Massachusetts Institute of Technology Political Science Department Working Paper No.

2012-5

Sharing the Pain: What Explains Public Opinion Towards International Financial Bailouts?
Michael M. Bechtel, ETH Zurich Jens Hainmueller, MIT Yotam M. Margalit, Columbia University

Electronic copy available at: http://ssrn.com/abstract=2032147

Sharing the Pain: What Explains Public Opinion Towards International Financial Bailouts?
Michael M. Bechtel ETH Zurich Jens Hainmueller Massachusetts Institute of Technology Yotam Margalit Columbia University
April 2012

Abstract
Why do voters agree to bear the costs of bailing out other countries? Despite the prominent role of public opinion in the ongoing debate in Europe, extant research oers little empirical insight about the factors that underlie attitudes towards contributions to international bailouts. We study voter attitudes towards bailouts in Germany, the country shouldering the largest share of the EUs nancial rescue fund. We develop a set of arguments that explain attitudes toward bailouts as a function of voters economic selfinterest, social dispositions, and partisan orientations. We test these arguments using observational and experimental data from two original national surveys. We nd that while the economic features of the bailout package itself strongly aect voters willingness to support the transfers, individuals own economic standing has limited explanatory power in accounting for their position on the bailouts. In contrast, social dispositions such as altruism and cosmopolitanism are robustly associated with support for the bailouts. Overall, our ndings suggest that the current divide in public opinion towards the bailouts is less along distributive lines between domestic winners and losers, but better understood as a foreign policy issue that pits economic nationalist sentiments versus other-regarding concerns and greater anity to other Europeans.
Michael M. Bechtel, ETH Zurich, Center for Comparative and International Studies, IFW C45.2, Haldeneggsteig 4, CH-8092 Zurich, Switzerland, mbechtel@ethz.ch. Jens Hainmueller, Department of Political Science, 77 Massachusetts Avenue, Cambridge, MA 02139, jhainm@mit.edu. Yotam Margalit, Department of Political Science, Columbia University, 702 IAB, Mail Code 3320, United States, ym2297@columbia.edu. *We thank Tobias Boehmelt, Jude Hays, Pablo Pinto, Frank Schimmelfennig, Juerg Vollenweider, and audiences at University of California San Diego. University of Pittsburgh, St. Gallen University, Yale, and the University of Zurich for helpful comments. We gratefully acknowledge nancial support by ETH Zurichs Cooper Fund. Stefan Brunner, Quynh Nguyen, and Bernhard Reinsberg provided valuable research assistance. The usual disclaimer applies.

Electronic copy available at: http://ssrn.com/abstract=2032147

I. Introduction The global nancial system is in the midst of its most severe crisis in the post-War era, with countries such as Greece, Portugal, Ireland, Spain and Italy facing possible sovereign defaults. To prevent such defaults and a subsequent economic collapse, other eurozone countries have been asked to contribute massive sums to a nancial bailout fund. Although most experts view such bailouts as necessary for averting catastrophe, many voters ercely oppose a large transfer of their tax contributions to prop up other countries public nances. Indeed, recent polls in Germany and France reveal a sharply divided public, with only about one in three voters willing to support the bailout program.1 What explains the variation in voters support for the bailouts? This paper provides the rst systematic analysis of the determinants of voter preferences on this major new form of international redistribution. Perhaps even more than on most issues of foreign economic policy, public opinion on the EU bailouts is widely regarded as a major constraint on government actions in dealing with the crisis. Facing a restive electorate and a daily onslaught against the bailouts in Germanys major tabloids, electoral considerations have allegedly led the German government to initially refuse, and later delay, the plan proposed in early 2010 for a bailout of Greece.2 Less than a year later, the German government struck down the proposal of introducing a Euro-bond, another policy that many experts embraced as eective yet faced strong opposition among the voting public.3 Given the deep unpopularity of the bailouts, some have expressed concern that public opinion could ultimately prevent governments in donor countries from taking the necessary steps to address the European debt crisis.4 The role of public opinion in dealing with the crisis has been further magnied by the agreement among eurozone countries that any coordinated rescue package must be approved by all member countries. The requirement for each country to ratify the bailouts means that public opposition to the these nancial
Mehrheit gegen st arkere Finanzhilfen f ur verschuldete EU-Staaten, Politbarometer Dezember II 2010, 12/17/2010; French People Oppose Second Greek Bailout, Ifop Poll Shows, Bloomberg News, 09/17/2011; Germany Backs Greece Aid, but at a Cost to Merkel, New York Times, 02/27/2012. 2 Verkauft doch eure Inseln, ihr Pleite-Griechen, Bild, 10/27/2010; Merkel Has Tough Sell on Aid to Greece, New York Times, 06/27/2011; Angela Merkel: Europes saviour or biggest problem?, The Guardian, 11/22/2011. 3 The euro is a macro-economic weapon of mass destruction it simply must be defused, The Telegraph, 11/19/2011. 4 Germany Backs Greece Aid, but at a Cost to Merkel, New York Times, 02/27/2012.
1

Electronic copy available at: http://ssrn.com/abstract=2032147

transfers, even in small member countries such as Slovakia or Finland, could potentially derail the entire bailout eort, with major ramications for the EU as a whole.5 Yet, despite the prominent role of public opinion in the ongoing discussions over a coordinated international response to the crisis, very little is known about the factors that shape public support toward such rescue packages. To explain voters position on this issue, we draw upon three strands of literature to derive a set of explanations for the formation of public attitudes on the bailouts. The rst account centers on the distributive impacts of the bailouts, the second on voters social dispositions, and the third on the role of partisanship and political information. We then test these alternative predictions using novel data from two large-scale national surveys that we conducted in Germany, the country shouldering the largest share of the EUs bailout program. The surveys, carried out in early 2012, included a broad range of detailed items pertaining to respondents objective and subjective economic conditions, as well as their attitudes towards various aspects of the crisis, allowing for direct tests of the dierent theoretical accounts. A central nding in our study is that economic self interest, estimated using a broad range of measures, explains little of the variation in individual preferences on the eurozones nancial rescues. In fact, dierences in the impact of the crisis on voters personal circumstances are only weakly associated with attitudes towards the bailouts. We present experimental evidence which shows that voters worry about the economic costs of the bailouts reects a sociotropic concern about the overall burden that Germany will shoulder, yet this concern that does not vary systematically as a function of individuals own position in the economy. In contrast, dispositions such as altruism and cosmopolitanism are found to be strong predictors of preferences over the bailouts. The strength and consistency of these empirical associations suggests that voters assess these nancial rescue transfers also along dimensions other than the expected consequences of the transfers on their own well-being. To a lesser degree, we nd that partisan orientations also help account for some of the variation in individual attitudes. Yet notably, support for the bailout exhibits a cross-cutting ideological pattern. Rather than a left-right divide, attitudes dier most signicantly between
5

Slovakia Threatens Euro Rescue Package, Der Spiegel, 09/20/2011.

Electronic copy available at: http://ssrn.com/abstract=2032147

supporters of centrist parties and extremist parties, whereby the latter are signicantly less supportive of the bailouts. For further robustness, we replicate the results based on stated attitudes towards the bailouts using a second, quasi-behavioral measure. This measure is based on survey respondents willingness to sign up to have a message with their personal details sent on their behalf to their legislator, expressing their position on the bailouts. The results using this arguably more costly measure are reassuringly similar. In addition, we examine the studys external validity by augmenting the results from the online survey with a second, nationally representative phone survey. Again, the results of the two studies reveal similar ndings. In sum, this study provides the rst systematic evidence on the factors that shape mass attitudes in the debate over the eurozone bailouts. More broadly, our ndings suggest that the way to understand the public divisions over the bailouts is less along distributive lines of domestic winners and losers, but rather as a foreign policy issue that pits economic nationalist sentiments versus other-regarding concerns and greater anity to other Europeans. II. Background: The Politics of the European Financial and Debt Crisis In December 2009, the entering socialist government of Greece announced that its predecessor had massively under-reported the countrys debt. The revised assessment revealed a debt of over e300 bn, amounting to 113% of GDP, nearly twice the eurozones limit of 60%. Despite the Greek presidents assurances that the country was not about to default on its debt, rating agencies downgraded Greek bank and government debt. Soon after, concerns about other debt-exposed EU economies, primarily those of Ireland and Portugal, led eurozone countries to consider action over Greeces debt. The idea of providing a nancial bailout for Greece met immediate resistance among publics in the region, but after contracted negotiations, the eurozone governments and the IMF agreed in March 2010 to provide a nancial safety net of e22 bn for Greece, a sum that was soon revised up to e30 bn. Yet the rapid deterioration of the crisis, including violent strikes in Athens, led negotiators to further increase the bailout package and by early May they announced a e110 bn nancial rescue for Greece. Driven by the fear that the crisis could spread to other indebted countries, and after lengthy

negotiations with considerable discord, European leaders agreed upon a temporary bailout fund of e440 bn, the European Financial Stability Facility (EFSF).6 Each eurozone country was required to individually ratify the agreement and contribute to the fund proportionally. Germany, the largest eurozone economy, would shoulder the largest share, approximately 27%. As the initial EFSF funds proved insucient to contain the unfolding crisis in Ireland (e85 bn bailout) and later Portugal (e78 bn bailout), eurozone nance ministers negotiated a new, permanent bailout fund, called the European Stability Mechanism (ESM), worth close to e500 bn. Yet, throughout 2011, and despite the agreement over the ESM, worries continued to grow about the ability of the larger eurozone economies, primarily Italy and Spain, to service their debts. Renewed calls for expanding the bailout fund faced stark resistance from Angela Merkel, the German Chancellor. Other proposals for protecting the eurozone, in particular the idea to issue Eurobonds that would enhance the eurozones lending capacity, were also met by strong resistance from the German government, arguing that such bonds could substantially raise the countrys liabilities in a debt crisis.7 At the same time, increasing dissent from members of her governing coalition who threatened to oppose the expansion of the bailout fund, raised questions about Merkels political ability to lead an eective response to the crisis. However, in October of 2011, European leaders announced a three-pronged agreement that included bank recapitalization, a haircut by private lenders to Greece of 50%, and the expansion of the bailout fund to e1 trillion.8 This fund will invest in, and insure against sovereign debt, and be used to guarantee bonds from countries such as Spain and Italy. While the long-term impact of this agreement remains in question, most observers agree that further bailouts are needed if the eurozones embattled economies are to overcome this crisis. The situation confronting the eurozone and the attempts to coordinate international transfers of such magnitude are an unprecedented phenomenon in the post-War era. The key question is whether skeptical publics in donor countries will provide sucient political backing to allow further bailouts. The answer is far from clear, as a sizable share of citizens in major
The EFSF had a notional e440 bn euros at its disposal, but could only lend out approximately e250 bn because of the amount it was required to maintain in order to keep its borrowing costs low. 7 See, for example, The Return of Madame Non: Why Merkel Remains Opposed to Euro Bonds, Spiegel Online, 11/24/2011. 8 The term haircut refers to the size of the loss that lenders suer on the loans and interest they are owed.
6

donor countries express strong opposition to the rescue funds.9 To understand the factors underlying the public sentiments on this issue, the next section lays out a number of theoretical perspectives that may account for variation in individual support for the bailouts. III. Public Attitudes on International Bailouts The literature on the determinants of public opinion on foreign economic policy has developed substantially in recent years in both volume and scope, but no study to date has systematically explored the factors that shape attitudes toward international nancial bailouts. Yet, while this issue represents largely unchartered territory, one may still gain insights from related work examining the determinants of voter preferences on other international economic interactions such as trade, immigration, and foreign aid. Although most scholars share the view that the knowledge of citizens about these policy issues is limited, there is less scholarly agreement regarding the factors that ultimately shape peoples position on these issues. In this analysis we draw on three strands of literature to derive a set of explanations for the formation of public attitudes on the bailouts. The rst explanation centers on economic self-interest as the main determinant of attitudes. The second focuses on traits like altruism and cosmopolitanism, and the third approach emphasizes the inuence of partisan orientation and political knowledge. A. Economic Self-interest A common claim in the political economy literature holds that voters preferences on economic policy are a function of the expected eects of the policy on their own economic standing. Put simply, individuals that personally expect to gain from a given policy will support it, while potential losers will oppose it. Several studies have recently extended this claim, which lies at the core of work on the politics of taxation and redistribution, to also explain preferences on foreign economic policies. Various studies argue that voters preferences on policies such as trade (Hays et al., 2005; Mayda and Rodrik, 2005; Scheve and Slaughter, 2001a; ORourke and Sinnott, 2001), immigration (Mayda, 2008; Scheve and Slaughter, 2001a), or even foreign aid (Milner and Tingley, 2011), are largely a function of individuals position in the economy
Mehrheit gegen st arkere Finanzhilfen f ur verschuldete EU-Staaten, Politbarometer Dezember II 2010, 12/17/2010; French People Oppose Second Greek Bailout, Ifop Poll Shows, Bloomberg News, 09/17/2011.
9

and their subsequent expectations regarding how the proposed policy would aect their future earnings. In this vein, Broz (2005) analyzes roll call votes on bills proposed to restrict nancial rescues by the United States Treasury Departments Exchange Stabilization Fund. He argues that members of Congress will support rescues if their district is likely to benet from them and will oppose a bill otherwise. Applying the economic self-interest logic to explain variation in support for international bailouts requires theorizing about how voters assess the likely impact of a bailout (or lack thereof) on their own economic situation. To do so, a useful point of departure is the fact that the massive bailouts under consideration need to be nanced somehow. In the short run, this can happen primarily in at least one of three ways: (a) by raising taxes; (b) by reducing spending on domestic programs; or (c) by borrowing. These channels are relevant mainly for how individuals perceive the impact of the bailouts on their income from wage labor and public spending programs. However, individuals can also earn income from asset ownership (Scheve and Slaughter, 2004). Therefore, a fourth route by which bailouts could aect citizens economic well-being is by helping avert a collapse in the value of nancial assets triggered by sovereign defaults. Each of these mechanisms has dierent implications for how voters might assess the desirability of the bailouts. We elaborate on each of these mechanisms in turn.
A.1. Transfers funded by Taxes

A prominent approach in the literature on preferences for domestic redistribution holds that the net beneciaries of tax-based transfers will support such a policy while the net contributors will oppose it (Alesina and Ferrara, 2005; Corneo and Gr uner, 2002). This logic, formalized in Meltzer and Richards (1981) model, predicts that support for redistributive policies will correlate negatively with income, since high-income individuals have to nance the bulk of the transfers to lower-income individuals. If high earners perceive international bailouts as transfers that will require nancing through higher taxation, then the same prediction derived from models explaining preferences on domestic redistribution should also apply in the case of international transfers: all else equal, we might expect that high-income individuals would exhibit greater opposition to the bailouts than low-income individuals.

A.2. Reduced Domestic Spending

Domestic redistribution and welfare transfers benet poorer domestic constituencies (Alesina and Ferrara, 2005). In contrast, international nancial transfers provide immediate benets to individuals in other countries but assist the native population only indirectly. Therefore, voters that expect these transfers to reduce the nancial resources available for domestic assistance programs are more likely to oppose contributions to international bailouts. This logic would suggest that, in contrast to the tax increase argument, individuals with a higher probability of dependence on social transfers low income, unemployed, or workers employed in struggling sectors may be the ones most antithetical to the bailouts.10 In this context, we could also note the signicant geographic variation in economic prosperity across Germany, which means that some regions (L ander ) are net receivers of federal transfers while other regions are net contributors, subsidizing the federal governments assistance to less well-o parts of the country. Another prediction, following on the same logic described above, would therefore be that support for the bailout, ceteris paribus, would be lower in net receiving regions, since residents might worry that international nancial transfers would come at the expense of transfers to their own region.
A.3. Borrowing

A third approach for nancing the transfers, at least in the short and mid-term, could be by borrowing. This approach may appear especially attractive to policymakers because it does not immediately impose the bailout costs on the domestic constituents as in the case of tax increases or reductions in domestic spending. However, borrowing implies that the burden of nancing the bailouts will fall mostly on the younger age cohorts, who will have to pay the lenders back at some point in the future. Thus, nancing bailouts through borrowing results in an intergenerational redistribution of the costs. If this mechanism underlies the way voters assess the bailouts, we can expect to observe a clear intergenerational divide: younger
One caveat might be when an individuals job insecurity directly stems from the deteriorating eurozone crisis, for example if she is employed in a sector that depends almost entirely on business with other countries in the region. Such workers are expected to express greater support for the bailouts. Our empirical analysis takes this potential channel into account.
10

individuals would oppose the bailouts more strongly than older individuals, as the former can expect to shoulder in the future much of the costs arising from the present nancial transfers to other countries (Fullerton and Dixon, 2010; Thomson, 1996; Johnson and Falkingham, 1988; Campbell, 1977).
A.4. Financial Asset Prices

A fourth route by which individuals might assess the impact of the bailouts on their own well-being is the likely eect of a nancial meltdown on the nancial assets they own, such as stocks. For example, Scheve and Slaughter (2001b) have argued that since economic policy aects house prices through regional housing demand, economic policy preferences may also depend on asset ownership. Indeed, they nd some evidence that a combination of home ownership and the house being located in a region with a larger share of employment in less internationally-competitive sectors is associated with greater opposition to trade liberalization. In the context of the current crisis, individuals that have invested in nancial assets such as stocks would stand more to lose from a market crash following sovereign defaults in the eurozone. Furthermore, citizens that own nancial assets are likely to have a greater incentive to be informed about the consequences of a sovereign default. Therefore, individuals with stock investments would be more likely to support an international nancial rescue program that would stabilize nancial markets. B. Social Dispositions and Traits Social norms are an increasingly prominent explanation in research on preferences for redistribution and the provision of public goods. A wealth of evidence from lab experiments demonstrates that non-selsh motivations such as altruism or concern for fairness aect individuals willingness to contribute to public goods and to charity (Vigna et al., 2012; Fehr and Gachter, 2000; Fehr et al., 2007; Fehr and Fischbacher, 2002; Andreoni, 1990). These ndings have also inspired work outside the lab, which analyzes the link between other-regarding concerns and preferences for domestic redistribution. Fong (2001) and Linos and West (2003), for example, focus on attitudes towards distributive justice and nd that the more an individual perceives

luck as a determinant of economic success, the more she supports redistribution. In a similar vein, others show that preferences for redistribution reect individual humanitarian values (Feldman and Steenbergen, 2001) and beliefs on whether the economic system ensures equality of opportunity (Alesina and Giuliano, 2009; Alesina and Angeletos, 2005). These ndings suggest that social dispositions such as altruism may be important also for explaining why some individuals support international nancial transfers while others oppose them. Altruistic individuals are likely to exhibit a greater concern for the well-being of others, and therefore are expected to be more favorably disposed to the contribution of funds for assisting other countries in crisis. Since bailouts are a form of international redistribution, another potential social dispositions relevant for this analysis is citizens degree of cosmopolitanism. By cosmopolitanism we refer to the extent to which people identify with groups of individuals that are distant from them, geographically or culturally. Prior research shows that a measure of whether individuals are more cosmopolitan or nativist helps account for their attitudes on economic openness, particularly on policies such as trade and immigration (Margalit, 2012; Hainmueller and Hiscox, 2006a,b; Manseld and Mutz, 2009; ORourke and Sinnott, 2001). Following this research, a cosmopolitan outlook, which is likely to increase individuals degree of interest in, and empathy towards, citizens in other countries, is expected to be strongly associated with greater support for nancial bailouts. C. Partisan Orientation In democracies, electoral accountability provides incentives for policymakers to respond to public opinion. However, citizens are typically ill-informed about complex policy issues and therefore tend to rely on informational shortcuts such as partisan cues when forming their opinions. Studies have documented that citizens preferences are informed by such cues as voters observe and then adopt the position that their preferred party or politician takes on a given issue (Lupia, 1994; Druckman, 2001). Clearly, nancial bailouts constitute a prime example of a complex policy issue on which citizens may be particularly inclined to rely on such partisan cues when forming their own stance. This argument suggests that we should observe pro-

nounced dierences in preferences over nancial bailouts among supporters of dierent parties, dierences that should reect the parties public position. Moreover, the literature suggests that reliance on partisan orientation is most pronounced among individuals who possess lower levels of information on the topic (Zaller, 1992). Therefore, an individuals partisan orientation should be a stronger predictor of attitudes toward the bailouts among individuals who are less politically informed. If voters attitudes are responses to partisan cues or reections of ideological beliefs, the question is voters of which parties are more likely to support the bailout. Borrowing from the literature on domestic redistribution, we may expect that a classical left-right distinction also applies to international redistribution, with parties on the right being more opposed to helping foreign countries and parties on the left being more in favor of redistribution. Some scholars also contend that the left-right distinction oers further systematic insight on variation in foreign policy preferences more generally: voters on the left hold stronger internationalist sentiments while voters on the right are more isolationist with respect to international engagement (No el and Th erien, 2008). By this view, support for the bailout should also be strongest among voters of parties on the left. One possible alternative might be that the eurozone bailouts are an issue that much like the debate over EU integration (Hooghe et al., 2002) separates the center from the ideological extremes: rather than a left-right divide, partisans on either the far left (i.e., voters of Linke ) or the far right (NPD ) would oppose the bailouts whereas voters of the centrist parties would be more supportive. IV. Data and Measurement A. Sample Our main analysis draws upon an original survey that we elded online to a sample of 5,000 German voters in early January of 2012. To assess the external validity of the results we also added our main attitudinal measures to a national telephone survey.11 The representative phone sample contained 1,000 interviews with voters that were recruited using random digit
The survey was elded between January 2-5, 2012. Respondents were recruited by Respondi, an international survey rm. The telephone survey was conducted by IPSOS. Both organizations are renowned survey rms.
11

10

dialing. The survey had a response rate of 45.1% and was elded during the same days as the online survey. Similar to most other surveys based on internet panels, the online sample is somewhat skewed towards younger, more educated, and male voters compared to the total voter population.12 The demographics of the phone survey sample are more closely representative of the general voter population, except for a slight skew toward more educated respondents. To address these imbalances, we use entropy balancing (Hainmueller, 2012) to re-weight the data from the online and phone survey to match the demographic margins from the voter population. We use the weighted samples for all subsequent analyses. (Table B.1 in appendix B shows the demographics of the unweighted and reweighted online and phone survey samples as well as the general voter population.) Further below we also provide a comparison of the attitudes expressed in the online and phone samples and nd that the participants provided comparable responses and exhibited a similar relationship between key covariates and attitudes towards the bailout (see Table 6 below). B. Measures of Attitudes towards Bailouts We use two main questions to measure attitudes towards the nancial bailouts. The rst question is worded as follows: Now we would be interested in your opinion about the current nancial and debt crisis. As you may know, the EU countries have agreed to establish a nancial rescue fund which can be used to make bailout payments to over-indebted EU countries. In general, are you in favor or against bailout payments for over-indebted EU countries? The answer categories ranged from 1 (very much in favor) to 5 (very much against). We refer to this measure as Against Bailouts. While this item measures general attitudes towards the bailout, we also use a second measure that more directly captures attitudes towards the size of the German contribution to the fund: Should Germany pay more or less money into the European nancial rescue fund, which is used to provide bailout payments to over-indebted EU countries? Do you think Germany should pay in much more, pay in somewhat more, pay in neither more nor less, pay in somewhat less,
12

The sample is well balanced geographically.

11

or pay in much less? We code the answer categories in the same direction from 1 (pay in much more) to 5 (pay in much less) and refer to this variable as Pay In Less.13

C. Predictors of Attitudes towards Bailouts

C.1. Socio-demographic and Economic Variables

We use a range of measures to capture the dierent factors that may predict attitudes towards nancial bailouts.14 The set of predictors includes a battery of socio-demographic variables that measure a respondents age (ve age groups), highest educational attainment (four groups), and gender. The analysis employs a variety of variables that seek to capture potential self-interested economic concerns that correspond to each of the mechanism described in the theoretical section. The rst variable measures respondents reported household income, divided into ve income groups.15 The second variable is an indicator for whether the respondent owns stocks or mutual funds. This variable is included in the analysis since individuals that have invested in nancial assets such as stocks would stand more to lose from a market crash following eurozone defaults. Third, we include a binary indicator that distinguishes between respondents residing in a state that is a net beneciary of regional transfers (L andernanzausgleich ) and respondents that live in a state that is a net contributor. We also include various trade dependence measures which are based on respondents industry of employment, classied at the two digit level of the ocial German sectoral classication. Using the ocial trade data we computed the export, import and overall trade dependence of respondents industry of
To rule out order eects, we randomly used two versions of these items that only diered in the order of the answer categories. In version 1, the order was pay in much more to pay in much less. In version 2, the order was pay in much less to pay in much more. For levels of agreement/disagreement as answer categories, version 1 ranged from very much in favor to very much against and version 2 ranged from very much against to very much in favor. 14 Appendix A provides variable denitions, the question wording, and/or data sources for all variables. 15 The regressions also include an indicator variable denoting respondents that did not report their potential earnings
13

12

employment. In including these trade-dependence measures, the expectation is that individuals whose job depends on business from overseas will express greater support for the bailout. Since trade dependence could include engagement with countries outside the eurozone, we add an additional variable that indicates whether the establishment in which the respondent works has very strong, strong, weak, or no business ties with companies or customers in other EU member countries (Broz, 2005). Finally, we include measures for respondents current employment status and dierentiate between employed, unemployed, and retired individuals.

C.2. Altruism and Cosmopolitanism

As an indicator of respondents level of altruism, we employ a quasi-behavioral measure constructed as follows: we informed survey participants that an Amazon voucher of 100 euros would be raed among the survey participants. Respondents were then told that they can select to donate some fraction of their (potential) winnings to a charitable donation of their choice; in case they win, the donated amount will be deducted from their voucher. Respondents who chose to donate some portion of their remuneration were then prompted to select their preferred option for charity out of a menu of 30 organizations, and indicate the fraction that they wish to donate. Almost one third of respondents (35%) chose to donate at least some portion of the earnings, and among those individuals the median donation was 50% of the total sum (with notable clustering also at 10%, 25%, and 100%). For our regression analyses we classify the donations into three groups: respondents who chose not to donate at all, those who chose a donation between 1 and 50%, and individuals that selected a donation greater than 50%. These categories split the respondents who chose to donate into two, almost equally-sized groups. We measure cosmopolitanism using a question from the Local Cosmopolitan Scale (Dye, 1963) that has been extensively used in survey research on this topic. It measures the extent to which respondents think that national and international happenings are more or less inter-

13

esting than events that occur within their local community. This measure taps directly into the original formulation of the concept, as dened in the seminal work of Merton, who described cosmopolitanism as an orientation towards ones local community as opposed to more encompassing groups at the global level (Merton, 1947, 1968).

C.3. Partisan Orientation and Political Knowledge

We measure respondents partisan orientation by asking for which party they would vote if federal elections were held on the upcoming Sunday.16 We also include two measures of political knowledge. The rst is a standard general knowledge question that asks respondents whether the PR or SMD vote determines the share of seats in the Bundestag; about 50% correctly answered this question, and we code it as a binary variable. The second knowledge variable is a measure of knowledge specically about the bailouts, in which we provide respondents with a list of 6 possible countries and ask them to identify the countries that received nancial assistance from the European bailout fund.17 Based on their responses, we code a binary variable that takes the value 1 for the 50% of the respondents who correctly classied at least two of the three countries that received bailout payments and did not mark any of the three countries that did not receive bailout payments or the do not know option.18

V. Results A. Attitudes towards the Bailout Table 1 presents the distribution of responses for our two main outcome measures in the online sample. A clear majority of respondents opposes transfers to the struggling eurozone
16 In a pilot study we also included an item that asks which party a respondent leans towards in the long-run. The correlation between this measure and partisan orientation is about .8, suggesting that they ultimately measure the same construct. Therefore, we can rely on the partisan orientation measure without losing too much information. 17 The list of countries is Portugal, Ireland, Greece, Slovakia, Netherlands, France, and do not know. 18 About 24% correctly classify all three, and 26% correctly classify two out of three. Using three correct answers instead of two categories for this measure does not aect the results.

14

countries, although about 24% are somewhat in favor and about 3% are strongly in favor of the bailouts. Opposition to the bailouts is even more pronounced when we ask about the German contribution to the bailouts specically: less than 5% of respondents think that Germany should pay in more, and about 25% think that the level of contribution should be kept as is. The rest support paying somewhat less (34%) or a lot less (33%). Taken together, these results indicate that nancial bailouts are unpopular among German voters. The distribution of responses in our online sample is comparable with those in the main snap polls conducted on this topic and reported in the media.19

B. The Correlates of Opposition to Bailouts

B.1. Economic Self-Interest

What accounts for the variation in public attitudes toward the bailouts? We rst turn to factors pertaining to respondents economic self-interest and evaluate the support for each of the potential channels by which these self interested concerns may aect opposition to the bailouts. Models 1 to 6 in Table 2 report the results for our Against Bailout variable, which captures general opposition to the bailouts, and models 7 to 12 report results for the Pay In Less variable, which measures individuals views on the size of Germanys contribution to the bailout fund. All models control for basic demographics (i.e., include education, age, and gender). The rst set of results pertain to the measure of income. We nd that compared to respondents with low income (the reference category), high-income individuals are on average slightly more in favor of the bailouts. This goes against the tax-hike argument specied earlier, which predicts a negative correlation between income level and support for the bailouts. It
According to a representative snap survey, 32% of the German citizens support nancial bailouts for overindebted EU countries and 62% oppose these nancial transfers, see Mehrheit gegen st arkere Finanzhilfen f ur verschuldete EU-Staaten, Politbarometer Dezember II 2010, 12/17/2010, see also the more recent results from another snap survey reported in Germany Backs Greece Aid, but at a Cost to Merkel, New York Times, 02/27/2012.
19

15

is also notable that the coecient on income is statistically insignicant with regards to the second outcome variable, which measures individuals opinion about the level of Germanys contribution to the bailout fund (models 7 to 12). Furthermore, as we report below, once the model controls for measures of social dispositions and political orientations none of the income measures reaches statistical signicance (see Table 3). Clearly, these patterns are inconsistent with the taxation argument. According to the domestic spending account, individuals that currently rely or expect to depend on domestic assistance programs should exhibit stronger opposition to the bailouts. The results, however, do not lend much support to this argument. Models 2 and 8 in Table 2 introduce the variable Transfers: Net Beneciary, which denotes individuals that live in a state that is a net recipient of regional transfers. The variable is statistically insignicant in both of the models. As another way to assess the potential impact of regional distributional dierences, models 3-6 and 9-12 include a full set of state xed eects. According to the domestic spending argument, individuals in states like Berlin or Brandenburg, which are net beneciaries of regional transfers, should oppose nancial bailouts more strongly than BadenW urttemberg, which is one of the main net contributors. However, the p-values reported in the bottom panel of Table 2 show that using F-tests we cannot reject the null hypothesis that the state xed eects are jointly insignicant in any of the models indicating that opposition to bailouts does not much vary geographically. To further evaluate the spending argument, models 4 to 6 and 9 to 12 include measures of the overall rate of employment change in the respondents industry of employment. The models also include measures of the industries degree of trade dependence. The results indicate that none of the trade dependence variables is signicantly correlated with the two measures of attitudes on the bailouts. The same holds for the self-reported Trade Ties measures (models 5 and 11), which are based on respondents own assessment of how important they perceive trade relationships with the EU are for their employer. Working in a declining sector is also weakly

16

associated with support for the bailouts, but the variables do not come close to statistical signicance once social and political controls are introduced (see left and middle panel in Figure 1 below). Similarly, respondents employment status is also a weak predictor of attitudes on the bailouts: compared to fully employed individuals (the reference group), the unemployed are not more opposed to extending nancial assistance to struggling EU countries (models 4-6 and 10-12 in Table 2). In sum, we nd little support for the hypothesis that individuals assess the bailouts as a function of the potential impact on reducing the resources available for domestic assistance programs. As noted, the borrowing argument predicts an intergenerational divide over the bailouts, with younger individuals those likely to bear most of the future costs expected to oppose the transfers more than older age cohorts. The results in Table 2 do reveal some dierences in attitudes across the dierent age groups. The strength of the association changes somewhat across specications, but is fairly consistent nonetheless. Yet notably, the greatest opposition to the bailouts is among middle aged individuals; support for the bailouts is, in fact, higher among the youngest age cohort (individuals between 18 and 29). This pattern is at odds with the borrowing argument, which predicts the opposite divide between young and old. Finally, of the mechanisms linking self-interested considerations to attitudes on the bailouts, we nd the greatest support for the fourth one, linking asset ownership to support for the bailouts. As the results reported in Table 2 indicate, individuals who own stocks tend to support the bailouts more than individuals who do not own such assets. A possible interpretation of this association is that voters who worry about the impact of a nancial meltdown on the value of their own nancial holdings see greater benet in German nancial transfers aimed at stabilizing eurozone economies. While plausible, we also acknowledge hat the eect associated with stock ownership is quite modest: only 29% of the respondents own any stocks, and among those who do, probability of support for the bailout is about 0.2 greater.20 Taken
Moreover, interpretation of this nding is problematic, as a further examination of the result shows that it is driven by a small group of individuals with low incomes who own stocks. In fact, when we reestimate the model and include an interaction term between income and stock investments, the estimate on the interaction
20

17

together these results suggest that attitudes towards bailouts are not sharply divided along distributional lines as potential domestics winners and losers of the bailout exhibit very similar levels of opposition. To account for the bulk of the variation in preferences among the broad German public with respect to the bailouts, one must look for additional explanations beyond arguments based on material self-interest. Finally, we nd that higher levels of education are associated with lower opposition to the bailouts. Yet, the interpretation of this nding remains quite ambiguous, as other research has demonstrated that education may capture individual characteristics other than the respondents position in the labor force, e.g., internalized cultural or social norms (Hainmueller and Hiscox, 2006b; Margalit, 2012).

B.2. Social Dispositions and Political Orientation

So far we have focused on the role of economic self-interest in shaping attitudes toward the bailouts. Next, we evaluate whether social dispositions and political factors are important inuences. The results reported in Table 3 show that altruism, as measured by an individuals assigned donation to charity, strongly correlates with support for the bailouts. Compared to individuals that choose not to donate, voters that donate between 1-50% (Altruism: Medium ) and voters that donate more than 50% (Altruism: High ) are much more likely to support the bailout for both outcome variables. This nding remains robust also when including the full set of economic controls from the benchmark specication (models 3 and 6 in Table 2) together with state xed eects. To facilitate an easier interpretation and comparison of the magnitudes of these correlations, the left panel in Figure 1 provides a graphical representation of the size of the coecients together with 95% condence intervals. The gure presents marginal eects from a linear probability model that includes all economic, social, and political variables (as well as state
term is negatively signed. This nding is at odds with the theoretical argument, since richer individuals tend to hold a much larger share of stock investments than poorer individuals.

18

xed eects) with opposition to bailouts as the binary outcome variable. This variable equals 1 if a respondent opposes bailouts and is 0 otherwise. As the gure shows, highly altruistic individuals have an .12 points lower probability of opposing the bailouts than individuals that are not altruistic; about a 18% increase over the baseline probability of opposing the nancial transfer of .67. As the middle panel in Figure 1 shows, the eects are of comparable magnitude with respect to individuals attitudes toward the size of the German contribution to the bailout fund. Figure 1 also reveals a strong association between cosmopolitanism and attitudes on the bailout. The left panel suggests that individuals with high levels of cosmopolitanism have about a .15 points lower probability of opposing to the bailouts, a drop of about 23% relative to the mean level of opposition. For respondents with very high levels of cosmopolitanism, opposition to bailouts is on average about .3 lower than for individuals with very low levels of cosmopolitanism, a decrease of about 44% relative to the baseline probability. The results are of similar magnitudes when respondents are asked about whether Germany should pay in more or less into the bailout fund (middle panel in Figure 1). Notably, the results on the measures of both altruism and cosmopolitanism remain robust even with the inclusion of all the economic controls and state xed eects. Models 3 and 6 in Table 3 include a set of political covariates to evaluate the importance of voters political orientation and political knowledge. The results show that individuals who would vote for the Linke, the most left-wing party in the German parliament, as well as supporters of far right parties like the NPD and the Republikaner, oppose the nancial bailouts signicantly more than CDU voters, the governing party whose voters form the reference category. As the left panel in Figure 1 shows, voters of the Linke are on average .1 probability points more likely to oppose the bailouts than voters of the CDU. This represents a 15% increase over the baseline probability of opposing the bailouts. The increase is twice as large for voters of the right-wing extremist parties, who are on average .2 points more likely to

19

oppose the bailouts. While voters of the liberal FDP, currently the CDU s minor coalition partner, tend to oppose the bailouts somewhat more than CDU voters, the coecient is only borderline signicant. While these results suggest that partisan orientation is a noteworthy predictor of preferences for bailout policies, they also show that the dierences in partisan preferences do not map onto to the traditional left-right divide. Rather, we nd that voters located on the two extremes of the ideological spectrum are signicantly more opposed to the nancial rescues than the voters of centrist parties. This result is consistent with previous work that has documented an inverted U-shaped relationship between parties positions on the left-right dimension and support for European integration (Hooghe et al., 2002).

B.3. Political Knowledge

How should one interpret the strong association between partisan orientation and attitudes on the bailouts? Do voters follow cues from their preferred partys leadership, or is the association due to a congruence in the views of voters and party leaders? One way to explore this question is to focus on the relationship between knowledge, partisan aliation and views on the bailout. As noted, a sizable literature has demonstrated that less knowledgeable individuals tend to rely more on partisan cues when forming opinions about political issues, since these cues allow them to economize on the costs of acquiring knowledge on complex issues. By this logic, one would expect that partisan orientation would be a stronger predictor of attitudes toward the bailouts among voters who are less politically knowledgeable. To evaluate whether this is indeed the case, Table 4 breaks down the results by respondents levels of knowledge. We focus on two types of knowledge: possessing general information on political matters and knowledge pertaining specically to the nancial crisis. Models 1 to 4 show the results for attitudes toward the bailouts and models 5 to 8 report results for individual opposition to Germany contributing more into the European bailout fund. If the

20

cue-taking logic applies to this case, the prediction is that the coecients on measures of partisan orientation should be larger among individuals with less political knowledge. The results, however, do not reveal a pattern that supports this prediction. In fact, the relationship between partisan orientation and opposition to the bailouts is in some case stronger among more knowledgeable individuals. For example, the coecient is 0.7 for extreme right voters (NPD and Republikaner ) who possess a low degree of general political knowledge (model 1 in Table 4), but about 1.3 for their more informed counterparts (model 2). The 95% condence interval of the former estimate [.36; 1.06] indicates that the positive relationship between voting for extreme right-wing parties and opposition to the bailouts is signicantly greater among more informed respondents. We observe similar patterns for voters of the liberal party (FDP ). These results thus suggest that partisan aliation may be associated with attitudes on the bailout due to a congruence in views about the bailout issue between parties and voters, rather than because voters are relying on cues from their partys leadership. Alternatively, voters might be expressing support for a given party because of its position on the bailouts.

VI. Robustness and External Validity In this section we evaluate both the robustness of our results and the extent to which they generalize to the broader German voter population.

A. Quasi Behavioral Measure of Attitudes toward Bailouts We begin by comparing the results obtained from the attitudinal measures analyzed above with the results obtained when using a quasi-behavioral measure of sentiments toward the bailouts (for a similar strategy, see Hainmueller et al., 2011). The goal is to evaluate if the results from the relatively costless attitudinal measures remain intact once we increase the stakes somewhat, by providing respondents with an opportunity to sign up to send a message

21

regarding their position on the bailout to their legislators. Respondents are notied that this message will include some basic personal information. The question wording is as follows: Currently there is much discussion about whether Germany should pay more or less money into the European nancial rescue package, which is used to help over-indebted EU countries. Should we inform the Members of Parliament on your behalf whether you want Germany to pay somewhat more or somewhat less into the European nancial rescue fund? This information notice would contain your name and residence location. Answer categories included: no, please do not send a notice to the MPs or please inform the MPs on my behalf that Germany should pay a lot less/somewhat less/neither less nor more/somewhat more/a lot more money into the European nancial rescue fund. About 65% of the respondents signed up to send a message to their MPs. Among those who signed up, almost 71% asked to send a message to their MP calling for Germany to contribute either less or much less. 24% would like to keep the amount at the current level and only 5% want to send a message calling for Germany to pay in more. We re-estimated all the models reported above using this quasi-behavioral outcome measure and report the main results in Table 5. The ndings again suggest that social values and political orientation are strong predictors of attitudes toward the nancial bailouts while economic factors perform relatively poorly. Model 1 in Table 5 indicates that upper-middle income earners prefer Germany to pay in more into the bailout fund than individuals with a very low income (reference category). As before, stock ownership is negatively correlated with opposition to greater contributions to the bailouts. Notably, both coecients are only borderline signicant in the fully-specied model (model 3) which includes state xed eects as well as economic and sociodemographic controls. Although we nd some evidence of an intergenerational divide in views on the bailouts, the nding that individuals close to retirement age (50 to 59 years) are more opposed to the bailouts than younger citizens goes against the logic of the borrowing model described earlier. The analysis of the quasi-behavioral measure again reveals the strong empirical association between social dispositions and support for the bailout. More altruistic individuals, as well as 22

those with a higher level of cosmopolitanism, are signicantly less likely to sign up on a message to their legislator opposing Germanys contribution to the bailout. To ease the interpretation of the eects, the right panel in Figure 1 displays the results graphically. The probability that highly altruistic individuals will sign up to demand a lower German contribution to the bailouts is about .2 points lower than among individuals that score lowest on the altruism measure, a shift of about 30% over the baseline probability. Thus, the eects are very similar to those from the attitudinal outcome variables.21 Finally, the results pertaining to the political variables also remains unchanged: voters of both extreme left and extreme right parties prefer to contribute signicantly less than CDU supporters, the reference group. The magnitudes of these correlations are in the order of .2 probability points. As before, the results indicate that higher levels of general political knowledge correlate positively with an individuals willingness to transfer larger funds to overindebted countries. Finally, we explore whether our results on the moderating role of political knowledge remain robust when using a quasi-behavioral outcome measure. Again, we nd the same patterns: the correlation between partisan orientation and opposition to Germanys contribution is stronger among the more knowledgeable respondents (results are reported in Table B.3 in appendix B). Taken together, these checks reveal that the results reported in the main analysis remain robust across a large set of dierent specications as well as when using attitudinal or quasibehavioral outcome measures.

B. External Validity So far, we report results based on data obtained through an online survey. How well do these ndings generalize to the broad voter population? To help answer this question, we
To further probe the comparability of these results with those obtained for the attitudinal measures, we re-estimated the models reported in Table 2 using the quasi-behavioral measure. Table B.2 in appendix B reports the results and again shows that of all the economic self-interest variables, only stock-ownership is a signicant predictor of individuals willingness to send an identied anti-bailouts message to their MP.
21

23

included the main outcome variables in another survey that was carried out over the phone by a dierent polling rm, and which recruited respondents via random-digit-dialing. Table 6 compares the responses to the main attitudinal questions obtained in the phone and in the online samples. For the two attitudinal measures, the rst two columns present the fraction of sample respondents that selected each of the answer options (both samples are weighted such that their education, age, and gender margins match those of the total voter population). As the table shows, the answer distributions across the two samples are comparable, although the respondents in the online sample were more strongly opposed to the bailouts. Another dierence is that more respondents selected the neutral category (neither in favor nor against) when interviewed over the phone. This can perhaps be explained by the greater diculty that respondents in phone interviews faced as they were asked to instantly process the question, as opposed to respondents in the online mode who had more time to think about each survey item before selecting their answer. As a result, respondents in phone surveys may have been drawn to select the mid-point category, a choice that reduced the possibility of providing a wrong answer. To correct for this potential dierence across modes, the last two columns compare the answer distributions across the two samples when we omit the neutral and the dont know categories. Indeed, with this correction the similarity in responses improves considerably, although some dierence still remain, as respondents in the phone sample were a little more likely to report being strongly rather than weakly opposed to the bailouts and were overall somewhat more likely to support the bailouts. We now explore whether the correlates of bailout preferences found using the online data are replicated when analyzing the data from the telephone sample. Table 6 reports the results from several regressions which model support for bailouts as a function of all sociodemographic and economic variables that are available in both the online and the telephone survey (weighted and unweighted). Each model includes a full set of interactions of all predictors with an indicator variable denoting whether the survey was carried out by phone or online. This test enables us

24

to assess whether the online results signicantly dier from those obtained in the telephone sample.22 Even without controls for any of the social and political variables, the analysis again nds limited support for the predictions based on an economic self-interest perspective. Furthermore, the p-values reported in the last row of the table reveal that when using F-tests we cannot reject the null of joint insignicance for the interaction terms, which indicates that the coecients from the online survey do not signicantly dier from those obtained in the telephone survey.23 Taken together, this suggests that the results obtained in the online mode do have a notable level of external validity.

VII. Economic Characteristics and Bailout Preferences: Experimental Findings So far, our results underscore the role of social dispositions and political orientation as signicant correlates of attitudes toward nancial bailouts. In contrast, factors pertaining to economic self-interest seem to matter relatively little. Yet, the weakness of economic selfinterest as a predictor of attitudes does not imply that voters take no account of economic considerations when forming preferences over the bailouts. For example, voters might support or oppose such international transfers as a function of sociotropic concerns, such as how burdensome they perceive the bailout contributions to be on the German economy as a whole. In that case, the question of interest is whether such sociotropic concerns vary across respondents as a function of their own position in the economy. To explore this question, we turn to examine: (a) whether the size of Germanys contribuWe unfortunately lack several important measures of social dispositions and partisan aliation, as budget constraints precluded introducing all these items in the telephone survey. Nevertheless, the analysis still enables us to explore whether the ndings replicate for standard economic predictors such as income or employment using the telephone sample. 23 The only two exceptions are the interaction between the variables Age: 50-59 and Phone survey mode in models 3 and 4 and the interaction between Phone and High School: Medium Tier in model 3. The negative signs of the corresponding coecients suggest that respondents in these categories are less opposed to Germanys contribution to the bailouts when interviewed in the telephone than in the online model.
22

25

tion to the bailout fund aects voters willingness to support the proposal; and (b) whether sensitivity to the cost of the program varies systematically across income groups. To do so, we embedded an experiment in the survey in which we asked respondents how likely they would be to vote in favor of or against Germany contributing eX bn to the bailout fund. The experiment randomly assigned respondents to one of four groups that diered with respect to the size of Germanys reported contribution to the bailout fund. In the rst treatment, the size of Germanys contribution was set at e123 bn. The other three treatments reported the contribution to be either e189 bn, e211 bn, or e418 respectively. Respondents were then asked whether they would vote in favor or against the bailout in a direct-democratic vote, with answer categories ranging from vote surely against to vote surely in favor.24 Random assignment in combination with a large sample ensures that all four treatment groups are comparable with respect to observable and unobservable confounders. Therefore, we can directly compare levels of support and opposition across the three groups and the results have a causal interpretation. Figure 2 plots the results from a linear probability model with opposition to the bailouts as the dependent variable. This variable is a binary indicator which equals 1 if a respondent disapproves of the bailouts and zero otherwise. We rst consider the results for all respondents, which are displayed at the very top of Figure 2. The graph indicates that opposition to bailouts among individuals indeed varies signicantly as a function of the size of Germanys contribution. As Germanys contribution rises from e123 bn, the reference category, to e189 bn, the probability of opposing the bailout increases by about .03 points. Since the average baseline probability of opposing the bailout is .52, this treatment eect represents a shift of about 6% over the mean level of opposition. If the size of Germanys contribution increases to e211 bn, which is the amount Germany has agreed to contribute, the probability of opposing the bailouts rises by about .06 points, which equals a 12% increase relative to the baseline probability. Finally, in the treatment group assigned the
The possible answers were: vote surely against, vote very likely against, vote probably against, neither/nor, vote probably in favor, vote very likely in favor, vote surely in favor. Other features of the bailout described in the experiment were also fully randomized and can thus be ignored in this analysis.
24

26

highest contribution (e418 bn), opposition to the bailout is on average 0.16 points greater than in the group of respondents that received the e123 bn treatment.25 This implies an increase of about 31% in the probability of opposition relative to the baseline probability. These results indicate that voters assign great signicance to the economic costs of the bailout, a nding that appears consistent with a sociotropic concern about the nancial burden placed on the countrys economy. Next, we examine whether voters with dierent economic interests, as measured by their income level, exhibit dierent degrees of concern about the economic costs of the bailout. If self-interested concerns underlie the sensitivity to the size of Germanys contribution to the bailout, one would expect that individuals with higher incomes, those likely to shoulder a larger share of the cost through a progressive tax system, will exhibit greater responsiveness to the experimental treatment. To evaluate this prediction, we separately analyze the treatment eects by income levels, distinguishing between individuals in low-, middle-, and high-income households. The results in the lower part of Figure 2 are quite striking in that the point estimates for each of the bailout treatments are remarkably similar across the dierent income groups. If we compare the eects of the e123 bn, e189 bn, and e211 bn treatments across income groups, we nd that they are virtually the same. In particular, high-income individuals are not more responsive to contribution size: even the extreme comparison between low-income and highincome individuals shows that the eects of the contribution treatments do not dier across the two groups. The small dierence across income groups is that middle-income earners respond somewhat more strongly to the e418 bn-treatment than high-income individuals. Clearly, these ndings do not sit well with the predictions of the two mechanisms laid out earlier, namely those focused on future tax increases or a reduction in domestic spending. However, the ndings remain consistent with the idea that citizens engage in sociotropic evaluations of
The German government has indeed recently considered increasing Germanys contribution up to e500 bn, see: Merkel erw agt nun doch gr osseren Rettungsschirm, Der Spiegel, 03/01/2012.
25

27

economic policy.

VIII. Discussion The nancial crisis and the political struggles with its consequences are widely regarded as having led, or contributed to, the recent fall of governments in Ireland, Italy, Greece, the Czech Republic, Romania, and Spain.26 Notably, the political fallout from the crisis appears not to be conned to countries in nancial distress; governments residing over relatively healthy economies also face internal strife, as they are asked to divert major public funds for unpopular nancial rescues of other countries. In making their case, some politicians have argued for the bailouts on grounds of its economic rationale; others have emphasized instead notions of fairness or of European solidarity.27 In this paper, we provided the rst systematic account of individual preferences over international nancial bailouts. We tested a set of theoretical accounts linking voter preferences on the bailouts to their position in the economy, with the aim of exploring whether and how the distributive consequences of these international transfers shape voters stance on this issue. A key nding of this study is that the various measures of citizens economic self-interests are only weakly associated with their position in the debate over the bailouts. The greatest support we nd for any of these potential accounts is between stock ownership and support for the bailout, yet even this relationship is small in substantive terms and only marginally significant in some of the specications. Taken together these ndings suggest that the main divide in public opinion towards the bailouts does not cut along distributive lines pitting domestic winners again losers. Economic self-interest simply does not appear to be the main prism by which German citizens assess the countrys contribution to the nancial rescue fund. At the same time, we do nd evidence of sociotropic concerns about the cost of the bailout
See, for example, Spain the fth victim to fall in Europes arc of depression, The Telegraph, 03/16/12; or Now Romania: Another European government falls amidst eurocrisis, Associated Press, 02/06/12. 27 See, for example, The Irish Times, 12/05/2011; Schmidt scathing on German role in crisis or Slovakias rebel isnt a nationalist: hes the hero of all discontented Europeans, The Telegraph, 03/11/2012.
26

28

to the German economy as a whole, concerns that appear unrelated to voters own economic standing. Moreover, the strongest predictors of voters stance on the bailouts are measures of key social dispositions, such as degree of altruism or cosmopolitanism. Thus, characteristics that reect individuals levels of generosity, empathy, and perhaps a broader conception of cultural anity, seem to provide a more promising guide to understanding voters positions on international nancial burden-sharing. We also nd a fairly strong empirical relationship between voters partisan orientation and their stance on the bailout. Notably, the key variation we observe across voters does not reect a left-right divide, but rather a division between centrist parties and those located at the two extremes of the ideological spectrum. Moreover, our analysis, which compares voters with high and low degrees of relevant knowledge suggests that this partisan pattern does not result from partisan cue-taking, whereby poorly informed voters form their attitudes based on the stated positions of their preferred parties. Possibly, the reverse causal link may operate, as many observers perceive shifts in the electorate towards parties because of their position on the eurozone crisis.28 However, more research is needed to clarify the role of partisan orientations in this context. These public divisions over contributions to international bailouts provide an important new arena for evaluating the usefulness of several theoretical approaches currently debated in the literature for explaining mass opinion on foreign economic policy. One pertinent question that arises from this studys ndings is how to evaluate the limited explanatory usefulness of accounts linking individuals own position in the economy to their attitudes on the bailouts? One possible response might be that the bailout issue may be too hard of a test of self-interested accounts of economic policy preferences. Given the complexity of the bailout scheme and the high degree of uncertainty regarding its potential usefulness, most voters are limited in their ability to assess the distributive eects of the bailouts and the likely impact on their own
In the Netherlands, for example, the far left Socialists have gained major new ground as a result of their staunch opposition to the bailouts, almost halving the support for the pro-bailouts, more centrist Labour party. See, for example, Dutch Socialists gain from eurozone woes, Financial Times, 01/31/2012.
28

29

well-being. Thus, only those individuals who have a clear nancial stake in the stabilization of the eurozone economies, e.g. stock-holders, are able to make the connection between policy and outcomes. Another possible approach for why the papers ndings regarding the economic explanations are not applicable to other domains could be the unique role of the media in framing the public debate over this specic issue. The coverage of the bailouts, especially among tabloid newspapers such as the Bild, which carried out a prolonged campaign against the bailouts, frequently used hyperbolic headlines with overt nationalistic themes, and thus potentially marginalized the issue of the bailouts domestic distributive consequences. While both arguments may have some merit, this studys ndings are in fact consistent with an growing number of analyses that demonstrate that narrow economic self-interest may not, after all, play a very prominent role in shaping voter attitudes toward various economic policies; other, non-material considerations may be at least as, or even more important factors (Hainmueller et al., 2011; Hainmueller and Hiscox, 2010, 2006b; Margalit, 2012). Indeed, a growing number of studies indicate that voters assess economic interaction with foreigners be it through trade, foreign direct investment or immigration along additional dimensions that go beyond the strict distributive consequences of the policies. For example, cultural considerations, ethnocentricism (Manseld and Mutz, 2009; Kinder and Kam, 2009; Sniderman et al., 2004; McLaren, 2002), or environmental concerns (Bechtel et al., 2012) appear to be important factors that predict individual attitudes toward economic openness. Thus, while our results pertain only to a specic policy debate, the accumulated evidence suggests that while economic self-interest has its role in analyzing individuals policy preferences, a stronger focus on other-regarding preferences, cultural anity, and political orientation would likely help scholarship to gain a more comprehensive understanding of individual attitudes toward foreign economic policy. By providing the rst systematic analysis of the key determinants of attitudes toward the bailouts, our results also raise questions about the sensitivity of the attitudes that voters

30

exhibit toward specic features of the bailout package. For example, if the contributions of the country (in this case, Germany) to the overall fund were decreased, or if the relative distribution of the burden across the eurozone members was dierent, to what extent would that sway voters to adopt a dierent position in the debate? Answers to such questions have obvious policy implications as they can clarify which features of a bailout package could make it more palatable to voters. We plan to explore these issues in future research. Finally, further research could explore the impact of the nancial crisis on voters attitudes toward the broader European integration project. The nancial crisis, in combination with the calls for increased burden-sharing among member states, may have fundamentally changed the way citizens perceive the costs and benets of European integration. As economic projections continue to reveal signs of a prolonged slump in the eurozone, voters may grow more open to the calls for changing the Unions structure, either by shedding some of its struggling members or by adopting the opposite approach, namely pursuing a tighter political union. Either way, in confronting public opinion, the bailout eorts that the eurozone countries will ultimately decide to take may determine the success and stability of the European integration project for many years to come. Our study is merely a rst step in improving our understanding of when and how countries respond to the nancial crisis and its profound long-term challenges.

31

References
Alesina, A. and Angeletos, G. (2005). Fairness and redistribution, American Economic Review 95(4): 960980. Alesina, A. and Ferrara, E. L. (2005). Preferences for redistribution in the land of opportunities, Journal of Public Economics 89(5-6): 897931. Alesina, A. and Giuliano, P. (2009). Preferences for Redistribution. NBER Working Paper 14825. Andreoni, J. (1990). Impure Altruism and Donations to Public Goods: A Theory of Warm-Glow Giving, Economic Journal 100(401): 464477. Bechtel, M. M., Bernauer, T. and Meyer, R. (2012). The Green Side of Protectionism: Environmental Concerns and Three Facets of Trade Policy Preferences, Review of International Political Economy forthcoming. Broz, J. L. (2005). Congressional Politics of International Financial Rescues, American Journal of Political Science 49(3): 479496. Campbell, A. (1977). Politics Through the Life Cycle, The Gerontologist 11(2 Part 1): 112117. Corneo, G. and Gr uner, H. P. (2002). Individual preferences for political redistribution, Journal of Public Economics 83(1): 83107. Druckman, J. N. (2001). The implications of framing eects for citizen competence, Political Behavior 23(3): 225256. Dye, T. (1963). The local-cosmopolitan dimension and the study of urban politics, Social Forces 41(3): 239246. Fehr, E. and Fischbacher, U. (2002). Why social preferences matter - the impact of Non-Selsh motives on competition, cooperation and incentives, Economic Journal 112(478): C1C33. Fehr, E. and Gachter, S. (2000). Cooperation and punishment in public goods experiments, American Economic Review 90(4): 980994. Fehr, E., Klein, A. and Schmidt, K. M. (2007). Fairness and contract design, Econometrica 75(1): 121 154. Feldman, S. and Steenbergen, M. R. (2001). The Humanitarian Foundation of Public Support for Social Welfare, American Journal of Political Science 45(3): 658677. Fong, C. (2001). Social preferences, self-interest, and the demand for redistribution, Journal of Public Economics 82(2): 225246. Fullerton, A. S. and Dixon, J. C. (2010). Generational Conict Or Methodological Artifact? Reconsidering the Relationship between Age and Policy Attitudes in the U.S., 19842008, Public Opinion Quarterly 74(4): 643673. Hainmueller, J. (2012). Entropy balancing for causal eects: A multivariate reweighting method to produce balanced samples in observational studies, Political Analysis 20(1): 2546.

32

Hainmueller, J. and Hiscox, M. J. (2006a). Educated Preferences: Explaining Attitude Toward Immigration in Europe, International Organization 61(2): 399442. Hainmueller, J. and Hiscox, M. J. (2006b). Learning to Love Globalization: Education and Individual Attitudes toward International Trade, International Organization 60(2): 469498. Hainmueller, J. and Hiscox, M. J. (2010). Attitudes toward Highly Skilled and Low-skilled Immigration: Evidence from a Survey Experiment, American Political Science Review 104(1): 124. Hainmueller, J., Hiscox, M. J. and Margalit, Y. (2011). Do Concerns about Labour Market Competition Shape Attitudes Toward Immigration? New Evidence from U.S. Workers. MIT Political Science Department Research Paper No. 2011-20. Hays, J. C., Ehrlich, S. D. and Peinhardt, C. (2005). Government Spending and Public Support for Trade in the OECD: An Empirical Test of the Embedded Liberalism Thesis, International Organization 59(2): 473494. Hooghe, L., Marks, G. and Wilson, C. (2002). Does Left/Right Structure Party Positions on European Integration?, Comparative Political Studies 35(8): 965989. Johnson, P. and Falkingham, J. (1988). Intergenerational Transfers and Public Expenditure on the Elderly in Modern Britain, Ageing and Society 8(2): 129146. Kinder, D. R. and Kam, C. D. (2009). Us Against Them: Ethnocentric Foundations of American Opinion, University of Chicago Press, Chicago, IL. Linos, K. and West, M. (2003). Self-interest, Social Beliefs, and Attitudes to Redistribution. Readdressing the Issue of Cross-national Variation, European Sociological Review 19(4): 393409. Lupia, A. (1994). Shortcuts versus encyclopedias: Information and voting behavior in california insurance reform elections, American Political Science Review 88(1): 6376. Manseld, E. D. and Mutz, D. C. (2009). Support for Free Trade: Self-Interest, Sociotropic Politics, and Out-Group Anxiety, International Organization 63(3): 425457. Margalit, Y. (2012). Lost in Globalization: Economic Integration and the Politics of Discontent, International Studies Quarterly forthcoming. Mayda, A. M. (2008). 101(3): 160163. Why are people more pro-trade than pro-migration, Economics Letters

Mayda, A. M. and Rodrik, D. (2005). Why Are Some People (and Countries) More Protectionist Than Others?, European Economic Review 49(6): 13931430. McLaren, L. M. (2002). Public Support for the European Union: Cost/Benet Analysis or Perceived Cultural Threat, Journal of Politics 64(2): 551566. Meltzer, A. H. and Richard, S. F. (1981). A Rational Theory of the Size of Government, Journal of Political Economy 89(5): 914927.

33

Merton, R. (1947). Selected problems of eld work in the planned community, American Sociological Review 12(3): 304312. Merton, R. (1968). Social theory and social structure, Free Pr. Milner, H. V. and Tingley, D. H. (2011). Who supports global economic engagement? the sources of preferences in american foreign economic policy, International Organization 65(1): 3768. No el, A. and Th erien, J.-P. (2008). Left and Right in Global Politics, Cambridge University Press, Cambridge. ORourke, K. H. and Sinnott, R. (2001). What Determines Attitudes Towards Protection? Some Cross-Country Evidence, in S. M. Collins and D. Rodrik (eds), Brookings Trade Forum 2001, Brookings Institute Press, pp. 57206. Scheve, K. F. and Slaughter, M. J. (2001a). Labor market competition and individual preferences over immigration policy, Review of Economics and Statistics 83(1): 133145. Scheve, K. F. and Slaughter, M. J. (2001b). What determines individual trade-policy preferences?, Journal of International Economics 54: 267292. Scheve, K. F. and Slaughter, M. J. (2004). Economic Insecurity and the Globalization of Production, American Journal of Political Science 48(4): 662674. Sniderman, P. M., Hagendoorn, L. and Prior, M. (2004). Predispositional Factors and Situational Triggers: Exclusionary Reactions to Immigrant Minorities, American Political Science Review 98(1): 3550. Thomson, D. (1996). Selsh Generations? How Welfare States Grow Old, The White Horse Press. Vigna, S. D., List, J. A. and Malmendier, U. (2012). Testing for Altruism and Social Pressure in Charitable Giving, Quarterly Journal of Economics 127(1): 156. Zaller, J. (1992). The Nature and Origins of Mass Opinion., Cambridge University Press, New York.

34

Tables Table 1: Attitudes Towards Financial Bailouts How much are you in favor or against bailout payments for over-indebted EU countries? strongly in favor 3.0% somewhat in favor 24.5% neither in favor nor against 10.2% somewhat against 40.1% strongly against 20.9% dont know 1.5% Should Germany pay more or less money into the European nancial rescue fund..? pay in much more 0.5% pay in somewhat more 4.0% pay neither more or less 25.4% pay in somewhat less 34.2% pay in much less 32.6% dont know 3.4%

Note: N=4,499. Results are weighted by the sample adjustment weights so that the education, age, and gender margins match the total voter population (see text for details).

35

Table 2: Predictors of Opposition to Bailouts: Personal Economic Interest

Model No. Outcome Female School: Medium Tier School: Highest Tier University/College Income: Low Income: Lower Middle Income: Upper Middle Income: High Own Stocks Age: 30-39 Age: 40-49 Age: 50-59 Age: 60+ Transfers: Net Beneciary (1) 0.04 (0.06) -0.00 (0.07) -0.40*** (0.08) -0.52*** (0.08) 0.06 (0.10) 0.05 (0.11) -0.12 (0.13) -0.22 (0.14) -0.24*** (0.07) 0.19*** (0.07) 0.24*** (0.07) 0.20*** (0.07) 0.00 (0.09) (2) 0.02 (0.06) -0.01 (0.07) -0.40*** (0.08) -0.51*** (0.09) 0.05 (0.10) 0.05 (0.11) -0.14 (0.13) -0.22 (0.14) -0.24*** (0.07) 0.19*** (0.07) 0.25*** (0.07) 0.20*** (0.07) 0.02 (0.09) -0.01 (0.07) (3) (4) Against Bailouts (1-5) 0.05 0.04 (0.06) (0.07) -0.00 -0.01 (0.07) (0.07) -0.39*** -0.34*** (0.08) (0.08) -0.50*** -0.50*** (0.08) (0.08) 0.05 0.00 (0.10) (0.10) 0.04 -0.01 (0.11) (0.11) -0.14 -0.19 (0.13) (0.13) -0.23* -0.28** (0.14) (0.14) -0.23*** -0.23*** (0.07) (0.07) 0.18** 0.08 (0.07) (0.08) 0.24*** 0.13* (0.07) (0.08) 0.22*** 0.11 (0.07) (0.08) 0.02 -0.09 (0.09) (0.13) (5) 0.02 (0.06) 0.02 (0.07) -0.30*** (0.08) -0.47*** (0.08) 0.04 (0.10) 0.03 (0.11) -0.14 (0.12) -0.23* (0.14) -0.21*** (0.06) 0.09 (0.08) 0.13* (0.08) 0.10 (0.08) -0.09 (0.13) (6) 0.04 (0.06) -0.01 (0.07) -0.34*** (0.08) -0.51*** (0.08) 0.00 (0.10) -0.01 (0.11) -0.20 (0.12) -0.29** (0.14) -0.23*** (0.07) 0.08 (0.08) 0.14* (0.08) 0.11 (0.08) -0.09 (0.13) (7) 0.09** (0.05) 0.02 (0.05) -0.28*** (0.06) -0.41*** (0.06) 0.03 (0.08) -0.08 (0.09) -0.06 (0.09) -0.11 (0.09) -0.20*** (0.05) 0.22*** (0.05) 0.20*** (0.06) 0.21*** (0.06) -0.04 (0.07) (8) 0.09** (0.05) 0.00 (0.05) -0.31*** (0.06) -0.43*** (0.06) 0.02 (0.08) -0.10 (0.08) -0.08 (0.09) -0.12 (0.09) -0.19*** (0.05) 0.22*** (0.06) 0.18*** (0.06) 0.19*** (0.06) -0.05 (0.07) 0.06 (0.05) (9) (10) Pay In Less (1-5) 0.10** 0.09* (0.05) (0.05) 0.03 0.01 (0.05) (0.05) -0.28*** -0.27*** (0.06) (0.06) -0.40*** -0.42*** (0.06) (0.06) 0.04 0.01 (0.08) (0.08) -0.08 -0.12 (0.08) (0.08) -0.06 -0.11 (0.09) (0.09) -0.12 -0.17* (0.10) (0.09) -0.20*** -0.19*** (0.05) (0.05) 0.20*** 0.15** (0.05) (0.06) 0.18*** 0.13** (0.06) (0.06) 0.22*** 0.17** (0.06) (0.07) -0.04 -0.04 (0.07) (0.11) (11) 0.08* (0.05) 0.02 (0.05) -0.25*** (0.06) -0.40*** (0.06) 0.02 (0.08) -0.10 (0.08) -0.09 (0.09) -0.13 (0.09) -0.18*** (0.05) 0.16** (0.06) 0.13** (0.06) 0.17** (0.07) -0.04 (0.11) (12) 0.10** (0.05) 0.01 (0.05) -0.27*** (0.06) -0.42*** (0.06) 0.01 (0.08) -0.11 (0.08) -0.10 (0.09) -0.17* (0.10) -0.20*** (0.05) 0.16** (0.06) 0.13** (0.06) 0.17** (0.07) -0.03 (0.11)

36

Sector Employment: Decrease Sector Employment: Increase Sector Employment: Strong Increase Status: Part-time employed Status: In Education Status: Unemployed Status: Retired Trade Ties: Weak Trade Ties: Strong Trade Ties: Very Strong Sector: Medium Trade Dependence Sector: Strong Trade Dependence Constant State Fixed Eects Joint Test: p-value Observations 3.59*** (0.10) 3.60*** (0.11) 3.62*** (0.14) 0.15 4,441

0.06 (0.11) 0.04 (0.10) 0.05 (0.10) -0.03 (0.07) -0.26 (0.19) -0.06 (0.11) -0.02 (0.13)

0.03 (0.11) 0.01 (0.10) 0.01 (0.10) -0.06 (0.07) -0.28 (0.20) -0.10 (0.11) -0.05 (0.13) -0.10 (0.08) -0.14 (0.10) -0.14 (0.12)

0.08 (0.20) 0.05 (0.19) 0.07 (0.19) -0.03 (0.07) -0.26 (0.19) -0.06 (0.11) -0.02 (0.13)

0.14* (0.08) 0.08 (0.08) 0.11 (0.08) -0.02 (0.06) -0.07 (0.15) -0.09 (0.09) -0.09 (0.10)

0.11 (0.08) 0.07 (0.08) 0.08 (0.08) -0.04 (0.06) -0.08 (0.16) -0.11 (0.09) -0.11 (0.10) -0.07 (0.06) -0.00 (0.07) -0.14 (0.09)

0.29* (0.17) 0.23 (0.16) 0.26 (0.16) -0.02 (0.06) -0.07 (0.15) -0.08 (0.09) -0.09 (0.10)

3.75*** (0.16) 0.13 4,441

3.78*** (0.17) 0.12 4,437

-0.05 (0.23) 0.13 (0.22) 3.74*** (0.23) 0.13 4,441

4.01*** (0.09)

4.01*** (0.09)

3.99*** (0.10) 0.20 4,362

4.02*** (0.12) 0.18 4,362

4.05*** (0.13) 0.14 4,358

0.20 (0.19) 0.24 (0.19) 3.86*** (0.19) 0.22 4,362

4,441

4,331

4,362

4,258

Note: OLS coecients shown with robust standard errors in parenthesis (*** p < 0.01, ** p < 0.05, * p < 0.1). Regressions also include dummy variables for Income: missing, State: Missing, Sector: missing, and Trade Ties: do not know respectively (coecients not shown here). Reference categories for the respective dummy variable sets are: School: Lowest Tier; Income: Very Low; Age: 18-29; Sector Employment: Strong Decrease; Status: Full-Time Employed; Trade Ties: None; Sector: Nontradables. Results are weighted so that the education, age, and gender margins match the voter population (see text for details).

Table 3: Predictors of Opposition to Bailouts: Social Values and Political Orientation

Model No. Outcome Altruism: Medium Altruism: High Cosmopolitanism: Low Cosmopolitanism: Medium Cosmopolitanism: High Cosmopolitanism: Very High Vote: SPD Vote: Greens Vote: FDP Vote: Linke Vote: NPD/Reps Vote: Other Political Knowledge: General Political Knowledge: Specic Female School: Medium Tier School: Highest Tier University/College Income: Low Income: Lower Middle Income: Upper Middle Income: High Own Stocks Age: 30-39 Age: 40-49 Age: 50-59 Age: 60+ Constant Other Economic Controls State Fixed Eects Observations (1) (2) Against Bailouts -0.25*** -0.21*** (0.08) (0.07) -0.42*** -0.38*** (0.09) (0.08) -0.32*** -0.33*** (0.11) (0.10) -0.42*** -0.42*** (0.11) (0.10) -0.49*** -0.44*** (0.11) (0.10) -0.85*** -0.74*** (0.13) (0.13) 0.05 (0.10) 0.07 (0.09) 0.26* (0.15) 0.40*** (0.14) 1.01*** (0.13) 0.59*** (0.08) -0.20*** (0.06) -0.10* (0.05) 0.05 -0.01 (0.06) (0.06) 0.04 0.09 (0.07) (0.07) -0.28*** -0.13* (0.08) (0.08) -0.36*** -0.22*** (0.08) (0.08) 0.01 0.09 (0.09) (0.09) 0.01 0.13 (0.10) (0.10) -0.14 -0.03 (0.13) (0.11) -0.19 -0.04 (0.13) (0.12) -0.18*** -0.12** (0.06) (0.06) 0.14** 0.11 (0.07) (0.07) 0.25*** 0.23*** (0.07) (0.07) 0.22*** 0.24*** (0.07) (0.07) 0.09 0.14* (0.09) (0.08) 4.08*** 3.75*** (0.16) (0.17) (3) (1-5) -0.21*** (0.07) -0.39*** (0.08) -0.34*** (0.10) -0.43*** (0.11) -0.45*** (0.10) -0.75*** (0.13) 0.06 (0.10) 0.08 (0.08) 0.27* (0.15) 0.41*** (0.13) 1.00*** (0.13) 0.59*** (0.08) -0.21*** (0.06) -0.10* (0.05) -0.01 (0.06) 0.08 (0.07) -0.11 (0.08) -0.23*** (0.08) 0.05 (0.09) 0.08 (0.10) -0.09 (0.11) -0.09 (0.12) -0.12** (0.06) 0.04 (0.08) 0.16** (0.08) 0.16* (0.08) 0.10 (0.12) 3.86*** (0.22) (4) Pay -0.17*** (0.06) -0.30*** (0.07) -0.33*** (0.08) -0.44*** (0.08) -0.41*** (0.08) -0.56*** (0.09) (5) (6) In Less (1-5) -0.14** -0.14** (0.06) (0.06) -0.27*** -0.28*** (0.06) (0.06) -0.33*** -0.35*** (0.08) (0.08) -0.44*** -0.44*** (0.08) (0.08) -0.37*** -0.38*** (0.08) (0.08) -0.47*** -0.49*** (0.09) (0.09) -0.03 -0.03 (0.07) (0.07) 0.01 0.01 (0.07) (0.07) -0.01 -0.03 (0.13) (0.13) 0.20* 0.22** (0.11) (0.11) 0.65*** 0.67*** (0.09) (0.09) 0.38*** 0.37*** (0.06) (0.06) -0.15*** -0.16*** (0.04) (0.04) -0.06 -0.06 (0.04) (0.04) 0.06 0.05 (0.04) (0.05) 0.06 0.04 (0.05) (0.05) -0.13** -0.13** (0.06) (0.06) -0.24*** -0.26*** (0.06) (0.06) 0.09 0.06 (0.07) (0.07) -0.00 -0.03 (0.08) (0.07) 0.02 -0.02 (0.08) (0.08) -0.00 -0.05 (0.09) (0.08) -0.13*** -0.13*** (0.05) (0.04) 0.17*** 0.15** (0.05) (0.06) 0.18*** 0.16*** (0.06) (0.06) 0.25*** 0.23*** (0.06) (0.06) 0.06 0.11 (0.06) (0.10) 4.22*** 4.06*** (0.13) (0.19)

0.10** (0.05) 0.04 (0.05) -0.22*** (0.06) -0.33*** (0.06) 0.03 (0.08) -0.08 (0.08) -0.05 (0.09) -0.10 (0.09) -0.17*** (0.05) 0.20*** (0.06) 0.20*** (0.06) 0.23*** (0.06) 0.03 (0.07) 4.40*** (0.12)

4,350

4,350

4,350

4,281

4,281

4,281

Note: OLS coecients shown with robust standard errors in parenthesis (*** p < 0.01, ** p < 0.05, * p < 0.1). All regressions include state xed eects and models 3 & 6 also include our full set of economic controls as in model 6, Table 2 (including employment status, trade dependence, and sector employment change). Regressions also include dummy variables for Income: missing, and State: Missing respectively (coecients not shown here). Reference categories for the respective dummy variable sets are: Cosmopolitanism: Very low; Vote: CDU; School: Lowest Tier; Income: Very Low; Age: 18-29. Results are weighted so that the education, age, and gender margins match the voter population (see text for details).

37

Table 4: Predictors of Opposition to Bailouts: Political Knowledge

Model No. Outcome Sample: (2) (3) (4) Against Bailouts (1-5) General Knowledge Specic Knowledge Low High Low High -0.33*** (0.11) -0.39*** (0.10) -0.36*** (0.12) -0.49*** (0.12) -0.46*** (0.13) -0.74*** (0.20) 0.08 (0.15) 0.13 (0.12) 0.10 (0.22) 0.23 (0.17) 0.71*** (0.18) 0.53*** (0.11) 0.02 (0.08) 0.05 (0.08) -0.18* (0.10) -0.36*** (0.11) 0.17 (0.11) 0.20* (0.12) 0.02 (0.13) 0.08 (0.19) -0.27*** (0.08) 0.21** (0.10) 0.23** (0.10) 0.27*** (0.10) 0.24** (0.12) 3.66*** (0.22) 1,906 -0.07 (0.09) -0.33*** (0.11) -0.29** (0.15) -0.32** (0.16) -0.44*** (0.15) -0.75*** (0.16) -0.01 (0.13) -0.04 (0.11) 0.41** (0.20) 0.54*** (0.19) 1.29*** (0.15) 0.61*** (0.10) 0.01 (0.07) 0.15 (0.11) -0.09 (0.11) -0.12 (0.11) -0.05 (0.14) -0.03 (0.14) -0.14 (0.16) -0.18 (0.15) 0.01 (0.08) 0.02 (0.09) 0.26*** (0.09) 0.23** (0.10) 0.01 (0.11) 3.56*** (0.24) 2,444 -0.22** (0.11) -0.41*** (0.09) -0.32** (0.13) -0.44*** (0.13) -0.44*** (0.13) -0.57*** (0.15) -0.09 (0.16) 0.04 (0.12) 0.09 (0.24) 0.47*** (0.14) 0.76*** (0.17) 0.45*** (0.10) 0.01 (0.07) 0.06 (0.08) -0.24** (0.09) -0.42*** (0.10) 0.15 (0.11) 0.16 (0.11) -0.06 (0.12) 0.22 (0.15) -0.33*** (0.07) 0.18* (0.10) 0.24** (0.10) 0.23** (0.11) 0.18 (0.13) 3.75*** (0.21) 1,917 -0.22** (0.09) -0.36*** (0.11) -0.35** (0.15) -0.39*** (0.15) -0.47*** (0.14) -0.86*** (0.17) 0.18 (0.13) 0.11 (0.12) 0.52*** (0.17) 0.39** (0.20) 1.31*** (0.16) 0.75*** (0.11) 0.01 (0.08) 0.08 (0.10) -0.11 (0.11) -0.17 (0.11) -0.01 (0.14) 0.07 (0.15) -0.09 (0.17) -0.23 (0.17) 0.02 (0.08) 0.04 (0.09) 0.23** (0.09) 0.24** (0.10) 0.12 (0.11) 3.51*** (0.24) 2,433 (1) (6) (7) (8) Pay In Less (1-5) General Knowledge Specic Knowledge Low High Low High -0.20** (0.08) -0.14* (0.07) -0.39*** (0.09) -0.49*** (0.10) -0.33*** (0.10) -0.45*** (0.12) 0.02 (0.10) 0.14 (0.11) 0.04 (0.15) 0.28** (0.13) 0.61*** (0.12) 0.45*** (0.09) 0.04 (0.06) 0.02 (0.06) -0.15* (0.08) -0.31*** (0.08) 0.08 (0.09) 0.01 (0.09) 0.06 (0.10) 0.00 (0.13) -0.21*** (0.06) 0.26*** (0.08) 0.25*** (0.08) 0.33*** (0.08) 0.22** (0.09) 3.97*** (0.16) 1,861 -0.08 (0.07) -0.35*** (0.08) -0.31*** (0.11) -0.39*** (0.12) -0.40*** (0.11) -0.52*** (0.12) -0.06 (0.09) -0.12 (0.08) -0.13 (0.24) 0.16 (0.16) 0.71*** (0.13) 0.32*** (0.08) 0.06 (0.06) 0.11 (0.09) -0.14 (0.09) -0.20** (0.09) 0.04 (0.11) -0.05 (0.11) -0.03 (0.12) -0.02 (0.12) -0.06 (0.06) 0.09 (0.07) 0.13* (0.07) 0.16** (0.08) -0.14 (0.09) 4.27*** (0.18) 2,420 -0.08 (0.08) -0.16* (0.08) -0.31*** (0.10) -0.38*** (0.11) -0.31*** (0.11) -0.45*** (0.13) -0.20* (0.12) -0.03 (0.10) -0.04 (0.16) 0.24* (0.13) 0.53*** (0.13) 0.34*** (0.09) 0.04 (0.06) -0.01 (0.06) -0.19** (0.07) -0.34*** (0.08) 0.15 (0.09) 0.09 (0.09) 0.02 (0.10) 0.16 (0.11) -0.26*** (0.06) 0.25*** (0.08) 0.26*** (0.08) 0.29*** (0.08) 0.14 (0.10) 4.05*** (0.17) 1,879 -0.22*** (0.07) -0.34*** (0.08) -0.38*** (0.11) -0.51*** (0.11) -0.47*** (0.11) -0.55*** (0.11) 0.11 (0.08) 0.09 (0.09) 0.05 (0.22) 0.21 (0.16) 0.75*** (0.12) 0.46*** (0.07) 0.09 (0.06) 0.12 (0.08) -0.14* (0.08) -0.23*** (0.08) 0.04 (0.10) -0.05 (0.11) 0.01 (0.12) -0.07 (0.11) -0.04 (0.06) 0.11 (0.07) 0.11 (0.07) 0.19*** (0.07) -0.01 (0.08) 4.17*** (0.17) 2,402 (5)

Altruism: Medium Altruism: High Cosmopolitanism: Low Cosmopolitanism: Medium Cosmopolitanism: High Cosmopolitanism: Very high Vote: SPD Vote: Green Vote: FDP Vote: Linke Vote: NPD/Reps Vote: Other Female School: Medium Tier School: Highest Tier University/College Income: Low Income: Lower Middle Income: Upper Middle Income: High Own Stocks Age: 30-39 Age: 40-49 Age: 50-59 Age: 60+ Constant State Fixed Eects Observations

Note: OLS coecients shown with robust standard errors in parenthesis (*** p < 0.01, ** p < 0.05, * p < 0.1). Models 1 & 2 and 5 & 6 use the subsample of respondents with low and high general political knowledge respectively; Models 3 & 4 and 7 & 8 use the subsample of respondents with low and high specic political knowledge respectively. All regressions include state xed eects. Regressions also include dummy variables for Income: missing, and State: Missing respectively (coecients not shown here). Reference categories for the respective dummy variable sets are: Cosmopolitanism: Very low; Vote: CDU; School: Lowest Tier; Income: Very Low; Age: 18-29. Results are weighted so that the education, age, and gender margins match the voter population (see text for details).

38

Table 5: Predictors of Opposition to Bailouts: Social Values and Political Orientation (Quasi Behavioral Measure)
Model No. Outcome Altruism: Medium Altruism: High Cosmopolitanism: Low Cosmopolitanism: Medium Cosmopolitanism: High Cosmopolitanism: Very High Vote: SPD Vote: Green Vote: FDP Vote: Linke Vote: NPD/Reps Vote: Other Political Knowledge: General Political Knowledge: Specic Female School: Medium Tier School: Highest Tier University/College Income: Low Income: Lower Middle Income: Upper Middle Income: High Own Stocks Age: 30-39 Age: 40-49 Age: 50-59 Age: 60+ Constant Other Economic Controls State Fixed Eects Observations 0.09 (0.06) -0.01 (0.06) -0.21*** (0.08) -0.37*** (0.08) 0.01 (0.09) -0.02 (0.11) 0.14 (0.11) 0.00 (0.13) -0.15** (0.06) 0.15* (0.07) 0.16** (0.08) 0.29*** (0.08) -0.00 (0.09) 4.35*** (0.14) (1) Petition -0.14* (0.07) -0.45*** (0.11) -0.25*** (0.10) -0.43*** (0.10) -0.39*** (0.10) -0.42*** (0.13) (2) (3) MP against contribution (1-5)) -0.12* -0.12 (0.07) (0.07) -0.41*** -0.43*** (0.10) (0.10) -0.25*** -0.25*** (0.09) (0.09) -0.41*** -0.40*** (0.10) (0.09) -0.35*** -0.35*** (0.10) (0.09) -0.34*** -0.34*** (0.13) (0.12) -0.04 -0.04 (0.09) (0.09) 0.04 0.06 (0.08) (0.08) -0.29* -0.30* (0.16) (0.16) 0.31** 0.31** (0.14) (0.13) 0.59*** 0.60*** (0.11) (0.11) 0.40*** 0.39*** (0.08) (0.08) -0.11** -0.13** (0.06) (0.05) -0.04 -0.03 (0.05) (0.05) 0.07 0.09 (0.06) (0.06) -0.01 -0.02 (0.06) (0.06) -0.15** -0.15* (0.08) (0.08) -0.31*** -0.34*** (0.08) (0.08) 0.06 0.05 (0.09) (0.09) 0.06 0.03 (0.10) (0.10) 0.20** 0.17* (0.10) (0.10) 0.11 0.08 (0.12) (0.12) -0.10* -0.10* (0.06) (0.06) 0.12* 0.10 (0.07) (0.08) 0.13* 0.11 (0.07) (0.08) 0.28*** 0.29*** (0.08) (0.08) 0.03 0.18 (0.08) (0.12) 4.15*** 4.06*** (0.15) (0.22)

2,695

2,695

2,695

Note: OLS coecients shown with robust standard errors in parenthesis (*** p < 0.01, ** p < 0.05, * p < 0.1). All regressions include state xed eects and models 3 & 6 also include our full set of economic controls as in model 6, Table 2 (including employment status, trade dependence, and sector employment change). Regressions also include dummy variables for Income: missing, and State: Missing respectively (coecients not shown here). Reference categories for the respective dummy variable sets are: Cosmopolitanism: Very low; Vote: CDU; School: Lowest Tier; Income: Very Low; Age: 18-29. Results are weighted so that the education, age, and gender margins match the voter population (see text for details).

39

Table 6: Comparison of Attitudes Towards Financial Bailouts in Online and Phone Sample Group: All Respondents Sample: Online Phone Outcome: Against bailout strongly in favour 3.0% 3.7% somewhat in favour 24.5% 28.7% neither in favour nor against 10.2% 21.6% somewhat against 40.1% 28.5% strongly against 20.9% 15.0% dont know 1.5% 2.5% Outcome: Pay in less pay in much more pay in somewhat more pay neither more or less pay in somewhat less pay in much less dont know w/o Neither Online Phone 3.3% 27.7% 45.3% 23.6% 4.9% 37.8% 37.6% 19.8%

0.5% 4.0% 25.4% 34.2% 32.6% 3.4%

0.9% 3.0% 34.6% 40.2% 19.1% 2.3%

0.7% 5.6% 48.0% 45.8%

1.4% 4.8% 63.6% 30.2%

Note: N=4,499 for the online sample and N=1,002 for the phone sample. Both samples are are weighted by sample adjustment weights so that the education, age, and gender margins match the total voter population (see text for details).

40

Figures Figure 1: The Correlates of Preferences for Financial Bailouts

DV: Against bailout

DV: Pay in less

DV: Petition MP against contribution

Female High School: Lowest Tier High School: Medium Tier High School: Highest Tier High School: University Age 1829 Age 3039 Age 4049 Age 5059 Age 60+ BadenWrttemberg Bayern Berlin Brandenburg Bremen Hamburg Hesse MecklenburgVorpommern Lower Saxony North RhineWestphalia RhinelandPalatinate Saarland Saxony SaxonyAnhalt SchleswigHolstein Thuringia HH Income: Lowest HH Income: Low HH Income: Lower Middle HH Income: Upper Middle HH Income: High Own Stocks Sector Employment: Large Decrease Sector Employment: Decrease Sector Employment: Increase Sector Employment: Large Increase Status: Fulltime employed Status: Parttime employed Status: In Education Status: Unemployed Status: Retired Trade Ties: None Trade Ties: Weak Trade Ties: Strong Trade Ties: Very Strong Altruism: Low Altruism: Medium Altruism: High Cosmopolitanism: Very Low Cosmopolitanism: Low Cosmopolitanism: Medium Cosmopolitanism: High Cosmopolitanism: Very High Vote: CDU/CSU Vote: SPD Vote: Greens Vote: FDP Vote: Linke Vote: Reps/NPD Vote: Other Political Knowledge: General Political Knowledge: Specific

q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q

q q q q q q q q q q q q q

q q q

q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q q

Predictors
q q q q q

SocioDemographics State Economic Interest Social Values Political Orientation

0.4

0.2

0.0

0.2

0.4

0.4

0.2

0.0 0.2 0.4 Change in Pr(Oppose Bailouts)

0.4

0.2

0.0

0.2

0.4

Note: The gure shows marginal eects from a linear probability model with opposition to bailouts as the binary outcome variable which equals 1 if a respondent (strongly) opposes bailouts and is 0 otherwise (neither, somewhat in favor, and strongly in favor). In the middle panel, the dependent variable (DV) is a binary indicator variable which equals 1 if a respondent wants Germany to pay in (much) less into the European bailout fund and is 0 otherwise (neither, pay in more, pay in much more). In the right panel, the dependent variable (DV) is a binary indicator variable which equals 1 if a respondent wants to send an identied message to the MPs saying that Germany should pay in (much) less into the European bailout fund and is 0 otherwise (neither, pay in more, pay in much more). Horizontal lines indicate .95 condence intervals. The unconditional baseline probabilities and numbers of observations are: Against bailouts = .57, N = 4, 899; Pay in less = .64, N = 4, 409; Petition MP against contribution = .68, N = 2, 761.

41

Figure 2: The Eects of Increases in the Size of Contributions to Financial Bailouts: Experimental Results

q q q

All Respondents

q q q

HH Income: Low

Germany's Contribution:
q q q q q q q

Eur 123B Eur 189B Eur 211B Eur 418B

HH Income: Middle

q q q

HH Income: High

0.00

0.05

0.10 0.15 Effect on Pr(Vote Against Bailout)

0.20

Note: The gure shows marginal eects from a linear probability model. The dependent variable, opposition to bailouts, is a binary indicator variable which equals 1 if a respondent (strongly) disapproves bailouts and is 0 otherwise (neither, somewhat approve, and strongly approve). Horizontal lines indicate .95 condence intervals. The unconditional baseline probabilities are: all=0.46, high income=0.45, middle income=0.46, low income=0.46. N = 4, 364.

42

Appendix A Variable Denitions

Against Bailouts : Measures opposition to/support for bailout payments for over-indebted EU countries. Question wording: Generally speaking, how strongly do you approve or disapprove of nancial bailouts for over-indebted EU countries? Answer categories: 1=very much in favor, 2=somewhat in favor, 3=neither/nor, 4=somewhat against, 5=very much against. Pay In Less : Measures opposition to/support for paying in less into the European rescue fund for over-indebted EU countries. Question wording: Should Germany pay in more or less money into the European nancial rescue fund used to help over-indebted EU countries? Do you think Germany should pay in much more, pay in somewhat more, pay in neither more nor less, pay in somewhat less, or pay in much less? Answer categories: 1=pay in much more, 2=pay in more, 3=neither/nor, 4=pay in less, 5=pay in much less. Petition MP against contribution : Question wording: Should we inform the Members of Parliament on your behalf whether you want rather more or rather less payments of Germany into the European nancial rescue fund? This information notice would contain your name and location. The answer categories were: 1=Inform the MPs on my behalf that Germany should pay a lot more, 2Inform the MPs on my behalf that Germany should pay a somewhat more, 3=Inform the MPs on my behalf that Germany should pay neither more nor less, 4=Inform the MPs on my behalf that Germany should pay somewhat less, 5=Inform the MPs on my behalf that Germany should pay a lot more. Respondents could alternatively choose the please send no message option. Vote against bailout : Question wording: How likely would you vote in favor or against Germany contributing eX bn to the bailout fund? The size of Germanys reported contribution X was chosen randomly (123, 289, 211, or 418). Answer categories: 1=vote surely against, 2=vote very likely against, 3=vote probably against, 4=neither/nor, 5=vote probably in favor, 6=vote very likely in favor, 7=vote surely in favor. Education : Measures respondents highest level of completed education. Converted into dummy variables that distinguish between high school-medium tier, high school-highest tier, and university/college. HH Income : Self-reported household income. Converted into dummy variables that identify low, lower middle, upper middle, and high. Own Stocks : Binary indicator variable that distinguishes respondents that reported to currently hold stocks (including index funds) and those that said that they do not hold stocks. Sector Trade Dependence : Sum of the value of exports and imports divided by the total value of production in respondents sector of employment (coded at the two-digit GP2009 level). Source: German statistical oce. Sector Employment : Change in annual employment (2009 vs 2010) in respondents sector of employment (coded at the two-digit GP2009 level). Source: German statistical oce. Sector: Nontradables : Binary indicator that distinguishes respondents working in the nontradable goods sector (based on a coding of industries at the one-digit WZ2008-level).

43

 Trade Ties : Measures respondents self reported level of trade ties with other countries. Question wording: How strong or weak are the business ties between your rm with rms and/or customers in other EU countries? Answer categories: 1=no ties, 2=weak ties, 3=strong ties, 4=very strong ties. Converted into binary indicators for weak, strong, and very strong trade ties. Reference groups is no trade tries. Transfers: Net Beneciary : Binary indicator that distinguishes between respondents living in a state that is a net beneciary of regional transfers (L andernanzausgleich ) and respondents that live in a state that is a net contributor. Altruism : Respondents degree of altruism as measured by the willingness to donate a share of a e100 voucher raed o among all survey participants. Question wording: We rae o a e100 Amazon voucher among all respondents. You can decide to donate a part of this voucher to a charity of your own choosing. If you win the voucher, your donation will be deducted from the value of your voucher. Would you like to donate a share of your voucher? Respondents that wanted to donate could then choose from a menu of 30 charities and indicate the amount they wanted to donate (allowing for any integer value between 0 and 100). Converted into binary indicators for medium levels (donation between 1 and 50%) and high levels of altruism (donation greater than 50%). Respondents that did not want to donate form the reference group. Cosmopolitanism : Measures extent to which respondents think that national and international happenings are more or less interesting as events that occur within their local community. Question wording: How strongly do you approve or disapprove the following statement? Although the media often reports about national and international events and developments, this news is seldom as interesting as the things that happen directly in our own community and neighborhood. Answer categories: 1=strongly disapprove, 2=somewhat disapprove, 3=neither/nor, 4=somewhat approve, 5=strongly approve. Converted into three binary indicators that identify respondents with low, medium, and high levels of cosmopolitanism. No cosmopolitanism forms the reference group. Vote: Party : Records which party respondents would vote for in German federal elections. Converted into binary indicator variables that distinguish between voters of the SPD, Greens, FDP, Linke, NPD/Reps, and Other. CDU voters form the reference category. Political Knowledge: General : Binary indicator that discriminates between respondents knowing that the PR vote determines the share of seats in the Bundestag, the German parliament and those that did not know this. Political Knowledge: Specic : Binary indicator that distinguishes respondents that correctly identify at least two countries that received money from the European bailout fund while not marking any of the countries that did not receive bailout payments or the do not know category. The list of countries is Portugal, Ireland, Greece, Slovakia, Netherlands, France, and do not know.

44

Descriptive Statistics Table A.1: Descriptive Statistics (un-weighted)

Variable Against Bailouts Pay In Less Pay in Less (Email to MP) High School: Medium Tier High School: Highest Tier University/College Age: 30-39 Age: 40-49 Age: 50-59 Age: 60+ Female Status: Part-time employed Status: In education Status: Unemployed Status: Retired HH Income: Low HH Income: Lower Middle HH Income: Upper Middle HH Income: High HH Income: Not reported Own Stocks Trade Ties: Weak Trade Ties: Strong Trade Ties: Very Strong Industry: Log Trade Dependence Industry: Change in Employment Industry: Not reported Industry: Nontradables Altruism: Medium Altruism: High Political Knowledge: General Political Knowledge: Specic Cosmopolitanism: Low Cosmopolitanism: Medium Cosmopolitanism: High Cosmopolitanism: Very High Vote: SPD Vote: Greens Vote: FDP Vote: Linke Vote: NPD/Reps Vote: Other Obs 4899 4409 2761 4499 4499 4499 4977 4977 4977 4977 5007 4522 4522 4522 4522 4510 4510 4510 4510 4510 5007 4513 4513 4513 5007 5007 5007 5007 4488 4488 5007 5007 4463 4463 4463 4463 5007 5007 5007 5007 5007 5007 Mean 3.39 3.89 3.93 0.32 0.29 0.28 0.21 0.26 0.19 0.09 0.40 0.16 0.14 0.06 0.09 0.25 0.24 0.15 0.11 0.12 0.29 0.18 0.14 0.13 0.39 0.03 0.26 0.92 0.17 0.17 0.50 0.50 0.25 0.26 0.29 0.14 0.17 0.15 0.02 0.07 0.03 0.31 Std. Dev. 1.18 0.93 0.96 0.47 0.46 0.45 0.40 0.44 0.39 0.28 0.49 0.36 0.35 0.24 0.29 0.43 0.43 0.36 0.31 0.32 0.46 0.38 0.34 0.34 1.34 0.04 0.44 0.27 0.38 0.38 0.50 0.50 0.43 0.44 0.45 0.34 0.38 0.36 0.14 0.25 0.16 0.46 Min 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -0.08 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Max 5 5 5 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 7.74 0.13 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

45

Appendix B Table B.1: Demographics of the Survey Samples (in %) Raw Weighted Group Voter Online Phone Online Phone Population Sample Sample Sample Sample High School Low Tier 43.8 10.9 18.2 43.7 43.8 High School Medium Tier 25.7 32.1 37.9 25.7 25.7 High School High Tier 14.5 29.4 17.4 14.6 16.0 University/College 16.1 27.6 26.5 16.1 14.5 Age 19-29 14.8 25.0 14.0 14.9 14.8 Age 30-39 14.9 20.5 14.9 14.9 14.9 Age 40-49 20.5 26.2 18.2 20.5 20.5 Age 50-59 17.4 19.1 14.9 17.4 17.4 Age 60+ 32.5 8.6 32.6 32.3 32.4 Female 51.1 39.6 53.3 51.1 51.1
Note: See appendix B for a description of the variables. Data on the voter population are obtained from the German statistical oce (http://www.destatis.de) for the year 2010.

46

Table B.2: Opposition to Bailouts: Personal Economic Interest (Quasi Behavioral Measure)
Model No. Outcome Female School: Medium Tier School: Highest Tier University/College Income: Low Income: Lower Middle Income: Upper Middle Income: High Own Stocks Age: 30-39 Age: 40-49 Age: 50-59 Age: 60+ Transfers: Net Beneciary State: Bavaria State: Berlin State: Brandenburg State: Bremen State: Hamburg State: Hesse State: Mecklenburg-Vorpommern State: Lower Saxony State: North Rhine-Westphalia State: Rhineland-Palatinate State: Saarland State: Saxony State: Saxony-Anhalt State: Schleswig-Holstein State: Thuringia Sector Employment: Decrease Sector Employment: Increase Sector Employment: Strong Increase Status: Part-time employed Status: In Education Status: Unemployed Status: Retired Trade Ties: Weak Trade Ties: Strong Trade Ties: Very Strong Sector: Medium Trade Dependence Sector: Strong Trade Dependence Constant Observations 4.02*** (0.11) 2,745 3.97*** (0.12) 2,675 3.98*** (0.14) 2,745 4.04*** (0.17) 2,745 4.04*** (0.19) 2,743 (1) 0.08 (0.06) -0.02 (0.06) -0.26*** (0.07) -0.41*** (0.08) -0.02 (0.10) -0.04 (0.12) 0.08 (0.11) -0.04 (0.13) -0.19*** (0.06) 0.19** (0.07) 0.19** (0.07) 0.27*** (0.08) -0.06 (0.09) (3) (4) (5) Pay In Less, Email to MP (1-5)) 0.08 0.09 0.11 0.10 (0.06) (0.06) (0.07) (0.07) -0.03 -0.03 -0.05 -0.04 (0.06) (0.06) (0.06) (0.06) -0.28*** -0.27*** -0.25*** -0.23*** (0.07) (0.07) (0.08) (0.08) -0.42*** -0.42*** -0.45*** -0.44*** (0.08) (0.08) (0.08) (0.08) -0.01 0.00 -0.02 -0.01 (0.10) (0.10) (0.10) (0.10) -0.03 -0.02 -0.05 -0.04 (0.12) (0.11) (0.11) (0.11) 0.08 0.11 0.07 0.08 (0.11) (0.11) (0.11) (0.11) -0.05 -0.05 -0.09 -0.06 (0.13) (0.13) (0.13) (0.13) -0.18*** -0.18*** -0.17*** -0.15** (0.06) (0.06) (0.06) (0.06) 0.17** 0.16** 0.11 0.11 (0.08) (0.07) (0.08) (0.08) 0.17** 0.15** 0.11 0.10 (0.08) (0.08) (0.08) (0.08) 0.26*** 0.27*** 0.24*** 0.24*** (0.08) (0.08) (0.09) (0.09) -0.06 -0.08 0.01 0.01 (0.10) (0.10) (0.15) (0.15) 0.08 (0.07) -0.01 -0.01 -0.02 (0.13) (0.13) (0.13) 0.14 0.13 0.10 (0.12) (0.12) (0.12) -0.01 -0.04 -0.06 (0.17) (0.18) (0.18) -0.43 -0.40 -0.41 (0.29) (0.28) (0.29) -0.05 -0.06 -0.06 (0.18) (0.18) (0.18) -0.11 -0.12 -0.13 (0.11) (0.11) (0.12) 0.11 0.11 0.12 (0.19) (0.18) (0.18) 0.01 0.00 -0.01 (0.12) (0.11) (0.12) 0.02 0.01 0.01 (0.11) (0.10) (0.10) 0.18 0.17 0.17 (0.14) (0.13) (0.13) -0.37 -0.35 -0.37 (0.23) (0.24) (0.23) 0.10 0.07 0.07 (0.13) (0.13) (0.13) -0.20 -0.17 -0.18 (0.21) (0.20) (0.21) 0.34** 0.32** 0.32** (0.15) (0.15) (0.15) 0.17 0.16 0.14 (0.16) (0.16) (0.16) 0.01 -0.00 (0.11) (0.11) 0.11 0.11 (0.10) (0.10) 0.08 0.07 (0.10) (0.10) -0.03 -0.04 (0.07) (0.07) -0.18 -0.19 (0.18) (0.19) -0.05 -0.07 (0.11) (0.11) -0.19 -0.21 (0.14) (0.14) -0.09 (0.08) 0.02 (0.09) -0.08 (0.11) (2) (6) 0.11* (0.06) -0.05 (0.06) -0.25*** (0.08) -0.45*** (0.08) -0.02 (0.10) -0.05 (0.11) 0.07 (0.11) -0.09 (0.13) -0.17*** (0.06) 0.11 (0.08) 0.11 (0.08) 0.24*** (0.09) 0.02 (0.14)

-0.01 (0.13) 0.13 (0.12) -0.03 (0.18) -0.39 (0.27) -0.05 (0.18) -0.12 (0.11) 0.11 (0.19) 0.00 (0.11) 0.02 (0.11) 0.17 (0.13) -0.34 (0.24) 0.08 (0.13) -0.15 (0.20) 0.33** (0.15) 0.16 (0.16) 0.10 (0.23) 0.20 (0.22) 0.17 (0.22) -0.03 (0.07) -0.18 (0.18) -0.05 (0.11) -0.21 (0.13)

0.13 (0.25) 0.12 (0.23) 3.93*** (0.25) 2,745

Note: OLS coecients with robust standard errors (*** p < 0.01, ** p < 0.05, * p < 0.1). Regressions include dummy variables for Income: missing, State: Missing, Sector: missing, and Trade Ties: do not know respectively. Reference categories for the respective dummy variable sets are: School: Lowest Tier; Income: Very Low; Age: 18-29; Sector Employment: Strong Decrease; Status: Full-Time Employed; Trade Ties: None; Sector: Nontradables. Results are weighted on education, age, and gender margins (see text for details).

47

Table B.3: Measure)

Predictors of Opposition to Bailouts: Political Knowledge (Quasi Behavioral

Model No. Outcome Sample: Altruism: Medium Altruism: High Cosmopolitanism: Low Cosmopolitanism: Medium Cosmopolitanism: High Cosmopolitanism: Very High Vote: SPD Vote: Green Vote: FDP Vote: Linke Vote: NPD/Reps Vote: Other Female School: Medium Tier School: Highest Tier University/College Income: Low Income: Lower Middle Income: Upper Middle Income: High Own Stocks Age: 30-39 Age: 40-49 Age: 50-59 Age: 60+ Constant State Fixed Eects Observations

(2) (3) (4) Pay In Less, Email to MP (1-5) General Knowledge Specic Knowledge Low High Low High -0.12 -0.16* -0.00 -0.26*** (0.10) (0.08) (0.10) (0.08) -0.17 -0.57*** -0.10 -0.57*** (0.11) (0.13) (0.10) (0.13) -0.29** -0.25* -0.27** -0.27** (0.11) (0.13) (0.13) (0.12) -0.52*** -0.25* -0.48*** -0.33*** (0.12) (0.14) (0.13) (0.12) -0.31*** -0.37*** -0.45*** -0.27** (0.12) (0.13) (0.13) (0.13) -0.34* -0.33** -0.33** -0.28* (0.20) (0.14) (0.16) (0.16) -0.13 0.07 -0.16 0.06 (0.13) (0.12) (0.14) (0.10) 0.13 0.01 0.06 0.10 (0.11) (0.10) (0.12) (0.10) -0.41** -0.25 -0.50*** -0.15 (0.19) (0.29) (0.19) (0.29) 0.31** 0.38** 0.33** 0.33* (0.15) (0.18) (0.13) (0.18) 0.35** 0.81*** 0.35** 0.93*** (0.17) (0.14) (0.15) (0.13) 0.29*** 0.61*** 0.28*** 0.60*** (0.10) (0.10) (0.11) (0.09) 0.09 0.05 0.09 0.07 (0.07) (0.08) (0.07) (0.08) -0.03 0.02 -0.11 0.12 (0.08) (0.09) (0.08) (0.09) -0.18* -0.13 -0.24*** -0.08 (0.10) (0.10) (0.09) (0.10) -0.43*** -0.25** -0.54*** -0.20* (0.11) (0.10) (0.10) (0.11) -0.04 0.14 0.02 0.10 (0.09) (0.16) (0.10) (0.15) -0.01 0.13 0.11 0.00 (0.11) (0.17) (0.11) (0.16) 0.04 0.31* 0.08 0.22 (0.11) (0.17) (0.11) (0.16) 0.05 0.19 0.18 0.04 (0.17) (0.18) (0.13) (0.18) -0.18** -0.01 -0.17** -0.07 (0.08) (0.07) (0.08) (0.07) 0.11 0.10 0.12 0.14 (0.10) (0.09) (0.11) (0.09) 0.11 0.14 0.13 0.13 (0.10) (0.10) (0.10) (0.10) 0.26*** 0.26*** 0.23** 0.30*** (0.10) (0.10) (0.11) (0.10) 0.04 -0.08 -0.01 0.06 (0.12) (0.11) (0.13) (0.10) 4.20*** 3.91*** 4.17*** 3.96*** (0.18) (0.24) (0.18) (0.21) 1,270 1,425 1,223 1,472

(1)

Note: OLS coecients shown with robust standard errors in parenthesis (*** p < 0.01, ** p < 0.05, * p < 0.1). Models 1 & 2 use the subsample of respondents with low and high general political knowledge respectively; Models 3 & 4 use the subsample of respondents with low and high specic political knowledge respectively. All regressions include state xed eects. Regressions also include dummy variables for Income: missing, and State: Missing respectively (coecients not shown here). Reference categories for the respective dummy variable sets are: Cosmopolitanism: Very low; Vote: CDU; School: Lowest Tier; Income: Very Low; Age: 18-29. Results are weighted so that the education, age, and gender margins match the voter population (see text for details).

48

Table B.4: Predictors of Opposition to Bailouts: Online vs Phone Survey

Model Outcome Sample Reported: Phone Age: 30-39 Age: 40-49 Age: 50-59 Age: 60+ Phone Age: 30-39 Phone Age: 40-49 Phone Age: 50-59 Phone Age: 60+ High School: Medium Tier High School: Highest Tier University Phone High School: Medium Tier Phone High School: Highest Tier Phone University HH Income: Low HH Income: Lower Middle HH Income: Upper Middle HH Income: High HH Income: Not reported Phone HH Income: Low Phone HH Income: Lower Middle Phone HH Income: Upper Middle Phone HH Income: High Phone HH Income: Not reported Female Phone Female Status: Part-time employed Status: In education Status: Unemployed Status: Retired Phone Status: Part-time employed Phone Status: In education Phone Status: Unemployed Phone Status: Retired Constant Observations p-value: joint test of interactions terms (1) (2) Against Bailouts (1-5) Weighted Unweighted coef se coef se -0.29 (0.32) -0.41* (0.25) 0.08 (0.08) 0.06 (0.06) 0.12 (0.08) 0.17*** (0.06) 0.09 (0.08) 0.07 (0.06) -0.11 (0.13) -0.04 (0.09) -0.01 (0.19) 0.03 (0.16) -0.01 (0.18) -0.11 (0.16) -0.28 (0.18) -0.25 (0.17) 0.08 (0.24) 0.05 (0.19) -0.02 (0.07) -0.05 (0.06) -0.38*** (0.08) -0.41*** (0.07) -0.57*** (0.08) -0.64*** (0.06) -0.17 (0.13) -0.12 (0.12) -0.03 (0.15) 0.03 (0.15) -0.07 (0.14) 0.03 (0.14) 0.01 (0.11) 0.02 (0.06) -0.02 (0.11) -0.03 (0.06) -0.27** (0.13) -0.13* (0.07) -0.21 (0.13) -0.14 (0.09) -0.06 (0.12) 0.05 (0.07) 0.23 (0.27) 0.34* (0.19) 0.04 (0.27) 0.16 (0.19) 0.07 (0.29) 0.02 (0.21) 0.18 (0.32) 0.24 (0.25) 0.23 (0.28) 0.20 (0.19) 0.06 (0.07) 0.05 (0.04) -0.07 (0.11) -0.03 (0.09) -0.03 (0.07) -0.03 (0.05) -0.32*** (0.10) -0.34*** (0.07) -0.02 (0.11) 0.02 (0.08) -0.03 (0.13) -0.01 (0.08) -0.02 (0.15) 0.04 (0.13) -0.09 (0.21) -0.08 (0.18) -0.20 (0.19) -0.14 (0.16) 0.01 (0.23) -0.03 (0.16) 3.72*** (0.12) 3.70*** (0.09) 5,392 5,392 0.66 0.35 (4) Pay In Less (1-5) Weighted Unweighted coef se coef se 0.08 (0.22) 0.00 (0.19) 0.16** (0.06) 0.06 (0.05) 0.13** (0.06) 0.10** (0.04) 0.15** (0.07) 0.05 (0.05) -0.05 (0.11) -0.09 (0.07) -0.16 (0.14) 0.01 (0.12) -0.15 (0.14) -0.13 (0.12) -0.38*** (0.14) -0.23* (0.13) -0.18 (0.19) 0.02 (0.15) 0.01 (0.05) -0.07 (0.05) -0.28*** (0.06) -0.35*** (0.05) -0.45*** (0.06) -0.52*** (0.05) -0.19** (0.09) -0.11 (0.10) -0.08 (0.11) 0.02 (0.11) -0.05 (0.11) 0.03 (0.11) 0.00 (0.08) -0.02 (0.05) -0.15* (0.09) -0.10* (0.05) -0.17* (0.09) -0.10* (0.05) -0.16* (0.10) -0.11 (0.07) -0.04 (0.09) -0.00 (0.06) -0.05 (0.18) -0.03 (0.15) 0.01 (0.18) -0.07 (0.15) -0.01 (0.20) -0.14 (0.16) 0.21 (0.23) 0.05 (0.20) 0.05 (0.18) -0.08 (0.15) 0.11** (0.05) 0.07** (0.03) -0.15* (0.09) -0.07 (0.07) -0.02 (0.06) -0.05 (0.04) -0.19** (0.08) -0.19*** (0.05) -0.06 (0.10) -0.01 (0.06) -0.10 (0.11) -0.03 (0.06) -0.08 (0.12) -0.08 (0.10) -0.14 (0.17) -0.17 (0.14) 0.05 (0.15) -0.05 (0.12) 0.12 (0.18) -0.15 (0.13) 4.10*** (0.10) 4.18*** (0.07) 5,309 5,309 0.15 0.25 (3)

Note: OLS coecients shown with robust standard errors in parenthesis (*** p<0.01, ** p<0.05, * p<0.1). Dataset pools observations from the phone and the online survey to examine if the eect of the predictors dier by survey model. Models 1 and 3 are weighted by the sample adjustment weights so that the education, age, and gender margins match the total voter population (see text for details); Models 2 and 4 are unweighted. Reference categories for the dummy variable sets are: HH Income: Very Low. The p-value in the last row refers to an F-test against the null that the coecients on the interactions terms are jointly equal to zero.

49