By Jane Green and Will Jennings. Jane Green is Professor of Political Science at University of Manchester (Academia.edu, Twitter) and Will Jennings is Professor of Political Science and Public Policy at University of Southampton (Academia.edu, Twitter). You can read more posts by Will Jennings here.
Cross-posted at the APSA British Politics Group Blog.
It has been described as one of the most important unanswered questions in political science: why governments regularly and predictably lose popular support over their time in office. Such is the regularity of ‘costs of governing’ that it appears that governments are simply ‘passive observers’ of their diminishing support, leading to the suspicion that these trends may be almost wholly independent of the performance of a government in office. The implication of costs of governing is stark: it seems to matter little what a government does in office, its decline in popular support is all but guaranteed. Here we consider the implications of costs of governing for the 2015 British general election and summarise our answer to the question of why governments experience these all-important governing costs.
The Implications of Governing Costs for 2015
The Conservatives began their period of government without a majority. What this means, of course, is that the party has to increase its popular support between 2010 and 2015 (and how that is translated into seats) to have a chance of winning a majority in 2015. David Cameron has to buck the ‘costs of governing’ trend if he is to win back support before 2015.
That isn’t looking likely. Vote intentions towards the Conservatives since June 2010 have followed the predictable pattern of governing costs that we identify in all countries for which regular polling data are available. The following two figures show (a) the decline in vote intention for the Conservatives since June 2010 (the average of all available polls for each month), and (b) the decline and curve that best fits the data for governing party support across 79 government lifecycles in 31 countries. The first figure plots vote intention for the Conservatives over the course of this parliament by month, the second plots vote intention over often much longer time periods by year.
(a) Conservative Party vote intention June 2010 – August 2014
(b) Governing party vote intention (79 governing periods, 31 countries)
The high level of support (or honeymoon) at the start of Conservative-Lib Dem government in 2010, and the loss of support thereafter, is consistent with the trend we find exhibited in the largest collection of cross-national over-time poll data it is currently possible to analyse.
UKIP are the beneficiaries
What is striking to us is that the only pattern in the last four years which doesn’t entirely fit our conventional expectations is the following. Whereas we would usually expect the major party of opposition to be the beneficiary of declining trust and support for the government, it is UKIP rather than Labour that appears to be capitalising on the costs of governing for the Conservatives (the Liberal Democrats’ support collapsed early in the parliament and has been flat lining around 10% or less since). The following figures display (c) vote intentions for Labour between June 2010 and August 2014, and (d) vote intentions to UKIP in the same period (again taking the average of all polls in both cases). We can see that although Labour received a boost to its support in the first 2-3 years of the Conservative-Lib Dem coalition – a pattern we would have predicted – it has steadily lost that support since 2013. UKIP, by contrast, has witnessed a significant gain in popular support over the same period.
(c) Labour Party vote intention June 2010 – August 2014
(d) UKIP vote intention June 2010 – August 2014
A reading of recent British public opinion data may be interpreted simply that Conservative voters are moving over to UKIP due to UKIP’s policy and rhetorical appeal, and latterly Labour voters too. But looking at these data in the context of cross-national and over-time trends in costs of governing suggest something more profound may be happening. The Conservative Party should have been expected to lose its support. That support could have gone to the Lib Dems, to Labour, to UKIP or to being undecided. It is a signal of the distrust in mainstream politics that the predictable costs of governing have resulted in rewards to UKIP. Labour would have been the beneficiaries under usual expectations but on the face of public opinion alone, the trends point to UKIP as the classic party of opposition. This is in a context whereby the Liberal Democrats cannot pick up those opposition party spoils. The anti-politics mood in Britain may be fundamentally shifting the winners and losers of some of the most important and conventional trends we are aware of in political science.
The Conservative Party may experience an uptick in support as we near the 2015 election. The tendency of some incumbent parties to experience an uptick can be seen in the modest U-shape curve in Figure (b) above. But any uptick to the Conservatives won’t reverse the fundamental trends that we highlight above.
Explaining Costs of Governing
Costs of governing are surprisingly poorly understood, despite their prevalence and their profound implications. The reason for this has been an absence of data on public perceptions of party and government performance. Our recent paper for the annual conference of the Elections, Public Opinion and Parties specialist group of the Political Studies Association sets out new explanations for the decline in governing party support using a unique data set we have collated on subjective performance evaluations of governing parties by British, American, Canadian, Australian and German voters. This draws on over 10,000 individual survey questions asked over as many as 65 years (a measure we call ‘macro-competence’). For more information see http://www.competence-politics.co.uk.
The first explanation for costs of governing concerns the initial honeymoon period; the high from which governing costs occur. We find that the early period of a new government is characterised by blame to the government’s predecessor; an effect that lasts around one typical election cycle (of 4-5 years). This means that Gordon Brown’s government will have been blamed for the first years of the new Conservative-Liberal Democrat coalition, with that effect unlikely to persist into and beyond 2015.
The second explanation concerns how blame accumulates to the incumbent government. As a government continues its period in office, blame begins to stick, and the effects of negative information stick more than positive information. As governments are seen as performing badly, we show that this has a significantly greater effect on vote intention than positive changes in perceived government performance or competence. This negative information accumulates over a government’s time in office. Mistakes, policy disasters and scandals remain in the minds of voters long after politicians have moved on. In our paper we reveal that the addition of a new negative change in governing party competence, and another new negative change, each has a unique effect. The final innovative theoretical (and evidence-based) expectation is that there is a saturation point in the effect of competence evaluations. Negative competence effects begin to weaken after ‘shocks’ accumulate above a certain level, as voters make up their mind that a government cannot be trusted – and their attitudes become fixed in stone. In the case of the present Conservative-Liberal Democrat coalition, our analysis explains how information about bad performance has been weighed more heavily than information about good, and how this information accumulates until the electorate no longer has trust in the government to deliver on its objectives. This is consistent with the gradual decline in Conservative Party support displayed above between June 2010 and August 2014. It is also notable that the costs of governing have happened for the Conservatives very quickly in relation to the amount of time they have actually served in office. The early ‘omni-shambles’ and the unpopularity of austerity measures may well have contributed to this, as well as their relatively low starting point at from May 2010.
 Here we paraphrase the observations of Stimson (1976) in his analysis of declines in presidential approval.