The benefits of disreputable psephological habits
This post originally appeared on Progress Online.
It is time to confess one of my disreputable habits. Every time local elections come round, I cannot resist sitting down by my computer and adding up the ward-by-ward results by constituency and working out who ‘won’, at least in the marginal seats that make the difference. The Fabian Society has published this, for the Labour target seats, in 2013 and 2014. The overall picture was quite similar in both elections – Labour making enough progress to put the party just short of an overall majority, but with the United Kingdom Independence party making strong gains in votes (and, more rarely, in seats) in areas where Labour depends on a white working-class and lower middle-class vote.
The sceptical reader may ask how relevant local election results are to what might happen in a general election, and the sceptical reader would have a point. Not all elections are truly comparable with each other, and I would be very wary of using European parliament election results to say anything about other elections. But electors, and the parties, seem to treat local elections a bit more seriously, and with careful handling they are a useful indicator.
But how, one may ask, can one read from a local election with 30-35 per cent turnout to a general election with 60-65 per cent turnout? The pragmatic answer is that it seems to work. The overall picture is, like a big opinion poll, a snapshot rather than a prediction. In both 2013 and 2014, the message was that Labour had a small overall lead in public opinion, while in 2012 the Labour lead was larger and in 2011 the two main parties were basically tied with each other. In each case, this confirms the broad message of national opinion polls, suggesting that general election voting intention and observed behaviour in local elections are not very different.
However, the turnout factor can be very important. In the May 1992 elections, when Labour supporters were demoralised by the loss of the general election the previous month, the Conservatives did extremely well because of differential turnout. Under Labour governments, it seems consistently harder to get Labour supporters to turn out in local elections, even in good years like 1966 and 1998. Local election voters are older and whiter than the average. Smaller parties do better in low-turnout elections because their supporters are usually more motivated to turn out than the more apathetic habitual voters of the main parties. The general election turnout in 2010, for instance, caused a number of Labour gains from the Greens in the local elections through higher turnout. Higher turnout will probably also mean that the Ukip share of the vote is much lower in 2015, even if (unlikely) none of their 2014 voters shifts allegiance, because the local electorate will be diluted by the less angry and pessimistic voters who come out only for general elections.
Some people do also vote differently in local elections. This is often reflected in the Liberal Democrats doing better in local elections because their campaigning techniques are well adapted to that environment; even people who dislike the party’s national stance sometimes think their councillors are particularly hard working. Watford is probably the most outstanding example of this. Independents also prosper more in local elections and their effect on the major parties’ votes can be wildly different in different localities. There are some parts of the country that, even on the same day as in 2010, vote differently between the two main parties in local and general elections. This may be because of personal votes for incumbents at local and national level (hence probably the vote in Birmingham Edgbaston for Gisela Stuart as member of parliament and the Conservatives locally in 2010), or different attitudes to council and national government. For some reason, Labour seems to overperform in council elections in Stevenage and underperform in Wirral, and there are examples scattered across the country to confuse the analyst. There are a few areas where running the council is a poisoned chalice and doing well in local elections sets a party up for defeat in subsequent elections; Torbay and Plymouth have behaved this way in the recent past.
There are also some more technical reasons why caution may be required. Parties may not stand full slates of candidates, and may be adversely affected by local splits and independent candidates. Even if the Liberal Democrats do not put up any candidates in local elections – as they did not in any ward in Tameside, for instance – they will manage to scrape together sufficient nominations to stand a parliamentary candidate. Local elections in many areas are frustratingly irregular, with some wards not up for election every time (Bristol has its own uniquely baffling electoral cycle), and to get a full estimate for a constituency one needs to model what would have happened in the missing wards. Sometimes wards are split between constituencies, and allocating bits to each side is guesswork (for instance, whether Labour ‘won’ Swindon North in 2012 depends on allocating votes in a split ward). It is sometimes impossible to call a result. While Harlow district council voted Labour, with Ukip a close second, the constituency includes a small rural area which did not have elections and which would have made the eventual result ‘too close to call’ between Conservative, Labour and Ukip.
Local elections are therefore far from infallible as guides to what might happen in general elections, even a year out from the poll (much to Labour’s relief in 2009-10). But the pattern is meaningful. Apparently bizarre Labour gains in the 1995 local elections, like Hove and Castle Point, foreshadowed parliamentary gains in 1997. The Conservatives’ strong results in marginals such as South Ribble, South Derbyshire and North-west Leicestershire in 2005-09 were also indicative of future parliamentary gains. While one should never despair or be complacent, it is clear that the path to a Labour victory in Enfield North (7.6 per cent swing to Labour since 2010) looks smoother than the path to winning Gloucester (0.2 per cent swing to Conservative). The local elections are also valuable guides to local campaigning strategy – often in 2014 the results and turnout in middle-class marginal wards were good, while low turnout and Ukip voting were prevalent in ‘council estate’ wards. Good campaigning between now and May 2015 can make a lot of difference.