individuals in sheltered occupations, higher levels of education are associated with partisan preferences for more leftist parties and less market liberal policies, and are not related to parties’ stances on welfare-state policies. This finding might reflect the fact that high-skilled individuals sheltered from offshoring are those who provide many of the services an advanced welfare state offers and echoes the argument that many left parties opened up for new, left-libertarian voter groups in the late 20th century (Kitschelt, 1988).

The results for the control variables are in line with our expectations. On the micro-level, we find that poorer, female, and older respondents, those living in urban areas, and the unemployed are more likely to vote for welfare-state supporting parties and less likely to vote for parties advocating market-liberal policies. Immigration skeptics are less inclined to prefer parties explicitly proposing welfare-state extension. On the macro-level, support for rightist and market-liberal parties tends to be higher in countries with higher unemployment rates, higher levels of FDI, and a higher number of electoral parties.

Our results are robust to a variety of modifications, all documented in the online appendix. One objection to our analyses is that offshorability is highly correlated with other forms of labor-market risk. The most prominent candidate here is routinization, because workers with routine jobs are most likely to lose their jobs in a deindustrializing world, and routine jobs also tend to be the ones that can most easily be offshored. Including routinization as a control variable does not change our results for offshorability, however. Furthermore, skill specificity of an individual’s occupation and whether she is a labor-market outsider also do not alter the effect of offshorability (see Table A1). Another objection is that the effect of offshoring might capture unobserved factors that underlie respondents’ occupational choices in a way that sorts them into offshorable and non-offshorable jobs. An obvious candidate here is risk aversion: Offshorable jobs are much more likely to be private-sector jobs that offer not only large rewards but also large risks. It is thus possible that more risk-averse voters choose safer jobs (say, as a teacher or nurse) than more risk-taking individuals. Unfortunately, the ESS surveys do not contain any questions that allow us to directly assess voters’ levels of risk aversion. We therefore rely on two proxies: The first one is a dummy for public-sector employment, which tends to be less risky than private-sector employment. The second proxy is respondents’ answers to the question whether they generally plan for the future or take each day as it comes, assuming that those who plan for the future are more risk averse than those who live by the day. Unfortunately, this question is only included in the 2006 survey, so that our sample is substantially reduced. Moreover, we control for the respondents’ ideological self-placement on the left-right scale. This is not our preferred specification because we believe that the ideological self-placement captures part of what we want to explain in the first place. The self-placement on the general left-right dimension should, however, be correlated with potential unobserved confounders that might drive our effects of offshorability. Our results are unchanged when we include these proxies, increasing our confidence that offshorability in and of itself does affect voters’ party choices. In addition, our results are robust to restricting the sample to active labor-market participants, expanding the sample to all retired and non-retired respondents, or including Eastern European countries.²⁴ Using alternative coding of skills and offshorability and including more micro- and macrolevel controls similarly do not change the conclusions we draw with regard to the conditional effect of offshoring.

Summing up, our results show that the gap in partisan preferences between low- and high-skilled individuals is larger among those working in offshorable occupations than among those individuals sheltered from the offshoring trend. Moreover, this gap opens up in the expected directions: Among the high skilled, those benefitting from offshoring show a stronger preference for parties with neoliberal policy positions than those in sheltered occupations, whereas those most threatened by offshoring (low-skilled workers in offshorable jobs) most strongly prefer parties advocating a generous welfare state.

Offshoring and Preferences for Party Families

In the next step, we turn to individuals’ voting preferences for specific party families. We focus on the five most common party families and expect that offshoring affects voters’ behavior for parties with a clear and vocal position on the economic dimension (left, liberal, and center-right parties), but to play a negligible role in explaining voters’ propensities to vote for parties that predominantly focus on non-economic issues (populist right and green parties). Table 3 presents the results of five multilevel probit regressions.

Table 3.

Determinants of Individual Preferences for Party Families.

	Left	Liberal	Center-right	Populist right	Green
Education years	−0.017***	0.019***	0.004*	−0.049***	0.049***
Education years	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Offshorability	0.159***	−0.052	−0.051	−0.018	−0.013
Offshorability	(0.04)	(0.07)	(0.05)	(0.08)	(0.08)
Education × Offshorability	−0.015***	0.009**	0.008**	−0.002	−0.003
Education × Offshorability	(0.00)	(0.00)	(0.00)	(0.01)	(0.01)
Income	−0.033***	0.040***	0.042***	−0.023***	−0.045***
Income	(0.00)	(0.00)	(0.00)	(0.01)	(0.00)
Female	0.054***	−0.053***	−0.030**	−0.214***	0.216***
Female	(0.01)	(0.02)	(0.01)	(0.02)	(0.02)
Age in years	0.029***	−0.016***	−0.019***	−0.004	0.008
Age in years	(0.00)	(0.00)	(0.00)	(0.01)	(0.01)
Age squared	−0.000***	0.000***	0.000***	−0.000	−0.000***
Age squared	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Unemployed	0.138***	−0.131***	−0.151***	0.132***	0.018
Unemployed	(0.03)	(0.04)	(0.03)	(0.05)	(0.04)
Urban resident	0.102***	−0.014	−0.187***	−0.017	0.221***
Urban resident	(0.01)	(0.02)	(0.01)	(0.02)	(0.02)
Antiimmigration (culture)	−0.060***	−0.004	0.048***	0.175***	−0.127***
Antiimmigration (culture)	(0.00)	(0.00)	(0.00)	(0.00)	(0.01)
Unemployment rate	0.021***	−0.056***	−0.010**	−0.011	−0.002
Unemployment rate	(0.00)	(0.01)	(0.00)	(0.02)	(0.01)
FDI stock	0.001***	−0.002***	0.001***	−0.001***	−0.001**
FDI stock	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Effective number of parties	−0.102***	−0.012	0.014	0.069***	0.107***
Effective number of parties	(0.02)	(0.02)	(0.02)	(0.02)	(0.03)
No. of respondents	52,629	52,629	52,629	52,629	52,629
No. of countries	18	18	18	18	18
ICC (rho)	.085	.629	.148	.534	.676
Panel SD (sigma)	0.305	1.303	0.416	1.069	1.444
R² (McKelvey/Zavoina)	.063	.023	.045	.123	.083
Prob > χ²	.000	.000	.000	.000	.000
BIC	66,685.14	31,539.77	61,652.81	17,296.23	22,420.72
Log-likelihood	−33,239.30	−15,666.61	−30,723.13	−8,544.84	−11,107.08

Multilevel probit estimates, standard errors in parentheses. FDI = foreign direct investment; BIC = Bayesian information criterion; ICC = intraclass correlation.

Level of statistical significance: *p ≤ .10. **p ≤ .05. ***p ≤ .01.

As expected, working in a potentially offshorable job significantly increases voters’ tendencies to vote for a leftist party and decreases the likelihood of voting for a liberal or center-right party among the low-skilled, although this effect is not statistically significant. The sizable and statistically significant interaction terms between job offshorability and education years for all of these parties demonstrate, however, that exposure to offshoring affects voting behavior, and does so especially among the high-skilled. Figure 3 plots the marginal effects of offshorability on party preferences at different skill levels.²⁵ In case of left parties (Figure 3A), working in a potentially offshorable occupation significantly increases the likelihood of a vote for everyone who enjoyed less than 8 years of full-time education. In contrast, respondents with 12 years of education and more are significantly less likely to vote for a left party when they work in an offshorable occupation. This shows that low-skilled individuals exposed to offshoring risks are particularly likely to vote for the traditional advocates of welfare-state expansion and redistribution, whereas high-skilled individuals in offshorable jobs are least likely to vote for these parties.

Figure 3.

Conditional effect of offshorability on voting for party families.

Our argument suggests that the latter group should instead vote for liberal or center-right parties, and our results support this claim. Figure 3B shows that offshorability significantly increases the propensity to vote for a liberal party for all individuals with at least 11 years of education. Somewhat unexpectedly, offshorability does not have a direct effect on liberal party preferences among the low skilled, possibly reflecting the fact that these parties are unattractive to less privileged voters in general. We find similar effects for center-right parties. Job offshorability has a negative, though insignificant effect on the voting propensity of low-skilled individuals for centrist parties (Figure 3C). As for the liberal parties, offshoring has a statistically significant effect on the voting behavior of those individuals with at least 11 years of education. Among this group, offshoring increases the propensity to vote for center-right parties, and this effect gets larger the more years of full-time education an individual has received.

The main prediction of our argument is that exposure to offshoring increases the vote gap between low- and high-skilled individuals for these three party families. To analyze the change in vote gaps, we calculate the first difference in predicted probabilities between low- and high-skilled respondents for those in occupations sheltered from and those in occupations exposed to offshoring, holding all other variables at their median.²⁶ We then use this information to calculate the percentage increase (or decrease) in the vote gap between these groups. Table 4 presents the results. For left parties, the difference in predicted probabilities between low- and high-skilled respondents working in nonoffshorable occupations is 7.75 percentage points, but is 14.42 percentage points if voters are exposed to offshoring. Exposure to offshoring thus almost doubles the vote gap between high- and low-skilled voters, and this difference in vote gaps is statistically significant. We observe similar effects for liberal and center-right parties. Concerning the former, the vote gap between high- and low-skilled respondents is 1.20 percentage points among those in sheltered occupations and 2.09 percentage points among respondents in exposed occupations, a sizable 74% increase. The same is true concerning center-right parties. Among those in sheltered jobs, high-skilled voters are more likely to vote for the conservatives than low-skilled voters, leading to a difference in voting propensities of 1.79 percentage points. Among those in offshorable jobs, the vote gap between high-skilled and low-skilled voters more than triples amounting to a difference in voting probabilities of 5.41 percentage points. Taken together, this suggests that job offshorability amplifies the difference in voting probabilities for center-right parties between low- and high-skilled individuals by 3.63 percentage points. Given that a multitude of factors influence individual voting behavior and people do not switch votes often, the effect of offshoring is thus comparatively large.

Table 4.

Substantial Effect of Offshorability.

	(I)Difference in voting probability between low- and high-skilled respondents in nonoffshorable jobs	(II)Difference in voting probability between low- and high-skilled individuals in offshorable jobs	(III) = (II – I)Change in voting probability between low- and high-skilled individuals due to offshorability	(IV)Change in voting probability between low- and high-skilled individuals due to offshorability in %
Left	7.75	14.42	6.67***	86.12
Liberal	−1.20	−2.09	−.89	74.36
Center-right	−1.79	−5.41	−3.63**	203.06
Populist right	1.19	1.11	−.07	6.18
Green	−1.49	−1.20	.29	19.50

Predicted probabilities are based on models reported in Table 3; control variables held at their mean. Low-skilled individuals have 8 (5th percentile), high-skilled individuals 20 (95th percentile) education years. A positive difference implies that low-skilled individuals are more likely to vote for the respective party families.

Level of statistical significance: *p ≤ .10. **p ≤ .05. ***p ≤ .01.

In contrast to the three party families examined so far, we do not expect systematic differences related to offshoring for those parties with a strong focus on noneconomic issue. The results presented in columns 4 and 5 in Table 3 support these expectations. Whereas the educational background has indeed a strong and statistically significantly effect on individuals’ voting preferences for these parties—education is negatively correlated with party preferences for the populist right and positively with those for green parties—exposure to offshoring has no such effect. For both party families, skill level also does not affect the relationship between offshorability and electoral preferences, evidenced by the small and statistically insignificant interaction terms. Figures 3D and 3E further demonstrate that offshorability has no consistent statistically significant conditional effect on voting propensity irrespective of education. As a result, the vote gap between high- and low-skilled individuals is almost identical for individuals working in non-offshorable occupations and those working in offshorable occupations (see Table 4). All in all, our findings show that offshoring does not increase the popularity of populist right and green parties. Rather, green parties attract high-skilled, and populist right parties appeal to low-skilled voters across the board. Especially with regard to the populist right, this suggests that voters’ subjective feelings of being threatened by globalization (de Vries & Hoffmann, 2016) may not necessarily reflect objective risks associated with offshoring.

Concerning the control variables, women and respondents living in urban areas are more likely to vote for the left and for green parties. Left parties and the populist right are particularly likely to attract votes from the unemployed. And liberal and center-right parties are more likely to attract high-income voters, whereas poorer respondents are more likely to vote for the left, the populist right, and the Greens. Unsurprisingly, populist right and center-right parties are attractive to respondents who are sensitive toward immigration on cultural grounds. Higher overall unemployment rates strengthen electoral support for parties of the left but depress the vote for liberal parties. The same holds for a country’s exposure to FDI. Interestingly, FDI seems to dampen the prospects of populist right and green parties.

Again, our results are generally robust to several robustness checks. The offshoring effect remains robust to the inclusion of other sources of labor-market risk (see Table A2) and generally robust to including the risk-aversion proxies public-sector job and planning propensity. Similarly, altering the sample size or including more control variables does not change this picture. Only in case of the liberal party family are the results sensitive to some model specifications. Furthermore, the interaction term loses its statistical significance for both the liberal and center-right parties if we include ideology. Results are also robust to modeling the simultaneous electoral choice among alternative party families using a multilevel multinomial model (see Table 6 in the online appendix).

Overall, these results provide strong support that offshoring has a significant demand-side effect for those political parties with a clear and salient position regarding economic policies of specific relevance with regard to the material effects of offshoring. This effect is particularly sizable for leftist and center-right parties, and prominent but slightly more sensitive to alternative specifications for the liberal parties. Nevertheless, our findings show that offshoring shapes individual voting behavior—but only for some parties. Offshoring does not play an important role in voters’ calculus in case of populist right or green parties, who privilege non-economic issues in their partisan agendas. This reinforces the argument that it is important to take seriously the fact that parties compete in multiparty systems; for some of which, but not for all, the globalization of production is likely to have electoral consequences.

Conclusion

Offshoring has become widespread in developed economies and has turned into an increasingly salient topic in public debates. What are the political consequences of this development? Our article has investigated this question with regard to electoral and partisan politics in multiparty systems. We argue and show that the material consequences of offshoring affect voters’ party preferences. However, this effect is far from being uniform: Not only is exposure to offshoring associated with significant differences in party preferences among high- and low-skilled individuals, offshoring also only matters for some parties, but not for others.

We argue that this variation is explained by two important insights: First, offshoring creates both winners and losers. The individual-level material consequences of offshoring vary significantly among individuals based on their skill level: High-skilled individuals in offshorable jobs benefit from the opportunities of offshoring, but low-skilled individuals with offshorable jobs face increasing labor-market risks. This translates into variation in party preferences. Second, especially in multiparty systems, political parties differ in their policy positions and the salience they attach to them. Some parties pursue policies that are particularly relevant for individuals affected, positively or negatively, by offshoring, whereas other parties emphasize policy fields for which offshoring only plays a minor role.

Taken together, this suggests that offshoring affects the voting behavior of some individuals (those exposed to offshoring) for some parties (those with a strong focus on socio-economic policy issues) but has no effect on others. Our analyses of the determinants of individual electoral preferences for policy positions and party families in 18 European countries confirm this hypothesis. Political parties advocating income redistribution and a strong welfare state (left parties) are particularly attractive to low-skilled individuals working in offshorable occupations. In contrast, parties with a more market-liberal policy profile (liberal and center-right parties) attract the beneficiaries of offshoring, namely high-skilled individuals in offshorable occupations. Finally, as parties emphasizing policies unrelated to the globalization of production, offshoring does not affect the electoral fortunes of populist right and green parties.

Our study speaks to two ongoing debates about the effects of globalization on domestic politics: First, the debate about the influence of globalization on voting behavior in general (for a summary, see Kayser, 2007). Several authors have argued that globalization reduces the importance of economic issues on vote choice (Hellwig & Samuels, 2007; Steiner & Martin, 2012). While this may be true in the aggregate, our results suggest that there is much more nuance in individual voting behavior. Globalization does not affect all voters in a uniform manner, but its consequences vary widely within the electorate. Moreover, by showing that these material consequences matter for voting decisions, our analysis challenges studies claiming that individual voting behavior and policy preferences are influenced mostly by nonmaterial issues rather than voters’ material self-interest related to globalization (Hellwig, 2008; Hellwig & Samuels, 2007; Mansfield & Mutz, 2013).

Second, our article contributes to the debate about the influence of globalization on party competition. A large literature shows that globalization affects partisan politics in developed countries (Ezrow & Hellwig, 2014; Garrett, 1998; Haupt, 2010; Kriesi et al., 2008; Swank, 2002). Nonetheless, researchers have lamented the lack of attention to how the effects of globalization on public opinion affect party competition indirectly (Ward, Ezrow, & Dorussen, 2011) and have emphasized the need for further research on globalization’s impact on political parties, particularly on parties of the center and right (Adams, Haupt, & Stoll, 2009). Our analysis shows that the impact of offshoring varies strongly across party families. Rather than voting in favor or against incumbents when exposed to globalization, voters consciously choose parties that cater toward their needs. Our findings are thus particularly relevant for the majority of countries characterized by multiparty systems. Importantly, our analysis shows that the objective material consequences of offshoring are not a salient issue for one party family that has been frequently characterized as catering to globalization losers—the populist right (Kriesi et al., 2008; Mughan et al., 2003). In contrast, we find that these parties appeal to low-skilled workers in general, irrespective of whether they work in occupations exposed to or sheltered from offshoring. Although voters of these parties report that they feel threatened by globalization (de Vries & Hoffmann, 2016), this suggests that policies limiting the globalization of production may not necessarily alleviate these voters’ problems. Leftist, liberal, and center-right parties advocate policies that specifically benefit the losers and winners from offshoring, although not necessarily those affected by other forms of globalization. For these parties, we observe a distinct electoral effect of offshoring as a specific type of globalization.

Overall, our findings underline the importance of distinguishing between specific types of globalization, their specific individual-level effects, and different types of parties. When this is taken into account, offshoring has clear and identifiable effects on voters’ electoral preferences and on party politics more generally.

Footnotes

Appendix

Table A2.

Determinants of Individual Preferences for Party Families—Robustness.

	Left	Liberal	Center-right	Populist right	Green
Education years	−0.012***	0.014***	−0.001	−0.033***	0.047***
Education years	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Offshorability	0.118**	−0.048	−0.030	0.005	0.030
Offshorability	(0.05)	(0.07)	(0.05)	(0.09)	(0.09)
Education × Offshorability	−0.014***	0.011**	0.007**	−0.008	−0.005
Education × Offshorability	(0.00)	(0.01)	(0.00)	(0.01)	(0.01)
Income	−0.027***	0.032***	0.039***	−0.020***	−0.044***
Income	(0.00)	(0.00)	(0.00)	(0.01)	(0.00)
Female	0.079***	−0.075***	−0.047***	−0.221***	0.212***
Female	(0.01)	(0.02)	(0.01)	(0.02)	(0.02)
Age in years	0.031***	−0.017***	−0.022***	−0.003	0.011*
Age in years	(0.00)	(0.00)	(0.00)	(0.01)	(0.01)
Age squared	−0.000***	0.000***	0.000***	−0.000	−0.000***
Age squared	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Unemployed	0.099***	−0.121**	−0.105***	0.090*	0.012
Unemployed	(0.03)	(0.05)	(0.03)	(0.05)	(0.05)
Urban resident	0.107***	−0.025	−0.191***	0.003	0.225***
Urban resident	(0.01)	(0.02)	(0.01)	(0.03)	(0.02)
Antiimmigration (cultural)	−0.064***	0.000	0.050***	0.170***	−0.123***
Antiimmigration (cultural)	(0.00)	(0.00)	(0.00)	(0.01)	(0.01)
Routinization	0.061***	−0.019*	−0.046***	0.086***	−0.051***
Routinization	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)
Skill specificity	0.074***	−0.088***	−0.087***	0.078***	0.001
Skill specificity	(0.01)	(0.02)	(0.01)	(0.02)	(0.02)
Outsider	0.032	−0.007	−0.090***	−0.030	0.130***
Outsider	(0.02)	(0.03)	(0.02)	(0.04)	(0.03)
Unemployment rate	0.023***	−0.054***	−0.012**	−0.023	−0.005
Unemployment rate	(0.00)	(0.01)	(0.00)	(0.02)	(0.01)
FDI stock	0.001***	−0.002***	0.001***	−0.002***	−0.001
FDI stock	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Effective number of parties	−0.104***	−0.014	0.021	0.061**	0.120***
Effective number of parties	(0.02)	(0.02)	(0.02)	(0.03)	(0.03)
No. of respondents	43,627	43,627	43,627	43,627	43,627
No. of countries	18	18	18	18	18
ICC (rho)	.086	.629	.155	.540	.655
Panel SD (sigma)	0.308	1.301	0.429	1.083	1.379
R² (McKelvey/Zavoina)	.076	.025	.056	.126	.091
Prob > χ²	.000	.000	.000	.000	.000
BIC	54,938.18	25,162.28	51,092.85	14,219.11	18,495.32
Log-likelihood	−27,351.57	−12,463.62	−25,428.91	−6,992.04	−9,130.14

Multilevel probit estimates, standard errors in parentheses. FDI = foreign direct investment; BIC = Bayesian information criterion; ICC = intraclass correlation.

Level of statistical significance: *p ≤ .10. **p ≤ .05. ***p ≤ .01.

Acknowledgements

We would like to thank Jim Alt, Michael Bechtel, Matthew Bergman, Christian Breunig, Brian Burgoon, Daniele Caramani, Irene Menendez, Liz Zechmeister, participants in seminars at UC San Diego, the University of Konstanz, and the University of Zurich as well as the three anonymous reviewers and the editors at Comparative Political Studies for helpful comments and suggestions. We are especially grateful to Linda Maduz, who helped us push this project forward in its early stages.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Tobias Rommel is a PhD candidate in political science at the Department of Political Science at the University of Zurich. His research focuses on the political consequences of economic globalization in both democracies and autocracies. In his dissertation, he investigates how foreign direct investment affects autocratic regime survival.

Stefanie Walter is a full professor for International Relations and Political Economy at the Department of Political Science at the University of Zurich. Her research focuses on topics related to globalization, financial crises, and public opinion with a particular interest in how distributive conflicts shape individual political behavior and national crisis politics. She is the author of Financial Crises and the Politics of Macroeconomic Adjustments (2013, Cambridge University Press).

References

Acemoglu

Autor

(2011). Skills, tasks and technologies: Implications for employment and earnings. In Card

Ashenfelter

(Eds.), Handbook of labor economics 4 (pp. 1043-1171). Amsterdam, The Netherlands: Elsevier.

Adams

Haupt

Stoll

(2009). What moves parties? The role of public opinion and global economic conditions in Western Europe. Comparative Political Studies, 42, 611-639.

Norton

(2003). Interaction terms in logit and probit models. Economics Letters, 80, 123-129.

Alesina

Perotti

(1997). The welfare state and competitiveness. American Economic Review, 87, 921-939.

Allan

Scruggs

(2004). Political partisanship and welfare state reform in advanced industrial societies. American Journal of Political Science, 48, 496-512.

Arzheimer

(2009). Contextual factors and the extreme right vote in Western Europe, 1980-2002. American Journal of Political Science, 53, 259-275.

Benoit

Laver

(2006). Party policy in modern democracies. London, England: Routledge.

Bernard

Jensen

Schott

(2006). Survival of the best fit: Exposure to low-wage countries and the (uneven) growth of U.S. manufacturing. Journal of International Economics, 68, 219-237.

Betz

H.-G.

(1993). The new politics of resentment: Radical right-wing populist parties in Western Europe. Comparative Politics, 25, 413-427.

10.

Blinder

(2006). Offshoring: The next industrial revolution? Foreign Affairs, 85, 13-28.

11.

Blinder

(2009). How many U.S. jobs might be offshorable? World Economics, 10(2), 41-78.

12.

Bormann

N.-C.

Golder

(2013). Democratic electoral systems around the world, 1946-2011. Electoral Studies, 32, 360-369.

13.

Bornschier

(2018). Globalization, cleavages, and the radical right. In Rydgren

(Ed.), The Oxford handbook of the radical right. Oxford, UK: Oxford University Press. doi: 10.1093/oxfordhb/9780190274559.013.11

14.

Brambor

Clark

Golder

(2006). Understanding interaction models: Improving empirical analyses. Political Analysis, 14, 63-82.

15.

Burgoon

(2012). Partisan embedding of liberalism: How trade, investment, and immigration affect party support for the welfare state. Comparative Political Studies, 45, 606-635.

16.

Camia

Caramani

(2012). Family meetings: Ideological convergence within party families across Europe, 1945–2009. Comparative European Politics, 10, 48-85.

17.

Chase

(2008). Moving Hollywood abroad: Divided labor markets and the new politics of trade in services. International Organization, 62, 653-687.

18.

Crinò

(2009). Offshoring, multinationals and labour market: A review of the empirical literature. Journal of Economic Surveys, 23, 197-249.

19.

Dancygier

Walter

(2015). Globalization, labor market risks, and class cleavages. In Beramendi

Häusermann

Kitschelt

Kriesi

(Eds.), The politics of advanced capitalism (pp. 133-156). New York, NY: Cambridge University Press.

20.

de Lange

. (2007). A new winning formula? The programmatic appeal of the radical right. Party Politics, 13, 411-435.

21.

de Vries

Hoffmann

. (2016). Fear not values. Public opinion and the populist vote in Europe. Gütersloh, Germany: Bertelsmann Stiftung.

22.

Dolezal

(2010). Exploring the stabilization of a political force: The social and attitudinal basis of green parties in the age of globalization. West European Politics, 33, 534-552.

23.

Drezner

(2004). The outsourcing bogeyman. Foreign Affairs, 83(3), 22-34.

24.

Emmenegger

Häusermann

Palier

Seelaib-Kaiser

(2012). The age of dualization: The changing face of inequality in deindustrializing societies. Oxford, UK: Oxford University Press.

25.

Ezrow

Hellwig

(2014). Responding to voters or responding to markets? Political parties and public opinion in an era of globalization. International Studies Quarterly, 58, 816-827.

26.

Feenstra

Hanson

(1996). Globalization, outsourcing, and wage inequality. American Economic Review, 86, 240-245.

27.

Feenstra

Hanson

(1999). The impact of outsourcing and high-technology capital on wages: Estimates for the United States, 1979-1990. Quarterly Journal of Economics, 114, 907-940.

28.

Garrett

(1998). Partisan politics in the global economy. New York, NY: Cambridge University Press.

29.

Grossman

Rossi-Hansberg

(2008). Trading tasks: A simple theory of offshoring. American Economic Review, 98, 1978-1997.

30.

Hainmueller

Hiscox

(2006). Learning to love globalization: Education and individual attitudes toward international trade. International Organization, 60, 469-498.

31.

Haupt

(2010). Parties’ responses to economic globalization: What is left for the left and right for the right? Party Politics, 16, 5-27.

32.

Head

Mayer

Ries

(2009). How remote is the offshoring threat? European Economic Review, 53, 429-444.

33.

Hellwig

(2008). Globalization, policy constraints, and vote choice. Journal of Politics, 70, 1128-1141.

34.

Hellwig

(2014). Globalization and mass politics: Retaining the room to maneuver. Cambridge, UK: Cambridge University Press.

35.

Hellwig

Samuels

(2007). Voting in open economies: The electoral consequences of globalization. Comparative Political Studies, 40, 283-306.

36.

Helpman

Itskhoki

Redding

(2010). Inequality and unemployment in a global economy. Econometrica, 78, 1239-1283.

37.

Helpman

Melitz

Yeaple

(2004). Export versus FDI with heterogeneous firms. American Economic Review, 94, 300-316.

38.

Hummels

Jørgensen

Munch

Xiang

(2014). The wage effects of offshoring: Evidence from danish matched worker-firm data. American Economic Review, 104, 1597-1629.

39.

Iversen

Soskice

(2001). An asset theory of social policy preferences. American Political Science Review, 95, 875-893.

40.

Jensen

Kletzer

(2010). Measuring tradable services and the task content of offshorable services jobs. In Abraham

Spletzer

Harper

(Eds.), Labor in the new economy (pp. 309-335). Chicago, IL: University of Chicago Press.

41.

Jensen

Quinn

Weymouth

(2017). Winners and losers in international trade: The effects on U.S. presidential voting. International Organization, 71, 423-457.

42.

Jones

(2001). Are educated workers really more productive? Journal of Development Economics, 64, 57-79.

43.

Kayser

(2007). How domestic is domestic politics? Globalization and elections. Annual Review of Political Science, 10, 341-362.

44.

Kitschelt

(1988). Left-libertarian parties: Explaining innovation in competitive party systems. World Politics, 40, 194-234.

45.

Kitschelt

(1994). The transformation of European social democracy. Cambridge, UK: Cambridge University Press.

46.

Kitschelt

(2007). Growth and persistence of the radical right in postindustrial democracies: Advances and challenges in comparative research. West European Politics, 30, 1176-1206.

47.

Kriesi

Grande

Lachat

Dolezal

Bornschier

Frey

(2008). West European politics in the age of globalization. Cambridge, UK: Cambridge University Press.

48.

Long

Freese

(2006). Regression models for categorical dependent variables using Stata. College Station, TX: Stata Press.

49.

Luttwak

(1995). Turbo-charged capitalism and its consequences. London Review of Books, 17(21), 6-7.

50.

Mankiw

Swagel

(2006). The politics and economics of offshore outsourcing. Journal of Monetary Economics, 53, 1027-1056.

51.

Mansfield

Mutz

(2013). Us versus them: Mass attitudes toward offshore outsourcing. World Politics, 65, 571-608.

52.

Margalit

(2011). Costly jobs: Trade-related layoffs, government compensation, and voting in U.S. elections. American Political Science Review, 105, 166-188.

53.

Marks

Hooghe

Nelson

Edwards

(2006). Party competition and European integration in the East and West: Different structure, same causality. Comparative Political Studies, 39, 155-175.

54.

Mayda

(2006). Who is against immigration? A cross-country investigation of individual attitudes toward immigrants. The Review of Economics and Statistics, 88, 510-530.

55.

Melitz

(2003). The impact of trade on intra-industry reallocations and aggregate industry productivity. Econometrica, 71, 1695-1725.

56.

Mudde

(2007). Populist radical right parties in Europe. Cambridge, UK: Cambridge University Press.

57.

Mughan

Bean

McAllister

(2003). Economic globalization, job insecurity and the populist reaction. Electoral Studies, 22, 617-633.

58.

Mughan

Lacy

(2002). Economic performance, job insecurity and electoral choice. British Journal of Political Science, 32, 513-533.

59.

Owen

(2017). Exposure to offshoring and the politics of trade liberalization: Debates and votes on free trade agreements in the U.S. house of representatives, 2001-2006. International Studies Quarterly, 61, 297–311.

60.

Owen

Johnston

(2017). Occupation and the political economy of trade: Job routineness, offshorability and protectionist sentiment. International Organization, 71, 665–699.

61.

Rabe-Hesketh

Skrondal

(2008). Multilevel and longitudinal modeling using Stata. College Station, TX: Stata Press.

62.

Rehm

(2009). Risks and redistribution: An individual-level analysis. Comparative Political Studies, 42, 855-881.

63.

Roberts

(2014). The offshore outsourcing of American jobs: A greater threat than terrorism. Retrieved from http://www.globalresearch.ca/the-offshore-outsourcing-of-american-jobs-a-greater-threat-than-terrorism/18725

64.

Rovny

(2013). Where do radical right parties stand? Position blurring in multidimensional competition. European Political Science Review, 5, 1-26.

65.

Rueda

(2005). Insider-outsider politics in industrialized democracies: The challenge to social democratic parties. American Political Science Review, 99, 61-74.

66.

Scheve

Slaughter

(2001). What determines individual trade-policy preferences? Journal of International Economics, 54, 267-292.

67.

Scheve

Slaughter

(2004). Economic insecurity and the globalization of production. American Journal of Political Science, 48, 662-674.

68.

Schmidt

(2010). Parties. In Castles

Leibfried

Lewis

Obinger

Pierson

(Eds.), The Oxford handbook of the welfare State (pp. 211-226). Oxford, UK: Oxford University Press.

69.

Schmidt

(2012). Partisan composition of governments in OECD democracies, 1945-2010. Heidelberg, Germany: University of Heidelberg.

70.

Smith

Rivkin

(2008). A replication study of Alan Blinder’s “how many U.S. jobs might be offshorable?” (Harvard Business School Strategy Unit Working Paper No. 08-104). Retrieved from http://www.hbs.edu/faculty/Publication%20Files/08-104_16594743-0ea1-4a32-899f-321aa198fdd5.pdf

71.

Spitz-Oener

(2006). Technological change, job tasks, and rising educational demands: Looking outside the wage structure. Journal of Labour Economics, 24, 235-270.

72.

Steenbergen

Jones

(2002). Modeling multilevel data structures. American Journal of Political Science, 46, 218-237.

73.

Steiner

Martin

(2012). Economic integration, party polarisation and electoral turnout. West European Politics, 35, 238-265.

74.

Swank

(2002). Global capital, political institutions, and policy change in developed welfare states. Cambridge, UK: Cambridge University Press.

75.

van Kersbergen

. (1995). Social capitalism: A study of Christian democracy and the welfare state. London, England: Routledge.

76.

Volkens

Lehmann

Merz

Regel

Werner

(2013). The manifesto data collection (MRG/CMP/MARPOR, Version 2013b). Berlin: Wissenschaftszentrum Berlin für Sozialforschung (WZB).

77.

Walter

(2010). Globalization and the welfare state: Testing the microfoundations of the compensation hypothesis. International Studies Quarterly, 54, 403-426.

78.

Walter

(2017). Globalization and the demand-side of politics: How globalization shapes labor market risk perceptions and policy preferences. Political Science Research and Methods, 5, 55-80.

79.

Walter

Maduz

(2009). How globalization shapes individual risk perceptions and policy preferences: A cross-national analysis of differences between globalization winners and losers (WCFIA Working Paper Series No. 09-0015). Retrieved from http://www.tinyurl.com/lzgh9l9

80.

Ward

Ezrow

Dorussen

(2011). Globalization, party positions, and the median voter. World Politics, 63, 509-547.

81.

Wren

Rehm

(2013). Service expansion, international exposure, and political preferences. In Wren

(Ed.), The political economy of the service transition (pp. 248-281). Oxford, UK: Oxford University Press.

82.

Zohlnhöfer

Obinger

Wolf

(2008). Partisan politics, globalization, and the determinants of privatization proceeds in advanced democracies (1990-2000). Governance, 21, 95-121.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

7.10 MB

0.00 MB

1.86 MB