Brexit Domino? The Political Contagion Effects of Voter-endorsed Withdrawals from International Institutions

This article examines the systemic implications of the growing popular backlash against international cooperation and analyzes how voter-endorsed attempts to withdraw from international institutions reverberate abroad. Observing other countries’ disintegration experiences allows voters to better assess the feasibility and desirability of such withdrawals. More positive withdrawal experiences encourage exit-support abroad, whereas negative experiences are likely to have a deterring effect. These contagion effects will be conditioned by the availability of information and voters’ willingness to learn. The article empirically examines this argument for the case of Brexit. It leverages original survey data from 58,959 EU-27 Europeans collected in six survey waves during the Brexit withdrawal negotiations and from a two-wave survey of 2,241 Swiss voters conducted around the first Brexit extension in spring 2019. It finds both encouragement and deterrence effects, which are bigger when respondents pay attention to Brexit and are dampened by motivated reasoning.


Introduction
International institutions have become increasingly contested in the past years. Institutions as diverse as the EU, the Paris Climate Agreement, or international courts have become salient and polarizing issues in national public debates. Efforts to not only slow down, but to reverse international integration have proliferated. The most prominent example of this phenomenon is Brexit, the United Kingdom's decision to leave the European Union (EU). But other examples include the US's 2020 withdrawal from the Open Skies Treaty, Burundi's 2017 and the Philippines' 2019 withdrawals from the International Criminal Court (ICC), or the increasing rate of investment treaty terminations (Huikuri, 2020). Although skepticism about the merits of international cooperation, exits from international treaties, or even dissolutions of international organizations are nothing new, the frequency with which they manifest themselves has increased in recent years .
The spread of non-cooperative, or even disintegrative, tendencies is widely seen as a threat to international institutions and international cooperation more generally. Increasingly, these tendencies are endorsed by voters through referendums or the election of parties and candidates who make noncooperation a centerpiece of their policy agenda. Against this backdrop, we need to better understand how such attempts to revert or undermine international institutions spread, how they can be contained, and which dynamics they produce in the international arena. Whereas there is vast research on the creation and functioning of international institutions and the integration process more generally, the causes, dynamics, and consequences of international dis-integration are not yet well understood (e.g., Jones, 2018;Schneider, 2017;Vollaard, 2014). A few studies examine under what circumstances states withdraw from international institutions (Helfer, 2005(Helfer, , 2017Shanks et al., 1996;von Borzyskowski & Vabulas, 2019), and when international organizations cease to function or even to exist (Crasnic & Palmtag, 2019;Eilstrup-Sangiovanni, 2020;Gray, 2018). But we are still only at the beginning of understanding how these processes are related, how they interact, how they spread, and how they can be contained.
This paper contributes to a better understanding of these questions by focusing on how voter-endorsed challenges to international institutions, one of the most extreme expressions of the popular backlash against international cooperation, reverberate internationally. I define such voter-endorsed challenges as instances in which one member state of an international institution attempts to unilaterally change the terms of or withdraw from an existing international institution on the basis of a strong popular mandate, such as a referendum vote or a successful candidate's key election promise (Walter, 2020b). Such voter-endorsed disintegration efforts have proliferated in recent years: Among the twenty referendums on international issues that were held worldwide between 2010 and 2019, for example, every second referendum was on an issue that either implied the withdrawal from an international institution or non-compliance with or renegotiation of the rules of existing international institutions (see De Vries et al., 2021).
Voter-endorsed withdrawal or renegotiation efforts are challenging for international institutions because they politicize questions of international cooperation and about the costs and benefits of international disintegration far beyond the country in which they originate. For example, after the Brexit referendum vote, euphoric Eurosceptics across Europe, from France's Marine le Pen to the Slovak People's Party-Our Slovakia, called for similar referendums in their own countries. Similarly, the leaders of Spain's Podemos or Italy's Five-Star-Movement celebrated Greece's 2015 referendum-based bid for a more generous bailout package, raising fears that it would spark similar demands in other Eurozone crisis countries. As one country successfully challenges an international institution, demands in other countries to follow this example are likely to grow.
However, such political contagion does not always occur. For example, the US withdrawal from the Paris Climate Agreement, a prominent promise of US President Trump's election campaign, has not sparked strong popular pressure for leaving the Accord in other countries. Faced with a threat to an existing international institution, voters in the other member states sometimes even respond by mobilizing in support of the institution. For instance, the 2016 presidential election that brought Donald Trump to power led to a marked uptick in support for European integration among Europeans (Minkus et al., 2018). Given that unilateral, voter-endorsed challenges to international institutions may thus both reduce or strengthen voters' support for international cooperation in other countries, they provide a fertile ground for studying the political contagion effects of such challenges.
I argue that voter-endorsed disintegration processes can create political contagion effects abroad because they inform voters in other countries about the likely economic, social, and political consequences of unilateral disintegration efforts. This allows them to assess more accurately whether and to what extent disintegration presents a viable and better alternative to membership in the international institution. The more successful another country's disintegration experience, the more it encourages voters in other countries from supporting a similar path for their own country, and vice versa. The strength of these contagion effects is shaped by how easy it is for voters abroad to obtain and process new information about the other country's disintegration experiences and by how willing they are to update their own priors with new information.
Empirically, this paper focuses on Brexit, arguably the most consequential unilateral withdrawal from an international organization to date. It examines how the UK's withdrawal process from the EU reverberated in the remaining EU-27 member states and in Switzerland, a country that is not an EU member state but has close ties with the EU. The analysis uses original survey data from 58,959 EU-27 Europeans collected in six survey waves during the withdrawal negotiations between the UK and the EU (July 2017-January 2020) to show how these negotiations influenced assessments about the consequences of Brexit for the UK and support for EU exit in the remaining member states. The analyses reveal both encouragement and deterrence effects: EU-27 respondents' evaluations of the UK's Brexit experience are strongly associated with support for an EU-exit of their own country, even after controlling for several dimensions of euroskepticism. Moreover, these effects are conditioned by how much attention respondents are paying to Brexit and the strength of their pre-existing attitudes. The analysis then turns to Switzerland and explores how Brexit, and particularly the dramatic events surrounding the UK's failed first attempt to leave the EU in late March 2019, shaped vote intentions in three actual EU-related referendums by using original survey data collected in two waves surrounding these events. It shows that Swiss voters became significantly more willing to cooperate with the EU after observing Britain's difficulties of leaving the EU as originally scheduled. These effects once more are stronger among more attentive respondents and weaker among respondents with strong prior beliefs.
Taken together, the evidence suggests that voter-endorsed withdrawals from international institutions do reverberate among the mass public in other countries. Although these effects are conditioned by the attention people are paying to these processes and motivated reasoning, I document both encouragement and deterrence effects. Overall, deterrence effects dominated during the Brexit withdrawal negotiations.

Is International Disintegration Politically Contagious?
International cooperation has been increasingly politicized in recent years (De Vries et al., 2021;Hutter et al., 2016;Zürn et al., 2012). The rise of populist nationalism (Copelovitch & Pevehouse, 2019) and the growing backlash against globalization more generally present a significant challenge for international institutions . Dissatisfaction with the functioning of an international institution does not automatically translate into a desire to leave that institution, however (Clements et al., 2014;De Vries, 2018). This is because it is genuinely difficult to correctly predict how one's country would fare if it left an existing international institution. It is not certain that a country will be better off outside the institution, and the benefit of leaving it depends on how good the national alternative is. Only if voters believe that their country would overall do better outside the institution will they be willing to risk "going it alone" (De Vries, 2018). Most studies to date suggest that voters imagine such a counterfactual situation by comparing their own country relative to others, such as indicators of their own country's economic performance (e.g., Gärtner, 1997;Hobolt & Leblond, 2009, 2013, or their satisfaction with their own national political system (e.g., Anderson, 1998;Ecker-Ehrhardt, 2012;Rohrschneider, 2002;Sánchez-Cuenca, 2000). But such national benchmarks are imperfect proxies so that voters face a lot of uncertainty about the likely consequences of a disintegration decision.
Another country's withdrawal from an international institution, therefore, provides voters with important information about the feasibility and the consequences of disintegration. There is growing evidence that voters take developments abroad into account when forming policy preferences for their own country (e.g., Böhmelt et al., 2020;Malet, 2019). Research on the diffusion of domestic regime contention has demonstrated that major events of regime contention in one country-such as the 1917 Russian revolution or the 1989 fall of the Berlin wall-can have powerful effects on political contention in other countries (Bamert et al., 2015;Hale, 2013;Weyland, 2009Weyland, , 2010. Such "iconic events" allow voters (and elites 1 ) to update their beliefs about the extent to which regime change is both desirable and feasible (Capoccia & Ziblatt, 2010). More generally, the challenge to established institutional structures can inspire people abroad to similarly challenge their own country's established structures (Weyland, 2009).
I argue that such dynamics not only have the potential to engender domestic regime contention, but also the contention of international regimes. Watching how another country's withdrawal process from an international institution unfolds allows voters to better calibrate the likely economic, social, and political consequences of disintegration, especially in a context where exits from international institutions are rare. Another country's disintegration experience thus provides voters with a powerful counterfactual that allows them to assess more accurately to what extent disintegration presents a viable and better alternative to membership in the international institution (De Vries, 2017;Walter, 2020b). This may encourage but also discourage similar attempts abroad (Solingen, 2012).

What Kind of Contagion? Encouragement and Deterrence Effects
Whether observing another country's disintegration experience encourages or deters voters abroad to support a withdrawal of their own country from an international institution ultimately depends on the nature of this experience. A positive disintegration experience is likely to make voters in other countries more optimistic about their country's prospects outside the international institution. This is likely to create an "encouragement effect" (also known as "demonstration effect", see Bamert et al., 2015;Hale, 2013;Weyland, 2009) that makes successful disintegration of one member state "socially contagious" (Pacheco, 2012) and increases public disintegration pressure abroad. In contrast, when the withdrawing country's disintegration experience is negative, voters equally update their priors and are likely to become more pessimistic about their own country's disintegration prospects (De Vries, 2017). The resulting "deterrence effect" should decrease voters' enthusiasm for a withdrawal of their own country from the international institution. Such dynamics are well-documented for separatism and secession on the national level (Coggins, 2011;Walter, 2006aWalter, , 2006b, and I argue that they exist on the international level as well. 2 How do voters abroad evaluate whether a country's voter-endorsed disintegration experience is good or bad? After all, the long-term and even medium-term consequences of disintegration take a long time to materialize and are difficult to assess early on in the process. The costs and benefits of having left the international institutions will thus only become obvious long after the voter-endorsed disintegration process has started. Nonetheless, observing how another country's voter-endorsed efforts to withdraw from an international institution plays out provides voters with important insights about the politics of the process, as well as the political difficulties and opportunities that are likely to arise along the way (Gilardi, 2010;Gilardi & Wasserfallen, 2019;Saideman, 2012). As these in turn influence the longterm consequences of disintegration, this enables voters to revise their expectations about the political and policy consequences of disintegration, and as such about its desirability and feasibility.
Perhaps most importantly, another country's disintegration experience reveals a lot of information about the likely behavior of the institution's other member states. This is a crucial aspect in the context of withdrawing from international institutions, because the costs and benefits of disintegration for the withdrawing states depend to a large degree on how the other member states respond (Walter, 2020b). When they accommodate the withdrawing state, for example by finding new, and potentially more generous, ways of cooperation, disintegration is likely to be advantageous for the withdrawing state. However, the other member states may also react in a hard, non-accommodating way, such as making few concessions in any withdrawal negotiations or shunning the withdrawing state in other areas of cooperation. This gives an international institution's remaining member states an important role in shaping the withdrawal process and outcome, especially when the withdrawing state depends on continued cooperation with its former partners. The more accommodatingly they respond, the more positive the disintegration experience is likely to be for the withdrawing state.
Many supporters of disintegration argue and expect that the other member states will accommodate their country if it decides to pursue renegotiation or withdrawal (Grynberg et al., 2019;Milic, 2015;Sciarini et al., 2015;Steenbergen & Siczek, 2017;Walter et al., 2018), and such optimism tends to make voters more willing to risk breaking apart from an international organization (Hobolt, 2009). 3 However, the response of the other member states is not easy to predict, because they face a dilemma about whether to accommodate or not to accommodate a withdrawing state: Accommodation is attractive because it allows the remaining member states to salvage at least some of the cooperation gains that the joint membership in the international institution conferred. This is why Brexiteers insisted, for example, that "German carmakers" would surely push the EU toward accommodation. However, such a response not only shifts the distribution of cooperation gains in the withdrawing country's favor, but also creates moral hazard and political risks, as other countries might be incentivized to pursue a similar path. Non-accommodation, on the other hand, is attractive because it minimizes these risks, but at the same time is usually costly for everyone involved because of the foregone gains from cooperation. The resulting accommodation dilemma Walter, 2020aWalter, , 2020b) makes it genuinely difficult for voters to correctly predict how their country would fare if it left an existing international institution.
Being able to observe how the other member states respond to the withdrawing country's disintegration request thus provides voters with important information about the likelihood of accommodation, and hence, about the likely long run costs and benefits of disintegration. 4 The actual behavior of the other member states during the withdrawal process and their willingness to cooperate with the withdrawing state on more generous terms in the future is especially informative, as this is where competing claims about the likelihood of accommodation are put to the test. The less accommodating the remaining member states are, the less positive the disintegration experience is for the withdrawing state. In some cases, countries even abort their disintegration bids when confronted with non-accommodation. 5 Nationalist-populist parties and politicians tend to justify their efforts to withdraw from international institutions or to renegotiate existing agreements in their favor with the argument that more assertiveness in international relations and more emphasis on their own country's interests rather than accepting compromise will increase their country's prosperity, national sovereignty, and democratic quality. A nonaccommodating negotiation stance by the other member states and the associated difficulties in the withdrawal negotiations should therefore make voters abroad more pessimistic about the withdrawing country's outlook and less willing to support a similar path for their own country. In contrast, a more accommodating negotiation stance should induce voters to update their expectations about the costs and benefits of voter-endorsed disintegration in a positive direction and to become more supportive of a potential withdrawal of their own country from a disliked international institution.

Who Updates? Attention, Dramatic Events, and Pre-existing Beliefs and Attitudes
The contagion pathway that I have sketched out so far depends on two assumptions: first, that voters actually receive the information about the other country's disintegration experience and second, that they are willing to update their priors. Neither of these assumptions are likely to be met by all voters at all times.
First, to the extent that voters update their expectations and policy demands based on new information, such learning depends on the availability of new information. However, not all voters pay attention to news about international events, and not all international events are broadly covered by the media. Information-based contagion effects should thus be stronger the more people are paying attention and the stronger the signal is. Important and dramatic events in a disintegration process such as the initial disintegration decision, the completion of withdrawal negotiations, or the actual withdrawal event are likely to receive more media coverage than the day-to-day politics of the disintegration process. The availability heuristic, a frequent decisionmaking shortcut (Tversky & Kahneman, 1973), means that people tend to attribute unusual evidentiary weight to such instances which they can easily recall, whereas a second heuristic, the representativeness heuristic, additionally induces individuals to believe in the generalizability of such events (Weyland, 2010). This makes vivid and widely publicized events particularly influential.
Second, not all voters are willing to update their priors, even when the evidence seems to point in an opposite direction. The large literature on motivated reasoning has shown that individuals may care more about arriving at a particular conclusion than about being accurate (Kunda, 1990). As such, voters' expectations and beliefs may not be shaped by a rational assessment of their own country's disintegration prospects, but may rather reflect voters' core beliefs and desired outcomes (Bisgaard, 2015;Kraft et al., 2015). It is difficult to change individual (mis-)perceptions with corrective information when people hold strong prior beliefs (Baekgaard et al., 2017;Gaines et al., 2007;Grynberg et al., 2019;Kertzer & Zeitzoff, 2017;Taber & Lodge, 2006). In addition, motivated reasoning suggests that information in line with a person's priors should reinforce her beliefs, whereas contradicting information is more likely to be discarded. This discussion suggests that the more pronounced voters' pre-existing attitudes are before the onset of the disintegration process, the less updating will occur. As a result, the encouragement and deterrence effects of observing another country's disintegration experience is likely to be weaker both among hard cooperation-sceptics and among staunch cooperation-supporters. In contrast, individuals with less strongly held beliefs about the merits of international cooperation and disintegration can be expected to be more susceptible to the new information provided by an actual disintegration process. 6

Brexit: Voter-endorsed Withdrawal from the EU
To study how another country's voter-endorsed withdrawal from an international institution reverberates abroad, I focus on Brexit: the UK's withdrawal from the European Union, endorsed by British voters in the 2016 Brexit referendum. The spillover effects of Brexit in other countries, especially the remaining EU-27 member states are large. Brexit changes the balance of power within the EU, puts the integrity of the Single Market at risk (Jensen & Kelstrup, 2019), and diminishes the EU's global standing (Bulmer & Quaglia, 2018). Especially a hard, let alone a "No Deal" Brexit, creates significant economic costs in the remaining member states (Hix, 2018). Perhaps the biggest concern has been, however, that Brexit may split the EU, leaving it disunited and ultimately at risk of further withdrawals. Not surprisingly, there has been considerable concern that Brexit might pose a serious, perhaps even existential, threat for the EU as a whole (Laffan, 2019).
Brexit is not just a fascinating and important case of voter-endorsed disintegration in and of itself, however, but also a case that lends itself particularly well for studying the possible political contagion effects of such processes. As the most consequential case of voter-endorsed disintegration so far, the reverberations of Brexit on public opinion in other countries are likely to be large. This turns Brexit into a most likely case for observing contagion effects. Not surprisingly policymakers, pundits, and academics have warned that Brexit might induce a domino effect and encourage voters in other countries to push for a withdrawal of their own countries from the EU as well (De Vries, 2017;Hobolt, 2016;Szczerbiak & Taggart, 2016). At the same time, Brexit is a useful case to study because the UK's disintegration experience has seen considerable ups and downs since the 2016 Brexit referendum. Right after the referendum, jubilant reaction of euroskeptics across Europe to the referendum outcome sparked concerns about significant encouragement effects, especially as it came at a time when European integration had become a heavily contested issue among European voters and elites (Hobolt & de Vries, 2016;Hooghe & Marks, 2009;Hutter et al., 2016). However, since then Brexit seems to have reduced, rather than increased, support for disintegration in the remaining member states (Glencross, 2019). Studying the contagion effects of Brexit in detail allows us to better understand whether and how these dynamics are related to Brexit, how and why this apparent turn from an encouragement effect to a deterrence effect came about, and how it is related to the European response to Brexit.
To shed light on these questions and the more general question of how another country's voter-endorsed disintegration experience reverberates abroad, this paper therefore examines the contagion effects of Brexit among the mass public in other European countries. For this purpose, it focuses on the period of the withdrawal negotiations between the UK and the EU, that were held between July 2017 and January 2020 and studies contagion effects in two different types of countries. A first set of analyses looks at how the Brexit withdrawal negotiations reverberated among voters in the 27 remaining EU member states. Political contagion effects are likely to be most pronounced within the EU, where euroskeptic political entrepreneurs not only have incentives to use the momentum that Brexit may generate for their own purposes, but where the effects of Brexit are also most immediately felt. Using survey data from a rolling cross-section of data collected in six survey waves from almost 60,000 respondents, this analysis exploits the considerable ups and down during the withdrawal negotiations and explores how individual expectations and support for disintegration have developed in response to these events over time.
A second set of analyses focuses on the effects of Brexit on public opinion in Switzerland, a third country with close ties to the EU, and a country with strong direct democratic institutions in which voters were voting on concrete disintegration proposals in referendums during and shortly after the Brexit withdrawal negotiations. These analyses explore how dramatic disintegration-related events matter for contagion processes by leveraging a two-wave survey design. In this design, some respondents were surveyed shortly before and some others shortly after one of the most dramatic and chaotic events during the withdrawal negotiations: the UK's near No-Deal-exit from the EU in late March 2019. The Swiss case study thus allows me to identify the effects of a key disintegration event on actual vote intentions.

Brexit Reverberations in the EU-27
To study possible contagion effects of Brexit in the EU-27, I examine how voters in the remaining EU member states evaluate Brexit, whether and how these evaluations vary with the ups and downs of the Brexit withdrawal negotiations, and how they are related to voter's support for a hypothetical EU-exit of their own country.

Research Design
I use original survey data from a rolling cross-section of data collected in six survey waves fielded in 6-month intervals during the Brexit withdrawal negotiations between July 2017 and December 2019. 7 The data were collected by placing questions on an EU-wide online survey omnibus (the "EuroPulse") that is regularly conducted by Dalia Research. In each wave, the sample consists of a census representative sample of approximately 10,000 working-age respondents (aged 18-65) from all EU member states per wave, 8 with sample sizes roughly proportional to their population size. I omit the data from UK respondents to arrive at a sample of respondents in the remaining EU-27 member states only. In order to obtain census representative results, the data are weighted based upon the most recent Eurostat statistics. 9 I use this data to analyze how Brexit has affected support for an EU-withdrawal among voters in the remaining EU member states.
The two main variables measure EU-27 respondents' evaluations of Brexit and their support for an EU-exit of their own country. Assessment of how well or badly Brexit is going for the UK is measured with the question "Five years from now on, do you think Brexit will make the UK (1) much worse off, (2) somewhat worse off, (3) neither better nor worse off, (4) somewhat better off, or (5) much better off?" EU-27 Europeans vary substantially in their assessment about how Brexit will affect the UK in the medium term. Overall, about 37% think that Brexit will make the UK somewhat or much worse off, whereas a good quarter of respondents thinks that Brexit will be a success for the UK. 10 Respondents' potential support for an EU-withdrawal of their own country is measured with answers on a four-point scale to the question "If [YOUR COUNTRY] were to hold a referendum on leaving the EU today, how would you vote?" Although overall, a clear majority of Europeans support remaining in the EU (40.2% state they would definitely and 25.1% probably vote to remain in such an exit-referendum), about a quarter of respondents would endorse an EU withdrawal (10.8% definitely and 13.4% probably intending to vote in favor of leaving the EU).
The argument suggests that contagion effects will be mediated by the extent to which voters actually receive the information about the other country's disintegration experience and their willingness to use this information. I operationalize the former by how much respondents are following news on Brexit 11 and the latter using their general opinion of the EU. 12 Respondents with more extreme opinions of the EU, both among Euroskeptics and among Europhiles, pay more attention to Brexit and there is considerable variation across all categories. 13 Because recent research demonstrates that Euroscepticism is a multidimensional concept (De Vries, 2018;Hobolt & de Vries, 2016), the analyses below additionally control for respondents' preferred future course for the EU ("The EU should return some power to national governments," "The division of power between national governments and the EU should remain as it is today," and "National governments should transfer more power to the EU"). Moreover, I control for age, gender, education, and whether the respondent lives in a rural or urban setting. I use weighted data and multilevel models to take into account that the data were collected in 27 different national contexts over six different waves.

Brexit Withdrawal Negotiations and Contagion Dynamics
To examine the contagion effects of Brexit in the EU-27, I begin by examining how the ups and downs of the Brexit withdrawal negotiations affected public opinion abroad. The chances of a successful Brexit waxed and waned considerably over the withdrawal negotiations. Both parties initially made considerable progress. But the negotiations began to run into serious difficulties in late 2018, as the EU side showed itself unwilling to accommodate many of the UK's demands. In Spring 2019 it became increasingly clear that the UK Parliament would not ratify the Withdrawal Agreement in time for the originally scheduled withdrawal date of March 29, 2019. Two chaotic and dramatic weeks followed, during which the EU twice extended the Brexit deadline in the last minute. These events eventually led to the resignation of the UK's prime minister, Theresa May, and the election of a new prime minister, Boris Johnson, who insisted that he would rather let the UK crash out of the EU than make any compromises. In the early fall 2019, a No-Deal-Brexit once more seemed like a distinct possibility. In the end, however, Boris Johnson negotiated a revised withdrawal agreement with the EU, which was widely perceived as a success for the UK, but was in fact an agreement mostly along the lines set out by the EU (Menon, 2020). The withdrawal agreement was ratified by both the UK and the EU side in late January 2020, paving the way for the UK's actual withdrawal from the EU a few days later.
These ups and downs are reflected in the value of the British Pound over the course of the withdrawal negotiations (Martini & Walter, 2020). The lefthand panel in Figure 1 shows the monthly changes in the GBP's average nominal exchange rate with a basket of four major currencies (EUR, USD, CHF, and JPY) relative to the beginning of the withdrawal negotiations in July 2017. 14 Reflecting the optimism of the early months of the withdrawal negotiations, the exchange rate initially appreciated but then lost in value as the negotiations, the became increasingly difficult. The uncertainty surrounding the failed ratification of the withdrawal agreement in the UK and the domestic political upheaval it produced were reflected in a sharp fall of the currency, which recovered, however, as the negotiations were successfully concluded in January 2020.
How did these ups and downs influence public opinion in the EU-27? The middle and right-hand panel in Figure 2 shows how evaluations of the UK's Brexit experience and support for EU-exit evolved over the course of this period. In the first half of the Brexit negotiations opinions were relatively stable, but respondents became more pessimistic when the negotiations began to run into difficulties in late 2018. 15 Especially after the UK's failed attempt to leave the EU in late March 2019, respondents began to view the UK's Brexit outlook significantly more negatively and became significantly less likely to support an EU-exit of their own country, even though support for EU-exit recovered somewhat after Boris Johnson had managed to secure a Brexit deal. Examining these dynamics more systematically, the first two columns in Table 1 show that these ups and downs (as reflected in the exchange rate 16 ) are strongly related to Brexit evaluations (model 1) and support for EU-exit (model 2). We can thus observe both encouragement and deterrence effects of Brexit, although overall, support for similar moves in the remaining member states has declined over the course of the withdrawal negotiations.
To examine the contagion effects of Brexit more systematically, I next examine how individuals' subjective evaluations of how well Brexit will play out for the UK are related to vote intentions in a hypothetical referendum on respondents' own country's EU membership. Of course, both respondents' desire to leave the EU and their expectations of how Brexit will play out for the UK are related to what respondents think about the EU more generally, and this general evaluation may also be influenced by the Brexit negotiations. 17 I address this problem by controlling both for respondents' general assessment of the EU and their views on their preferred future course for the European Union.
Model 3 in Table 1 shows the results of a multilevel analysis that examines the correlates of supporting one's own country's EU-exit in a hypothetical EU-exit referendum. Controlling for respondents' views of the EU, the analysis shows a substantively and statistically significant effect of respondents' Brexit evaluations on their propensity to support a withdrawal of their own country from the EU. We can see both a deterrence and an encouragement effect at play: Those who think that Brexit will leave the UK much better off in the medium-term are more likely to support an EU-exit of their own country, whereas those who assess the UK's Brexit experience more negatively are less likely to support such an exit. These effects are sizable especially as they persist after controlling for two different dimensions of EU-related opinions-general EU opinion and more specific opinions about a possible EU reform-both of which also have strong and statistically significant effects on support for EU-exit. 18 In sum, this analysis supports the argument that voter-endorsed disintegration efforts in one country reverberate in the international institution's other member states. A positive disintegration experience encourages voters abroad  to support a similar path for their own country, whereas a negative experience has a deterring effect.

Mechanisms: Attention and Motivated Reasoning
Turning to the contagion mechanisms, I have argued that contagion effects depend on whether voters actually receive the information about the other country's disintegration experience and whether they are willing to update their priors. I begin by exploring the informational pathway. For this purpose, I focus on the extent to which the attention respondents are paying to Brexit shapes their propensity to update their priors when new information becomes available. Ideally, this would be done using panel data, but in the absence of such data I once more leverage the temporal variation in the Brexit negotiations to assess this question. To examine how respondents with different levels of information responded to ups and downs of the Brexit withdrawal negotiations, models 4 and 5 in Table 1 interact the GBP exchange rate as a proxy for how well or badly Brexit is going for the UK with the attention variable that measures how much attention respondents report to be paying to Brexit. Figure 2 shows the marginal effects of the exchange rate, conditional on respondents level of attention paid to Brexit. The analyses show that the effect of the UK's Brexit experience is conditioned by how much attention they are paying to Brexit, but also that respondents are more willing to update their priors on how well Brexit is going for the UK than to change their support for  Whereas the exchange rate is not related to inattentive voters' Brexit evaluations, well-informed voters become significantly more optimistic about the UK's post-Brexit prospects when the exchange rate appreciates, and vice versa. It also makes respondents more likely to support an EU-exit of their own country, although the conditional effect of attention is weaker here. Nonetheless, the analysis demonstrates that-in line with the predictions of the informational mechanismthe contagion effects are stronger the more attention voters abroad are paying to another country's disintegration experience.
Turning to voters' willingness to update, the last step in the analysis examines to which extent motivated reasoning conditions the contagion mechanism. As discussed above, contagion effects will be shaped not just by the availability of information, but also by the extent to which the withdrawing country's experience squares with voters' pre-existing attitudes and their priors about the likely consequences of leaving an international institution. To explore this aspect in more detail, I interact respondents' subjective assessment of the UK's Brexit experience with their general views on the EU. 20 Figure 3 shows the marginal effects of different evaluations of the mediumterm Brexit effects on support for EU-exit, by respondents' general opinion of the EU. 21 The analysis once more confirms the existence of both encouragement and deterrence effects: Those who think that Brexit will have positive effects on the EU are more likely to support an EU-exit of their own, and vice versa. However, it also shows that these effects are largest amongst respondents who do not hold very strong opinions about the EU. In contrast, they are much more subdued among those with strongly entrenched opinions of the EU. Among strong Euroskeptics and strong Europhiles, especially expectations about the UK's Brexit experience that run counter to their priors (and that should thus, in a Bayesian updating framework, have particularly strong effects), have the smallest and often not even statistically significant effects on their propensity to support an EU-exit of their own country. Rather than update their beliefs, they seem to react more strongly to developments that confirm their priors. These findings suggest that motivated reasoning indeed conditions the contagion effects of voter-endorsed disintegration processes abroad.
Taken together, the analysis of the EU-27 survey data suggests observing voter-endorsed disintegration efforts in one country can have significant ripple effects throughout the other member states of the international institution. Whether these ripple effects deter or encourage further disintegration, however, depends on how the disintegration experience plays out for the withdrawing country, on the attention voters are paying to the other country's disintegration process, and on their willingness to update their beliefs and attitudes based on this information.

Brexit Reverberations in Switzerland
In a second set of analyses, I examine possible contagion effects of Brexit on voters in Switzerland. This case study complements the EU-27 analyses in three ways. First, Switzerland is not an EU member state, but a country with close ties with the EU, institutionalized in a dense network of bilateral treaties. As such, it allows me to study to what extent voter-endorsed disintegration extends beyond the other member states of an international institution grappling with the withdrawal of one of its members.
Second, the Swiss case allows me to examine how Brexit reverberates among voters abroad who actually vote on withdrawing from an international treaty. As a direct democracy, Swiss voters often vote on foreign policy issues, including on proposals to withdraw from or not complying with major existing international treaties. Some recent examples are the 2014 "Against-mass immigration" and "ECOPOP"-initiatives (both directed against the free movement of people treaty with the EU), the 2016 "Implementation-initiative" (which, if accepted, would have mandated non-compliance with the European Human Rights Charta), and the 2018 "Self-determination-initiative" (which, if accepted, would have mandated that the government renegotiate and ultimately terminate international treaties found incompatible with domestic referendum votes). Moreover, whereas the EU-27 respondents surveyed for the analyses above necessarily evaluated the issue of EU-exit as a hypothetical scenario, because no EU-27 government was actively pursuing EU-exit during the withdrawal negotiations, Swiss voters actually had to decide on how to vote in three referendums on Swiss-EU relations scheduled during or shortly after the Brexit withdrawal negotiations: two disintegration referendums and one referendum on deepening Swiss-EU relations. The first Swiss disintegration referendum, scheduled for May 2019 and hence shortly after the UK's first and failed attempt to leave the EU, put the country's reformed weapons' law to a vote. The reform had become necessary because the EU had amended its Weapons Directive in 2017, a move that by law required Switzerland, as a member state of the Schengen and Dublin agreements, to reform its own weapons law in accordance with this new EU directive. Switzerland's Schengen and Dublin memberships allow the country to benefit from the abolishment of border checks among the Schengen member states and cross-border cooperation in the fight against crime and asylum policy. However, the treaty also stipulates that failure to implement new EU-wide regulations that fall under the Schengen/Dublin treaties leads to an automatic termination of Switzerland's membership in these treaties after 6 months. This de facto turned the vote on the weapons law reform into a referendum about withdrawal from the Schengen agreement. The second disintegration referendum was the September 2020 popular vote on the "limitation initiative," an initiative launched by the Eurosceptic Swiss People's Party (SVP). The limitation initiative required the Swiss government to renegotiate and, if unsuccessful, to withdraw from the Swiss-EU treaty on the free movement of people. Because of the so-called "guillotine clause," however, withdrawal from one of the seven core Swiss-EU bilateral treaties results in the termination of all seven treaties. A voter-endorsed withdrawal from the free-movement-of-people Treaty thus had the potential to end the existing framework for Swiss-EU relations. Finally, Switzerland and the EU have negotiated a new institutional framework agreement that deepens the relationship between the two parties. After 4 years of negotiations, an agreement was reached in December 2018, but the Swiss government then refrained from signing it because of domestic opposition and asked the EU for renegotiations of a few key points. These negotiations thus occurred alongside the Brexit withdrawal negotiations, and the final agreement will have to be ratified by a popular vote (an integration referendum) to enter into force. The Swiss case thus provides a context in which voters' decision to support disintegration is not just a theoretical question, but one with real-world consequences.
Third, the Swiss case allows for a more immediate analysis of how dramatic events matter for the diffusion of regime contention abroad (Weyland, 2010). It exploits the fact that data were collected in two cross-sectional survey waves before and after two of the most chaotic weeks of the Brexit negotiations: the events surrounding the UK's first, failed attempt to leave the EU on 29 March 2019 and the first Brexit extension in early April 2019. Although the Brexit negotiations between the UK and the EU-27 have been difficult from the start, these difficulties were put into glaring light during this episode. As the end of British EU membership drew nearer and was increasingly likely to end with a chaotic No-Deal-Brexit on March 29, 2019, attention across Europe was focused on Brussels and London. For weeks, European headlines, including Swiss media, had Brexit as their front-page news. The "Brexit chaos" that unfolded in the weeks surrounding March 29 put the trade-offs and difficulties associated with leaving the EU into the spotlight and made it glaringly clear that Brexit was not going as well as many Brexiteers had promised and that the EU was not as accommodating as they had predicted. This suggests that these events should have a deterring effect on Swiss voters' propensity to pursue a similar path. The two-wave survey design allow me to examine this hypothesis empirically.

Research Design
The analyses are based on an original two-wave rolling cross-section survey, that was designed to cover the critical phase around the UK's original withdrawal date on March 29, 2019. A first wave of 1622 respondents was surveyed between March 13 and 28, 2019. After a 2-week break during which British and UK policymakers struggled to find a viable way out of the Brexit impasse and during which a chaotic "no-deal-Brexit" had become a distinct possibility, the EU extended the Brexit deadline to October 31, 2019. Immediately after this decision, a second wave of 836 respondents was surveyed (fieldwork April 12-18, 2019). 22 The dependent variables in the analyses are Swiss vote intentions for the three upcoming EU-related referendums, as well as a hypothetical referendum on terminating Switzerland's bilateral treaties with the EU (the Swiss equivalent to the EU-exit question in the EU-27 analyses). Vote intentions are measured on a four-point scale (certainly for, probably for, probably against, certainly against) and recoded in a way so that higher values always denote support for disintegration, or opposition to compliance, or more/continued cooperation. Respondents who state that they plan not to vote are treated as missing. 23 The main independent variable is a dummy variable that records whether respondents were surveyed after the chaotic Brexit events in April 2019. In addition, I use a variable measuring how respondents evaluate the medium term (5 years) overall effects of Brexit on the UK, using the same wording as in the EU-27 survey. Like in the EU-27, these assessments vary widely. With 31% expecting an overall positive effect of Brexit on the UK, Swiss respondents are slightly more optimistic about Brexit than the EU-27 Europeans. Nonetheless, as in the EU-27, the largest group (37%) thinks that the UK will be somewhat or much worse off because of Brexit. As in the EU-27, these expectations are likely to be strongly related to respondents' general opinion of the EU. I therefore once more control both for respondents' general opinion of the EU, as well as for voters' intention to vote for the Swiss People's Party (SVP) in the next elections. This euroskeptic, populist-right party has supported all disintegration referendums held to date and is strongly opposed to any closer ties with the EU. 24.7% of respondents in the weighted sample say that they would vote for the SVP, which is close to the 25.6% vote share the SVP achieved in the parliamentary elections later in the year in September 2019. To control for the level of awareness of Brexit-related events, a variable on how much attention respondents are paying to Brexit-related news is included. I also control for respondents' risk propensity (Steenbergen & Siczek, 2017). In terms of sociodemographics, I control for education, age, gender, unemployment, and language in which the survey was answered.

Brexit Contagion Effects on Swiss Vote Intentions
How did Swiss vote intentions in upcoming EU-related referendums evolve in response to the dramatic events surrounding the first Brexit extension in April 2019? Figure 4 compares these vote intentions immediately before (March 2019) and immediately after (April 2019) the first major Brexit chaos and subsequent extension of the Brexit deadline by the EU. For ease of comparison, it shows the share of Swiss voters who said they were planning to probably or certainly vote in favor of disintegration-that is, a vote against the weapons' reform, in favor of the limitation initiative, and in favor of terminating the bilateral treaties (a hypothetical vote), as well as a vote against a deepening of cooperation in the form of the institutional framework agreement. These analyses show that within only 2 weeks, the Swiss public became significantly more cautious about withdrawing from Switzerland's existing bilateral agreements with the EU. They also became more positive toward complying with existing rules (the reforms intended to ensure Swiss weapons law's compliance with Schengen rules) and toward deepening cooperation with the EU (the institutional framework agreement). For the institutional framework agreement, a politically highly salient issue at the time of the survey, and for the hypothetical referendum on withdrawing from the bilateral treaties, these differences are statistically significant. 24 Because Brexit dominated the news and no other major events occurred in Switzerland during these two weeks, we can be reasonably confident that this deterrence effect is indeed attributable to observing the UK's difficulties in implementing Brexit. 25 Table 2 examines these results in a more systematic fashion and shows the results of OLS regression analyses for vote intentions on the four different referendums. These analyses not only confirms that support for disintegration or non-cooperation dropped across the board over this period. It also underscores that as in the EU-27 analysis, we can observe both encouragement and deterrence effects: Even after controlling for respondents' general opinion of the EU and vote intentions for the SVP, Switzerland's most Euroskeptic party, Note. Dots show share of respondents planning to vote in favor of disintegration or against compliance or a deepening of cooperation with 90% and 95% confidence intervals.   we see that those who think that Brexit will be a boon for the UK are much more likely to vote for disintegration or against cooperation in a Swiss referendum on EU relations than those who think that Brexit will turn out badly for the UK. This evidence suggests that voter-endorsed disintegration negotiation processes not only reverberate among the remaining member states, but also in other countries for whom the withdrawing country's disintegration experience provides important lessons.

Witnessing Dramatic Disintegration Events Abroad: Attention and Motivated Reasoning
How exactly do dramatic disintegration events such as the Spring 2019 Brexit chaos reverberate abroad, and when are such reverberations muted? I once more focus on the availability of information and the willingness to update. To examine the informational mechanism, I focus on the amount of attention respondents are paying to Brexit, whereas I use a dummy that is coded as one if the respondent states that she is planning to vote for the populist-right, Euroskeptic Swiss People's Party (SVP) in the next parliamentary elections to capture the strength of prior beliefs and motivated reasoning. Much like the Brexiteers in the UK, the SVP has been arguing that the EU needs Switzerland more than Switzerland needs the EU and has therefore insisted that the EU would negotiate and accommodate Switzerland if only Switzerland pursued an uncompromising negotiation strategy with the EU. The belief that the EU will ultimately negotiate and accommodate some Swiss demands for limiting free movement of people is widespread among Swiss voters (Armingeon & Lutz, 2019: Figure 8), and especially so among SVP voters (Sciarini et al., 2015, p. 274). Given these strongly held prior beliefs, the motivated reasoning mechanism suggests that the dramatic Brexit events in spring 2019 should have much less of an effect on SVP voters than on other respondents. Figure 5 shows the results of two sets of regression analyses that interact the April 2019 dummy with the attention paid to Brexit and the SVP-dummy, respectively. 26 The marginal effects show that the dramatic events of spring 2019 dampened support for non-cooperative referendum votes especially among those paying high levels of attention to Brexit, in line with the informational mechanism. In contrast, SVP voters remained largely unaffected by these events (if anything, their support for non-cooperation was strengthened), suggesting once more that motivated reasoning is dampening contagion effects among voters with strongly held beliefs.
In sum, the Swiss analysis further corroborates the argument that voterendorsed disintegration processes such as Brexit reverberate abroad, especially in moments of highly dramatic action. These reverberations extend beyond an international institution's member states, affect vote intentions in actual upcoming referendums, and are enhanced by informational effects but dampened by motivated reasoning.

Conclusion
What does the increase in instances of voter-endorsed withdrawals from international institutions mean for the stability of international cooperation? Are such instances likely to spread, as voters abroad are encouraged to push for similar paths for their own countries? Or do these instances deter voters in other countries from pursuing such strategies because they highlight the difficulties and trade-offs associated with disintegration?
This paper has examined these questions by studying the reverberations of Brexit on public opinion in the 27 remaining EU member states and in Switzerland. The findings suggest that voters abroad watch how voterendorsed disintegration processes unfold and draw their own conclusions from observing this experience. When another country's disintegration efforts are perceived as successful, this encourages voters abroad to equally pursue a less cooperative strategy. However, when the experience is negative, this deters voters abroad from supporting a similar strategy for their own country. These encouragement and deterrence effects are stronger the more information voters have about the other country's disintegration experience and the more this experience is covered by the national media. Moreover, they are also strongly shaped by their prior beliefs and attitudes. Not every voter is willing to update her beliefs in the light of new and countervailing information. The evidence from Switzerland underlines that even drastic displays of the difficulties of withdrawing from international institutions can have a limited effect on those voters' most skeptical of international cooperation.
While these findings are important for our understanding of the Brexit process and the dynamics it is creating in Europe, they also have implications for our understanding of the popular backlash against international cooperation more generally. Many nationalist populist policy proposals are based on optimistic assumptions about the viability of a more assertive, nationalist foreign policy. As more and more of these policies are implemented and put to a reality test, one question is whether these policies will diffuse across countries, with potentially serious consequences for the contemporary cooperation-based global order, or whether they will be checked by voters and elites who observe the consequences of these policies abroad and evaluate them negatively. The findings in this paper suggest that a potential for a disintegration cascade exists, but is far from a forgone conclusion. Just like the secession of one region does not necessarily lead to the disintegration of states and just like democratization in one country does not necessarily spread to other autocratic states, the diffusion mechanisms are strongly context-dependent.
Rather, my findings suggest that the governments of the other member states of international institutions facing unilateral withdrawals or renegotiation bids have an important, if difficult, role to play in this process (Walter 2020b). When negotiating new, post-disintegration arrangements they need to balance the lure of maintaining cooperation gains with the risk of encouraging further disintegration pressure by being too accommodating. These dilemmas are particularly pronounced when the disintegration bid has been endorsed by voters and is hence highly politicized and when other member states are toying with the idea of leaving the same international institutions. Yet disintegration attempts may also reverberate beyond the circle of member states, as the evidence from Switzerland suggests, when they are initiated by elites rather than voters, and even within the same country over time. To the extent that other countries' disintegration experiences allow voters abroad to learn about the merits of international cooperation and possible upsides of leaving, it is easily conceivable that these reverberations extend far beyond the affected international institution and its members. Exploring under which conditions we can expect such more far-reaching contagion effects, which roles political elites and governments play in exacerbating or countering such effects, and to what extent and how such contagion effects matter on the systemic level, are important avenues for future research. 16. I use the 3-months average (survey month and two preceding months). 17. Pearson's correlation coefficient between respondents' opinion of EU and Brexit evaluation is 0.22 (and 0.38 between Brexit evaluations and support for EU-exit). Nonetheless, considerable variation exists in Brexit evaluations both among euroskeptics and europhiles (see Figure A2a and A2b, Supplemental appendix). 18. As is to be expected, the substantive effects are considerably larger when I do not control for euroskeptic/europhile attitudes. 19. I use a dummy variable that identifies respondents who would definitely vote to leave the EU. The interaction effects are even weaker when I use the continuous vote intention variable. 20. For full results see Table A6 in the Supplemental Appendix 21. The model interacts Brexit evaluations and EU opinion dummies, with the EU-exit vote intention variable as dependent variable. For full results see Table  A6, Supplemental Appendix. 22. The sample was built using quota sampling of Swiss citizens. I additionally use weights in the analyses. The survey was carried out by Infratest dimap. Replication materials and code can be found at . 23. Descriptive statistics for all variables are listed in Table A7, Supplemental Appendix 24. At the time of the survey, the referendum campaign on the weapons law referendum had not yet fully started and the fact that a rejection of the reform would lead to an automatic termination of Switzerland's Schengen membership had not yet been widely discussed. 25. The other EU-related topic that was salient in that period in the Swiss discourse and media were the consultations about the new Swiss-EU framework agreement. Because criticism of the agreement dominated the debate, this would rather push opinions of the EU in the opposite direction, however. 26. Full results are shown in Table A8 in the Supplemental Appendix.