Jump to ContentJump to Main Navigation
Electoral ShocksThe Volatile Voter in a Turbulent World$

Edward Fieldhouse, Jane Green, Geoffrey Evans, Jonathan Mellon, Christopher Prosser, Hermann Schmitt, and Cees van der Eijk

Print publication date: 2019

Print ISBN-13: 9780198800583

Published to Oxford Scholarship Online: January 2020

DOI: 10.1093/oso/9780198800583.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (oxford.universitypressscholarship.com). (c) Copyright Oxford University Press, 2022. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use.date: 01 July 2022

Turbulent British Politics

Turbulent British Politics

An Explanation

Chapter:
(p.27) 3 Turbulent British Politics
Source:
Electoral Shocks
Author(s):

Edward Fieldhouse

Jane Green

Geoffrey Evans

Jonathan Mellon

Christopher Prosser

Hermann Schmitt

Cees van der Eijk

Publisher:
Oxford University Press
DOI:10.1093/oso/9780198800583.003.0003

Abstract and Keywords

This chapter sets out a novel theoretical approach for understanding electoral change. This approach focuses on the mechanisms and consequences of electoral shocks in the context of electoral volatility. The chapter sets out the reasons for focusing on shocks, a definition of electoral shocks, and the mechanisms through which shocks affect voting behaviour. We argue that the consequences of electoral shocks depend on how politicians respond to those shocks and compete around them, and how they are politicized in the wider media environment. The chapter explains why the gradual destabilization of the party system means that electoral shocks are having increasingly dramatic consequences.

Keywords:   shocks, events, elections, political change, partisan dealignment, critical elections, salience, competence, party identification

British politics has become less predictable and the party system less stable. More voters are switching their vote choices than ever before. This switching has led to dramatic changes in election outcomes and dramatic changes in political party support over very short time periods. What can account for this instability in British politics, and what does this instability tell us about the outcomes of the two most recent British general elections?

Our explanation focuses on the long-term and short-term antecedents of electoral choice.

We start with a foundation of the gradual and long-term changes that have made voters more likely to switch their support to different parties: the changing long-term context of volatility. Chapter 4 deals with these changes in greater detail and explains why they have come about. Here we explain why the gradual destabilization of the party system means that shocks can result in extraordinary political consequences; from the largest vote share for parties other than Labour, the Conservatives, and Liberal Democrats in 2015, to the largest two-party vote share since 1970 in 2017. However, there are still features of the system that provide stability, not least the advantages enjoyed by the major parties in retaining voters and the majoritarian electoral system.

The remainder of this chapter focuses on the mechanisms and consequences of electoral shocks in the context of increasing electoral volatility. We set out the reasons for focusing on the impacts of shocks, a definition of electoral shocks, and the mechanisms through which shocks affect voting behaviour. Our approach to the study of electoral shocks recognizes that the potential impacts of shocks are multifaceted and can have far-reaching, system-wide effects. Shocks do not have inevitable consequences, and are not defined by their consequences. They create political opportunities, and their consequences depend on how politicians react to them and compete around them, and how they are politicized in the wider media environment. The chapter also explains why the gradual destabilization of the party system means that electoral shocks are having increasingly dramatic consequences.

This account of electoral change departs from a focus on single explanatory factors or votes for a particular political party. It rejects a false choice between (p.28) bottom-up demand-based explanations and top-down political supply-oriented explanations. Political outcomes are more complex than this. Understanding the changing nature of the party system calls for an explanation that is broad and multifaceted. We need to understand how the structures and incentives that underpinned the stable party system have become weaker over time, what the consequences of that weakening are, how they account for instability in British politics, and what the prospects are for a return to greater stability in the future. This chapter sets out an explanation that considers relatively short-term but complex factors—electoral shocks—and situates them and their effects in the context of long-term gradual political change. We explain how these long-term and short-term factors interact: the impact of electoral shocks can be accentuated when party attachments are weak.

3.1 The long-term trend in volatility

In Chapter 2 we showed how sharp and trendless fluctuations in aggregate level volatility have been accompanied by a long-term and gradual increase in individual level volatility over the last five decades. The long-term trend reflects an electorate that is more fluid and potentially more responsive to the choices provided by political elites and parties.

The gradual rise in individual-level volatility can be explained by the combination of two other long-term trends, the evidence for which is set out in Chapter 4.

The first long-term trend is the gradual and sustained reduction in the strength and number of people identifying with political parties over time—a process widely referred to as ‘partisan dealignment’ (Särlvik and Crewe 1983; Dalton 2000). As people have become less attached to political parties—both in terms of the number of party identifiers in the electorate and the weakening strength of the attachment of those who still identify with a party—the ballast in the party system has been steadily eroded. Partisan attachments have the effect of stabilizing the party system. Conversely, partisan dealignment has the effect of destabilizing the party system. A system with strong social identities built around parties tends to reproduce itself for several reasons. First, in an electorate with strong party attachments, a higher proportion of voters have a default vote choice that they are likely to revert to at each election. In other words, voting can have a habitual element (Plutzer 2002), with voters demonstrating a ‘homing instinct’ towards a foundational party over consecutive elections (Butler and Stokes 1969b). Second, party identifiers tend to socialize their children into their own partisanship (Zuckerman, Dasovic, and Fitzgerald 2007), so that successive generations have some of the distribution of partisan leanings to their parents. This gives the system stability over the long term, such that there is ‘memory’ in (p.29) partisanship, inherited over time. Third, partisan identification acts as a perceptual filter, which means that voters interpret political information through the prism of their pre-existing political commitments (Lodge and Taber 2013), making the system less vulnerable to disturbances, such as negative performance or policy change. Partisans are more likely to reject information that conflicts with their prior beliefs, by rejecting the message or the source (Zaller 1991; Zaller 1992), and attribute responsibility for problematic outcomes to other political actors or institutions (Tilley and Hobolt 2011). Indeed, some scholars see this filter as the primary way in which partisanship affects vote choice (Bartels 2002). Weaker party identifiers and those with no party identity, on the other hand, are more likely to consider other party choices because they are open to information from other parties. As partisan dealignment means there are now more people in Britain with weak attachments to political parties, or no attachment to political parties at all, this partisan dealignment creates the conditions for the system to become unstable because more people are available to switch between parties.

The second long-term factor is the increase in the proportion of votes won by parties other than Conservative or Labour between 1950 and 2015, which we refer to as fragmentation.

As the electorate has become increasingly ‘dealigned’, minor parties have seized opportunities to gain votes from a more fluid and available electorate. As we showed in Chapter 2, the share of the vote won by the two major parties declined steadily from the 1960s to 2015. However, 2017 saw a sharp increase in the two-party share of the vote and a drop in fragmentation. The fragmentation of votes in British elections has directly affected volatility because, compared to the major parties, smaller parties struggle to retain voters from one election to the next. We demonstrate the extent to which this is the case in Chapter 4. The fact that major parties are more likely to retain their voters successfully between elections is the reflection of a number of important and continued stabilizing features of British politics which create inertia and help maintain the party system over time, albeit with some major disturbances. Increasing returns to electoral success stem from the institutionalized advantages enjoyed by established larger parties including high start-up costs for new parties, funding disparities, differential media coverage, and advantages bestowed by the electoral system (Pierson 2000). Success leads to success, especially in the British majoritarian electoral system, as larger vote shares are rewarded with a disproportionate number of seats. This in turn provides a strong disincentive to ‘waste’ votes by supporting smaller, less viable alternatives (Duverger 1954). Major parties also have a significantly greater likelihood of being able to form a government, creating further incentives for voters to support them at the expense of minor parties (Green, Fieldhouse, and Prosser 2015). Additionally, major parties enjoy informal advantages of an established support (p.30) base. By virtue of having more supporters and partisans, major parties benefit from habitual voting and the passing down of partisan preferences from generation to generation through political socialization, as discussed above. Similarly, this greater support base provides advantages in terms of interpersonal influence (Huckfeldt and Sprague 1995), mobilization (Rosenstone and Hansen 1993,) and normative pressures (Fieldhouse and Cutts 2016). Smaller parties have to overcome all of these built-in advantages that favour their larger rivals in order to attract voters and retain them in subsequent elections. Even those voters who are convinced that a smaller party is viable in one election are likely to grow disillusioned when success does not materialize, and then switch back to a more viable option in subsequent elections.

Taken together, these factors make it difficult for minor parties to attract and retain voters, and therefore to sustain their support from one election to the next. Because of the structural disadvantages facing minor parties, increased votes for minor parties in one election tends to lead to an increased level of volatility in subsequent elections, as smaller party voters are more likely to switch parties. In Chapter 4 we show that this effect is substantial, accounting for a large portion of the increase in individual-level volatility since 1964.

Partisan dealignment and fragmentation change the system in more ways than just the total amount of switching between elections. A dealigned electorate is one that has greater potential to respond to stimuli and political disturbances—the storms and headwinds of politics—because there are fewer stabilizing factors for these stimuli to overcome. The system has less inertia. Similarly, because minor party voters already have a high probability of switching between elections, a larger number of minor party voters means a larger pool of voters who are more vulnerable to the effect of shocks and other stimuli. This might be seen as normatively desirable, creating a closer connection between political actors and the mechanisms of electoral accountability. However, a more tumultuous and unpredictable electoral environment may have its own risks, particularly where parties miscalculate the likely outcomes of their policy offerings and where voters find it harder to anticipate the likely outcomes of collective voting decisions.

Our analysis in Chapter 4 suggests that volatility cannot be completely explained by partisan dealignment and fragmentation, however. This raises the question of how else we might explain the increase in vote-switching and the dramatic changes in British elections over a short period of time. An increase in the willingness of voters to switch parties is not sufficient to explain dramatic changes in support for particular parties at particular elections. To understand these outcomes, we need an explanation that can account for the choices that the more volatile electorate makes in a particular election. Furthermore, we need to explain how the long-term destabilization of the electorate provides the context for such shorter-term dynamics to have greater effects.

(p.31) 3.2 Electoral shocks

In economics, international relations, and public policy studies, systems undergo sharp changes in outcomes in response to shocks. In the absence of shocks, systems are expected to function in a relatively stable and incremental way. Economists state that system shocks ‘interrupt and disrupt the process of economic growth and development’ (Martin 2012, 3), leading to long-lasting societal implications and changes in public policy (Rodrik 1999). In international relations, system-level shocks can be necessary prerequisites for changes to otherwise intransigent tensions. ‘A political shock is a dramatic change in the international system or its subsystems that fundamentally alters the processes, relationships, and expectations that drive nation-state interactions’ (Goertz and Diehl 1995, 31). In public policy, major policy change comes in bursts in response to pressure accumulating, external events, or shocks, known as a process of ‘punctuated equilibrium’ (Baumgartner and Jones 1993). Shocks change policy paradigms and can have permanent, wide-reaching consequences (Hall 1993). These ideas can be very usefully introduced to the study of party systems and elections.

In electoral politics, we propose that electoral shocks are the disturbances that have the ability to lead to substantial and dramatic increases in vote-switching, and therefore to changes in the party system. They may alter the political system in the short-term, and potentially the long-term—cutting through ‘normal’ political ebbs and flows, loyalties, and levels of public inattention to politics. They are not, however, defined by their consequences.

Electoral shocks are unavoidable, high-salience changes or events that can prompt large sections of the population to update their political evaluations and party preferences. This is in contrast to more stable, ‘normal’, or uneventful periods in politics in which voters have fewer reasons to update their partisanship and when new information can be more readily rationalized into pre-existing beliefs. Shocks cannot be as easily avoided via partisan selection mechanisms of information and social networks, or through partisan rationalization. This is consistent with work showing that major economic shocks alter the relationship between partisanship and economics. In stable economic periods, economic evaluations are more likely to be endogenous to party preferences, driven by party attachments and voting behaviour. In times of exogenous economic shocks, however, partisanship is more likely to be updated in response to economic changes (Chzhen, Evans, and Pickup 2014).

A definition of electoral shocks

A electoral shock has the potential to be recognized even by people who might otherwise rationalize, ignore, or attribute responsibility to someone else (Green and (p.32) Jennings 2017). Electoral shocks are not the minor everyday routine political happenings such as misdemeanours, resignations, speeches, or announcements. A major event or discontinuity must be highly salient and relevant to party choice and competition to have the power to cut through traditional loyalties, inattention to politics, and cause dramatic political change. Shocks can cause people to re-evaluate their political preferences in a way that everyday politics should not. This means that shocks can lead to substantial volatility, shifting people out of their habitual voting behaviour.

Electoral shocks might not always have major consequences in terms of abrupt system-wide electoral change. They could occur within a very stable system that is resistant to the effects of external shocks, they could reinforce rather than cut across stable patterns of electoral choice, and their effects might be dampened by the failure of political actors to capitalize on them. If the concept of electoral shocks is to be theoretically and analytically useful, they must not be defined by their consequences. However, electoral shocks are necessary conditions for abrupt system-wide changes to occur alongside the broader context of a system less constrained by party loyalties.

In terms of their defining characteristics, electoral shocks: 1) represent a sharp change to the status quo outside the normal course of politics; 2) are highly salient and noticeable over prolonged time periods, and 3) are relevant to party politics.

We now elaborate on these properties in greater detail.

1. Electoral shocks are an abrupt change to the status quo. They are not necessarily exogenous to the party system, but they are more than simply the outcomes of normal everyday politics. They represent a significant and often unanticipated change.

Electoral shocks are extraordinary political events or changes, representing a departure from the status quo. Many such events might be described as ‘exogenous’, originating outside of the political system. In reality, however, most events are not entirely exogenous. For example, wars, economic recessions, and major political crises usually have origins inside the political system, reflecting the coming together of a complex array of decisions made by political actors. However these events frequently transpire because of external factors or contingencies that could not be foreseen. They are therefore not the inevitable outcome of the usual, more predictable pattern of policymaking and politics. Such events may be considered electoral shocks. In contrast, where events or decisions are within the normal gamut of party politics we should not consider it a shock. This also helps explain why we define shocks as abrupt changes to the status quo. Because their origins are at least partly exogenous it is possible to determine when they occurred: there is a discrete point in time at which a shock takes place, and whilst the effects of an electoral shock are likely to be prolonged, they will also be immediate.

(p.33)

2. Electoral shocks are manifest over prolonged time periods and are highly salient: they have the potential to be noticed and recognized even by people who do not have much interest in politics, and by people who might otherwise select into information that fits their partisan beliefs and preconceptions. Electoral shocks are very difficult for voters and politicians to ignore.

Most political ups and downs are little recognized by the public. As hard as it may be to believe for people who are fascinated by politics, most people do not know the outcomes of major political negotiations, who is in the cabinet, and many do not understand the major policy shifts of political parties, what parties stand for, or what parties are focusing on in their election campaigns. Some people may be able to recognize party leaders in only a superficial and cursory way, relying on rumours and the occasional story to form an impression. However, there are other moments in politics and in public life that are inescapable. They permeate public discourse and reach beyond the Westminster bubble. Such moments have the potential to shape public opinion, even fundamentally so. For something to effect a major change in the public, it has to be recognizable and more than a ‘blip’ in the public’s consciousness. As a result of being both substantial and persistent, shocks reach large numbers of people and they have the potential to create significant shifts in electoral behaviour. That is to say, they register in public opinion. They are also impossible for parties to ignore. Under normal conditions, political actors can frequently choose which issues to emphasize and which to ignore, usually opting to draw attention the issues that they ‘own’ (Petrocik 1996). However, electoral shocks are sufficiently salient that political actors are forced to engage with them even if they are electorally disadvantageous (Mader and Schoen 2018). Electoral shocks are highly salient and noticeable, and have the potential to change partisan attachments, party support, and to cut through public discourse in a way that regular events do not.

3. Electoral shocks are politically relevant and have the potential to change how parties are perceived and therefore to (re)shape the party system.

Something could happen in a country that is hugely significant, and also extremely salient, but it might not be political or relevant to party choice. However, for a shock to be relevant to a party system, it has to be party political in nature, or potentially party political in nature, enabling political parties to compete around it and for vote choices to be swayed on the basis of it. Electoral shocks must, then, be changes that have the potential to impact on the party system because they affect how voters evaluate or feel about different parties.

To see the need for these three requirements, consider the following events that lack each one of the three criteria, and which we would not therefore classify as electoral shocks.

(p.34) We could have a situation where there was an event or change that represented a sharp change from the status quo and was political, but failed to become salient. The establishment of the UK Supreme Court might be such an example. This was an important and abrupt constitutional change that almost entirely failed to register with the general public and had no apparent party political impact. It became temporarily salient in the media in 2017 when the Supreme Court ruled that an Act of Parliament was required to authorize withdrawal from the European Union, but in 2017 it did not remain noticeably salient over a sustained period or cut through public inattention to politics. In this period the Supreme Court did not penetrate the public consciousness and consequently its establishment was not an electoral shock.

We could also have a situation where a political event or change was highly salient but does not represent a sharp change in the status quo. Such events reflect the normal in-and-outs of regular party politics and do not fundamentally alter how parties are perceived in terms of what they stand for and how competent they are. For example, Theresa May’s ill-considered 2017 manifesto commitment concerning the funding of social care which was labelled a ‘dementia tax’ falls into this category. It could also apply to popular policies such as new commitments to increasing NHS funding, to changes in party leaders, and to election campaigns. In each, the event or development is political and salient but does not represent a sharp change to the status quo. There are other examples that might be more borderline in terms of definition. While the choice of party leaders reflect changes within a political party, and often bring shifts in party policy, they can usually be considered a direct consequence of everyday party politics. The rise and fall of Margaret Thatcher were highly salient and politically relevant, and marked a clear change in policy direction. However, it is arguable whether Thatcher brought about—as opposed to reflected—fundamental changes in the nature of British politics. Similarly, the election of Jeremey Corbyn to Labour’s leadership in 2015 had substantial effect on electoral politics in Britain. While the unusual circumstances of his election1 mean that this might plausibly be considered a shock, the circumstances which enabled his victory originated from within the Labour Party. His election to the leadership should therefore be considered part and parcel of normal party politics and of the regular shifts in policy and political representation that entails. A change in leader or policy might have very large effects on electoral outcomes, but these are better understood through existing frames of analysis such as spatial and valence politics. This highlights that the abrupt change criteria does not merely mean that an event or decision changes something, but that the change is atypical and does not arise from the normal course of politics.

(p.35) A third example would be a situation where an event was highly salient and represented a sharp change in the status quo but does not link to party politics. A clear-cut example of such an event would be if England (eventually!) won the World Cup. Such an event would be unavoidably salient and, at least in terms of national self-image, might be a large change from the status quo. But no party’s fortunes are highly linked to the England team’s performance, so it would not be an electoral shock.2 A more borderline example would be the death of a monarch. This would again be hugely salient and represent a large change in the status quo for the national experience. However, while there is an obvious political element to the monarchy, the current party system and state of public opinion provides no immediate way for the death of the monarch to translate into electoral consequences. The death of a monarch is a political event but not a party political event.

In reality, of course, the three components of shocks are usually overlapping. Furthermore, the archetypal shock clearly and unambiguously fulfils all three criteria, but shocks can vary in their size and significance. Change becomes salient because it is so significant; the larger and more consequential the change, the more it becomes worthy of media attention and public attention. Public salience provides an imperative for political parties to compete on something that matters to large groups of voters, with the potential to become party political. Shocks may arise because of political decisions—or at least be painted as the responsibility of politicians—being, therefore, inherently party political in nature. However, events should not be considered shocks unless they exhibit all three characteristics; each are necessary conditions but none are sufficient on their own.

Our definition of shocks requires a higher threshold of change than has been used in the existing political science literature. Researchers of presidential and prime ministerial approval have used the term ‘shock’ to denote a wide range of events that leads to an interruption to a time-series, with simple and direct subsequent effects at the ballot box (Mueller 1970, e.g.; Kernell 1978; Ostrom and Simon 1985; Nadeau et al. 1999). In the US, Kernell (1978) demonstrated the effects of the Korean War, the Vietnam War, the economy, Watergate, and international ‘rally’ events on presidential approval. In the UK, the Falklands War, the poll tax, the Exchange Rate Mechanism (ERM) currency crisis, Major’s reselection as Conservative prime minister, the Iraq War, have all been defined as shocks (Clarke and Lebo 2003; Green and Jennings 2017; Green and Jennings 2012). It is certainly true that some shocks in the existing literature would qualify as electoral shocks under our definition, for example, major economic crises and recessions, and the Watergate crisis.

(p.36) It is important to reiterate that shocks are not defined by their electoral outcomes. Even if a shock moves voters, it may not affect aggregate outcomes depending on what other flows happen in that election, which could counteract the aggregate level outcome of shocks. A shock may be responded to successfully by the parties in the system, political entrepreneurs may fail to mobilize effectively in response to the shock, or voters may be too attached to their current parties to be moved by the shock. We expand on the contingent nature of shocks in section 3.6 of this chapter. But first, we illustrate the concept of electoral shocks with the five examples of the electoral shocks which shaped the outcomes and voting behaviour in the 2015 and 2017 British general elections. We discussed, above, the properties of shocks, and some cases that we would not categorize as shocks. Here we summarize five shocks in recent British politics that do possess the properties of electoral shocks.

3.3 Five electoral shocks in recent British politics

The following examples illustrate considerable variation in their nature but each meets our definition. We will demonstrate in later chapters (5, 6, 7, 8, and 9) how each of the following shocks has fundamentally reshaped the outcomes of the recent 2015 and 2017 British General Elections, the ways in which they have done so, and also how the underlying trends in destabilization have combined with these electoral shocks to create dramatic outcomes.

1. The EU immigration shock

The first shock we consider is the rapid rise and sustained level of immigration to the UK, particularly following the 2004 accession of ten new EU member states whose citizens suddenly enjoyed freedom of movement to the UK. EU migration reflected a sharp change from the status quo because it very substantially increased the flow of migration into Britain, representing a step-change in immigration policy. It also disrupted the traditional pattern of thermostatic policymaking and public opinion that had previously characterized UK immigration policy (Jennings 2009). When public concerns about immigration had been high in the past, governments tended to tighten immigration rules to bring numbers down. Despite Conservative promises to do so again, however, freedom of movement within the EU prevented any meaningful steps to reduce migration while remaining in the EU. In essence, the UK government was unable to control EU immigration or manage its speed, which became politicized in public and political debate. This electoral shock was highly salient due to the corresponding rise in media coverage of immigration, and in turn immigration routinely topped the list (p.37) of most important issues stated by the public. Immigration from the EU was party political in nature due to: 1) the existing party competition over immigration and the willingness of the Conservative Party to use the issue against the Labour government; 2) the linkage of immigration to the long politicized issue of EU membership which had spawned two new challenger parties in the 1990s (UKIP, and formerly the Referendum Party) and; 3) the political campaigning of a charismatic radical right leader, Nigel Farage, who was well placed to capitalize on the EU immigration issue at the head of the already ascending party of UKIP.

2. The Global Financial Crisis

The second shock that reshaped British politics was the 2007–8 Global Financial Crisis, and the subsequent Great Recession. An economic crisis inevitably involves a change in the status quo as it directly affects the lived experience of the population in significant and far-reaching ways. It also led to a sharp, long-lasting, and important shift in policy—austerity—which was justified on the grounds of reducing the level of national debt and the government’s budget deficit. The Global Financial Crisis and Great Recession were hugely salient, both through media coverage at the time, the strong sense of outrage over the failure of banks, the unaccountability of financial institutions that were ‘too big to fail’, and their direct effects on the economic well-being of British citizens. These concerns were also salient politically, providing the backdrop and justifications for political competition, blame attribution, and policy shifts. An economic crisis is inevitably linked to party politics because the economy is one of the most important ways that an incumbent government is judged, and also because of the ways in which all parties responded. The crisis was clearly an electoral shock in the 2010 election, but its impact persisted through to 2015, as we demonstrate in Chapter 6.

3. The Conservative and Liberal Democrat coalition

The third shock that shaped British politics was the Conservative and Liberal Democrat coalition formed in the wake of the 2010 election that delivered a hung Parliament. The coalition represented a sharp departure from the type of government most British voters had ever experienced, and a major disruption in the image of the Liberal Democrats, and hence to the nature of party choices available to voters. The Liberal Democrats had primarily been seen as an anti-Conservative, centre-left party, with the majority of its supporters falling attitudinally closer to Labour than the Conservatives (Russell and Fieldhouse 2005). Most people who expected a coalition would have expected the Liberal Democrats to govern more naturally with the Labour Party. In the space of one week, the Liberal Democrats (p.38) abandoned this position, enabling a government led by a party that the majority of its supporters strongly opposed. The coalition also fulfils the salience and party political elements of the shock definition. A coalition government was not hugely unexpected in 2010. A coalition formed of the Conservatives and the Liberal Democrats, however, was a surprise. It was a huge news story, announced with great fanfare and sustained for five years of media attention. By virtue of being unusual and a different type of government to what people were accustomed to, it was certainly noticeable to the general public, and Nick Clegg, the Liberal Democrat leader, was a high-profile deputy prime minister. In short, if you knew one thing about the Liberal Democrats during this time period, it was likely to be that they had entered into a coalition with the Conservatives.

4. The Scottish Independence Referendum

The fourth shock that shaped British politics happened in Scotland. Since the advent of devolution in 1999, the SNP had been making inroads with Scottish voters in their campaign for Scottish independence. This culminated in a referendum on independence in September 2014. On the face of it, the outcome of the referendum was not a major change: the pro-independence ‘Yes’ side achieved 45 per cent of the vote, lost the referendum, and Scotland remained part of the UK. However, the referendum itself was a major departure from normal Scottish politics. It united the two major parties, Labour and the Conservatives, behind a common cause (the Union); it placed a high-stakes electoral choice into a binary decision, and in so doing demonstrated and deepened an existing schism in the Scottish electorate. It was unquestionably salient and unavoidable in Scotland, as well as throughout the UK (to a lesser extent). It permeated political and cultural life in Scotland for a prolonged and intense period, and it culminated in an exceptionally high level of participation, with an 84.6 per cent turnout. Finally, the independence referendum was clearly a party political issue given that the parties all took positions on independence on one side or the other, making it easy for voters to link their positions on independence to party politics.

5. The Brexit referendum and outcome

The final shock should come as no surprise: the 2016 EU referendum and the vote for Brexit. As with the Scottish independence referendum, the EU referendum exposed and accentuated visceral divides. The EU referendum was clearly a party political event: the referendum was called by the Conservatives to head off internal dissent and compete against UKIP, although the referendum campaign did not fall along party lines. The vast majority of Conservative, Labour, and (p.39) Liberal Democrat leadership opposed Brexit (although Labour’s leader, Jeremy Corbyn, was naturally more Eurosceptic), as did the bulk of their MPs. Similarly, most of the smaller parties—the SNP, Plaid Cymru, and the Greens—lined up on the pro-Remain side. Only UKIP was united as a party behind the Leave campaign. There were, however, significant divisions within the two main parties: the official Leave campaign was headed by prominent Conservatives MPs, including Boris Johnson, and a small number of vocal Labour MPs launched a ‘Labour Leave’ campaign.

The referendum campaign was divisive and highly salient, but so was the aftermath of the referendum. The Conservatives, with a new leader who had supported Remain (Theresa May), almost universally threw themselves wholeheartedly behind Brexit (with some notable but more peripheral exceptions). Labour maintained a position of studied ambiguity, adopting a softer, more critical version of Brexit, following a brief period of infighting between Corbyn and the Parliamentary Labour Party which was resolved in Corbyn’s favour. The Liberal Democrats, the SNP, and the Greens continued their opposition to Brexit. Suddenly, finding itself without a clear purpose, UKIP was engulfed in internal conflict. Since the outcome of the referendum, the trials and tribulations of the government’s negotiations with the EU dominated the news cycle. The EU referendum and its aftermath are inescapable, enduring, and very high in salience. Brexit is perhaps the most vivid and stark example there is of an electoral shock.

Each of these examples is of special interest to us because of their proximity to the election outcomes in 2015 and 2017. However, this does not mean that earlier shocks did not have any impact on previous elections, or that shocks have not always been important drivers of volatility and election outcomes. What is particularly important now, as we discuss presently, is that shocks are taking place within an electorate that also exhibits a greater propensity for volatility and sensitivity to shocks. That combination of underlying volatility and shocks is necessary to understand the outcomes of recent elections, the increasing level of switching and system volatility, and the vulnerability of the British party system to future disruption.

3.4 How electoral shocks work

We identify three mechanisms through which shocks drive electoral behaviour. This is not to say that the way in which shocks shape electoral behaviour can be easily divided into separate categories. They typically work through a combination of multiple interacting mechanisms. However, these categories help us theorize more clearly about the combinations of ways in which shocks can shape election outcomes.

(p.40) Salience

Parties and voters have policy positions, or potential positions, on a wide range of issues, but only a few of these are crucial in any particular electoral choice. However, a shock can change which of these issues a voter attaches weight to and how clearly they perceive the positions of the parties on the issue. The ‘non-ignorability’ of shocks makes them cut through other concerns and attachments, changing voters’ views of what is important in politics. Any major high-profile disruption to politics has the potential to alter the relative importance of issues (or issue dimensions) to electoral choice.

The effects of salience on the factors that voters prioritize in their vote choice are well established on a wide range of issues. When issues become more salient voters tend to become better informed about those issues and about the parties’ positions on them (RePass 1971). As a result, those issues are given greater weight in a voter’s calculus (Green and Hobolt 2008). Voters have a wide variety of views on issues that make it hard for any party to simultaneously satisfy their preferences on all fronts. Parties therefore find it advantageous to downplay certain issues or maintain constructive ambiguity over what their position is (Somer-Topcu 2015). When an issue is highly salient, however, this ambiguity is harder to maintain: voters will make greater efforts to obtain information about the issues they see as most important. The media will also make greater efforts to pin down a party’s position on a salient issue and communicate this to voters, while opposing parties may take the opportunity to convey negative information about their competitor’s position on the issue. Ambiguity is a harder strategy to follow when an issue is highly salient, although as we discuss in Chapter 9, this is arguably the strategy adopted by Labour in 2017 with respect to Brexit. Consequently, when an issue is salient, parties are more likely to adopt the priorities of the electorate (Budge and Farlie 1983) and manage policy positions and campaigns accordingly.

Several of our shocks work through salience to an important degree. The sustained high levels of media coverage about immigration acted as a salience shock which increased the weight of immigration attitudes in the vote calculus. The shock had further reverberations, of course, when the Conservative Party responded to the rise of UKIP by promising to call a referendum on EU membership if they won the 2015 election (which they then did). The Scottish independence referendum was a salience shock insofar as it increased the weight that voters placed on their views on Scottish independence in their vote choice in 2015 and, to a lesser extent, in 2017. However, as noted above, salience shocks also affect the clarity of positions taken by parties: by making the position of parties clearer on the issue of independence, where some—most notably Labour—might have preferred to have maintained ambiguity to avoid splitting their base over the issue. As with the Scottish independence referendum, the EU referendum represented a salience boost to the issue of EU membership and the result of a vote to (p.41) leave the EU guaranteed that this salience would remain high and hugely politically significant for years to come.

Party image

Shocks may fundamentally alter the image of a political party regarding who and what it represents, altering the effective choices available in the party system. We discussed above how voters may seek and gain greater clarity on party positions when shocks make issues and evaluations more salient. Shocks can also reshape the way a party is perceived with respect to what it stands for and which groups’ interests it represents. This can happen via a number of different routes, each of which causes a sudden and fundamental change to a party’s image and its purpose to voters. First, by revealing or emphasizing a new or previously obscured position on an issue, a shock may alter what a party is perceived to stands for with respect to issues and ideology. Second, and relatedly, any such change may lead to a change in who a party represents with respect to the social and demographic interest groups it favours. A party image shock can therefore involve a shift in either the social identity of a party or the political identity (or both). For example, Green et al. (2002) described how the enfranchisement of black Americans through the Voter Rights Act of 1965 led to their incorporation into the Democrat Party, bringing about a change in the social imagery of both the Republicans and the Democrats that ultimately led to realignment (Carmines and Stimson 1989). Third, a shock may shift the position of voters on issues independently of (or prior to) any change in the position of political parties, changing the issue distance between different groups of voters and the political parties. A major event, such as the Fukushima nuclear plan disaster, can lead to a change in public opinion on that issue which may in turn lead to a change in the position of political parties (Meyer and Schoen 2017).3 The extent to which parties respond strategically to such a shift will determine the extent to which such an event affects party support (Mader and Schoen 2018).

In the case of Britain, we can draw on two examples detailed later in this book. First, the choice of a Conservative-led coalition shifted the image of the Liberal Democrats as a vehicle for preventing Conservative rule, and linked the Liberal Democrats to a Conservative-defined agenda. The coalition therefore acted as a party image shock to the Liberal Democrats. It was specifically the coalition with the Conservatives that made the change so dramatic: a coalition with (p.42) Labour would not have been as sharp a break with the status quo because it would not have changed the way that the Liberal Democrats were widely perceived. The coalition therefore changed many voters’ perceptions of what the party represented. Second, the Scottish independence referendum changed the image of the Labour Party in Scotland. As noted above, while Labour had previously been widely regarded as a primarily working-class, left-wing party, the independence campaign highlighted their identity as a party of the Union, alongside the Conservatives. Many voters, especially those who favoured independence, began to redefine Labour in these terms, leading—as we show in Chapter 8—to mass switching to the SNP.

Competence

Electoral shocks can affect parties’ reputations for sound judgement and management. Green and Jennings (2017) outline three consequences of ‘competence shocks’ for public opinion and electoral choice: competence shocks contribute to the loss of long-standing party reputations on issues, or their ‘issue ownership’ (Petrocik 1996), parties can suffer a deterioration in their perception of competence overall, and competence can become more electorally relevant. Competence shocks cut through otherwise stable party reputations on issues, causing the public to evaluate the ability of parties to handle and deliver on different issues, including their traditional issues. This is consistent with the ways in which the immigration shock worked (as described in Chapter 5) and the Global Financial Crisis (Chapter 6). The rise in immigration acted as a competence shock against the Labour government—and then the Conservative government—when both were unable to fulfil their promise to reduce net migration to the ‘tens of thousands’ once they took office. The outcome of the EU referendum had a supplemental effect of removing a lever for Conservative competence on immigration because freedom of movement within the EU meant that they previously had very limited policy tools to reduce EU immigration. The Global Financial Crisis took place under the Labour government and was widely seen as an indictment of Gordon Brown’s economic management and Labour’s economic policy more generally, particularly—and interestingly—after the 2010 General Election. The damage to Labour’s reputation persisted under Ed Miliband, placed Labour in a strategic bind over economic policy in 2015, and created competition over blame for the deficit that went on to damage Labour’s performance in the 2015 General Election.

Shocks provide a change in the relevance of different criteria by which voters might make their decisions. Simply, competence shocks make competence more salient to political choice, differentiating parties, making competence a more urgent criterion for voters, and by providing political opportunities to compete around (p.43) handling and competence. Concerns around immigration—and the handling of immigration—became more salient to electoral choice following the immigration shock, providing UKIP, and latterly the Conservatives, with votes following the Brexit referendum. The effects of Labour’s competence on the economy were also enhanced as the parties competed around this problem—and the necessity of different policy responses—throughout the period between 2010 and 2015.

3.5 Shocks as political opportunities

The impacts of electoral shocks depend not only on how they affect voters but also on the responses of political actors. In other words, they are a function of political supply and demand. This means that the outcomes of shocks are not determined solely by the shock, but also by the different ways in which political actors respond to them. Shocks can also have an effect by constraining the policy options open to a party. For instance, a terrorist attack might force a party to change foreign policy in a way it may have preferred to avoid, or a currency crisis might require a government to interfere with financial flows in a way that angers key supporters.

Shocks provide an imperative for parties to compete around newly salient issues. Not only are shocks impossible for voters to ignore, they are also impossible for parties to ignore. However, shocks do not simply create a burden that parties must bear: they also create new opportunities for strategic actors to alter their electoral fortunes. Consider the way in which the EU referendum affected the outcome of the 2017 General Election because the Conservatives—despite choosing the Remain-supporting Theresa May as their next leader—pursued a policy of hard Brexit, promising to leave the single market, and thereby avoiding the commitment to the principle of freedom of movement. This response, combined with the Conservatives’ increased perceived competence on immigration, changed the pattern of party choice in 2017, leading many 2015 UKIP voters to defect to the Conservatives.

Shocks affect party strategies and electoral behaviour in ways that are not anticipated, and they also create unpredictability and uncertainty among politicians about how to respond, leaving them struggling to understand why the rules of politics have shifted beneath them. In policymaking, a shock or a crisis creates a sense of urgency and unusually rapid responses, less reliance on experts and more on ideology, but also a high degree of uncertainty (see Fischer, 2015). Such events are often highly technical and complex, meaning the ramifications cannot always be understood, increasing the potential for mistakes (Grossman 2015). Shocks are, by definition, unusual, differentiating them from the normal events of politics for which there are precedents, providing political actors invaluable experience of how to best handle them. While shocks provide new strategic (p.44) challenges to politicians, the increased demand to do something means a range of actions are deemed to be politically necessary. By way of example, Labour and the Conservative pledges to cut net migration can be viewed as actions that were politically necessary, but which backfired given their inability to meet them. Ed Miliband’s policy difficulties around austerity were a response to a successful effort to blame Labour for the financial crisis, and especially the level of national debt, and the demand to respond in some way for Labour’s previous period of government. Consider again the EU referendum and the Brexit outcome. It might have appeared, on the face of it, that all the Conservatives needed to do in 2016 was immediately get behind the Brexit project and win over the majority of UKIP voters to increase the party’s majority under Theresa May. Yet under the surface, the churn in the electorate, Labour’s strong performance, combined with the salience of Europe following the EU referendum, meant that a fundamental change happened in 2017 that was not widely foreseen. The outcome was, in part, a product of the Brexit effect (a rise in votes for the two major parties) but it also demonstrated the unanticipated consequences that come with shocks (the loss of the Conservative Party’s majority).

If a new issue or dimension becomes salient as a result of a shock, and if parties provide new choices to voters when competing on that dimension, realignment of electoral choice on that issue or dimension could be the result. For this to come about, the impact of the shock must be sufficiently long-lasting and strong enough to overcome the inertial forces we described above. Electoral volatility means that any durable pattern of electoral choice may be unlikely. Without an electoral shock and without differentiated responses from political actors, elections may still produce patterns of support which deviate from normal alignments as a result of regular political competition (Evans and Norris 1999). However, it is also unlikely that an electoral realignment will occur in the absence of an electoral shock.

3.6 Shocks within a volatile system

Earlier in this chapter we highlighted the combination of two destabilizing electoral forces in British politics: the long-term and gradual increase in underlying volatility caused by party-dealignment and party system fragmentation, and the impact of electoral shocks which create additional election specific peaks in switching. Although the elections of 2015 and 2017 were almost certainly unusual in terms of the large number of electoral shocks that preceded them, it is not at all clear that shocks have become more frequent over time more generally. Even if shocks themselves are not more likely to occur now than in the past, there are a number of reasons to expect that the potential impact of shocks on election outcomes has grown over time as a result of the weakening of partisan attachments and party system fragmentation.

(p.45) First, shocks determine which parties gain and which parties lose from vote-switching. Therefore, insofar as shocks favour some parties over others, as the underlying level of volatility increases, the parties profiting from a shock stand to gain more at the expense of the losing parties. Even if the overall level of volatility was unaffected by a shock, the impact might be expected to be greater when the baseline level of volatility (the amount of voter switching we would expect on the basis of long-term trends in partisanship and fragmentation) is higher. In other words, because shocks influence the direction of vote-switching, higher levels of volatility can produce larger electoral swings. Second, as the underlying level of volatility increases, the additional churn created by electoral shocks becomes more likely to lead to marked shifts in electoral outcomes in terms of seats, since parties have a smaller base of voters on which they can rely. In contrast, when the baseline level of volatility is low, any additional vote-switching caused by shocks is less likely to bring about dramatic political change. Third, because unattached voters and smaller party voters have a greater propensity to switch, under most electoral conditions shocks might be expected to affect voters who do not identify with a party—or identify only weakly—more than those that have a strong attachment.4 In other words, party identification may provide some insulation from electoral shocks. This is consistent with the theories of partisan identification which argue that party identifiers are less likely to be swayed by new information or more likely to interpret that information in a way favourable to their own party (Lodge and Taber 2013). An example of the insulating effect of party identification is illustrated in Chapter 9. EU referendum vote had a stronger effect on 2017 vote choice among those identifying strongly with a major party compared to weak or non-identifiers. As suggested above, this insulating effect of party identification may not always hold because some shocks can cut through the stabilizing effect of party identification leading voters to update their relevant attitudes, evaluations, and even partisanship (Green, Palmquist, and Schickler 2002). For example, whilst the collapse of Liberal Democrat support was greater among non-identifiers and weak identifiers, the desertion rate of strong identifiers relative to weak/non-identifiers was similar to that seen in previous elections, but with both strong and weak identifiers defecting at a higher overall level (Chapter 7).

It is important to note that the argument that unattached voters and smaller party voters have a greater propensity to switch does not mean they switch randomly, or willy-nilly. It is only because of some other stimuli that cause them to change their party choice. In normal times this might be driven by changes in voters or parties, such as voter preferences, party offers, or party competence. When a political shock occurs, volatile voters are most likely to react. (p.46) Certainly, there is sufficient academic evidence to conclude that a dealigned electorate will be more likely to respond to political stimuli, and therefore more likely to exhibit responsiveness and switching in response to shocks. As noted earlier, weaker identifiers are less likely to rationalize information, and less likely to exhibit bias in the way they deal with evaluating politics and party policy positions. As a result, weaker attachments to political parties have been found to increase the impact of economic voting when electorates exhibit. Strong partisans are less affected by the ups and downs of economic performance, whereas weak party identifiers are much more so (Kayser and Wlezien 2011) and weaker partisans are more influenced by both issue positions and issue competence (Weßels et al. 2014). As a result, unattached voters have a broader choice set of political parties that they are willing to consider voting for. Moreover, people who have switched parties previously have already demonstrated a greater willingness to consider different and diverse political messages, policy positions, leaders, and priorities. This shortens the cognitive leap required to vote for a rival party.

We described above how increasing returns to electoral success (or positive feedback effects) mean that established parties enjoy institutionalized advantages. Hence, party systems are normally expected to persist, according to the freezing hypothesis of Lipset and Rokkan (Lipset and Rokkan 1967). As Pierson (2000, 258) explained,

Key historical junctures produced major political cleavages. These political divisions became organized into political parties. Once they have surmounted initial start-up costs and fuelled processes of adaptive expectations, these parties are reproduced through time, which generates ‘frozen’ party systems.

These institutionalized advantages, not least the majoritarian electoral system, still have a powerful effect on bringing the party system back towards equilibrium. This is apparent through the fact that major parties still retain substantially higher proportions of their voters between elections compared to minor parties. Nevertheless, partisan dealignment and fragmentation have weakened the ability of the system to remain in a stable equilibrium in the face of electoral shocks. Electoral volatility in tandem with electoral shocks counters the forces that stabilize the party system. Depending on the balance of forces, rapid change may occur as a result of shocks, because of the underlying vulnerability of the wider system.

3.7 The future of volatility and electoral shocks

It might be tempting to think that volatility could increase inexorably if the fragmenting and dealignment processes that lead to volatility are path dependent, and if shocks inevitably lead to further dealignment and fragmentation. However, (p.47) this is not what is in fact happening, as we will demonstrate in the remainder of this book. Certainly, an obvious outcome of fragmentation and dealignment is the weakening of the in-built advantages of major parties, and further partisan dealignment could be seen as an inevitable consequence of new cohorts entering the electorate without the partisan identification of their parents. Importantly, however, shocks can lead to disruptions to these processes, as well as the acceleration of existing processes. In other words, shocks have the capacity to reduce fragmentation and increase partisan attachments, as well as the reverse.

Consider the outcome of the 2017 General Election, which delivered the largest two-party vote since 1970 and an abrupt pause, stop, or reversal (we cannot yet know which) in the fragmentation that had taken place up to 2015. The 2017 General Election was still a very volatile election. The total level of switching was very high between 2015 and 2017, though not as high as between 2010 and 2015. However, the amount of switching between the Conservatives and Labour between 2015 and 2017 was the highest on record. The defragmentation of the party system between 2015 and 2017 was just as much the result of volatility as was the fragmentation of 2015.

There are good reasons to think that volatility might now drop somewhat in a future election, since mainstream parties are more likely to retain their voters—even if those voters are previous minor party voters—and if more people identify with parties. There was a small hint of such an increase in identification in 2017, but nothing of sufficient magnitude or strength to suggest a major stabilization of support. The period after 2017 was so fraught with political difficulties on all sides that it is also easy to see the trend in partisan dealignment continuing, and while fragmentation has reversed, the electorate still has far lower levels of partisanship than in earlier decades. This means that the impacts of future shocks could be particularly substantial. What we also know, of course, is that a future shock or multiple shocks could happen. Indeed, Brexit is very likely to continue to reshape the British party system, whatever its outcome. It is shaping up to be a very substantial electoral shock. We consider a number of different ways in which the future of British politics might become more stable or unstable in the concluding chapter of this book, considering the possibility of different political identities around Brexit and early evidence for volatility looking ahead.

For now, it is important to note that our explanation of volatility and shocks does not imply that the British party system is on a one-way journey to greater fragmentation. Volatility does not inevitably lead to fragmentation and the processes that have led to volatility are not necessarily irreversible. However, a destabilized party system and more volatile electorate brings with it considerable uncertainty about the future of British elections and politics. It will be particularly responsive—and therefore less predictable—in its response to future shocks. Those shocks seem an inevitable part of the future of British politics.

(p.48) 3.8 Conclusions

Election outcomes cannot be understood without an appreciation of the long-term context in which they take place, and without understanding the major events that precede them. Voters do not decide how to vote in a vacuum of major events and changes that alter the political and electoral calculus. The volatility of the 2015 and 2017 elections did not come completely out of the blue: these tumultuous elections followed a steady increase in vote-switching in British elections that has, until now, been unexplained. In this chapter we explained this volatility by the long-term processes of partisan dealignment and the increasing votes for ‘other’ parties up to 2015 (fragmentation), both of which we go on to describe in detail in Chapter 4.

Once we account for the long-term trend in volatility and the destabilization of the British party system that results from it, it was not preordained that the vote flows would look anything like the way they did in the elections in 2015 and 2017, or that they would be the size they were. We need an explanation that can account for sharp and sudden changes in electoral volatility and outcomes. In this chapter we set out our explanation combining long-term volatility in the British electorate with the importance and mechanisms of electoral shocks; the major, salient political changes to the status quo that may cause substantial shifts in electoral choice.

Shocks work via mechanisms of salience, changes to party images, and perceptions of competence. They provide political opportunities and create an imperative for parties and political actors to respond, meaning that the effects of shocks may be absorbed, or may also create major and long-lasting changes in the electorate, and to elections. They may even lead to ‘critical elections’, where the basis of electoral choice alters such that realignment takes place between parties and voters. However, our theory of electoral shocks departs from critical election theory (Key 1955; Burnham 1970). Most importantly we do not dichotomize elections into those which are critical and those which are not, nor do we assume that shocks only occur as a precursor to such critical elections. Electoral shocks have the potential to sharply change the patterns of support for political parties at any election, the magnitude of the effect being determined by the nature of the shock, the response of political parties, and the underlying volatility in the electorate.

As such, an explanation based on electoral shocks incorporates explanations based on the competence of political leaders, the functioning of the economy, party and voter positions on issues, the salience of particular issues, and the representation of social groups. A theory of electoral shocks does not make other theories of electoral behaviour redundant: it provides the context to understand why and when they matter. Our explanation of electoral behaviour is, then, very different to one that applies one theory or model to all elections. Not all elections are the same and not all elections can be explained by the same sets of factors.

(p.49) Electoral shocks have the potential to cause dramatic electoral outcomes under some conditions more than others. We argue that the increasing underlying volatility in the party system in Britain has enabled shocks to become particularly important in the elections that we are studying. The ballast that once maintained a more stable party system has weakened, making the system more vulnerable to the impacts of electoral shocks. Volatility and shocks could lead to the fragmentation of the party system, and they could lead to the defragmentation of the party system. They help to explain the seemingly unpredictable and tumultuous nature of recent British politics and elections.

Notes:

(1) Corbyn was elected leader in 2015 by a large majority of the membership vote. However, he had struggled to muster up the thirty-five nominations he required from the Parliamentary Party following Ed Miliband’s resignation. Contemporary reports suggest he managed to secure sufficient nominations with the support of some who wished to ensure a contest that represented the full spectrum of voices in the party but did not expect him to win.

(2) Some studies have claimed spillover competence effects of seemingly unrelated issues such as shark attacks and sporting outcomes (Achen and Bartels 2016; Healy, Malhotra, and Mo 2010). However, these results have not been replicated in subsequent studies (Fowler and Hall 2018; Fowler and Montagnes 2015) and are not likely to be major drivers of electoral outcomes.

(3) Attitudes of the electorate tend to move quite slowly, whereas parties can change positions quite abruptly, especially in face of a shock. The analyses of shocks presented in this book suggest that shocks are more likely to affect how attitudes and values are aligned with party choice rather than by instigating a change in attitudes.

(4) The direction of this effect is complex and depends on the baseline level of volatility, the size of the shock, and the strength of the effect of party identification on vote choice.