by Stephen Fisher

Janan Ganesh in today’s FT is right to point to trends in public sector employment as an import factor in understanding the political future of Britain. He notes that, “The public sector headcount has fallen by about 1m since the last election to 5.4m, which is where it was before Labour’s expansion commenced in 2000. […] Meanwhile, private-sector employment has risen and the self-employed account for a record 15 per cent of the labour force. The confluence of fiscal policy and economic trends is creating a looser, more individualised economy. This does not guarantee that the electorate will become less receptive to collectivist ideas, but it requires no feat of imagination to see how it might.”

As well as the electorate as a whole becoming less collectivist with a smaller state there is also a more simple mechanism at play. For obvious reasons, private sector workers have long been more likely to vote Conservative and public sector workers more Labour. The more of the former and the fewer there are of the latter the easier it will be for the Tories to win elections.

The gap between public and private sector workers grew in 2010 as it became clearer that public spending cuts would be larger under a Conservative than under a Labour government. In our analysis of the constituency results, John Curtice, Robert Ford and I noted, ” Constituencies containing a relatively large number of people working in the public sector proved to be relatively loyal to Labour. In constituencies where over 28% of those aged 16-74 who are economically active are employed in public administration, education or health, Labour’s vote fell on average by 4.7 points. In contrast, where that proportion was less than 22%, the party’s vote fell by no less than 8.6 points. As we might have anticipated, constituencies in which unemployment had increased most are typically those in which a relatively small proportion work in the public sector, but even if we take that into account, it is still clear that Labour’s vote fell less where the public sector plays a larger role in the local economy.15 As a result of this pattern a new political gap seems to have opened up in Britain’s electoral geography. In 2005 Labour’s vote was on average almost as high (37.2%) in seats with relatively low proportions of public sector workers as it was in seats with a relatively large proportion (37.5%). There is now a four point difference in the level of support Labour secures in the two kinds of constituency.

The public-private electoral divide looks like it will be similar if not wider this May. Among respondents to the latest (autumn 2014) wave of the British Election Study (BES), including those no longer working, Labour had a 13 point lead over the Conservatives in vote intention among public sector workers, but the Tories had a 2 point lead over Labour among employees of private sector organisations, and an 11 point lead among the self-employed. Overall there is a 17 point difference in the Con-Lab lead between the public sector workers (-13) and those employees and self-employed in the private sector (+4).

The equivalent figures for the 2010 vote among the same respondents were a 2 point Labour lead among public sector workers, and a 13 point Conservative lead among private sector workers, making a 15 point difference. This figure is not statistically significantly different than the 2014 figures

The public and private sector it is bigger than that for many more often talked of social cleavages, such as gender, generation or religion. In the same BES data the Conservative lead over Labour among men was 2 points while that among women was -6, a gender gap of 8 points. Similarly the Con-Lab lead among those over 65 was 4 points but -6 among those under 30; a difference of 10. Also there was a 2 point Con-Lab lead among Christians while the equivalent figure for those of no religion was -8; again a 10 point difference.

The public-private divide now seems to be of similar magnitude to the social class divide which used to be the famous bedrock of British politics. According to the latest YouGov poll, the Conservative lead over Labour is at 5 points among those from social grades A, B and C1, and -11 for those in C2, D and E groups. The difference of 16 points is close to that of 17 points quoted above for the corresponding pubic-private gap. Analysis with the NS-SEC classification in the BES data tells a similar story. But naturally the gap between the very top and bottom groups of a detailed class schema are much bigger than those between middle and working classes broadly defined.

The public-private divide is still however surpassed by the regional and ethnic divides. The Con-Lab lead among BES respondents in the four southern most regions of England was 6 points, but that in the northern three regions of England was -22, a difference of 28 points. But widest of all is the ethnic divide where the Labour lead over the Conservatives was a massive 38 points greater among non-white voters compared with whites.

Of all the social divisions in electoral behaviour, the public-private divide is perhaps least discussed given its magnitude. This especially surprising given it is politically malleable and pertinent to the foremost policy issue of the day: fiscal austerity.

Thatcher tried to increase the social base for the Conservatives with privatisation and council house sales designed to widen share and home ownership. But as Anthony Heath and his colleagues argued in Understanding Political Change the success of her project was muted because shares in newly privatised industries and council houses were both bought disproportionately by those already more ideologically inclined towards the Conservatives. There is unlikely to be such a strong ideological bias among those who have left a public sector job (or might otherwise have gained one had the opportunity been there). So perhaps the Osborne and Cameron programme of reducing expenditure on public services will have more success in expanding the pool of “natural Tory voters” and further weakening Labour’s social base.