Political Change in England and the 2017 General Election

The Political Quarterly

Tilting Towards the Cosmopolitan Axis? Political Change in England and the 2017 General Election

WILL JENNINGS AND GERRY STOKER

Abstract The general election of June 2017 revealed a continued tilting of the political axis in England that has been long in the making. This was not a Brexit ‘realignment’—in that the vote is better seen as a symptom of a longer-term divide that is emerging between citizens residing in locations strongly connected to global growth and those who are not. In this analysis, we explore constituency-level patterns of voting in England between 2005 and 2017. Over this period, Labour’s vote share has tended to rise in urban areas (that is, major cities), with younger and more diverse and more educated populations often working in ‘cosmopolitan’ industries, whereas the Conservative vote has tended to increase in less densely populated towns and rural areas, with older and less diverse populations. Signiﬁcantly, Labour has also increased its vote in constituencies with a higher share of ‘precariat’ and emerging service workers—somewhat at odds with characterisation of a party that has lost the ‘left behind’. To the extent that changes in electoral support for the Conservatives and Labour are linked to the Brexit vote, the relationship far predates the referendum vote and should be expected to continue to reshape British politics in future. Keywords: Brexit, globalisation, social class, cosmopolitan politics, precariat, 2017 general election, realignment

A NEW cleavage is emerging that is funda- the EU further highlighted this growing mentally changing politics in the early dec- divide, with urban-metropolitan areas tend- ades of the twenty-ﬁrst century. The divide ing to vote Remain and regional-coastal is between citizens residing in locations areas and post-industrial towns voting strongly connected to global growth and Leave. those who are not. In geographical terms, it This article on the general election of June is between those from the densely populated 2017 extends this focus on the long-term cosmopolitan and metropolitan centres of trends driving change in British politics by the emerging knowledge and creative econ- making three arguments: omy and those who live beyond that world in suburban communities, post-industrial 1 That this was not a Brexit election—or a towns, and coastal areas. In our article pub- Brexit ‘realignment’—in that the vote is lished in The Political Quarterly in early 20161 better seen as a symptom the longer-term we highlighted differences in outlook among geographical bifurcation of politics; less people living in urban-metropolitan areas revenge of the ‘Remainers’ and more a compared to those residing in more periph- continuing battle of mobilisation between eral coastal areas, with the former generally cosmopolitan and non-cosmopolitan areas. more positive about the EU, equal opportu- 2 That within cosmopolitan and non-cosmo- nities for ethnic minorities and gays and les- politan areas there are complex ecologies bians, the impact of immigration and change of occupations and social groups that cre- from the past more generally. Later in 2016, ate opportunities for a variety of party the referendum on Britain’s membership of strategies. Emphasis on the authoritarian

© The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017 Published by John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA 1 reflex of the ever-shrinking traditional growth in these areas. Citizens in these areas white ‘working class’ (classified based on are not automatically members of a global manual and routine occupations) has elite or without national or community feel- distracted from the changing nature of ing. They just happen to be located in, or socio-economic status and the precarious have been drawn to, areas where knowledge situation of a younger generation of voters and creative economic forces are strong. who tend to be more educated and cos- ‘Non-cosmopolitan’ isourlabelhereforthe mopolitan in outlook, but at the same mixture of coastal, post-industrial and sub- time, increasingly squeezed by falling real urban locations that are not so strongly con- wages and by intergenerational inequali- nected to these dynamic economic forces. ties (such as through increased debt, reduced job security and fewer opportunities for home ownership). Electoral change over more than a 3 That the tilting of the political axis, as observed in the outcome of the 2017 gen- decade eral election, reflects long-term trends. On The cosmopolitan tilt of politics in larger this occasion, the Labour party made metropolitan centres and economically gains in cosmopolitan areas while sustain- dynamic university towns is reflected in the ing some support outside those areas. The growth of support for the Labour party in Conservatives made inroads in less urban those areas. Figure 1 depicts the vote share areas made up of older and less educated for the two main parties in our fifty urban- populations (who tend to be more socially metropolitan constituencies in 2005 and conservative), but lost further ground in 2017, with the Labour vote share plotted on big cities and university towns. The party, the y-axis and the Conservative vote share or parties, that will be most successful in on the x-axis. It is striking that in these sorts future will be those able to navigate this of areas the frontier of Labour support has bifurcation of British politics and the moved outwards over time (despite endur- demands of citizens. The effectiveness of ing a setback in the national swing against it parties’ electoral strategies will have con- in the 2010 election). Even as the party’s sequences for the policy choices taken by electoral fortunes have fluctuated nationally, future governments—which in turn may it has made advances in more cosmopolitan reinforce, or ameliorate these forces driv- areas. At the same time, the trajectory of that ing electoral change. frontier has pivoted so that its support is increasingly concentrated in constituencies where its vote share is highest, potentially In this analysis, we use constituency results accentuating the bifurcation of Britain’s elec- and demographics to explore aggregate-level toral geography (such that the party’s gains patterns in the vote in England.2 We start by are largest in existing electoral strongholds). focusing on the election result in two sets of Figure 2 plots the vote share for the two fifty constituencies identified using the parties in our fifty regional-coastal con- method described in our 2016 article, arche- stituencies. In these areas, in contrast, the types of ‘Clactons’ and ‘Cambridges’ char- Conservatives have seen the most substantial acterised in Jeremy Cliffe’s essay in The gains in their support between 2005 and Economist on ‘England’s cosmopolitan future’,3 2017. Together, these patterns confirm our before proceeding to examine the changes in findings from 2016, of two subsets of Eng- England’s electoral geography between 2005 land, moving in different directions, regard- and2017againstarangeofdemographic less of the national electoral tide. The notion measures.4 There is a tricky issue of termi- of ‘uniform national swing’ has limited rele- nology here which we have found difficult vance for these groups of constituencies. to resolve. By using the term ‘cosmopolitan’ Another way of demonstrating this long- for certain areas of England, we are not stat- term trend is contrasting the difference ing a normative preference, but rather trying between Labour’s share of the vote in these to capture the global orientation and connec- urban-metropolitan constituencies and regio- tion to the current dynamic of economic nal-coastal constituencies over time. This is

2WILL J ENNINGS AND G ERRY S TOKER

The Political Quarterly © The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017 Figure 1: Labour/Conservative vote share in urban-metropolitan constituencies in 2005 and 2017

Figure 2: Labour/Conservative vote share in regional-coastal constituencies in 2005 and 2017 shown in Figure 3, alongside the gap ﬁve points higher. By way of contrast, there between urban-metropolitan constituencies was little change in the difference between and the average English constituency. The the Conservative vote in the two types of ﬁgure reveals a growing gap in the Labour area over the period between 2005 and 2015 vote between the two types of place, more (not shown here), but the gap widened sub- than doubling between 2005 and 2017 from stantially in 2017 (by over ten points). around ten points to over twenty points. The Finally, to consider electoral change in dif- increase between the 2015 and 2017 elections ferent sorts of places, in Figure 4 we plot was only small, however, relative to the change in vote share for all constituencies over change over the preceding decade. Notably, this period against their population density too, Labour’s support in these urban- (measured by the number of persons per hec- metropolitan areas was slightly lower than tare in the 2011 census). This reveals a pro- the average constituency in 2005, but is now nounced pattern whereby the Labour vote

T ILTING T OWARDS THE C OSMOPOLITAN A XIS?3

© The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017 The Political Quarterly Figure 3: Difference between Labour’s vote share in urban-metropolitan versus regional- coastal areas and all English constituencies, 2005–2017 share has tended to rise in urban areas (that is, others, of painting a picture of the former as major cities), whereas the Conservative vote a globalist, liberal metropolitan elite and has tended to increase in less densely popu- those ‘left behind’ in other areas as rooted in lated towns and rural areas.5 To understand nostalgia and community. The former are the general election of 2017, then, it is crucial characterised as people from ‘anywhere’ to recognise the longer-term tilting of electoral (due to their lack of connectedness to place politics in England—where the advances of or identity) and the latter as being from the parties have differed by place. ‘somewhere’. The argument is that it is the somewheres who need to be better under- The complex social ecology of the stood and, it would appear, be sympathised with given the normative undertones of new political divides these labels. This is the language used in There is a danger in some commentary on David Goodhart’s6 The Road to Somewhere the divide between cosmopolitans and and its tone is emphasised in the review of

Figure 4: Change in vote share (2005–2017) against population density of constituencies

4WILL J ENNINGS AND G ERRY S TOKER

The Political Quarterly © The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017 the book by Matthew Goodwin7 that focuses and the gym, and playing sports. Crucially, on ‘two different groups increasingly pitted cosmopolitan areas also contain a proportion against each other. On one side are the lib- of those classes that are not doing so well, eral, socially mobile and university-educated including emergent service workers and the “people from Anywhere”, who subscribe to precariat. The emergent service category an “achieved” and cosmopolitan identity. includes many employed as bar staff, chefs, On the other side are conservative, margina- entertainers, customer service workers and lised “people from Somewhere”, who sub- in other relatively low paid positions. They scribe to a roots-based conception of national have reasonable social capital and high identity and cherish ways of life that have levels of emerging cultural capital. The pre- been lost or are under threat.’ However, this cariat similarly score low on economic cultural demarcation is far too crude in that capital but also on social and cultural dimen- it fails to recognise that considerable social sions, yet as cleaners, care workers, van dri- capital and identity is to be found in cos- vers, retail cashiers, security guards and mopolitan areas where communitarian poli- caretakers, for example, they are an integral tics can be developed via alternative roots part of the cosmopolitan areas. The only two and social networks (or may be constructed classes that would be relatively absent from in online communities). Moreover, this char- many cosmopolitan areas are the traditional acterisation misses the diversity of socio- working class and the established middle economic statuses of the populations of class in the new class structure identified by cosmopolitan areas—which are often, espe- Savage et al. (2013). Significantly, while cos- cially in cities, characterised by high levels of mopolitan areas are home to high numbers inequality (an important point that our origi- of graduates in terms of these new social nal article in The Political Quarterly over- classes, they are not only found in the elite looked). and established middle classes, but also in A better starting point to understanding some proportion of the five other classes. the complexities of social change in Britain is Patterns of voting in the 2017 general elec- provided by Mike Savage, Fiona Devine and tion confirm the relevance of these new cate- colleagues who have mapped a more varie- gories, at the same time as highlighting the gated and fragmented social class structure persistence of traditional political divides for Britain that mixes economic, social and which are far from dying out just yet. These cultural dimensions.8 These go beyond tradi- are summarised in Table 1, where we report tional conceptions of social class premised the direction of (significant) correlations on the divide between manual and profes- between selected demographics and Labour’s sional occupations, which have become lar- vote share in 2017 (in the first column) and gely outdated by Britain’s shift to a service- change in its vote share between 2005 and based economy and to mass higher educa- 2017 (in the second column), by parliamen- tion, where around thirty per cent of young tary constituency. A number of important people go to university. Savage et al. iden- patterns and trends are evident. First, sup- tify seven classes and provide data on the port for Labour is higher in areas with a types of occupations associated with those greater proportion of people employed in groupings and also their geographical loca- routine occupations, below the national liv- tions. Cosmopolitan areas—big cities, univer- ing wage and with no qualifications. How- sity towns—have (in London especially) the ever, there has been a downward trend in its highest concentration of a very small elite vote in these areas over the past decade. class—with high income, wealth and social Interestingly, while Labour’s vote is substan- and cultural capital. However, these areas tially higher in constituencies with a higher are also home to large numbers of the tech- proportion of precariat and emergent service nical middle class and new affluent workers workers (for example, carers, cleaners, sales who are of reasonable economic standing, assistants, security guards)9 there has been good (if limited) social networks and moder- no decrease in its vote over time as observed ate cultural capital, especially focused on for the traditional (that is, routine, manual) ‘emerging’ cultural activities such as social working class. Additionally, Labour’s vote network sites, video games, going to gigs share has risen slightly more in areas that

T ILTING T OWARDS THE C OSMOPOLITAN A XIS?5

© The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017 The Political Quarterly Table 1: Vote share (2017) and change in vote share (2005–2017) for Labour by demo- graphics11

Vote share, 2017 Change in vote share, 2005–2017

Routine occupations + – Employment in manufacturing (2001–2011) – + Precariat + . Employment in cosmopolitan sectors . + Jobs density . + Below national living wage + – Increase in mean real pay (2011–2016) . – Social renters ++ Poor health (self-reported) + . No qualiﬁcations + – Graduates – + Younger voters (18–44 year olds) ++ Ethnicity: white British –– Population density ++

Note: ‘+’ or ‘–’ indicates a positive or negative pairwise correlation (Pearson’s r)p<0.05, N = 530. have seen a fall in real pay over the last five of its vote in cities with high job density, years. This suggests that the decline of among graduates and younger voters, and in Labour as a party of lower strata of social more ethnically diverse areas. However, it class—of the ‘left behind’—may have been retains an important underlying structure of over-stated, or over-simplified at very least. traditional forces—those in routine occupa- Alongside this, the results show that tions and with no qualifications, below the Labour has made gains in constituencies with national living wage, in social housing and more people employed in ‘cosmopolitan’ in poor health. sectors (that is, finance, science, public administration, education, health, arts and recre- A partial realignment? ation),10 and with greater jobs density (that is, the number of jobs in an area divided by the During the election campaign, and well resident population aged between sixteen and before, there was talk that on the back of the sixty-four years old). While it still tends to Brexit vote the Conservatives would strike receive a lower share of the vote in con- deep into Labour’s heartlands, with predic- stituencies with more graduates, the trend is tions that Labour would lose seats like Bols- for graduate support for Labour to rise. This over that the party had held for over fifty suggests that Labour has growing appeal to years. The Conservatives did achieve swings voters who are central to the modern British towards it in parts of the North and Mid- economy—retaining support of the precariat lands (often where the UK Independence while making advances with graduates. At Party (UKIP) had polled highly in 2015), but the same time, it receives a higher level of the election saw them lose constituencies like support from social renters and those report- Canterbury—which the party had held since ing ‘bad health’. Perhaps most important of 1874. In Figure 5(a) and 5(b) we plot changes all, Labour’s share of the vote is higher in in the Labour and Conservative vote share constituencies with a younger population against the Leave share of the vote in the EU (aged between eighteen and forty-four years referendum, using the constituency-level esti- old), and lower in areas with less ethnic mates produced by Hanretty.12 This reveals diversity (and a higher proportion of white that that the Conservatives did indeed tend British population). Both these trends have to make gains in areas where there had been increased over the past decade. a higher Leave vote in the EU referendum, Together, these results highlight the cos- while Labour tended to make gains in areas mopolitan tilt of Labour’s vote—with growth where there was a higher Remain vote.

6WILL J ENNINGS AND G ERRY S TOKER

The Political Quarterly © The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017 Figure 5: (a) Change in Labour/Conservative vote share (2005–2017) and Leave vote. (b) Change in Labour/Conservative vote share (2015–2017) and Leave vote

Significantly, however, the correspondence is To the extent to which the revolt on the more pronounced between the Remain/Leave right among working class whites matters, it vote and change in vote share over the is inextricably interlinked with the crystallis- longer-term between 2005 and 2017, in Fig- ing of a socially liberal, leftist outlook ure 5(a), compared to the change between among younger and more educated voters 2015 and 2017, in Figure 5(b), where the tra- that is tilting the axis of contemporary poli- jectory of the slopes of the (polynomial) line tics. As social class develops a more com- of best fit are much flatter. This suggests that plex pattern, age and education have to the extent that the Brexit vote was associ- become the defining dimensions of political ated with voting patterns in the 2017 election, competition. Crucially, these are interlocking it was more due to its being a symptom of the factors—since increasing numbers of young long-term social and political changes that people have enjoyed access to a university preceded it—rather than being the focus of an education, while the fragmentation of the immediate Brexit realignment of English, and structure of social class between traditional British, politics itself. (routine/manual) and new (precariat/

T ILTING T OWARDS THE C OSMOPOLITAN A XIS?7

© The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017 The Political Quarterly emergent services) has blurred the nature of These reveal that support for Brexit was class voting. associated with an increase in the Conserva- Our final step is, therefore, to model tive vote and a decrease in the Labour vote change in the vote share of both Labour and between 2005 and 2017. The coefficient the Conservatives between 2005 and 2017 as (À0.731, p<0.001) is twice as large for a function of each of the demographics dis- Labour, indicating a stronger relationship cussed above. The results are reported in between the forces that underpinned Brexit Table 2, and reveal how complex and partial and long-term change in Labour’s vote. To the electoral shift that has occurred is. We the extent Brexit is associated with a realign- first present the results of ordinary least ment of English politics during this period, squared (OLS) regression models that the consequences have been most profound include the Leave vote percentage in the EU for the Labour Party—but far predate the referendum, in the first and third columns. referendum.

Table 2: OLS regression model of change in Labour/Conservative vote share (by constituency), 2005–2017

Change in Conservative Change in Labour vote, 2005–2017 vote, 2005–2017

Leave vote (%) 0.357 À0.731 (0.055)*** (0.071)*** Population density 0.010 0.014 0.043 0.034 (0.020) (0.020) (0.025) (0.028) Routine occupations (%) 0.277 0.218 À0.243 À0.124 (0.113)* (0.117) (0.145) (0.159) Change in manufacturing employment (2001–2011) À0.554 À0.593 0.227 0.306 (0.130)*** (0.135)*** (0.167) (0.183) Precariat (%) À0.048 À0.217 0.104 0.451 (0.135) (0.138) (0.174) (0.188)* Employment: cosmopolitan industries (%) À0.178 À0.289 0.067 0.293 (0.072)* (0.073)*** (0.093) (0.099)** Jobs density À0.145 0.026 0.498 0.149 (0.369) (0.383) (0.475) (0.520) Below national living wage (%) 0.027 0.035 À0.009 À0.025 (0.035) (0.037) (0.045) (0.050) Change in mean pay (2011–2016) 0.026 0.015 0.012 0.036 (0.032) (0.033) (0.041) (0.045) Social renters (%) 0.109 0.102 À0.148 À0.134 (0.048)* (0.050)* (0.062)* (0.068) Bad health (%) À0.727 À1.009 1.596 2.175 (0.357)* (0.369)** (0.460)*** (0.501)*** Degree (%) À0.023 À0.695 0.075 1.449 (0.221) (0.203)*** (0.285) (0.276)*** No qualiﬁcations (%) À0.255 À0.204 0.028 À0.076 (0.162) (0.168) (0.209) (0.229) Aged 18–44 (%) À0.255 À0.547 0.460 1.057 (0.153) (0.152)*** (0.197)* (0.206)*** Ethnicity: white British (%) 0.168 0.177 À0.155 À0.174 (0.023)*** (0.024)*** (0.029)*** (0.032)*** Aged 18–44 (%) x degree (%) 0.008 0.014 À0.017 À0.029 (0.004) (0.004)** (0.005)** (0.006)*** Intercept À16.042 36.088 50.863 À55.784 (12.867) (10.425)*** (16.570)** (14.176)*** N 530 530 530 530 Adjusted R-squared 0.55 0.51 0.59 0.50

*p<0.05; **p<0.01; ***p<0.001 (standard errors in parentheses)

8WILL J ENNINGS AND G ERRY S TOKER

The Political Quarterly © The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017 The remainder of our analysis focuses on towns, with younger and more educated those models that exclude the Leave vote, populations—at the same time as shoring since it is itself correlated with many of the up its support among the new working class, other demographics considered here. The the precariat and emergent service classes estimates presented in the second and fourth (those employed in insecure and short-term columns of Table 2 provide important jobs, living alongside professionals occupa- insights on the partial realignment of British tions in economic hubs). The Conservatives politics during this period. Notably, Labour’s have made gains in less ethnically diverse vote increased considerably in constituencies areas, and in former manufacturing areas, with a higher proportion of younger voters though these advances were not enough to and graduates, while the interaction of those prevent seat losses to Labour in the 2017 measures is negative and significant—indi- general election. To what extent these forces cating that there was a saturation effect, with of electoral change will persist into the Labour making smaller advances in areas future in part depends on the policy choices with both more educated and younger popu- taken by government and the strategies of lations (understandably given these sub- parties. populations overlap). Labour’s vote also Parties need to develop strategies that increased in areas with higher levels of avoid simplistic responses to this bifurcation employment in what we classified earlier as of politics that pit one side against the other ‘cosmopolitan industries’. At the same time, —keeping in mind that there remains a sub- however, Labour also increased its vote in stantial centre ground in British politics, constituencies with a higher share of pre- even if it seems elusive at present. There has cariat workers and people with (self- been considerable debate among leading reported) poor health. This is somewhat at Labour figures on how to bridge the gap odds with characterisation of a party that between the ‘Two Englands’.13 Conservative has lost the ‘left behind’. At the very least, it attempts to address these dynamics have fiz- highlights that Labour has gained support zled out since modernisation efforts during among a particular lower strata of social the earlier days of Cameron’s leadership,14 class—those who are particularly vulnerable and may be further hampered by Brexit. in the modern economy. One type of strategy concerns the issue of In contrast, the Conservatives have made ‘cultural backlash’. In their ground-breaking electoral gains in areas with lower levels of study, Ford and Goodwin15 found that sup- ethnic diversity (that is, with a higher pro- port for the radical-right UKIP was associ- portion of white British population) and ated with economic pessimism and distrust those areas that have experienced declines in of government, but most of all, concern manufacturing employment in recent years about immigration and expression of —explaining their advances in parts of the national identity, in particular among the North and the Midlands. The proportion of older white working-class. These voters have people employed in routine occupations is been marginalised, it has been argued, by only weakly associated with increases in the the dominance of a university-educated, pro- Conservative vote (p=0.063), however. This fessional middle-class elite, whose priorities suggests that the class basis of the realign- and outlook define the mainstream. Inglehart ment of English politics is partial—or at least and Norris16 argue that growth of support not effectively captured by these aggregate- for populist parties more generally is due to level measures. a cultural backlash among those citizens who feel that the world has been moving away from their values. Conclusions An alternative strategy focuses on The general election of June 2017 revealed increases in social and economic inequality— the continued tilting of the political axis in such as popularised in the work of Piketty17 England. As part of this longer-term geo- —in particular circumstances where ‘the graphical bifurcation of British politics, wealth of an elite has grown apace while the Labour has made the largest gains in cos- wealth of the majority has stagnated’.18 mopolitan areas—big cities and university Bifurcation here is attributed as a function of

T ILTING T OWARDS THE C OSMOPOLITAN A XIS?9

© The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017 The Political Quarterly growing, often generational, divides in levels relative to the attitudes of their parents and of home ownership, and access to other grandparents at a similar age.20 This is not forms of wealth—where, interestingly, afflu- incompatible with the social liberalism of ent London fares poorly in terms of net younger cosmopolitans, but suggests that the wealth,19 and the increasing strain on public waters of British politics may be increasingly services after years of austerity. From this tricky for parties to navigate with a compass perspective, the clue to the economic dimen- designed for traditional left-right politics sion of the ‘cultural backlash’ is its particular organised around economics and the size of prevalence among the working class. the state. Broadly, in the 2017 election campaign the Given the changing nature of its support Conservatives under Theresa May’s leader- base, voter registration could present chal- ship sought to exploit the first diagnosis by lenges for Labour in future. Data from the offering restrictive policies on immigration, 2015 British Election Study shows that those the reassertion of national sovereignty aged under forty-five are less likely to be through commitment to leaving the EU and registered to vote or are more likely to be a general sense of ‘turning the clocks back’ registered at a different address. These (with signature policies such as grammar higher rates of mobility are a feature of the schools and a free vote on fox hunting). populations of cosmopolitan areas. How to Labour under the leadership of Jeremy Cor- mobilise disparate social groups may require byn adopted the logic of the second diagno- multiple campaign strategies. Social media sis by focusing its manifesto on the rhetoric may be effective in mobilising some prospec- of redistribution and in particular arguing tive supporters (especially the young), but that the winners from globalisation—wealthy not others. Meanwhile, traditional door-to- individuals and companies—should pay door canvassing may no longer provide more in tax and be prevented from evading accurate indications to parties of the local their tax liabilities (alongside symbolic poli- state of play. The result on 8 June seemingly cies focused on younger voters such as the took both parties by surprise, meaning they abolition of student fees). Ultimately, the often had not deployed resources in seats result suggested that May’s attack on ‘citi- where the result was close. zens of nowhere’ backfired badly, as the The landscape of political and electoral Conservatives lost seats that were home to campaigning appears to be changing. Yet some of its traditional supporters, the edu- while the general election of June 2017 hints cated, professional middle-classes. at a tilting of the political axis, it would be In the longer term, the problem for both premature to assume that a realignment of the parties is that aligning themselves with British politics is inevitable. either group of voters brings electoral risks. For the Conservatives, becoming the party of Notes Brexit and the shrinking (and ageing) tradi- 1 W. Jennings and G. Stoker, ‘The Bifurcation of tional working class is arguably not a Politics: Two Englands’, The Political Quarterly, straightforward path to future electoral suc- – ‘ ’ vol. 87, no. 3, 2016, pp. 372 82. cess. Indeed, an emphasis on the left behind 2 In this analysis we focus on England specifi- has distracted from the inequality and eco- cally, although would expect that the dynamics nomic distress within cosmopolitan areas— of bifurcation that we observe apply similarly privileging the traditional white working in Wales. Scotland and Northern Ireland have class at the expense of the growing precariat their own distinctive territorial politics, beyond and emergent service class—for whom social that which can be considered here. 3 J. Cliffe, ‘England’s cosmopolitan future: Clac- and political inequality is just, if not more, ’ acute. For Labour, relying on the enthusiasm ton versus Cambridge , The Economist,6 September 2014. of younger voters and support from cos- 4 Our data on constituency election results are mopolitan areas in future elections may also ’ from the Electoral Commission and Pippa Nor- be risky given the party s ambiguous posi- ris’ 2010 British General Election Constituency tion on Brexit, and the rightward shift Results Release 5.0 dataset. Our data on con- among younger generations on socio- stituency demographics are from the Office for economic and authoritarian dimensions— National Statistics.

10 W ILL J ENNINGS AND G ERRY S TOKER

The Political Quarterly © The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017 5 Note that the vote share of both Labour and mean real wages between 2011 and 2016 the Conservatives in England increased (Annual Survey of Hours and Earnings); social between 2005 and 2017 (from thirty-seven per renters (housing tenure; QS403UK), percentage cent to forty-three per cent and thirty-five per reporting ‘bad health’ or ‘very bad health’ cent to forty-five per cent respectively), which (QS302EW); no qualifications (KS501UK); grad- is why the average change for both parties is uates (degree level or above; KS501UK); greater than zero in Figure 4. younger voters (per cent of eighteen to forty- 6 D. Goodhart, The Road to Somewhere: The Pop- four year olds resident in constituency; ulist Revolt and the Future of Politics, London, C QS103EW); ethnicity (white: English/Welsh/ Hurst & Co., 2017. Scottish/Northern Irish/British; KS201EW); 7 M. Goodwin, ‘Shocked by populism? You population density (number of people per shouldn’t be. The revolt described by David hectare; KS101EW). Goodhart “The Road to Somewhere” has been 12 C. Hanretty, ‘Areal interpolation and the UK’s decades in the making’, The Financial Times,29 referendum on EU membership’, Journal of Elec- March 2017. tions, Public Opinion and Parties, 2017, https://d 8 M. Savage, F. Devine, N. Cunningham, M. oi.org/10.1080/17457289.2017.1287081. Taylor, Y. Li, J. Hjellbrekke, B. Le Roux, S. 13 T. Hunt and A. Lockey, ‘English radicalism Friedman and A. Miles, ‘A new model of social and the annihilation of the ‘progressive class? Findings from the BBC’s great British dilemma’’, The Political Quarterly, vol. 88, no. 1, class survey experiment’, Sociology, vol. 47, no. 2016, pp. 116–25; B. Phillipson, ‘The conditions 2, 2013, pp. 219–50. for Labour’s previous successes are falling 9 This measure is constructed using data on occu- apart. Where do we go from here?’ The New pations from the Office of National Statistics, Statesman, 18 November 2016; L. Nandy, ‘Who with categories selected based on the schema speaks for Britain’s towns in a post-Brexit developed by Savage et al. (2013), ‘A new world?’ The Times, 7 March 2017. model of social class?’: ‘road transport drivers’; 14 P. Kerr and R. Hayton, ‘Whatever happened to ‘other drivers and transport operatives’; ‘sales Conservative party modernisation?’ British Pol- and customer service occupations’; ‘caring, lei- itics, vol. 10, no. 2, 2015, pp. 114–30. sure and other service occupations’; ‘elementary 15 R. Ford and M. Goodwin, Revolt on the Right, cleaning occupations’; ‘elementary security London, Routledge, 2014. occupations’; ‘elementary sales occupations’. 16 R. Inglehart and P. Norris, ‘Trump and the 10 This measure is constructed from ONS data on populist authoritarian parties: the silent revolu- employment by sector (QS605UK), using these tion in reverse’, Perspectives on Politics, vol. 15, categories: ‘financial and insurance activities’; no. 2, 2017, pp. 443–54. ‘professional, scientific and technical activities’; 17 T. Piketty, Capital in the Twenty-First Century, ‘public administration and defence; compulsory Cambridge, MA., Harvard University Press, social security’; ‘education’; ‘human health and 2014. social work activities’; ‘arts, entertainment and 18 F. Devine and M. Sensier, ‘Class, politics and recreation’; ‘other service activities’. the progressive dilemma’, The Political Quar- 11 These data are from the ONS/Nomis: per cent terly, vol. 88, no. 1, 2017, p. 30. of population employed in routine occupations 19 C. D’Arcy and L. Gardiner, The Generation of (semi-routine and routine occupations, plus Wealth: Asset Accumulation Across and Within never worked and long-term unemployed; Cohorts, London, Resolution Foundation, 2017. QS607UK), change in per cent of population 20 See M. Grasso, S. Farrall, E. Gray, C. Hay, and employed in manufacturing between 2001 and W. Jennings, ‘Thatcher’s children, Blair’s 2011 (categories ‘mining and quarrying’ and babies, political socialisation and trickle-down ‘manufacturing’; QS605UK); jobs density; per- value change: an age, period and cohort analy- centage below national living wage (Annual sis’, British Journal of Political Science, 2017, Survey of Hours and Earnings); change in https://doi.org/10.1017/s0007123416000375.

T ILTING T OWARDS THE C OSMOPOLITAN A XIS?11