Is it worth campaigning in the Summer?
This post originally appeared on Progress Online.
The ‘dog days’ or ‘silly season’ are nearly upon us. The prime minister has had his reshuffle, restoring the sensible pattern of John Major’s premiership that new ministers should be given the summer to read themselves into the new job and know what they are talking about by the time full-scale politics resumes, and that the team presented at conference is the one that will be in office afterwards. The Commons broke up on 23 July, and the Lords did so yesterday. The political year ends with public opinion looking rather like it has done for months on end – Labour consistently but rather narrowly ahead of the Conservatives, Ukip doing very well and the Liberal Democrats embarrassingly badly. The government is modestly unpopular, and people think Labour’s heart is in the right place, but the Conservatives still have their strengths in key issue areas like leadership and the economy. What can we expect to happen to public opinion during the summer lull?
The conventional wisdom is that governments tend to improve their opinion poll ratings during the summer. Warm weather and holidays are supposed to make people feel better about life in general, and the government’s performance benefits from a more indulgent and rosy perspective. More concretely and cynically, politics is often about mishaps and it is easier to look competent when there is nothing much going on and the usual scrutiny mechanisms are disarmed, with parliament in recess and much of the media class camped out in Tuscany along with the politicians.
While plausible in theory, there is no decent evidence that governments actually do improve their fortunes during the summer. Your correspondent, with the invaluable assistance of Mark Pack’s spreadsheet of opinion poll results, has trawled through 50 years of voting intention polls to prove this point.
The table shows the net swing to the main government party from the main opposition party over the summer, based on the differences between their mean voting intention measured in July and September (i.e. approximately covering the mid July to mid September lull).
The mean pro-government swing between the average poll in July and the average poll in September is a measly 0.2 per cent, which given that polling figures are imprecise anyway effectively means that there is no summer effect whatsoever. Looking only at poll figures from the first 15 years of this series (1964-79) it looks as if there might once have been something in the theory, possibly, but the most recent years show small movements with no consistent pattern. Summers, as anyone of my generation will recall, were more idyllic back then, anyway. The apparent end of the pattern might represent change, or it may be that the larger numbers of polls conducted in recent years and more sophisticated methodology, both smooth out sampling variations that caused spurious patterns in earlier years. Recent significant movements are mostly explicable with reference to political events that took place in mid-September, rather than over the summer itself – Black Wednesday (1992) and the fuel crisis (2000) caused downward movements in government party fortunes in the second half of September, while the Lehman collapse accounts for the bulk of the Labour recovery in 2008.
There have been a handful of summers which have involved significant changes in third party voting intentions; in 1973, 1990, 2000 and 2003 their fortunes have risen, while in 1987 there was a drop caused by the acrimonious end of the Alliance.
If we can dismiss the ‘government parties improve over the summer’ hypothesis, how about an alternative; perhaps the summer news vacuum produces a drift back towards traditional allegiances, and a tendency for whichever party is in the lead to lose ground? This also seems plausible, but the evidence is no better than for the alleged pro-government swing. While 29 out of the last 50 summers have shown a pro-government swing, only 28 have regressed towards parity between the parties (17 have seen a party expand its lead, and in four the lead has changed between parties).
Labour is making a number of campaigning moves this summer, including Ed Miliband’s reflections on the trivialisation of politics and the promise of a people’s question time, and stepping up the party’s message on the NHS. On several occasions in the past, parties have tried to campaign over the summer recess with mixed and usually disappointing results. The prototype was probably Jeremy Thorpe’s hovercraft tour along the coast between the 1974 elections, which was a deflating experience. One of the splashiest was the Conservatives’ 1990 ‘Summer Heat on Labour’ campaign led by Kenneth Baker, featuring a wilting rose. While there was a small pro-Tory swing over the summer, there are other factors (such as the crisis over the Iraqi invasion of Kuwait in August) that can explain it. The next summer, 1991, saw Labour attempt to make hay but its message was too diffuse and the news dominated by foreign affairs (the Soviet coup in particular) and some successful image-building by John Major. What had seemed to be momentum in Labour’s favour in May and June ended up dissipating by September.
While summer campaigning might have a disappointing history, and the easiest proposition to argue with the poll figures is certainly that things do not change over the summer, it is not quite possible to say that parties should not bother with campaigning at all between the start of the parliamentary recess and the conference season. One cannot know what might have happened had parties not run summer campaigns at all. In the current circumstances, rather resembling 1991, there are good reasons for not allowing the government to dominate what public attention there is during August. In 1991 Labour tried to cover an agenda including Europe, the environment, the constitution, the economy, social welfare and burnishing Neil Kinnock’s image. In 2014, the party should probably just pick two (NHS and Ed Miliband’s image seem like the most useful possibilities). It is also strategically desirable to enter the conference season with the dominant message of the polls being ‘Labour has stubborn opinion poll lead’ rather than ‘Tories overtake’. Social media and internet access to news sources may also mean that summer campaigning is more effective than in the past. Labour should not expect miracles from the summer campaign; it is probably a good time for message testing and preparing the ground rather than making dramatic pronouncements.