Public opinion, at least in terms of voting intention, has been unusually stable during this parliament. There has barely been any change in the main parties’ ratings since the middle of 2013, when Labour’s lead over the Conservatives narrowed and Ukip support subsided a bit from its peak just after the county council elections in May. But for the Liberal Democrats the line has been completely flat since the end of 2010 when tuition fees alienated many of their remaining left-of-centre supporters who had stayed despite the formation of the coalition. The polls at the start of 2014 have been completely consistent with the pattern of the last three years – Liberal Democrat support is between eight and 13 per cent. There are some ‘house effects’ from each pollster: Populus is consistently the most generous and the lowest numbers tend to come from YouGov. There appears little that the scandals over Mike Hancock and Chris Rennard can do – so far – to drive their ratings any lower.
Every month, a detailed analysis of polling trends appears on the Nottingham University website – the work of Rob Ford and his co-authors at Polling Observatory. Every month, Liberal Democrats – including reality-based ones such as Stephen Tall – object to how the Polling Observatory numbers treat their party. For instance, Liberal Democrat support in the Nottingham data series for January 2014 is recorded at 7.78 per cent, which is less than any opinion poll has found. The reason for this is that in adjusting for house effects and variations in sample size and design, the 2010 election is the baseline:
To estimate ‘house effects’, we make use of the 2010 election result – our model treats the 2010 result as a reference point for judging the accuracy of pollsters, and adjusts the poll figures to reflect the estimated biases in the pollsters figures based on this reference point.
As Dennis Kavanagh and Philip Cowley note in the Nuffield study of the 2010 general election, polling error was highly unusual when compared to its predecessors. The final polls of an election campaign usually overestimate the Labour share of the vote (with particularly bleak effects in 1992) but in 2010 Labour did better than the final polls. All the pollsters overestimated the Liberal Democrat share, too. YouGov, despite producing low figures for the Lib Dems now, overstated the Liberal Democrats by more than average in May 2010. The first televised party leader debates and the brief eruption of ‘Cleggmania’ created a particularly unstable polling environment during the 2010 campaign even if the eventual result (and an exceptionally accurate exit poll) suggested that opinion ended up pretty much where an orthodox election campaign would have left it.
In most elections the Liberal Democrats increase their support during the short campaign period. But this is not a law of nature. A major part of the increase in support comes from the party being given more media coverage rather than being marginalised, and some also comes from tactical voters being mobilised in constituency campaigning. Being in government may well mean that the Liberal Democrats do not gain this campaign boost – the last time that it failed to appear was 1987, following a parliament in which there was much discussion of three-party politics and favourable coverage of the Alliance.
The Nottingham average, to this observer, does seem unduly bearish about the Liberal Democrat share of the vote; adjusting it down because of the way polls behaved in 2010 may be building factors that were peculiar to that election into a longer-term model. They may not – we shall see.
It is always troublesome to translate Liberal Democrat votes into seats. The party can have elections where it gains seats despite losing votes (as in 1997) and vice versa (as in 2010). But, even if the party struggles back up to 15-17 per cent, there is no way it will not result in a loss of seats. Local election results suggest that it will be near-impossible for the party to hold a number of metropolitan seats against Labour (Manchester Withington being the most drastic example).
Scotland is also a huge problem. Although their polling support is down by the same 12 points or so there as in England, they were already at rock bottom in half of Scotland in 2010 (losing 12 points would give them negative votes in 30 Scottish seats). The drop must be concentrated in the seats where they polled reasonably well in 2010. They are in serious danger of losing up to 10 of their 11 Scottish seats.
The crucial thing for the Liberal Democrats is whether differentiating themselves from the Conservatives, the visceral anti-Toryism of progressive voters, and Ukip, can save the suburban and south-western seats they hold against the Tories. Last year I ventured that the Liberal Democrats would get 35-40 seats. That still feels about right, but a bit of a closer look at the history and at Scotland makes me think that 35 is more likely than 40.
Lewis Baston is senior research fellow at Democratic Audit and a contributing editor to Progress. He has profiled all Frontline 40 seats and reviews the polls each month here as part of the Campaign for a Labour Majority. He tweets @lewis_baston
Progressive centre-ground Labour politics does not come for free.
Our work depends on you.