Thursday 17 April 2014

Can past elections help us predict the next?


Historians often don't much care for 'learning from the past'. As A.J.P. Taylor once said: all politicians learn from history is how to make new mistakes.

But as the next UK General Election hoves into view, just a year or so away now, we've got so much data to hand that it's impossible not to look at previous contests for any hint that they might help predict the outcome.

In the relatively long-run past, as we've said before, there's definitely a tendency for polling numbers to shift back towards the incumbent government as polling day approaches. Between 1950 and 1970, that equated to a shift of between five and six points towards the governing party in the year before any General Election. But there were wide disparities in that figure, and some governments in the past (Clement Attlee's in 1950, or Jim Callaghan's in 1979) actually saw the electoral tide go out on them. You can cut it a number of other ways, and one blog has recently said that we should expect the gap between government and opposition to halve, and then add 3.5 percentage points to the governing party's score. Use either of these methods, and you're projecting a complicated result: the Conservatives winning most votes, but Labour winning most seats in the House of Commons. You can slice the numbers up differently and push the Conservatives up a bit, so that they might continue as a minority administration, or in a relatively unhappy coalition with the Liberal Democrats. But the lesson is clear: expect the Conservatives to win most votes, but that the fight for seats to be quite a close-run thing.

Not that time is the only variable we should look at. News that real incomes have started to rise, as unemployment recedes, will cheer politics' blue camp no end. Rising wages are usually a harbinger of re-election, as they were for President Obama in 2012. They had fallen on his watch, but they began to catch up with prices just in the nick of time for the Democrats. Just like David Cameron's long-hoped-for living standards recovery taking place in the UK now, US incomes started to rise about twelve months or so before the election (opens as PDF - see page ten); the President won a convincing, if slightly narrow, victory.

We need to be very cautious about all this. Not every election is like others. The data on the various parties' vote shares is now completely different, and counted in a totally different manner, than it was throughout most of the post-war period. Even if we do draw direct comparisons with past battles, 'swingback' towards the Government may be greatly limited by the fact that most of the Conservatives' problems are not with voters defecting to Labour, giving them a net gain of plus two every time one returns, but to the United Kingdom Independence Party, which might not win a single seat in 2015. Given Britain's productivity problems and deep-seated changes in the workforce, wage increases are also likely to be pretty weak over the next year - at one per cent or so, even if we take the Government's preferred Consumer Price Index at face value. That's a lot slower than the real household income gains, of maybe three or four per cent, made in the US across late 2011 and in 2012.

The lessons from contemporary history are clear: the battleground favours the re-election of Mr Cameron (above), even if that is likely to be as a continuing minority Prime Minister. A combination of forced choice between the Government and the Opposition, and rising real wages, both suggest as much. But a deeper dive into the numbers shows that the contest might be very, very tight indeed - coming down to just a handful of seats. Hold onto your electoral hats.