Compensation and Compliance: Sources of Public Acceptance of the U.K. Supreme Court's Brexit Decision

Abstract

The perception that a high court's decision is binding and final is a crucial prerequisite for its ability to settle political conflicts. Under what conditions are citizens more likely to accept controversial judicial rulings? Mass acceptance is determined, in part, by how rulings are framed during public debate. This paper takes a broad view of the strategies and actors that influence the discursive environment surrounding judgments, calling attention to hitherto unexamined determinants of mass acceptance. We theorize that third parties can boost acceptance by pledging compliance, and that courts can moderate opposition by compensating losers. We also look at how populist attacks on judiciaries, common in contemporary democracies, affect acceptance. We test these propositions using a survey experiment conducted in the aftermath of the UK Supreme Court's Brexit decision, the most salient judgment handed down by this court to date. The paper moves the literature on courts and public opinion beyond the United States, and presents evidence backing largely untested assumptions at the heart of models of judicial behavior regarding the benefits of crafting rulings with an eye on the preferences of key audiences.

On January 24, 2017, the highest court in the United Kingdom handed down a decision in what the Guardian called “the most important constitutional case ever to be heard by the Supreme Court.”¹ In June of the previous year, a private claimant had thrown a lower court into the muddy waters of the Brexit saga, demanding that Theresa May's Conservative government be forced to seek parliamentary approval before triggering the process that will eventually culminate with Britain's exit from the European Union—approval the Prime Minister had hitherto insisted was not necessary. After these judges ruled against the government, tabloids accused them of being “enemies of the people.”² The Supreme Court subsequently upheld the judgment, and was attacked and praised in equal measure by the two sides of the debate. Under what conditions are citizens more likely to accept controversial rulings such as this one as final and binding? Do the characteristics of court outcomes as well as the reaction of important institutional players condition acceptance?

Judicial politics scholars define acceptance as the decision to acquiesce to a judgment, “cease opposition and get on with politics” (Gibson et al. 2005: 188). Acceptance speaks to the ability of judges to function as arbiters, “settle political conflicts, or at least make it more difficult for opposition to continue to mobilize” (Gibson and Caldeira 1995: 466). As “the least dangerous branch,” courts rely on the belief that they can legitimately hand down authoritative, binding, and final decisions as “a critical indirect path towards compliance” (Caldeira 1986; Nicholson and Hansford 2014: 621; Tyler 1984). Importantly, should judges upset powerful legislative or executive actors, mass acceptance of rulings provides an important source of leverage because it complicates efforts by political entrepreneurs wishing to engineer backlash against judicial institutions (Helmke and Staton 2011; Vanberg 2001, 2005).³ But even if backlash in the form of outright disobedience or proposals to weaken judicial prerogatives does not materialize in the wake of controversial rulings, mass acceptance is still important for the long-term development of courts' legitimacy. Acceptance of specific rulings is part of the running tally that leads citizens to build a “reservoir of goodwill” toward the judiciary, which is a form of institutional loyalty that “embodies the notion that failure to make policy pleasing in the short-term does not necessarily undermine the basic commitment to support the institution” (Gibson et al. 2005: 189). At a more basic level, studying mass acceptance allows us to see how the public views one of the key decision makers in the political system and its prerogatives. For example, is the public willing to entertain, contrary to what is stipulated in law, that Supreme Court decisions are not final and binding?

Acceptance of specific judgments is shaped by the way in which judicial behavior is framed during the public debate (Gibson and Caldeira 2009; Gibson et al. 2005; Hoekstra 1995; Nicholson and Howard 2003). Framing strategies encourage “readers or listeners to emphasize certain considerations above others when evaluating” the issues at stake (Chong and Drukman 2007: 367). This paper contributes to this body of work by exploring the effect of competing frames on the likelihood that citizens accept controversial rulings as final. We take a broader view of the strategies and actors that influence the discursive environment surrounding rulings, and call attention to determinants of acceptance hitherto ignored by scholars of courts and public opinion.

First, we argue that signaling that judgments are not zero-sum decisions, but effectively compensate losers, boosts levels of acceptance. It is well established that judges facing difficult choices tend to rule strategically, and often avoid inflicting losses on powerful actors or compensate them to prevent outright rejection of judgments (Helmke 2005; Knight and Epstein 1996; Staton and Vanberg 2008). An implicit assumption is that by framing outcomes in ways that showcase these compensations, judges and their allies can modify attitudes toward rulings among affected groups. We present what, to our knowledge, constitutes the first test of the microfoundations underlying this influential proposition.

Second, the paper looks beyond the characteristics of judgments to assess how the responses of nonjudicial actors affect acceptance. We contend that compliance pledges by salient political players are likely determinants of mass acceptance. The literature associates compliance with rulings by presidents or legislatures with greater judicial authority and legitimacy (e.g., Kapiszewski and Taylor 2013). Our design affords a unique empirical evaluation of this thesis, testing whether information about major political parties' willingness to comply has a positive effect on judges' ability to command acceptance of rulings among the mass public.

Third, we focus on the effect of accusations of bias on acceptance. Studies of the U.S. Supreme Court, which dominate this body of work, document the negative implications of framing judicial decision making as partisan as opposed to strictly legalistic (Gibson et al. 2005; Nicholson and Hansford 2014; Nicholson and Howard 2003). Unfortunately, this approach is not productive in contexts where judges are not divided along party lines. To expand the scope of this line of inquiry, we explore the effects of a different kind of frame that also signals partiality, or behavior that is inappropriate for the judicial role, but is a more common line of attack in contemporary democracies: populist depictions of courts as rigged in favor of the establishment.

The paper reports the results of a survey experiment with four treatment groups conducted days after the U.K. Supreme Court's 2017 Brexit decision. In addition to testing novel hypotheses, the study moves the literature beyond the U.S. context. This is important simply because we currently know very little about the public opinion side of judicial politics elsewhere (cf. Caldeira and Gibson 1995; Gibson et al. 1998). Our focus on the U.K., where the Supreme Court was only established in 2009, has the potential to offer valuable comparative perspectives. Like the U.K. Supreme Court, most high courts around the world, including some in other consolidated democracies, do not share with their U.S. counterpart a long history of public salience and engagement with important issues of the day. When courts are suddenly pushed into the political maelstrom they are much less likely to have amassed a reservoir of goodwill that protects them from harsh criticism, backlash, or other threats to their institutional integrity (Gibson and Caldeira 2009). It is precisely in these contexts where reactions to landmark rulings such as the Brexit decision should prove most consequential for the consolidation of a court's authority to render decisions that are perceived as authoritative and binding, making the study of the sources of acceptance all the more urgent. Indeed, these could become formative moments that strongly shape how citizens relate to the institution for the foreseeable future.

In line with our expectations, framings that signal the U.K. Supreme Court's attempt to compensate losers help boost acceptance of the ruling among those who benefit from compensations. Specifically, we show that reminding respondents that the court did not agree with all plaintiff demands, thus minimizing disruptions to the Brexit process, moderates acceptance among Brexit supporters and other relevant subgroups. Similarly, compliance pledges by two important institutional interlocutors—the Conservative and Labour parties—enhance acceptance of the ruling. But the mechanisms driving this relationship appear to differ depending on who pledges compliance. In the case of the Conservative Party's compliance pledge, changes in acceptance levels are limited to co-partisans, whereas Labour's pledge has a much wider impact. Finally, the populist framing has no observable effects on acceptance.

These results have implications for our understanding of how the behavior of a much broader set of actors beyond judges themselves, including governments and political parties, affects the authoritativeness of decision making by apex courts. We show that although judges are indeed able to broker acceptance of their judgments among relevant constituencies by crafting rulings in specific ways (i.e., compensating losers), they are sometimes at the mercy of the reactions of actors over whom they have little control (i.e., compliance pledges). The paper thus breaks new ground by presenting micro-level evidence in support of assumptions made by scholars of judicial behavior who claim that judges rule strategically to condition responses by politicians and the public. In this sense, the results show that the payoffs of pleasing different audiences can vary significantly.

Conditioning Acceptance: Compensation, Compliance, and Establishment Bias

Academic work on courts and public opinion is heavily dominated by studies of the U.S. Supreme Court. Research finds that the U.S. Supreme Court enjoys high levels of institutional loyalty or “diffuse support” (Easton 1975). The influential “positivity bias” thesis indicates that the court has been able to amass a “reservoir of goodwill” that acts as a shield, ensuring that even highly polarizing rulings fail to dent diffuse support (Gibson and Caldeira 2009). An important source of this favorable predisposition is “exposure to the legitimizing symbols” that courts project (Gibson et al. 1998: 356), which help cultivate the perception that judges are different from other decision makers, and lead citizens to consider the Supreme Court as a legitimate arbiter of constitutional conflicts (Casey 1974; Hibbing and Theiss-Morse 1995). Crucially, diffuse support builds up over time as judges hand down salient rulings satisfying nonoverlapping constituencies (Baird 2001). This points to the unlikely presence of “positivity bias” in contexts with younger or low-salience courts (Gibson and Caldeira 1995).

Despite high levels of diffuse support, translating the legitimacy of the U.S. Supreme Court into acceptance of specific rulings, or approval of the policies they sanction, is not straightforward (Bass and Thomas 1984; Grosskopf and Mondak 1997; Hoekstra 2003; Mondak 1990). Indeed, the ability of the court to deploy its reservoir of goodwill to command acceptance is often limited by the preferences of key constituencies (Hetherington and Smith 2007). Moreover, research shows that a series of rulings seen to go against the interests of a group can gradually erode diffuse support among its members (Caldeira 1991). For example, jurisprudential setbacks on civil rights dented support for the U.S. Supreme Court among African Americans (Gibson and Caldeira 1992). In sum, the “reservoir of goodwill” is neither a silver bullet nor bullet proof.

More recent scholarship goes a step further, showing that acceptance and approval of rulings is highly conditional on the frames deployed by commentators when discussing particular judgments. It is well established that the discursive context surrounding decisions rendered by public officials affects levels of acceptance. In order to make sense of complex policy/legal discussions, individuals take cues from peers (MacKuen and Brown 1987), political actors (Lenz 2009), and the media (Ladd and Lenz 2009; Levendusky 2013). Individuals differ in their susceptibility toward this new information depending on their level of political interest and knowledge (Zaller 1992); whether information confirms partisan priors (Carsey and Layman 2006; Gaines et al. 2007); and whether there are credible options (Groenendyk 2012). Actors involved in the public debate exploit this susceptibility by “emphasizing a subset of potentially relevant considerations,” which lead “individuals to focus on these considerations when constructing their opinions” (Druckman 2004: 672). The frames deployed tap onto latent predispositions, altering levels of satisfaction with a policy outcome. As Jacobson (2000): 751) puts it, “framing effects occur when different presentations of an issue generate different reactions among those who are exposed to that issue.”

In the case of the U.S. Supreme Court, Nicholson and Howard (2003) look at how different portrayals of Bush v. Gore (2000) conditioned acceptance. They find that when frames emphasize the partisan nature of the decision, mass acceptance suffers. Similarly, Nicholson and Hansford (2014) conclude that signaling the partisan identity of the members of the Supreme Court majority responsible for a ruling increases acceptance among individuals who are co-partisans, especially on issues in which the public is not strongly polarized along the Democrat/Republican divide. This is consistent with the argument that framing effects are stronger when individuals hold more ambivalent views on a topic (Alvarez and Brehm 2000). Other aspects of the decision-making process also play a role in conditioning acceptance. Zink et al. (2009) find that when individuals are told that decisions are the product of a unanimous vote, or that they respect precedent, thus indicating a more legalistic approach, acceptance increases. As Gibson et al. (2005: 188) argue, “the court profits when it grounds its decisions in law.” These studies highlight that courts' internal dynamics can be exploited to cultivate good public relations.

The focus of “framing” scholarship on the consequences of legalistic depictions of courts' decision-making processes is certainly applicable outside the United States. In fact, judges across the board are generally at pains to construct and preserve a “myth of legality” (Bybee 2010; Scheb and Lyons 2000) because their authority as arbiters suffers when the public does not perceive judicial reasoning and procedures to be impartial or appropriate for the judicial role. In an effort to build on this research agenda and enhance its comparative intent, in what follows we propose hypotheses about hitherto unexamined framing efforts that are also likely to affect mass acceptance of judicial decisions in a variety of contexts. These are related to the framing of outcomes as opposed to processes, and to the reactions of courts' institutional interlocutors rather than to judicial behavior itself. We refer to these as compensation and compliance effects, respectively. Qua sources of variation in acceptance, compensation, and compliance framings are theoretically interesting because they point to a broader menu of legitimation strategies available to high court judges in addition to “the myth of legality.” Furthermore, we adapt the partisan versus legalistic hypothesis at the heart of “framing” scholarship to make it more portable to contexts where, unlike in the United States, judges are not obviously divided along party lines. To do so, we explore the effect of populist depictions of judiciaries as biased in favor of the establishment.

Compensation Effects

Citizens care about procedural fairness and the extent to which legal reasoning as opposed to partisan or other considerations dominate judges' thinking (Tyler 1984, 2004), but they also care about results (Ford 2012). This, however, does not imply that courts are always bound to alienate losers. Judicial cases need not be zero-sum games in which one party wins everything, leaving the rest festering resentment. In fact, judges facing difficult choices often rule strategically, departing from (or watering down) their preferred outcomes to please several audiences simultaneously (Epstein and Knight 1998; Epstein et al. 2001; Ferejohn 1998; Helmke 2005). Marbury v. Madison (1803), a famous case in which the U.S. Supreme Court chose not to inflict losses on the most powerful party in the dispute, but carved a way for judges to curb legislative and executive authority in the future, is the paradigmatic example of this kind of behavior (Knight and Epstein 1996).

Common tactics to temper losses include resorting to particular forms of legal argumentation that narrow the scope of outcomes (Baum 2006), affording bureaucrats handsome margins of appreciation for the implementation of court-mandated policy changes (Spriggs 1997; Staton and Vanberg 2008), or granting victories to different sides in the same case or across related cases (Ginsburg 2013). Even when these concessions do not fundamentally affect the core policy implications of a ruling, and therefore fail to provide an airtight insulation against backlash from losers, they can still weaken the rationale for potential retaliations. Moreover, although the overall net effect of concessions could be neutral, increasing acceptance among the main losers and eroding it among putative winners, these tactics can help preserve a court's authority among politically powerful losers, a crucial concern in these situations.

Balancing losses under pressure by making concessions is no small feat, and effectively communicating these nuances is not either. The implications of a ruling are sometimes only relevant for the plaintiffs, which makes the latter task easier. But politically charged cases almost always affect the values and interests of much larger constituencies. In these cases, public commentary may fail to capture the complex distributional consequences of the judgment. Pundits may focus on the main issue at stake, identifying clear losers and fostering polarization, as opposed to carefully zooming in on the fine print designed to ameliorate a sense of loss. Research shows that courts and their allies are painfully aware of the problem of effectively communicating their conclusions, and know that important details can be obscured when passions run wild. To address this risk, judges often devise clever public relations strategies (Staton 2010).

Although these compensation efforts and the strategic motivations that inspire them are well documented, we know little about whether or not they are in fact effective in affecting levels of acceptance of rulings among the targeted constituencies. In order to address this gap in existing knowledge, we test whether effectively signaling the presence of compensatory measures boosts acceptance of judicial decisions. Although the overall net effect may be zero, losers are thus given additional reasons to accept the outcome.

1. Hypothesis 1: (Compensation Effect): Acceptance of a judicial decision among losers increases when they become aware of the presence of compensations or concessions to their side.

Compliance Effects

The framing effects literature on courts and public opinion usually focuses on the attributes of judges and the rulings they craft. But the reactions of other political actors to important decisions are also likely to contribute to the process whereby citizens develop their own views about these policy outcomes. In particular, compliance or compliance pledges send signals that help shape a court's authority.

The most important choice that bureaucrats, governments, and political parties face vis-à-vis apex court rulings is whether or not to accept judicial decisions as final, and comply. Judges care about compliance “because it conforms to their understanding of who they are and what law is” (Huneeus 2014: 441). The foundations of judicial authority shake every time a court hands down a ruling, and lacking the means to coerce enforcement, waits for the reaction of the relevant institutional interlocutor. Judges are indeed wary of the prospect of noncompliance, a real possibility in politically charged cases, and often rule in ways that are expected to minimize the risk (Botero 2018; Vanberg 2001, 2005).⁴

Compliance communicates and builds judicial authority by making it clear that abiding by court rulings is not optional, and most importantly, that key actors “feel responsible for following the directives of legitimate authorities” (Tyler 2004). In the words of Kapiszewski and Taylor (2013: 821), “compliance with a court's rulings may reflect and increase its legitimacy – thereby encouraging further assertiveness and enhanced compliance.” By contrast, a court that is systematically ignored, and is therefore perceived as inconsequential, could lead citizens to conclude that rules are indeed optional and is OK to sidestep them. Furthermore, it may convince potential plaintiffs to pursue other avenues of policy contestation, failing to “cease opposition and move on with politics”.

One possibility is that compliance pledges by political actors send generic signals regarding the authority of the court and the importance of abiding by its rulings no matter what, thus boosting overall acceptance levels. In this scenario, compliance affects acceptance via an “institutional” mechanism. The identity of the complier does not condition the relationship; all that matters is the status of the complier as a relevant institutional interlocutor who thus sets a visible example of the appropriate way to react to rulings.

We know, however, that citizens weigh information differently depending on the source. In particular, it is well established that partisanship is a powerful cue that helps voters evaluate policy, leaders, and electoral options (Campbell et al. 1960; Jacobson 2007). If you identify with a party, you are likely to use the public stances of its leaders as a heuristic to decide your own position on an issue, especially if priors are weak or uninformative (Arceneaux 2008; Kam 2005; Van der Brug 2001). Via established communication channels or their affective ties with supporters, parties are able to convey a sense of what is the appropriate thing to do or think. In fact, when party cues are provided, citizens sometimes ignore the substance of the policy to decide what their views on a topic are. Looking at the case of Britain, for example, Lenz (2009) shows that when parties change positions on issues, and partisans learn the new position of their party, they adjust their own stances accordingly.

Since the actors who hold the keys to compliance with judicial rulings tend to be elected officials belonging to different parties, there is therefore a second possible mechanism via which compliance pledges affect mass acceptance. In this alternative scenario, compliance pledges by partisan actors are likely to boost acceptance, but especially so among their co-partisans. Indeed, if the “partisan” mechanism is very strong, compliance pledges by any given party could even be counterproductive, as they could erode acceptance among those who identify with an electoral rival.

Studies of judicial behavior often see judges as strategically seeking to secure compliance in order to avoid falling into disrepute. In so doing, they assume that there is a positive relationship between compliance and their authority among the public. To our knowledge, however, neither the relationship nor the mechanisms driving it have been empirically tested. We therefore propose and test the following hypothesis:

1. Hypothesis 2. (Compliance Effect): Compliance pledges by important political actors increase acceptance of judgments, especially among citizens whose partisanship matches the complier's partisan identity.

Populist Effect

Despite our focus on outcome compensation and the signals sent by compliance pledges, it is undeniable that the characteristics of courts' decision-making processes, especially depictions of judges as biased arbiters, play an important role in shaping acceptance of judgments. As mentioned, scholarship on the U.S. Supreme Court has found that when decisions are framed as grounded in partisanship and not in law, acceptance suffers. Over time, these considerations have become more salient among U.S. voters because the partisan affiliations of judges have acquired greater visibility (Nicholson and Hansford 2014). But this feature of the U.S. Supreme Court does not translate well to other contexts where judicial appointments and the profile of judges themselves tend to be less politicized, and voting blocs inside supreme or constitutional courts are not defined by identifiable partisan divides. For example, U.K. Supreme Court judges are appointed by a nonpartisan selection committee. Scholars or commentators are unable to read decisions, let alone predict voting patterns, based on judges' partisan affiliations (Hanretty 2012).⁵

In order to explore whether depictions of a court's decision-making criteria as “biased” can indeed undermine acceptance of rulings beyond the U.S. context, we must therefore consider additional forms of prejudice. The rise of populism in contemporary advanced democracies offers one such possibility (Ford and Goodwin 2014; Mudde 2013). Populism is a “thin-centered ideology that considers society to be ultimately separated into two homogeneous and antagonistic groups, ‘the pure people’ and ‘the corrupt elite,’” (Mudde and Rovira 2012: 8). In Europe, for example, populist parties blame the “establishment” for neglecting issues at the top of voters' agendas including the decline in manufacturing, rising ethno-religious diversity, and the deepening of European integration (Cutts and Goodwin 2014; Evans and Tilley 2017; Jennings and Stoker 2017). It is not hard to see why the judiciary has become a favorite target for populists, who accuse courts of epitomizing all that is wrong with the liberal, self-serving elite. In addition to fiercely denouncing the counter-majoritarian influence of judicial institutions, these groups are periodically enraged when courts hand down rulings that apply human rights legislation to protect the rights of immigrants or the due process rights of terrorist suspects. The will of the people “cannot be limited by anything, not even constitutional protections, that is, vox populi, vox dei” (Mudde 2010: 1175). The quote in our opening paragraph, in which a popular tabloid accused the judges in the United Kingdom of being “enemies of the people,” is a clear example of these dynamics at play.

Populist framings deployed in everyday political discourse, which remind voters of the struggle between the elite and the people, have gained a lot of traction in countries such as the United Kingdom, France, the Netherlands, or Germany. Populist parties and tabloids successfully tap onto the perception that political institutions are stacked against the interests of the common man. Indeed, experimental evidence shows that adding populist elements in speeches increases politicians' popularity, especially among poorly educated voters (Bos et al. 2013). To our knowledge, however, no work has explored the effects of this kind of discourse on attitudes toward judicial decisions. We hypothesize that accusations of pro-establishment bias levied against judges are likely to reduce acceptance of rulings. This should be especially true among those who identify with parties of the radical right.

1. Hypothesis 3. (Populist Effect): Depictions of judges as biased in favor of the elite or the establishment decrease acceptance of judgments.

Case and Methods

The U.K. Supreme Court was established in October 2009 to replace the Appellate Committee of the House of Lords. Although the latter had exercised some form of supreme judicial authority since the late nineteenth century, the idea of a “Supreme Court” was new to the British public. The decision to replace the Law Lords with a tribunal based at the opposite end of Parliament Square was made by Tony Blair in 2003. The Prime Minister's proposal was motivated by growing concerns about the lack of independence of a court housed in one of the two legislative chambers. These arrangements were clearly out of sync with the standards of judicial independence set by the European Court of Human Rights. Moreover, the trend toward devolution that began in the 1990s exacerbated the need for an independent body able to adjudicate interjurisdictional conflicts effectively. The new Supreme Court was therefore tasked with overseeing issues arising under the 1998 devolution Acts (Le Sueur 2004).

The other major legal innovation in recent British history was the adoption of the Human Rights Act in 1998, which incorporates rights stipulated by the European Convention on Human Rights into domestic law. As a result of this development, first the Law Lords and later the Supreme Court saw their dockets flooded with cases that raised issues of critical public importance. The visibility of Britain's top judges increased dramatically. For example, between 2009 and 2013, one third of the Supreme Court's docket consisted of human rights cases, some of which made headlines (Dickson 2013: 14, ft. 72). The court upheld human rights claims in cases concerning points of law such as the legality of freezing the assets of terrorist suspects; the admissibility of evidence obtained during secret hearings in cases that compromise national security; or the rights of sex offenders.⁶

Despite the Supreme Court's heightened political relevance, scholars indicate that it can be hardly considered an activist tribunal. This “may just be an aspect of the judges' more general restraint, or it may reflect a specific reluctance to develop human rights jurisprudence too quickly, given the populist reaction there might be to so doing” (Dickson 2013: 51). The Supreme Court is indeed mindful of its public image. For instance, it makes a point of distributing case summaries to the media in order to control the narrative, and frequently convenes large panels to decide on tricky human rights cases so that it can “present a more powerful and united front to the outside world” (Dickson 2013: 4).

Adding to its ever more complex docket, in November 2016 the Supreme Court was asked to review a decision rendered by the High Court of Justice of England and Wales, which had ruled the Government was required to seek parliamentary authorization to formally start the country's withdrawal negotiations with the European Union.⁷ The government appealed the lower court's decision in part arguing that it was free to sidestep Parliament because the British people had themselves authorized talks with Brussels in the June 2016 referendum. The Supreme Court was thus forced to make a decisive intervention in the Brexit saga, assuming the obvious risks involved in dealing with such a controversial matter.

Backlash came instantly following the High Court ruling. Some politicians, including members of the populist United Kingdom's Independence Party, fiercely attacked the judges for meddling with the will of the people as expressed in the referendum. According to these critics, judges sought to slow down Brexit by adding veto points to an already arduous process. After the government appealed the decision, the tabloid press trained its eyes on the Supreme Court. The emphasis was almost invariably on the unelected status of the judges and their pro-establishment bias. For example, the anti-European Daily Mail, which has an estimated readership of around 4.5 million,⁸ developed the so-called Europhile rating which highlighted the elite educational background of Supreme Court judges and their formal and informal ties to the European Union.⁹ Six out of 11 judges received four or five stars out of a maximum of 5, indicating their establishment bias.

The Supreme Court finally ruled against the government on January 24, 2017.¹⁰ Among the court's most important institutional interlocutors, the leadership of the ruling Conservative Party expressed disappointment but chose not to protest, and like the opposition Labour Party, it pledged to comply with the decision by holding a vote in Parliament. After defeating a series of amendments passed by the House of Lords, the government's bill was approved unamended in a second vote in the House of Commons on March 17, 2017, attracting support from both sides of the aisle. Importantly, the Labour Party, consisting of mostly pro-Remain MPs, contributed with votes that ensured the passage of the legislation.

We are interested in assessing whether different framings of the decision shape public acceptance of the judgment as binding and final.¹¹ Studying reactions to this particular ruling has two central methodological advantages. First, the unusual salience of the case strengthens the face validity of the experimental setup: it allows us to avoid the use of survey questions that refer to either hypothetical rulings, or legal battles with which respondents are not familiar. Second, it is a “hard” case to detect framing effects due to the highly polarizing nature of Brexit. When respondents hold strong and clear views on an issue it is much harder to alter baseline opinions with experimental manipulations.

To test the three hypotheses outlined previously, YouGov administered an online survey of 3287 British voters on our behalf. The survey went live on January 26, just two days after the court's decision was made public, and responses were collected during a four-day period.¹² The instrument featured a between-subjects experiment with one control and four treatment vignettes containing different information about the Brexit decision. All vignettes were immediately followed by a question intended to measure acceptance of the ruling as the final word on the subject. Like similar studies of the U.S. Supreme Court, the dependent variable captures the extent to which voters are willing “to cease opposition and get on with politics,” not whether voters like or dislike the judgment (Gibson et al. 2005: 188; Nicholson and Hansford 2014):

“Do you think that the decision ought to be accepted and considered to be the final word on the matter? Or do you think that there ought to be an effort to challenge the decision and get it changed?

1. 1. I strongly believe the decision ought to be accepted.

2. 2. I somewhat believe the decision ought to be accepted.

3. 3. I somewhat believe there ought to be an effort to challenge the decision.

4. 4. I strongly believe there ought to be an effort to challenge the decision.”

There are two aspects of the case that require some elaboration. On the one hand, the Supreme Court's decision was final and could not be appealed. Moreover, this is a one-off case, thus a similar fact pattern is unlikely to be found in other cases, making it difficult to challenge the overall legal principle established in the ruling. Combined, these two characteristics limit the possibilities available for an attempt to legally challenge the decision. We therefore expect that acceptance of the ruling will be driven not only by a recognition that the Court can legitimately settle political conflicts, but also by an understanding that from a formal point of view this is a final and binding decision. This means that we expect high baseline levels of acceptance, which might generate ceiling effects, thus making the design conservative in identifying the impact of treatments that are expected to boost acceptance of the ruling. On the other hand, the Supreme Court's decision comes as a qualification of the popular will expressed in the Brexit referendum. Respondents might thus refrain from accepting the court's decision not only on the basis of their own preferences vis-à-vis Brexit but also because they might perceive the ruling as an effort to defy the popular will. This could translate into reluctance to accept the ruling even by respondents who voted for Britain to remain in the European Union. These considerations can interact in interesting ways generating variation in public opinion toward the ruling that does not necessarily match perfectly with attitudes toward Brexit. With this in mind, we examine the impact of our four experimental frames.

Each version of the experiment was randomly assigned to the participants in the study. Respondents in the control group were shown the text below, which contains neutral information about the ruling:

[CONTROL] “On Tuesday the Supreme Court of the United Kingdom decided that the government cannot begin negotiations to leave the European Union until Parliament votes to authorise Brexit.”

The treatment vignettes emphasized additional aspects of the ruling. To explore compensation effects (Hypothesis 1), we relied on an interesting feature of the judgment. In addition to demanding a vote by the parliament in Westminster, some of the plaintiffs asked the Supreme Court to rule that the assemblies of Scotland, Wales, and Northern Ireland ought to be given a voice in the process. The government, and Brexit supporters more generally, feared that the participation of regional parliaments could block the start of withdrawal negotiations indefinitely because Brexit was less popular in the regions, especially in Scotland. By agreeing to the need for the Westminster Parliament to authorize the beginning of departure talks, but at the same time rejecting the view that every legislative assembly in the country ought to have a say in the process, the Supreme Court effectively compensated the losers.¹³ The compensation framing reads as follows:

[TREATMENT 1] “On Tuesday the Supreme Court of the United Kingdom decided that the government cannot begin negotiations to leave the European Union until Parliament votes to authorise Brexit. It also ruled that the regional parliaments of Scotland, Wales and Northern Ireland do not need to be consulted.”

To measure compliance effects (Hypothesis 2), we leveraged the compliance pledges made by the Conservative and Labour parties in the aftermath of the ruling.¹⁴ We use two separate vignettes that vary the identity of the complier in order to explore whether compliance affects mass acceptance via a mechanism that is partisan in nature. If what matters is the institutional status of the complier, not its partisan identity, we should observe similar effects across treatments, and similar effects across partisan groups among respondents:

[TREATMENTS 2 & 3] “On Tuesday the Supreme Court of the United Kingdom decided that the government cannot begin negotiations to leave the European Union until Parliament votes to authorise Brexit. The Conservative government has pledged to comply with the Court's decision // The Labour opposition has pledged to comply with the Court's decision.”

Finally, the populist framing called attention to judges' elite educational background:

[TREATMENT 4] “On Tuesday the Supreme Court of the United Kingdom, consisting of judges educated at elite universities such as Oxford and Cambridge, decided that the government cannot begin negotiations to leave the European Union until Parliament votes to authorise Brexit.”

There are several reasons why we chose this manipulation. First, in the United Kingdom, degrees from universities such as Oxford and Cambridge are an informative marker of elite status. In fact, those who criticize the self-serving and out of touch nature of political elites often make reference to the fact that most of them attended these institutions of higher education, where they met each other, and in the populist imaginary, began to collude and build the networks that enabled their future political success. Second, elite educational background is one of the few elements of the populist attack arsenal that can be objectively measured. Judges' attitudes toward the European Union, for example, are harder to gauge in order to craft a plausible and nondeceptive framing. Third, we believe this is a subtle indication of pro-establishment bias that is less prone to generating strong demand effects and trigger confirmatory bias.

Results

As noted in the previous section, elites quickly decided to “cease opposition and move on with politics” once the Supreme Court handed down its decision. By contrast, according to our survey, the public exhibited significantly higher levels of variation in acceptance, giving us an excellent opportunity to explore framing effects. We find that only 45 percent strongly believe that the decision ought to be accepted. The majority displays varying levels of skepticism. One out of four respondents (26 percent) somewhat believes that the decision ought to be accepted, whereas the remaining 29 percent believes either strongly or somewhat that an effort should be made to challenge it.¹⁵ While the instrument does not allow us to know what voters had in mind when they stated support for a “challenge,” it is fair to say that for a nontrivial percentage of the British electorate the Supreme Court is not perceived as the final arbiter, capable to settling this important political question.¹⁶

We begin the analysis by testing Hypothesis 1, which postulates that compensations curb rejection of the judgment among losers.¹⁷ The first treatment reminded voters that although the Supreme Court forced the government to obtain Westminster's approval before launching negotiations with Brussels, it minimized the number of additional veto points by concluding that regional parliaments need not be consulted. We expect this compensation to boost acceptance among Brexit supporters (Leavers), who are the main losers in the case.¹⁸ As shown in Figure 1, those who voted to remain in the European Union in 2016 (Remainers) are on average more likely to accept the ruling than Leavers, regardless of the experimental condition. In addition, the court's decision to give voice only to the Parliament in Westminster does not have any effect on the overall acceptance of the judgment among Remainers (76 percent accepting the court's decision in both conditions). For Leavers, by contrast, this aspect of the ruling does operate in a compensatory manner, increasing support for the decision: whereas acceptance reaches 56 percent (95 percent confidence interval [CI]: 52.2-60.4) among Leavers in the control condition it escalates up to 64 percent (95 percent CI: 59-70) among those in the treatment group.¹⁹

Figure 1. Compensation Effect According to EU Referendum Vote.

Figure 2 provides a closer look at the magnitude and statistical significance of these differences. The first entry denotes the difference in the probability of accepting the ruling as final among Remainers after being told that the regional parliaments will not be asked to vote on this matter. The difference is practically zero. By contrast, the same difference among Leavers, shown in the second entry, is 8.6 percent and statistically significant (p < .05). The resulting difference-in-differences estimate, shown in the third entry of the graph, is 8.4 percentage points (s.e. 4.0 percentage points, p < .05), indicating a statistically significant compensation effect among the losers, compared to the winners.

Figure 2. Compensation Effect among Remainers, Leavers, and the Difference between the Two.

An alternative explanation for this finding could be that losers (Brexiteers) are also less positive toward regional parliaments and thus more willing to accept the ruling when reminded that these parliaments will not have a say, irrespective of what is their verdict on Brexit. In other words, instead of being compensated, losers under the compensation condition might simply express their taste for the marginalization of regional parliaments.

We address this potential confounder by delving into the mechanism underlying our hypothesis. Compensation implies that voters are aware that the regional parliaments could pose additional obstacles to a speedy authorization of talks with Brussels. This assumption, however, does not underpin the marginalization of regional parliament's hypothesis. Accordingly, although for the compensation hypothesis we would expect the effect to increase with political knowledge, no such heterogeneity is expected according to the alternative story. Using education as a proxy for political knowledge, we distinguish between respondents with (42.29 percent) and without (57.71 percent) university degree.²⁰ We repeat the same exercise as in Figure 2, separately for each education group. Full results are shown in the Online Appendix (Table B2). Here, we focus on the difference-in-differences estimate, which denotes the difference in the effect of the compensation treatment between Leavers and Remainers. Consistent with our expectations, the differential treatment effect is 5 percent among the first group (95 percent CI: −5 to 16 percent) and 15 percent among the second group (95 percent CI: 3-27 percent). In other words, among non-university degree holders, the compensation treatment does not seem to increase support for the ruling among Leavers significantly more than among Remainers. Instead, among individuals with a university degree, the compensation effect is significantly higher for Leavers than for Remainers, as the compensation mechanism would lead us to expect.²¹

That said, it is still possible that the two mechanisms complement each other. The treatment, in this case, is expected to vary in terms of magnitude according to whether compensation meets taste for devolution. To see whether this is the case, we use as a proxy respondents' place of residence. We expect respondents to react differently to the treatment depending on where they live. This is because of the relatively higher salience of devolution preferences outside England. In particular, for Leavers residing in Scotland and Wales, the compensation effect could be offset by the court's refusal to allow their parliaments to have a say during the Brexit process. Similarly, Scottish and Welsh Remainers are unique in that they are negatively affected by the compensation in two ways: they do not get to add more veto points to try to prevent Brexit and their regional representatives are sidelined from the debate. The treatment should therefore curb support for a ruling they otherwise like.

To explore these observable implications of the compensation hypothesis, we replicated the analysis comparing respondents from England and those in Scotland and Wales.²² Figure 3 shows that although the effect of the treatment is positive and significant among English Leavers, it is not statistically different from zero among Brexit supporters who reside in Scotland and Wales. Although the large standard errors surrounding the latter estimate could be a function of the small number of Scottish/Welsh voters in the sample, the result lends plausibility to our conjecture. Also as expected, the treatment has a large negative and significant effect among Scottish/Welsh Remainers, who are double losers from the compensation.

Figure 3. Compensation Effect in England and Scotland/Wales.

We now turn to the compliance hypothesis. The first panel in Figure 4 shows the result of pooling respondents in the two relevant treatment groups, and comparing them to those in the control group: being told that either of the two parties has endorsed the ruling boosts acceptance by approximately 6.5 percentage points (95 percent CI: 2.2-10.7). The second panel considers the treatment groups separately: the effect of the Labour pledge (8.4 percent, 95 percent CI: 3.8-13.4) is double in size than the effect of the Conservative pledge (4.2 percent, 95 percent CI: −1 to 9.1). These results suggest that compliance pledges boost acceptance, but what drives the effect?

Figure 4. Combined and Disaggregated Compliance Effect.

In what follows we test whether, consistent with the partisan mechanism, the observed change in acceptance is reinforced among supporters of the party making the pledge or, as the institutional mechanism would suggest, pledges are able to lend legitimacy to judicial outcomes across the board by setting a visible example of the appropriate way to respond to rulings. The analysis distinguishes between Labour and Conservative supporters by looking at vote choice in the May 2015 General Election.²³ The full results are presented in the Online Appendix (Table B3).

To illustrate these results, Figure 5 plots the effect of the two treatments among Conservative and Labour voters. Each entry denotes the difference in the probability of accepting the ruling between those receiving a party pledge and those in the control condition. The first panel looks at 2015 Conservative Party voters. We see that reminding this group that the Conservative government has pledged to accept the ruling leads to higher acceptance of the verdict (dashed line). Specifically, support increases by 13 percentage points (95 percent CI: 5.4-20.6). A similar result, but smaller in magnitude, emerges for the Labour Party's pledge among Labour 2015 voters (second panel, solid line): when these voters are reminded that their party has agreed to respect the decision, mass acceptance increases by 8 percentage points (95 percent CI: 0.4-15.5).

Figure 5. Compliance Effects by Partisan Group.

The smaller effect of Labour's pledge among Labour supporters makes sense. Although Conservative supporters were split on the Brexit issue (in our survey 56 percent of Conservatives voted Leave and 42 percent voted Remain), Labour voters broke decisively in favor of staying in the European Union (in our survey, 66 percent voted Remain). This means, for example, there were fewer Labour Leave voters who needed an endorsement from their party as an extra push to accept the Supreme Court's decision.

At face value, these results lend some support to the idea that compliance effects are driven by a partisan mechanism. At the same time however, they point to a paradox: whereas the overall effect of Labour's compliance pledge is stronger than that of the Conservative Party, the effect of Labour's pledge among Labour voters is more modest than the effect of the Conservative pledge among Conservative supporters. This suggests that perhaps there is more to the effect of Labour's pledge than partisan endorsement: Labour's pledge may operate through a more institutional mechanism, having wider implications.

In order to further explore this conjecture, we look at the effect of both treatments on the views of those who support the rival of the party making the pledge (Conservative voters for Labour's pledge, and Labour voters for Conservative's pledge). This effect is shown in the second entry for each partisan group in Figure 5. We see that when Conservatives are told that the Labour party is willing to comply with the ruling (first panel, solid line), they are more likely to accept the judgment: acceptance increases by 8.9 percentage points (s.e. 3.9 percentage points, p < .05). Conversely, when Labour supporters are told that the Conservative government is willing to comply (second panel, dashed line), acceptance of the court's decision remains practically unaffected.²⁴ The wider effect of the Labour Party's pledge is further confirmed by looking at the main effect of the treatment, as shown in Table A1 in the Online Appendix. This coefficient represents the change in the probability of accepting the judgment among respondents who voted neither for Labour nor the Conservatives. We see that even among this heterogeneous group, mentioning the intention of the Labour party to comply with the decision boosts acceptance by almost 10 percentage points (95 percent CI: 3.3-16.2). No such effect is found for the Conservative party, which seems to operate in a more polarizing fashion: its compliance pledge increases acceptance only among partisans.

Although there is some evidence that both the institutional and partisan mechanisms are at play, we are not in a position to fully adjudicate between these competing interpretations. The broader impact of the Labour treatment, suggestive of a more institutional mechanism, is particularly intriguing. There are several possible explanations. First, it could be the case that the Conservative Party is simply more polarizing. Second, voters could perceive Labour's pledge to comply as more costly and therefore more credible qua signal of the importance of accepting judicial decisions as final. Indeed, the parliamentary debate following the court's decision proved extremely divisive for Labour, whereas the Conservative Party emerged relatively unscathed.²⁵ Finally, the observed difference in effects could hinge upon the different role that opposition parties play in democracies more generally. The loyalty of opposition groups toward the regime, its institutions, and policy outputs is perhaps more informative when voters are making up their minds about whether or not to accept decisions rendered by specific actors.

Finally, we explore the effects of the populist treatment. Mentioning the judges' elite educational background should increase rejection of the judgment as final by activating populist sentiments. The average treatment effect is, however, indistinguishable from zero: 0.019 (s.e. 0.025), suggesting no such effect. But is this because of heterogeneous treatment effects canceling each other out? An obvious source of heterogeneity comes from the division over Brexit. One could argue that although for Leavers the treatment is a sign of establishment bias, Remainers may see it as a sign of expertise. We find no support for this conjecture. The results are also the same when we look at differences in socioeconomic status and support for populist parties.²⁶ Three explanations come to mind. First, but unlikely, populist framings could indeed be innocuous. Second, the elite status of members of the judiciary could be already priced in when respondents evaluate rulings. In other words, the treatment is not telling people anything they do not already consider when assessing judgments. Third, the treatment may be too weak. When populists attack liberal democratic institutions they normally adopt a much more caustic turn of phrase.

Conclusion

This paper explored several sources of variation in acceptance of judicial decisions as authoritative, binding, and final. The analysis went beyond the usual focus on the consequences of depictions of judicial decision making as either partisan or legalistic. We tested conjectures about the potential of outcome compensations offered to losers and compliance pledges made by relevant third parties to boost mass acceptance. We find that courts can indeed increase acceptance among losers by offering compensations, and that the reactions of important institutional interlocutors affect whether or not citizens accept rulings. The results support assumptions made by scholars of strategic judicial behavior regarding the benefits of crafting rulings with an eye on the preferences of key audiences. But they also underscore that acceptance is also at the mercy of factors over which judges have less control, and may be instrumental and wavering rather than principled and committed.

Our argument about compensation and compliance effects is likely to be applicable across a wide range of judicial decisions. When it comes to consequential judgments involving distributional conflicts, complex policies, and debates about the content of fundamental rights, there is almost always a state actor (bureaucrat, government agent, etc.) that has to comply. Their pledges can in all cases be significant determinants of mass acceptance. A fruitful avenue for future research would be to explore the authority-conferring effects of a multiplicity of actors, including incumbent versus opposition parties, as well as nonpartisan institutional interlocutors. In this sense, our paper suggests that not all forms of compliance are created equal since Labour and Conservative party pledges behave quite differently.

Similarly, in complex cases, judges usually have the possibility to hand down balanced decisions, either out of strategic, ideological, or purely legalistic considerations. This means that the study of compensation effects can be productive elsewhere. To be sure, the characteristics of compensations will vary from case to case, but even in cases that involve specific individuals or groups, compensations relevant to a larger audience are possible because decisions tend to speak to broader value clashes and set precedents. For example, a court could limit the reach of a controversial jurisprudential innovation on fundamental rights by limiting the extent to which it recognizes a new right, or by putting forward a narrow set of obligations that the state has to fulfill in relation to that new right. Our study supports the view that when courts make such compensations, effectively signaling them to voters can boost acceptance among losers.

In line with our effort to push the literature on courts and public opinion beyond the U.S. case by putting forward portable theoretical propositions, we also analyzed whether highlighting nonpartisan forms of judicial bias can also affect mass acceptance. We find no evidence that framing rulings using a discourse commonly deployed by populists to attack judiciaries undermines acceptance. It is certainly too soon to conclude that high courts are immune to these populist attacks. Future research should explore how the language deployed by the radical right affects the authority of courts to render binding decisions using a variety of additional framings. For example, experimental vignettes could highlight the antimajoritarian nature of judicial institutions more directly. Similarly, information about deference to or defiance of international court precedents, which have a clear association with the values of the “liberal establishment,” could activate populist sentiments. Finally, the power of populist discourse is likely to vary depending on the issues at stake and the preexisting levels of polarization. In this sense, the intensity of sentiments surrounding Brexit set a relatively high bar for the hypothesis.

The paper also offers lessons about the relationship between the U.K. Supreme Court and the public. The data suggest that commanding acceptance of key rulings is not straightforward. Although the court was able to survive the noise surrounding the Brexit case, there is significant variation in the extent to which citizens are willing to see the institution as a legitimate final arbiter of the most fundamental political conflicts. This suggests that legitimacy-building efforts should not be off the table. In particular, offering compensations and securing compliance emerge as potentially productive strategies for future cases. Despite the extremely unusual polarizing nature of Brexit, these frames clearly boosted acceptance of the judgment, which suggests that in less politically charged cases, these strategies are also bound to work in the court's favor. Finally, it is certainly too soon to know whether the Brexit ruling constituted a formative moment that decisively shaped how citizens relate to the institution for the foreseeable future. However, experiments could prove useful for more immediate evaluations of the impact of the Brexit case on the Court's legitimacy. For example, researchers can assess whether framing questions about judicial legitimacy or diffuse support with a reminder of the court's Brexit decision significantly moderate responses.