Why popularity impacts upon campaign effects: Evidence from the 2015 General Election.By Justin Fisher on 19 June 2018
Over the last twenty-five years or so, there has been repeated evidence that shows that the campaign efforts of political parties at constituency level lead to electoral payoffs. Broadly speaking, the more intense their campaigns, the better parties perform. This is true not only in Britain, but also in many other democracies. But, we also know that the effects are not always the same across all parties in every election. This has typically been explained as a function of resource distribution. If parties put most effort into target seats rather than safe ones, or ones where they have very little chance of success, they should perform better electorally. However, the context can vary from election to election. In other words, parties will not be equally popular in every election. What happens, for example, if a party finds itself being unusually unpopular or popular relative to the previous election? Such circumstances are outside the control of campaign managers, but can have a dramatic effect of the effectiveness of campaigns.
In a new article by David Cutts, Edward Fieldhouse, Bettina Rottweiler and myself, we test the impact of electoral context using the principle of what we call popularity equilibrium. The idea behind this is straightforward – campaigns will tend to deliver more electoral benefits where parties are not unusually popular or unpopular. If parties are unusually unpopular, the impact of their campaigns is less likely to be decisive (since voters will be less receptive) and equally, for more popular parties, the campaign is less likely to impact on voters’ decisions, since they will already have decided to vote for them.
The 2015 British general election provides an ideal way to test this idea. While the popularity of the Conservative and Labour parties remained broadly unchanged between 2010 and 2015, that of the Liberal Democrats changed dramatically – not only did the party’s poll ratings fall significantly, it also performed consistently badly in second-order elections. Under these conditions, we might therefore expect the Conservative and Labour campaigns to have been effective, but that those of the Liberal Democrats would be less so, since as an unpopular party, it would have particular difficulty in promoting its message to non-receptive voters. We test this idea in two ways – first by assessing the overall impact of the parties’ constituency-level campaigns and second by looking at whether campaign effects differed depending on the identity of the rival party.
As expected, Conservative and Labour campaigns both produced positive effects. Had the Conservatives not campaigned as well as they had, it would have cost the party at least an estimated nine seats – enough to deny the Conservatives a majority. Without Labour’s efforts, the estimated loss would have been an estimated twelve seats. For the Liberal Democrats, their campaign efforts were equally successful, however – boosting their share of the electorate by 1.8%, though not enough to save any seats. How do we explain this? In part, Liberal Democrat campaigns delivered payoffs because the party focused its resources not on marginal seats (which it knew it would lose on account of its national-level unpopularity) but on safe ones. This might suggest our arguments about popularity equilibrium were flawed. However, our theory gains more credence when we look at the impact of rival parties. Here, we find that Conservative and Labour campaigns were much more effective against Liberal Democrat incumbents than against each other. Meanwhile, Liberal Democrat campaigns were most effective in the party’s own seats. In seats held by other parties, their campaigns were notably less effective.
Thus far, we can say that popularity did have some impact on the Liberal Democrat campaigns overall. Indeed, the Liberal Democrats might have done even worse if the campaigns hadn’t been so focussed on the party’s safer seats. But we can also assess if the principle of the popularity equilibrium works at the constituency level, since this will vary for all parties across the country. Using the British Election Study, we can assess the parties’ levels of popularity in every constituency just before the campaigns commenced. What we find is that popularity did indeed impact upon campaign success across the board – campaigns were less effective where parties were less popular and where they were particular popular. In effect, there was an optimal level of popularity where parties’ campaigns were most effective.
Overall, these analyses show that in 2015, party popularity was a key influence on the electoral impact of campaigns at both national and constituency levels. Where parties were unusually popular or unpopular, their campaigns tended to be less electorally effective. Of course, in this and other elections, parties are not blind to their own levels of popularity and seek to compensate these effects if they can. The principal way in which parties can try and offset the impact of a change in equilibrium (when they become significantly less popular) is through the effective distribution of campaign resources. The Liberal Democrats were able to do this to an extent and remarkably, did achieve some electoral success. All of which neatly illustrates that while campaign effects are important, their impact may be contingent on the electoral context and parties’ response to it.
The full article 'The Impact of Electoral Context on the Electoral Effectivness of District-Level Campaigning: Popularity Equilibirium and the Case of the 2015 British General Election' is available to read now in Political Studies.
Justin Fisher is a Professor of Political Science and Director of the Magna Carta Institute at Brunel University London. He is Principal Investigator of the ESRC-funded project Constituency Campaigning at the 2015 and 2017 British General Elections and is lead editor of The Routledge Handbook of Elections, Voting Behaviour and Public Opinion. He tweets at @justin_t_fisher.
Image by David Holt