Labour’s new majority

The shifting ethnic composition of Britain has major implications for our political parties, suggest Trevor Phillips and Richard Webber in the latest in their series on Britain’s changing demography.

Following our last Demos Quarterly piece on the subject of ethnic voting there has been some speculation as to whether the Conservative party can ever win an outright majority in a future general election without raising its share of the non-white minority vote above its current 16 per cent. The question has gained new salience following the recent UKIP surge. But our research suggests, unexpectedly, that ethnocultural diversity poses even more profound questions for the politics of the centre-left.

The cultural difficulty for the Conservatives in attracting visible minority voters has been widely articulated, particularly by Lord Ashcroft. The Liberal Democrats currently face larger, existential problems. Labour has been accused of complacency; but we believe that an emerging cultural division presents a new – and immediate – dilemma for Mr Miliband and his team. Can the party respond to the UKIP assault on its (largely white) heartlands in the North, whilst holding on to its support amongst ethnic minority voters in the South? Conventional Labour thinking assumes that minority defection will always be a marginal phenomenon.

There is some support for this view in the magisterial analysis of the 2010 election by Anthony Heath and his colleagues, responsible for the Ethnic Minority British Election Study (EMBES). They show that in between 1997 and 2010, the Labour vote overall fell from 47 per cent to 29 per cent of White Britons, whilst the fall amongst ethnic minorities was proportionately much smaller – from 82 per cent to 68 per cent. Amongst minorities the EMBES measure of party allegiance – ‘partisanship’ – to Labour in fact rose from 68 to 75 per cent during the period, in contrast to a fall in white Labour partisanship from 43 to 36 points.

Thus, whilst the lower minority vote in 2010 may suggest a long-term decline in Labour support, it is more likely to represent a short-term expression of disappointment with the Blair/Brown administration. In practice, the 68 per cent of minority voters who opted for Labour in 2010 may come to be seen as the nadir of the party’s support in this group, rather than a trend away from Labour.

Heath and his colleagues also found that minority voters who defect from Labour are less likely than others to switch to another party; rather, they tend simply not to vote at all. This finding is supported by the constituency analysis made by John Curtice and his team. They showed that the average fall in Labour’s share of the vote in seats with a large minority population was much smaller (0.7 points on average) than the average constituency fall in Britain as a whole (6.5 points).

A recent report by Policy Exchange, A Portrait of Modern Britain, has added to the picture by presenting new demographic data. This report points to the arrival of what we’ve previously described as ‘superdiversity’ in the UK – a multi-ethnic society with high numbers of non-white Britons from many different ethnocultural backgrounds. The size of the visible minority population of working age is already around 5 million people – somewhat more than the 3-4 million votes currently predicted for UKIP in a general election.

The visible minority electorate will grow more rapidly than any other demographic group in the next thirty years. Superdiversity therefore raises several important challenges to the conventional wisdom. Will minority voters desert Labour as some become more prosperous and move away from their original areas of settlement? Will new groups of immigrants and their children behave similarly to previous waves? And how will settled groups respond to the dispersal of minority communities into largely white suburbs?

We’ve carried out some preliminary, broad, calculations, particularly focusing on the effect of demographic change on the composition of the Labour vote. This tells a very different story to the conventional narrative, and suggests that even if a ‘policy’ tilt towards UKIP may benefit Labour in the short-term (in time for the 2015 general election), there will be a huge price for Mr Miliband and his successors to pay in every subsequent general election for several decades.

Our conclusions are driven by three demographic and cultural factors, each of which will definitely feature in the UK during the next forty years:

The significance of the first of these factors is underestimated by most commentators, mainly because they usually fail to connect it to the second. But this pattern is part of a common European trend: for example, Germany’s net immigration rose to a record 437,000 in 2013; the Turkish minority has historically opted for the SPD, the Greens and other left-wing parties by a 70 to 20 majority – virtually identical to our own Labour/Conservative split amongst ethnic minorities. However, the UK (particularly London) remains the most striking example of visible minority population growth in the EU. Figure 1 shows the rise in minority share of the population in seven countries in Europe and North America, with the UK surpassing the USA by mid-century (though the USA figure does not include African-Americans).

Against this background, the three trends set out above produce some surprising outcomes, all other things being equal:

  • First, London’s recent Euro-election results were less a rejection of UKIP, and more a sign of an electorate sharply divided on ethnocultural lines: two out every three visible minority voters voted Labour, whilst two out every three white voters cast a ballot for the Tories or UKIP.
  • Second, UKIP outpolled Labour by almost two to one amongst white British voters in the capital.
  • Third, by 2050, a majority of Labour’s popular vote in the UK will be cast by visible minority voters. This ‘browning’ of Labour could take place even earlier, by 2035, based on projections from Oxford University.
  • Fourth, if Scotland leaves the UK, the browning of Labour will take place yet sooner – perhaps as soon as 2025, and certainly by 2035. This is well within the lifetime of the current Labour leader. It would be the first time in a European or American election that a potential governing party depended on ethnocultural minorities for a majority of its votes.
  • Fifth, brown Labour could enjoy a settled majority in a Scot-free UK. In 2010 the balance in the popular vote in England and Wales was Con 10.5m, Lab 9m; in 2050 we estimate that it could be Con 11.8m, Lab 12.3m.
  • Sixth, evidence from other European countries suggests that Labour cannot rely indefinitely on being the only destination for minority votes; both its SPD colleagues and the Greens in Germany have lost almost a third of their Turkish minority supporters in the past decade, the main beneficiaries being the centre-right CDU.
  • Seventh, the House of Commons suffers a major democratic deficit; if we compare the number of minority MPs with the number of minority voters we find that currently there are three times as many minority electors for each minority MP as there are electors for each MP overall.
  • Eighth, that the ethnic ‘penalty’ for minority Labour voters is twice as large as that paid by minority Conservative voters.

London and the 2014 Euro-elections

The recent Euro-elections in the capital gave a clear signal of the enormous changes that are already taking place in the electorate. In the wake of this poll there was some discussion about whether London was developing a cultural distance from the rest of the country. Our work suggests that the answer to this question is ‘yes’. However, the explanation lies less in the city’s supposed social liberalism, than in its demographic transformation.

Today, 87 per cent of people in the UK identify themselves as ‘white British’. In the eight largest British cities outside London that figure is slightly smaller – around 83 per cent. In the capital, by contrast, just under 45 per cent of the population identify as white British, with almost 40 per cent ticking the box of one visible minority group or another. Broadly, the ethno-cultural demography of the UK overall resembles the ethno-cultural demography of London thirty years ago – but it’s not standing still. Most demographers predict that the national picture in another thirty years will look something like London looks now.

To mangle LP Hartley, London isn’t a foreign country; it’s the future – and they do things differently there. The prospect is what might be described as an ‘American’ future in which politics are colour-coded, and where, by 2050, ethno-cultural preferences become a dominant factor in politics.

It might be argued that these ethno-cultural preferences are simply a proxy for socio-economic divisions; but we can find no socio-economic classification that sufficiently explains the weight and persistence of Labour support amongst visible minorities. This overwhelming preference has remained virtually impervious to economic circumstances, and has crossed generations. For example, voters in the largest ethnocultural minority group in London, (Indian-heritage voters) consistently prefer Labour by a sizeable majority, in spite of being more highly educated and wealthier than most other groups, including white Londoners. These political preferences are clearly largely cultural, not economic. (The same might be said for UKIP voters, as suggested by Ford and Goodwin).

In the May 2014 contests, the striking contrast between London and the rest of the country suggests that the predominance of visible minority voters in the capital is already having a decisive electoral effect. Labour success in Croydon, Redbridge and Merton indicates that the minority dispersal into London’s middle ring suburbs predicted by our recent Demos research is already taking place; former Tory marginals are turning into Labour strongholds.

At least part of the explanation lies in significant increases in visible ethnic minority shares of the population: Croydon (up 50 per cent); Redbridge (up 32 per cent); and Harrow, now over 60 per cent ethnic minority, and home to many affluent but Labour-loyal Indian voters, who in 2011 already constituted over a quarter of the borough’s population.

Labour’s ‘surprise’ victory in Hammersmith and Fulham looks far less surprising in the light of dramatic demographic change over the past decade. The white British proportion of the borough’s population fell from 58.0 per cent to 44.9 per cent; the visible minority share of the population rose by 45 per cent – from 20.2 per cent to 29.3 per cent – an increase of nearly half in a group which consistently votes Labour.

However, an even more significant political finding emerges from our initial analysis of the London results, suggesting that UKIP’s performance in London was not as abject as reported. The vote shares divide as below, between Labour, Conservatives, UKIP and others. We have adopted the normal assumptions about the share of minority votes for the largest parties: Conservatives 16 per cent, Labour 68 per cent, derived from the authoritative British Election Study. According to the Electoral Commission’s latest Winter tracker, there is no difference in registration rates directly attributable to ethnicity.

Set against the background of the capital’s ethnic composition (White British 45 per cent, visible minority 40 per cent, others 15 per cent) a striking pattern emerges. White Londoners exhibit markedly different voting behavior to non-white Londoners. As a result, UKIP was just one point adrift of the Conservatives amongst white British voters; and Mr Farage’s party actually beat Labour amongst white Brits, probably by a ratio of as much as two-to-one.

Table 1: London Euro-elections 2014: vote shares by ethnicity

% of total vote % share of vote from visible minority electors % share of vote from White British electors % share of vote from Other White electors
Labour 37 26 10 1
Conservative 23 3 18 2
UKIP 21 3 17 1
Others 19 9 9 1
TOTALS 100 41 54 5

The chart below illustrates with stark simplicity the degree of ethnocultural segregation demonstrated in the May 22 election, with the contrast being even more striking when the Conservative and UKIP shares are combined, perhaps a better guide to future mayoral and parliamentary elections. In effect, two out of every three visible minority voters voted Labour (26 per cent of 41 per cent), reflecting the party’s typical share of the minority vote. What is more surprising, however, is that two out every three white voters cast a ballot for the Tories or UKIP (35 per cent of 54 per cent).


The signs of this cultural divide were clearly signalled in the 2012 Mayoral election. Boris Johnson and Ken Livingstone each won about a million votes. Within that total of 2 million, about 0.8m will have been visible minority voters; but these votes were not distributed equally. The vast majority would have gone to Labour. Even assuming that Livingstone underperformed his party and attracted only 60 per cent of the minority vote, that still meant that 0.5m – over half of Labour’s votes – would have come from non-white Londoners. A future Conservative mayoral candidate is unlikely to carry Johnson’s cross-party appeal, or to dilute significantly the ethnic minority distaste for the Conservative brand.

The Future

In 2011, the total visible minority population in England and Wales was around 8 million people, including 5.4m adults. Scotland and Northern Ireland accounted for a further 200,000 visible minority adults – a total of 5.6 million UK-wide. We have made a series of cautious assumptions to allow us, on the basis of these data, to estimate the size of the visible minority share of the electorate and its distribution between the Conservative and Labour parties (see endnote). We have predicted all future outcomes as though the overall party shares achieved in 2010 remained constant. Any changes in overall share would have some effect on our projections, but outside of a complete collapse of one of the major parties, or the emergence of a completely new electoral force, the impact would be marginal.

Estimating conservatively, we have derived a figure for the total number of visible minority voters who cast a ballot in the UK, according to the formula:

  1. Total = 5.6 million (adults) x 0.9 (90% registration) x 0.65 (65.1% turnout) = 3.28 million votes cast

Of these, according to the British Election Study, 68 per cent would have voted for the Labour party. The total UK Labour vote fluctuates, but in 2010 approached 9m electors; hence the visible minority contribution to Labour’s vote in 2010 would have been around 2.2m or at least 24 per cent of the total. The equivalent figures for the Tories on a total of 10.5m would be 524,000, or 5 per cent.

If we turn to the Policy Exchange picture, it’s likely that division of the minority electorate between major parties will have altered only modestly by mid-century. We know from history that voting patterns amongst ethnic groups change very slowly, if at all. Memories are long and inheritance is strong. So, we believe that we can rule out anything but the most marginal changes in the 68:16 distribution over the next forty years.

There are various ways to calculate the likely effect of population growth by 2050. The simplest is to assume that the bulk of the population (63 million) then will behave pretty much as they do now, and then to work out what the new voters who make up the increased (80 million) population are likely to do. In this context, what we find is that the principal effect of the growth of the visible minority population lies in its impact on the Labour party’s popular vote.

For the purposes of this calculation, we are taking the projection of UK population growth used by Policy Exchange, provided by Philip Rees at Leeds University, of about 17m by the middle of the century. It’s estimated that 80 per cent of recent growth has come from immigrants, minorities and their children. Even if this relative high minority share of the growth slows to, say, 75 per cent, the additional population in 2050 would be divided roughly between 13m visible minority and 4m whites.

If we make the (unlikely) assumption that the age structure of the minority population remains as skewed towards the young as it is currently, this implies at least an additional 8.8 million minority adults (13 million x 0.675). We calculate, therefore, that there will be a further 5.1 million visible minority voters divided as follows: 3.6 million to Labour, and 0.8 million to the Tories, 0.7 million to others.

We calculate that there would also be an additional 3m white voters, yielding roughly 1m for each of the two largest parties. Thus by 2050 Labour could expect to draw a UK core vote of around 13.6 million, the Tories a UK core vote of 12.3 million. Both parties will see an increase in the number of their voters from visible minority communities (to 5.8m and 1.3m); the graph below demonstrates why these voters will be of far greater significance to Labour – compare the gentle rise in the Tory line with the steep Labour gradient.


If Scotland Goes: The Rest of the UK

What happens if Scotland leaves the Union? Using the Policy Exchange projections, Labour loses about a million votes of its 2010 total, a tiny proportion of whom are visible minority voters. The 2010 numbers now become, for the ‘rest of the UK’: Labour’s popular vote around 8m, the visible minority vote around 2.1m. In a general election without Scotland in, say, 2017, Labour would depend on visible minorities for more than a quarter of its core vote; of course, that proportion would fall if the party won extra votes from whites, but it would still be very significant in size.

The independence of Scotland, combined with asymmetric population growth could lead to further unexpected consequences. It is widely supposed that the rest of the UK would become a semi-permanent Conservative fiefdom. In fact, the truth is probably the opposite. Using the projections outlined above for UK vote shares in 2050 (Labour 13.6m, Conservatives 12.3m) we can estimate the effect on each party of the loss of their Scottish voters. This projection is based on the 2010 result.

In 2050 Labour would ‘miss’ 1.3m Scottish voters and the Conservatives would ‘miss’ 0.5m Scottish voters; thus giving Labour a core vote of 12.3m and the Conservatives a core vote of 11.8m. Labour would lose more votes than the Conservatives after a Scottish departure – but not enough to prevent the development of a semi-permanent Labour advantage in England and Wales.

In future contests, as the proportion of the whole electorate that is non-white changes, ‘brown’ Labour would become the default state of the non-Scottish party. In 2050, we would expect to see a Labour vote of 12.3m and a minimum minority share of 4.3m. On the Oxford projection this takes place by 2030; and by 2050, more than three out of every four Labour votes in the rest of the UK would be cast by visible minority voters.

Potential Party Strategies

What does all this mean for the parties? Are there ways in which the Conservatives could limit the decline in their overall vote share, for example? Perhaps. First, they could employ a race-conscious electoral strategy that would acknowledge that they won’t win many more minority votes any time soon, and adopt a less ambitious goal – to shed their image as a party hostile to minorities. In time they might then hope to nibble away at Labour’s advantage.

Alternatively, the Tories could adopt a more radical strategy by trying to peel off some minority groups. One version of this approach would be to copy Mrs Merkel’s systematic integration programme; or even more radical, to take a leaf out of the Canadian Tories’ playbook. Stephen Harper, the Canadian PM, has raised the share of the Tory vote from immigrants from 9 per cent in 2000 to 31 per cent in 2011; even more astonishingly, the Canadian Tories have become more popular amongst immigrants (42 per cent) than amongst the electorate at large.

The Canadian government has understandably put this down to their strenuous efforts to welcome new immigrants, but it may in fact owe more to a conscious policy over the past couple of decades of admitting more highly-educated, wealthier immigrants from countries with little memory of previous Conservative hostility to immigration (see David Goodhart in this issue for more).

For Labour, the bright patch on the horizon is that if the party can manage to keep its share of the minority vote intact and also attract some more white voters, it could create a permanent Labour advantage. But this is a far from straightforward balancing act. At the heart of Labour’s problem lies its response to the question of immigration, now regularly cited as the first or second most important issue for voters.

In the short term, Labour has adopted a policy response combining a series of apologies for underestimating the level of A8 migration in the early part of the century, with a promise to crack down on exploitation of low wage Labour by employers. But Labour’s effort to change voters’ minds on immigration has so far had little impact. One hypothesis is that defecting Labour voters are less concerned by the political and economic aspects of immigration, than by its cultural implications. The available evidence from Ford and Goodwin is that UKIP’s popularity has been due principally to the resonance of its emotional message.

But Labour’s range of responses to this ‘cultural’ question is complicated by the fact that its minority and white supporters hold utterly contradictory views. Anthony Heath and his team interviewed Labour voters in the wake of 2010 for the British Election Study, and discovered that there was virtually no difference of perception on most topics between white and non-white Labour partisans. The one exception was immigration. Amongst whites, the outgoing Labour administration had a negative approval rating of -36, compared to a positive rating of +10 amongst ethnic minority Labour supporters. To put it in context, this 46-point gap between black and white Labour loyalists was almost as large as that between white Labour partisans and white Conservative supporters (who scored the Brown government at -86%).

However, the inference from the behavior of SPD voters in Germany is that a failure to square the demographic circle could squander the natural centre-left advantage we’ve described above. A decade ago, the SPD could count on the support of 60 per cent of Turkish minority voters. In 2009, that support fell to 50.3 per cent; by 2013, it had dropped to 42.9 per cent – a loss of nearly a third of the minority vote in ten years. The Greens, the second most popular choice amongst ethnic Turkish voters, saw a similar level of defection over the period, falling from 31.9 per cent to 21.6 per cent. The major beneficiaries were Mrs Merkel’s CDU, up from 11.4 per cent to 20.3 per cent. It is unclear why this has taken place, but the defection of Muslims in 2005, post-Iraq war, and the unlikely victories of George Galloway should serve as a serious warning to Labour.

Parliamentary Representation

However, even if Conservative activism coupled with Labour’s own ambivalence were to result in the desertion of some minority voters from Labour, it could take a some time for the loss to register nationally. So Labour ‘complacency’ can – to some extent – be justified.

But our analysis still raises at least one major question for Labour: how long can it remain a largely non-white party with a largely white leadership? Looking at the party’s parliamentary representation, for example, which has grown from 4 minority MPs out of 229 (1987) to 16 out of 258 (2010), its leaders might feel some unease at the fact that whilst 24 per cent of Labour’s voters are non-white, just 6.2 per cent of its MPs are from visible minorities.

It is difficult to imagine how, at this rate of progress Labour will ever have a parliamentary party that truly represents its supporters. See Table 2 below, which uses the most modest projection we have developed and is based on the number of seats won in 2010 by the two major parties.

Table 1: Notional parity targets for visible minority MPs

vm share of Labour vote (%) Labour MPs based on 258 seats (%) vm share of Conservative vote based on 306 seats (%) Conservative MPs (%)
2010 24 62 6.6 20
2020 29 75 8.0 24
2030 34 88 9.3 28
2040 39 101 10.6 32
2050 45 116 12.0 37

There were 16 minority Labour MPs and 11 minority Conservative MPs elected in 2010. Thus in this parliament, the parties achieved 26 per cent (Lab) and 55 per cent (Con) of their notional parity targets – a clear democratic deficit. We can express this as a simple metric that measures the reward (or penalty) gained (or paid) in ethnic terms by each elector, ie how many electors does it take to produce an MP of any kind, compared to the number of electors it takes to put an MP who shares their ethnicity into office?

On average each UK Member of Parliament is drawn from a constituency of 68,175 electors, of whom around 44,400 cast a ballot. If we compare the total number of minority voters (3.3 million) with the total number of minority MPs (27) it is clear that it takes far more minority voters to put a minority MP into the House of Commons – 122,200 to be exact – than the average number of voters for each MP; in short, there is a democratic ‘penalty’ for minority voters.

But these figures are not the same for each party, as becomes clear if we compare the reward/penalty ratios for minority and white voters for the Labour and Conservative parties. In 2010 Labour’s 2.2 million minority electors were rewarded with just 16 minority Labour MPs – an average 137,500 per MP. The equivalent number for white Labour MPs and white electors was 28,100, an ethnic penalty of about four to one. For the Conservatives (0.7 million visible minority electors), it took 63,600 minority voters to produce a minority MP, and just 33,200 white voters to produce each white MP, an ethnic penalty of about two to one.

The remarkable finding here is that in 2010, minority Labour voters paid twice the democratic penalty suffered by minority Conservative voters. It does not take too much foresight to see that if the number of minority electors continues to rise without a corresponding change in the House of Commons, this striking disparity could become an unbridgeable gap, particularly for Labour.

If we take even the most modest projection of the visible minority Labour vote by 2050 (4.5 million), the penalty for Labour’s minority voters will rise to untenable levels; notionally, it would need over 280,000 minority electors to put each of the current 16 minority Labour MPs into office. The comparable figure for white Labour electors (8.5 million) and white Labour MPs (242) would be 35,123 – an ethnic penalty of about eight to one.

So can Labour dodge this demographic bullet? A focus on attacking UKIP may enthuse some minority Labour supporters, but it probably won’t add to their numbers, and will certainly alienate some potential white voters. So, mathematically, the options are limited – to close the gap, Labour needs either fewer minority voters or more minority MPs. A wholesale defection by minority voters to the Conservatives or Lib Dems is unlikely, and naturally, unwanted by Labour.

In the longer term, a sustainable approach may be indicated by Labour’s success in overcoming the paucity of female MPs. Though the party remains prone to some degree of what might be called ‘silverback syndrome’ – a gender diverse rank-and-file alongside a substantively male top leadership – the gender balance of the parliamentary party has changed radically during the past twenty years. Tackling the ethno-cultural equivalent – ‘snowy peak syndrome’ – may require measures which are just as radical as all-women shortlists, such as a UK version of the American ‘Rooney Rule’; for example, by requiring all shortlists and selection panels to include at least one person of colour.

This too will take time to deliver. In the immediate future, Labour’s best short-term response may lie in striking symbolic (and actual) steps – for example, changes in the complexion of the party’s top ranks. Is it too fanciful to suppose that an essential signpost to the future for Labour lies, not in raucous competition with the Conservatives and UKIP over immigration numbers, but, perhaps, in the shape of a potential Chancellor Chuka Umunna or Foreign Secretary Rushanara Ali?


We have made the following assumptions about the visible minority vote, all of which deliberately underestimate its size and growth, to arrive at a coefficient that relates the size of the visible minority population to the Conservative and Labour vote shares:

  • Adult minority population = total minority x 0.675, assuming no change in age profile from 2011
  • visible minority registration rate of about 90% (Electoral Commission estimates no difference in current registration rates between whites and BME)
  • turnout of about 65% (EMBES found no significant difference in white/minority rates in 2010)
  • total visible minority vote = no of adults x 0.9 x 0.65
  • 70% of minority vote goes to Labour (based on 68% estimate by British Election study for 2010, a poor year for Labour); 16% to Conservatives
  • increases in the popular vote of Labour and Conservative parties in line with the overall rise in population : Labour [(80/63) x 9] 11.4m, Conservative [(80/63 x10.5] 13.3m