3

Electoral behaviour and inequality

Having outlined the broad parameters of the 2014 election and examined the recent increase of social and economic inequality in New Zealand, our next step is to establish why and how inequalities might be connected to vote choice. To this end, this chapter introduces the main theoretical models that inform our empirical analyses, and how they may assist in addressing our research question.

The models we identify are intrinsic to the understanding of electoral politics and resonate with both political discourse and practice in New Zealand and elsewhere. For example, in September 2016, New Zealand Labour leader Andrew Little rejected the advice of former Labour prime minister Helen Clark (1999–2008) that parties on the left must ‘command the centre ground’ in politics, describing it as a ‘pretty hollow view’. Instead, he proposed that Labour should build ‘a coalition of constituencies’ (Sachdeva 2016). Clark’s understanding of politics is of an ideological contest between left and right. National Party prime minister John Key has made similar statements about ‘hugging the centre’. Andrew Little apparently sees politics as about constructing a coalition of social groups that can be persuaded to favour his party.

Clark’s position is based on the assumption that people’s policy preferences across an ideological dimension from left to right are shaped by their relative economic positions in society. To win, a party or coalition of parties must capture the support of the people in the centre: the median voters. The theory presents a rational calculus of voting: each voter compares the expected utility (or value to them) of having party A in government with the expected utility of having party B in government. The utility differential determines the voters’ party choice. Those in a lower socio-economic position who would gain most from redistribution of incomes are expected to support left-wing parties that pledge to tax the rich and transfer income to those who are poor. Those in higher socio-economic positions will vote for parties that say they will allow those with wealth and higher incomes to keep them. Those in the centre weigh up which party, left or right, will be closer to their interests. Parties of the left and right will therefore tend to moderate their ideological appeals to try to capture the votes of those in the centre. This model puts inequality between individuals and families in a capitalist democracy at the heart of its assumptions.

In contrast to this economic model, the social psychological model focuses on psychological affinities and identities. Defenders of this model argue that politics is fundamentally about identities that construct group affinities, not about ideologies (Achen and Bartels 2016); very close to Little’s idea of a ‘coalition of constituencies’. A group affinity—party identification—is said to be the main factor behind the behaviour of voters. The central role of partisanship is presented through the so-called funnel of causality, which presents a chain of events that contributes to the voters’ eventual party choice. It distinguishes factors that generate the basic conflicts of interest within society, such as economic structure, historical patterns and social divisions, from proximate factors that more directly influence people’s party choices: issues, candidates, election campaigns, the influence of friends and the media (Dalton 2014). As we discuss in more detail below, these models provide us with different ways of understanding the relationship between social and economic inequality and voting behaviour.

The economic model

Anthony Downs’ book An Economic Theory of Democracy (1957) applied the theoretical tools of economics to electoral behaviour. Voters were assumed to be instrumentally rational and self-interested. People vote in order to maximise their ‘utility’, defined as whatever they might want. In practice, though, Downs assumed that most voters wanted to increase their incomes, and would therefore vote for parties and politicians that promised to do so.

Downs started from the idea that voter preferences and party positions fall across the left–right dimension. Those on the left prefer government to redistribute income and use various other means to ensure the social security and wellbeing of all members of a society—they take a position favouring equality. By contrast, those on the right prefer a limited role for government. They want to protect private property and allow a free market to generate wealth and deliver income accordingly to those who work for it. They thus take a position likely to generate inequality, accept a degree of inequality and resist more than minimum efforts to reduce it.

To win an election in a two-party system, a party must attract the vote of the median voter; that is, the person in the middle if all voters stood in a line from left to right. The job of the two major political parties is to convince those people that hug the centre to shift one way or the other. Their best strategy for doing so is to present party policies that are attractive to the median voter. In a multi-party system, the logic is a little different—particularly for smaller parties closer to the left or right extremes. But if there are two larger parties competing to lead a government, it is expected that the median voter remains their main target. A party staying close to the median voter should have a good chance of being elected and re-elected, and large office-seeking parties will therefore tend to stay close to the centre (Adams 2012; Meyer 2013).

Several commentators have asserted that the National Party under the leadership of John Key has been extremely successful at winning median voters. National is thought to have been pragmatic by some, or an opportunistic office-seeker by others (Hooton 2016; Hosking 2014a; Nagel 2012). But National resisted getting on the inequality bandwagon during the 2014 election, despite the salience of the issue during the campaign. Only one new initiative was promised of potential benefit to those on low incomes: an extension of free medical care to children up to 13 (from the age of six) and a small boost to early childhood education funding (National Party 2014a). Nevertheless, despite this lack of largesse, National has been identified (Nagel 2012) as mathematically closer than Labour to the median voter since 2008. As shown in Chapter 1 (Figure 1.1), this is confirmed by public perceptions of the two parties’ positions. Only a few months after the 2014 election, National’s centrism or alleged opportunism re-emerged, and the government announced that to reduce ‘child material hardship’ it would increase welfare benefits (New Zealand Government 2015). One can take this as another sign of Key’s ‘middle-of-the-road’ strategy, potentially making the government less vulnerable to concern about inequality and poverty in the future. But this still does not entirely explain the failure of inequality to shift votes to the left at the 2014 election.

Early extensions of the Downs theory also fail to provide an explanation for this puzzle. A self-interested median voter should support redistribution of income in conditions of economic inequality where the average income tends to be higher than the median. Political parties aiming for median voter support, including those on the right, should adopt policies that reinforce this position (Meltzer and Richard 1981). This may have worked in the 1960s and 1970s. From the 1980s and 1990s onwards, growth in the size of government and the extent of social expenditure has flattened, with significant decreases in some countries. Political parties on the right have shifted further to the right on both economic and social policy, and parties on the left have tended to follow them. Inequality has increased, but median voters are not responding as the Meltzer-Richard (1981) extension of Downsian theory predicts.

One reason given for this apparent contradiction lies with increasing wealth through the ownership of assets, noted in Chapter 2. When house prices rise, this may increase the apparent wealth of those on median incomes and encourage them to borrow and spend more, and vote for parties of the right that support unregulated, free market wealth accumulation (Ansell 2014). It may be that many people think less in terms of their immediate interests and more in anticipation of what their incomes are likely to be over their lifetime. The prospect of upward mobility may dispose those on low incomes to vote for parties on the right because they anticipate higher incomes in future, and they may therefore behave like traditional right wing party voters (Alesina and La Ferrara 2005; Benabou and Ok 2001). Another way of defining those who see themselves as likely to be upwardly mobile is as ‘aspirational’ (Smith, Vromen and Cook 2006: 10–12; Simms and Warhurst 2005). Others may be more pessimistic about their economic futures and thus attracted to parties on the left. They may feel insecure economically and thus also be attracted to the left because the welfare states provide insurance against bad luck or unexpected adverse events such as job loss or illness (Iversen and Soskice 2001; Lupu and Pontusson 2011; Moene and Wallerstein 2001; Mughan 2007; Rehm 2016).

Another modification of Downs suggests that rather than voting for parties that are closest or most proximate to their position, people will vote according to the ‘direction’ of their position (for reviews and the various extensions of this literature see Adams 2001; Merrill and Grofman 1999). Consider a left–right dimension from 0 to 10, and a person who scores 5.1, just marginally to the right. If the left and right parties have positioned their policies at 4.5 and 6 respectively, the most proximate party to the person in question is the one on the left (a difference of 0.6). However marginally, that person is on the right, and might therefore vote for the right wing party directionally regardless of its more distant proximity (0.9) when compared with that of the left party. This means that large parties that are more extreme than the median voter can continue to attract moderate people on their side of the left–right dimension, even if those moderates might have policy preferences closer to the party on the other side.

There is another reason why moderate or median voters might vote for parties that are more right wing than themselves. Median voters may prefer a moderate redistribution that would benefit them but would resist high levels of redistribution that would require them to pay higher taxes. In two-party systems, the risk of a redistribution that might harm those on median incomes will dispose people in that situation to vote for the right. In a multi-party system, a significant centre party might be expected to defend the interests of middle-income voters and resist high levels of redistribution, thus capturing their votes. Thus coalitions of centre-left and centre parties may establish and defend systems of moderate redistribution (Iversen 2005; Iversen and Soskice 2006).

Downs’ model assumes that the acquisition of information is one of the costs of voting. Most voters lack political knowledge, an assumption confirmed by empirical research (for example, Delli Carpini and Keeter 1996). When deciding to vote, people acquire information from trusted sources, such as people they know, as well as from the communications media or the political parties or candidates themselves (Lupia 1994; Lupia and McCubbins 1998; Page and Shapiro 1992; Popkin 1991). Once people select a political party they believe is close to their views, often on the basis of limited information, they may use it as a source of further information, reinforcing their likelihood of giving that party their vote. People may have party identifications, but they can be updated as a result of various events and new cues, and can change over the long term (Fiorina 1981).

Those with the least economic resources and with most to gain from a reduction in inequality are also known to be the least knowledgeable and informed about politics (Grönlund and Milner 2006). Consequently, they may be more likely than those who are well-informed to vote for a party that is not the closest to their own position. It may also be that these people choose not to vote at all (Brady, Verba and Schlozman 1995; Smets and van Ham 2013). In most countries, including New Zealand, turnout has declined over time, and the trend continues with only temporary upswings (Vowles 2010, 2014a, 2015a, forthcoming). Low turnout tends to be more prevalent amongst those on low incomes, without wealth or assets (Solt 2008), with lower levels of education and the young (Franklin 2004). The Downs model provides a clear explanation for these groups not voting: the costs of making an informed vote are higher, as they tend to have lower levels of political knowledge than those who are older, more educated and middle class. With increasing levels of inequality, the young, less educated and the poor tend to feel that government policy is tilted against them, and they have little confidence that their vote will make a difference or result in any sort of benefit (Griffin and Newman 2013; Mahler 2008). We can label this a ‘disempowerment thesis’ (see Chapter 11). Research in other countries shows that economic adversity may depress turnout when social protection is less generous. Where beneficiary incomes are increasingly means-tested and access to these becomes more stigmatised, those most affected may withdraw from politics, rather than holding government accountable (Pacek and Radcliff 1995; Radcliff 1992; Rosenstone 1982).

Some suggest that if more people of lower socio-economic status voted, this would affect election outcomes. In particular, left-leaning parties would benefit from higher turnout (for example, Fowler 2013; Mackerras and McAllister 1999). Others argue that those on low incomes would benefit from higher turnout because both left and right leaning governments would pay more attention to addressing their needs. When low-income voters fail to turn out, it becomes easier for governments to tolerate inequality (Boix 2003; Husted and Kenny 1997; Mueller and Stratmann 2003).

According to Downs, people vote according to the benefits they can expect from tax and social policies. Beginning with Donald Stokes (1963), critics have argued that people are more likely to vote according to a set of shared interests that they might have in common: a strong and growing economy, for example. Alternatively, people might vote according to their perceptions of the competence of the government and its leaders in delivering public services and handling matters of the day. The focus thus becomes not the issue itself but which party is most likely to deliver growth, good public services and stable government. This vein of theory has become known as the ‘valence’ or performance model of electoral choice (Clarke et al. 2011). Within a broad economic model of voting behaviour, this ‘valence’ or performance theory contrasts with Downs’ ‘positional’ theory that is founded on his assumption that issue and ideological positions are based on different sets of economic interests that mostly fall across the left–right dimension.

Many others have taken up the idea that people might vote according to the economic performance of a country, returning governments to power under conditions of economic growth and dismissing them during a recession. While some have empirically confirmed models of retrospective economic voting (Fiorina 1981; Lewis-Beck and Stegmaier 2013), more generally, the findings have been mixed and contested. Clarity of government responsibility for the economy varies across countries because of cross-national differences in political institutions, and whether governments are made up of single parties or coalitions (Anderson 2007). It has been argued that voters use immediate past records of economic performance as an indicator of probable future performance. Expectations are therefore the driving force rather than wishes to reward or punish (Duch and Stevenson 2008). When voters are asked to rank election issues, they frequently identify ‘management of the economy’ as the most important issue driving their vote choice (Bean and McAllister 2012).

Despite the richness of research within the Downsian tradition, it has been subject to continual criticism (led most famously by Green and Shapiro 1994; see also Hug 2014; and, most recently, Achen and Bartels 2016). The criticism mainly targets the ‘thin’ assumptions on which Downs’ theory is based: that people can be assumed to be rational, at least up to a point; that political competition can be usefully simplified in a simple one-dimensional model; and that people with low knowledge can be successfully ‘cued’ to behave politically in accord with their interests (Kaye 2015). In sum, the model is theoretically elegant, helps us to think about aspects of electoral politics clearly and has identified many useful lines of inquiry. It is a good place to start, but not a place where we can comfortably conclude our inquiries.

The social psychological model

A second theoretical model through which we can examine voting behaviour and the relevance of inequality to such behaviour is the social psychological model, recently given new impetus by its recasting into a revived ‘realist’ theory of democracy (Achen and Bartels 2016; also Lenz 2012). The antecedents of the social psychological model are a little older than those of its economic rival. The earliest election studies in the United States in the 1940s began with assumptions that voters were akin to consumers and parties sold themselves as products. This was an approach that Downs took up and developed more successfully later. But researchers found that voters were too strongly loyal to their parties to behave as consumers in a market should or would (Lazarsfeld, Berelson and Gaudet 1944). This led to a shift in emphasis and a turn to psychology and group theories of politics (Berelson, Lazarsfeld and McPhee 1954). This underpinned the development of the theory of party identification, most clearly articulated in The American Voter (Campbell et al. 1960). Partisanship came to be seen as a central explanation for voting behaviour, conceptualised as a psychological affinity with a political party.

The social psychological model begins with the sociological idea that political parties both respond to and shape social cleavages and identities that might reflect class, language, ethnicity or religion, depending on the society in question (Lipset and Rokkan 1967). As such differences often reflect social and economic inequalities, this vein of theory also has implications for our analysis. Around those social cleavages and identities, people develop loyalties to political parties that are constructed and reinforced by social networks such as churches, trade unions, families and the political parties themselves. People’s apparent ideological and issue preferences are shaped for them by the discourses of the groups with which they identify. Individual preferences often shift to be consistent with group norms or changing party policies.

Such loyalties are not absolute and might be overcome temporarily by an effective or charismatic leader, attractive promises or a strong performance by another party. Votes may also change because of a shift away from a tired and exhausted incumbent government that has been in office for a long time. Yet, according to this model, the majority of people are expected to eventually default back to older loyalties; they have a ‘homing’ tendency. The social psychological model combines both sociological and psychological influences on voting into the so-called funnel of causality (Campbell et al. 1960). It distinguishes distant influences from more concrete, immediate and direct influences on voting behaviour (Dalton 2014). Partisanship and party identification are influenced by more distant factors, such as economic structure, social divisions, group identities and value orientations. Partisanship then influences issue opinions, candidate images and, both directly and indirectly, the actual voting act. This pattern of influence is itself influenced by various other factors, including campaign activity, media, friends and family, and economic and political conditions.

The central core of the social psychological model is its idea of party identification as rooted in loyalty and identity, and as possibly inherited through the family. It is emotional or ‘affective’ rather than rational or cognitive, as is the case in the Downsian economic model. However, the idea of party identifications as being stable over time has turned out to be the weakest claim of the psychological approach (Dalton and Wattenberg 2000; Dalton 2013). Voting choices from one election to the next have also become more volatile over time in many countries, including New Zealand, although perhaps not as consistently as some of the literature suggests (Dassonneville and Hooghe 2011; Mainwaring and Zoco 2007; Mair 2005; Vowles 2014b).

The continued strength of the psychological approach lies mainly in its claim that rationality does not provide sufficient explanation for political behaviour. Research into political psychology bears this out. For example, people are more affected by the prospect of loss than of gain and tend to be risk-averse (Kahneman and Tversky 1979). As a consequence, people may need to be more convinced to vote against a government than to vote in favour of it, providing a micro-level explanation for the old adage that ‘governments lose elections, oppositions do not win them’. Perceived government failure or disruptive policy changes may generate fears of loss. If voters are considering voting against a government, a perception of a viable alternative that reduces fears of loss will increase their probability of doing so.

Daniel Kahneman (2011) tells us that human beings have two modes of response to stimuli: a ‘fast’ mode by which we respond rapidly by instinct or emotion; and a slower mode, where we think about consequences. Much political behaviour takes place in the fast mode, in which voters are likely to be influenced by a recent experience or interaction that primes their response. At another time, or in a different context with different cues or influences, they might behave, vote or express their opinions differently. This reinforces the early findings of Philip Converse (1964), who analysed panel surveys that included people who had been interviewed three times over two to three years, with many of the questions repeated. He found very little stability in opinions, except among the most politically knowledgeable. Of even greater interest, he found evidence of ‘non-attitudes’, with people expressing opinions when asked trick questions about nonexistent legislation. Developing this line of inquiry, John Zaller (1992) has shown how mass opinion often follows cues from political elites (see also Lenz 2012; McAllister and Bean 2006). This phenomenon is affected by political knowledge. Those with lower levels of political knowledge receive fewer cues than those with higher levels of knowledge, but the former are more likely to follow them.

As discussed earlier, exponents of the economic model also accept that most people have low knowledge about politics, but argue that the cues they receive shape their preferences in ways consistent with what their preferences would be if they were more knowledgeable. In other words, voters’ preferences have an indirect rational foundation. But psychological research also finds that people are selective in the cues to which they respond, and people with higher levels of knowledge are not immune from bias. We can interpret many responses to questions about public opinion on the basis of ‘they would say that, wouldn’t they’. Even the most knowledgeable voters are prone to make choices that they may not normally make had they considered them further. Instead, they stay with pre-conceived ideas and assumptions and are prone to reject evidence that refutes them (Kuklinski et al. 2000; Kuklinski 2007; Lodge and Taber 2013; Nyhan and Reifler 2010).

People may therefore resist updating their party identification despite new information that challenges it. A strong emotional identification with a political leader may blind people to negative aspects of a president’s or prime minister’s leadership. Perceptions of an effective and likeable leader are more important than the economy and can play into a more general feeling that a government is competent. Perceptions of competent leadership almost certainly affect how people assess the performance of the economy and the government’s responsibility for that performance. Even perceptions of the state of the economy can be affected by party choice, depending on circumstances (Chzhen, Evans and Pickup 2014; Evans and Andersen 2006; Evans and Pickup 2010; Pickup and Evans 2013). Where the economy is strong, people who vote for opposition parties may declare that the economy is weak, as we shall see in Chapter 4. There are very few if any simple one-way causal pathways between perceptions and behaviour.

This all begs the question of where voters get their information about the performance of their country’s economy, as at least some of the sources of that information may be biased, and voters are often selective in their media consumption. In addition, governments in countries such as New Zealand can only have marginal effects on the performance of their economies, because of small size, global exposure to trade and investment and greater vulnerability to ‘shocks’. Using cues generated by the communications media, rather than responding to absolute changes in the economy, people may compare their country’s economic prospects to those of other countries that may be close trading partners or neighbours. For example, economic growth in New Zealand is frequently compared to or ‘benchmarked’ with economic growth in Australia, much as the strength of the dollar, tax and wage rates, and house prices relative to income are often compared. The role that the media plays in communicating these perceptions has been described as ‘pre-benchmarking’ (Kayser and Peress 2012). Achen and Bartels (2016) discuss the example of shark attacks on New Jersey beaches that adversely affected the local economy and later influenced voting in a presidential election. Droughts in American farming states can lead to the defeat of governors at the next state election. However, sometimes voters can evaluate the effectiveness of a government response to a shock, and in this case incumbent politicians can be reasonably held to account (Gasper and Reeves 2011). The impact of exogenous shocks can also work in reverse. In New Zealand, it has been suggested that the National Government was able to win key seats in Christchurch after the February 2011 earthquake in part because the government-funded recovery program helped both the local community and the national economy, demonstrating the government’s apparent skills in economic and crisis management (Young 2011).

The timeframe within which judgements about economic performance are made is also considered important. British evidence indicates that people do not respond so much to moderate or small short-term shifts in the economy but acquire perceptions over longer periods or when an economic shock is particularly strong (Chzhen, Evans and Pickup 2014). But if voters’ perceptions are based on short-term evaluations of strong shocks, they may be influenced by a temporary upswing or downturn that may not reflect a longer-term trend. In the United States, the evidence is different: while Democratic presidencies tend to be associated with better outcomes for people on lower and middle incomes, short-term economic upswings have coincided with crucial elections faced by Republican incumbents who have usually been re-elected (Bartels 2008).

The attribution of responsibility for a large ‘shock’ may be significantly misplaced. For instance, in Britain, the Labour Party was in government when the global financial crisis (GFC) of 2007 occurred. The Conservative Party, while in opposition between 1997 and 2010, had been if anything more strongly committed than the Labour Party to the ‘light-handed regulation’ of the financial sector that made the crisis possible. After the crisis hit, the Labour Government under Gordon Brown responded with an economic stimulus and avoided the deep economic downturn that many had feared. Brown played a large role in successfully arguing for a coordinated international stimulus. After Labour left office, the Conservative and Liberal Democrat Coalition’s austerity policies almost led Britain back into recession, its recovery since has been sluggish, its government finances remain in deficit and its government debt high. Nonetheless, Labour has been widely blamed for the effects on the economy of the GFC (Cowley and Kavanagh 2015; O’Hara 2015).

By contrast, Australia’s economy began to boom after 2003, not because of good economic management but because of China’s insatiable demand for natural resources. Consequently, the Howard, Rudd and Gillard governments could ride out the worst of the GFC (Kelly 2014). Nonetheless, voters saw Labor as poor economic managers (Koukoulas 2016), and when Abbott’s Liberal–National Coalition government came to power in 2013, it introduced policies of austerity. The latter did not lead to recession, but nor did they ameliorate Australia’s economic downturn. Using this sort of evidence, Achen and Bartels (2016) have been able to make a strong case that much voter evaluation of governments is ‘myopic’ if not ‘blind’. If these pessimistic findings hold true generally, voters who might benefit from or simply wish to see more redistributive government policies are unlikely to recognise and reward the politicians or the parties promoting them.

Positional politics in two dimensions

In Chapter 1, we introduced not just the traditional economic left–right ideological dimension, but also a second social dimension that partly cuts across it. Political conflicts around that second dimension may help to explain why preferences about equality or inequality, which might be expected to be shaped by the left–right dimension, may move in other directions. This dimension can be described as representing the ‘new politics’, as opposed to the ‘old politics’ of the left and right. The idea of the new politics emerged in political science and sociology in the late 1960s, when new divisions concerning cultural issues began to emerge in advanced western democracies. Based on the psychology of Abraham Maslow, Ronald Inglehart (1977, 1990) theorised the dimension as one of materialism versus post-materialism. Post-materialism is said to be found among people who have been born into recent generations in the developed democracies and brought up under conditions of material and physical security. Taking these conditions for granted, their political aspirations are said to become focused on enhancement of human freedom and creativity and the protection of the natural environment for aesthetic reasons. A significant literature has emerged, some of it developing the concept, but there has also been much criticism (for example, Davis and Davenport 1999; Duch and Taylor 1993). Subsequent developments also weaken the theory; material and physical security are not as fully assured in the developed democracies as may have been assumed in the 1970s when the theory emerged. Moreover, protection of the environment is not merely an aesthetic concern; for many, climate change is a serious problem that threatens physical and economic security in the present and the not-too-distant future.

Nonetheless, post-materialism has contributed to an analysis of politics in advanced democracies that moves beyond the left–right dimension of the ‘old politics’ (Dalton 1996; Poguntke 1987). In New Zealand and elsewhere, the political space is now generally modelled in two dimensions: the economic left–right dimension and the ‘new politics’ dimension, which can be taken to represent materialist versus post-materialist values; or, more simply, socially conservative against more liberal attitudes.

Indeed, a stronger theoretical basis for the second dimension may lie in orthodox psychology. The idea of a second dimension beyond left and right was anticipated in the 1950s by psychologist Hans Eysenck, and in his terms it was based on personality. It distinguished between the ‘tough-minded’ and the ‘tender-minded’ (Eysenck 1954). An even more influential strand of theory developed even earlier and came out of efforts to understand the appeal of Fascist and Nazi movements: the idea of the ‘authoritarian personality’ (Adorno et al. 1950). The concept of authoritarianism fell out of favour in subsequent decades, but in revised form has recently become popular again, particularly among political psychologists (Altemeyer 1988; Stenner 2005; Whitley 1999).

Indeed, materialist and post-materialist attitudes tend to correlate strongly with attitudes along a continuum from libertarian to authoritarian (Kitschelt 1994). This cleavage centres on questions of morality and freedom, and cross-cuts the traditional ‘old politics’ economic cleavage. It divides society according to levels of education, rather than economic class (Bornschier 2010; Flanagan and Lee 2003; Inglehart 1984; Houtman 2003; Van der Waal, Achterberg and Houtman 2007). Authoritarian attitudes can be enhanced by fear, by perceptions of insecurity and by experiences of rapid change that generate uncertainty. Authoritarians tend to be conformist and live structured and disciplined lifestyles. By contrast, libertarians tend to be critical and question orthodoxies, and to have lifestyles that are less structured and less anchored in ongoing commitments.

The rise of authoritarian populism in parts of Europe and more recently in the United States has placed the concept even more on the map (Lubbers, Gijsberts and Scheepers 2002; Mudde, 2007; Norris 2005; Rydgren 2007; Van den Berg and Coffé 2012; Van der Brug, Fennema and Tillie 2000). Some theorise this development is also associated with globalisation, a phenomenon that tends to threaten both material security and provoke cultural concerns associated with immigration (Kriesi et al. 2008). The two-dimensional structure of public opinion in New Zealand is not a new phenomenon and was comprehensively mapped after the 1993 and 2002 elections (Vowles et al. 1995: 100–21; Vowles 2004b). In this book, we operationalise the authoritarian/libertarian dimension, particularly in our discussions of the Green Party, New Zealand First and the Conservative Party.

The existence of these two cross-cutting dimensions to positional politics has significant political implications to be explored in later chapters. New politics issues include the rights of minorities or relatively powerless groups of various kinds: religious, ethnic and gendered, as already discussed in Chapter 2. They also include immigration, an increasingly sensitive topic in many countries.

New politics draws heavily on identities that can cut across the traditional debates between left and right that have obvious implications for social and economic inequality. So-called ‘wedge politics’ (Hillygus and Shields 2008) can mobilise cultural issues around race (Iyengar and McGrady 2007) and immigration (Newton 2008), detaching voters from their traditional loyalties and sidelining debates about the country’s high levels of inequality. Frustrations borne out of the stagnation of low and middle incomes in the United States have been mobilised on a racial and cultural basis, as Donald Trump’s election to US presidency bears out only too well. In Australia, refugees arriving by boat made control of the country’s borders a highly salient political issue, driving a ‘wedge’ between the Labor Party and many of its traditional supporters (Jupp 2012, 2015; Ward 2002). In Britain, similar concerns about border control featured strongly in the successful ‘Brexit’ campaign; immigration policy divided the Opposition Labour Party internally, alienating it from many of its traditional working-class voters (Dennison and Goodwin 2015: 171).

In New Zealand, from the other side of that dimension, ‘new politics’ support briefly helped the Lange-Douglas Labour Government of the 1980s retain votes in 1987, particularly the popularity of the government’s nuclear-free policies (Vowles 1990: McAllister and Vowles 1994). But in more recent years, critics have argued that Labour’s apparent focus on ‘identity politics’ has placed another wedge between the party and its traditional voters (Maddison 2006), a wedge that has been exploited by Labour’s opponents and may help explain its lack of electoral success in recent elections. Chapter 6 takes up these claims. New Zealand’s geographic location relieves it of most of the problems of border control faced elsewhere. Immigrants tend to be accepted, but a high rate of influx in recent years has made immigration a potential wedge issue, potentially drawing some voters away from their previous loyalties. Chapter 8 takes up this issue in more depth.

Finally, it is worth observing that the underlying assumptions of the ‘new politics’ model are derived from psychology, but differ significantly from those of the social psychological or group theory. Just as the economic model assumes that people have ‘interests’ that are largely independent of the ways in which elites construct the political world, new politics theorists assume that people have underlying values that are derived from psychological dispositions relatively independent of elite influence. By contrast, the social psychological model is based on the assumption that interests and values are mostly constructed by a combination of group interactions and elite cues. In practice, most researchers in electoral politics accept that it is reasonable to entertain the possibility that some interests and values may be relatively independent of group and elite influences, while others may not—these are ultimately empirical questions.

Conclusion

The economic voting model starts from the idea that people are guided by self-interest and therefore vote for the party that promises to promote their interests. The main difficulty they face is the cost of information needed to make that calculation. In contrast to the cognitive approach of the economic voting model, the social psychological model is based on social group identities and, in particular, partisanship and party identification based on psychological affinities. Social group affinities can give cues to provide people with information that helps them make decisions, but there are no guarantees that these decisions will serve their interests. Partisanship and other group identities are also major sources of ‘cognitive bias’, the tendency of people to resist new information that challenges their ideas and assumptions. Even if people care about inequality, affecting themselves or others, they may not necessarily respond politically as one might expect. The ‘new politics’ or ‘values’ approach identifies a second dimension to positional politics that may also cut across debates about inequality, distracting voters and diverting their economic concerns into cultural channels.

This chapter has compared and contrasted the three approaches to thinking about voting choice, and we consider all three to be of value. Despite decades of criticism, the ideas of left and right continue to resonate in political discourse. They still stand for state-led redistribution of resources on the left and preferences for free markets on the right. In the wake of the GFC and the great recession, with increasing international concern about inequality, we can hardly say they are irrelevant. Neither can we ignore the existence of the new politics dimension that may channel voting behaviour in new directions. The normative ideal of a democratic citizen who listens, learns and participates, and whose preferences are fundamentally ‘rational’, is not one that we should lightly dismiss. Knowledge of political psychology alerts us to the limitations of that vision, and the way democratic ideals can be perverted, often at the hands of less than scrupulous elites.