By the morning of May 8 2015, the British public had returned a majority Conservative government for the first time in 23 years. No-one saw it coming. A debate has since raged on the accuracy of the polls, and will no doubt continue (see the British Polling Council enquiry here). Just for the record, once the margin of error was taken into account, the polls were not that
bad. More to the point, on the election day itself, there would have been a multitude of relatively hidden dynamics that took the edge off the accuracy of the polls. As the British public have increasingly distanced themselves from the two party system and indeed from political parties themselves (Stoneman 2008), political scientists are under greater pressure to refine their vote choice forecast models to adapt to a more ‘volatile’ British electorate (see Whiteley and Clarke’s interesting piece in The Conversation here). Now, justification cannot be given to the complexity of such models here, but what can be offered is a plausible hidden dynamic that might have just caught the forecasters off-guard. To keep things simple, in what follows I will focus on English votes, as Scotland and to a lesser degree Wales had their own unique electoral dynamics to deal with.
“It was the right thing to do”
While UKIP were the biggest winners in terms of increase in overall share of the vote (not seats), up nearly 10% on their 2010 performance, the evidence suggests that there would not be a uniform swing from Conservative to UKIP, as plenty of ‘old’ Labour voters were switching support to Farage’s army too. UKIP, it would seem, would put a break on both of the main parties surging ahead. The Liberal Democrats, on the other hand, were torn to bits. By losing 16% of the share of the vote in England, a real question mark surrounds the Liberal Democrat voters of 2010. After successfully building its reputation as a coherent centre-left ‘progressive’ party, and after the giddy heights of Cleggmania and 20%+ poll ratings in 2010, by the morning of 8 May 2015 an army of 6 million Liberal Democrats in England were suddenly massacred to 2 million.
The over-riding dynamic that sent English voters into unchartered waters, then, was the sudden appearance of 4 million confused and weary Lib Dem voters. These poor souls had their political faces dragged through 5 years of a centre-right coalition government. “It was the right thing to do. It was the RIGHT thing to do”. You have to keep saying it to believe it. So, stumbling over the English hillsides they come. 4 million Liberal Democrats looking for political salvation. What isn’t clear just yet is the degree to which these poor souls actually voted or transferred their vote elsewhere (we would need estimates from panel data for that, which alas I have not seen yet). And if they did stray elsewhere, surely the next logical step is to vote Labour, right? This would have happened if the ‘word on the street explanation’ was true, that is, the Liberal Democrats were bashed for abandoning their centre-left position. But many constituencies demonstrated large swings in the vote from the Liberal Democrats to the Conservative and to UKIP, more so than expected. If an angry mob of predominantly centre-left voters gave the party a bashing for not being ideologically straight with them, then some of them went about demonstrating their anger in a rather odd way.
An incorrect assumption about the relationship between political parties and voters is that voters simply choose their preferred parties based on exogenously defined preferences (by exogenous, I mean preferences that are developed outside of and beyond the party system itself). The relationship between the two is more ‘endogenous’ than that (a fancy way of saying they are somehow necessarily related or bound together). This captures a political phenomenon called ‘reciprocal voting’ (see for example Page and Jones 1979). This idea applied to the 2015 election is that in the past 5 years, something has changed about the average Liberal Democrat voter, and that this change was caused by the party moving into a centre-right coalition. On this reading, political parties can ‘drag’ their voters into different ways of thinking about politics. This might be due to factors such as shifting the political agenda to make other topics more salient, or by voters ‘adapting’ their rationality to justify their previous vote choice.
One of the observations that the mechanism of reciprocal voting could seemingly help to explain are the constituencies where there were larger than expected swings of the vote from Liberal Democrats to the Conservative Party or UKIP. Quite a few exist in London and the South-West. There are two factors which would have made Liberal Democrats more susceptible to such a shift:
(1) Liberal Democrats have relatively weak party identification compared to Labour and Conservative voters. This lack of strength in party identification makes them more inclined to switch voting preferences and think, “sure, why not give them a go”. They’re liberal, after all.
(2) The second factor relates to the direction in which Liberal Democrats might swing. While associated with having centre-left leanings (indicated by the “Democrat” part of the name, derived from ‘Social Democrat’), instinctively some party supporters are very much traditional liberals in their belief in the individual and a healthy scepticism of the state. Indeed, there is now talk to rename the party back to ‘The Liberal Party’. This meant that Clegg et al. between 2010-2015 were not out of place in espousing a centre-right policy programme. In many ways, especially with the coalition’s drive towards localism, there were many classic ‘liberal’ ideas being implemented.
These two factors combined meant that Liberal Democrats were always going to be less loyal in their support, and perhaps more forgiving to the instincts of parties like the Conservatives and UKIP with their emphasis on a smaller state and individual economic freedom. Add in the possibility of reciprocal voting, and it becomes a little clearer how the Lib Dem vote, having collapsed spectacularly, was transferred to parties who were ostensibly very different in ideological terms. Unintentionally, by going into a centre-right coalition, the Liberal Democrat leadership could well have awoken the old liberal spirit of some of its voters.
 All good quality random probability surveys produce estimates of the ‘true population value’, that is, the number we are trying to measure – in this instance, percentages of voting intentions for political parties. Repeating the same survey will provide different estimates of this value, which we call ‘random sampling variation’. The ‘margin of error’ covers the probable range of estimates of the true population value.
Please note: Blog entries reflect the personal views of contributors and are not moderated or edited before publication. However, we may make subsequent amendments to correct errors or inaccuracies