THERE WILL BE NO LABOUR LANDSLIDE
David Carlton says forget the opinion polls and concentrate instead on the new volatility of the British electorate
IN the general election of 1997, pundits attempting to forecast the result could consult three principal indicators: opinion polls; recent parliamentary by-election results (pre-eminently the one at Wirral South); and local government elections and by-elections held during the previous year, as meticulously recorded and interpreted by the elections centre at the University of Plymouth headed by Colin Railings and Michael Thrasher. Those who relied exclusively on these three indicators would have been in luck, for all three suggested an overall Labour majority approaching 200 seats or more. Yet almost all the 20 leading pundits invited to join Reuters' panel of experts' took another course, and hence their final average prediction was for an overall Labour majority of no more than 92 seats, whereas the actual outcome was 179. As Lord Lipsey commented, 'The "experts" were inaccurate to a degree that, in a closer election, could have led them to predict the wrong result.'
Even those who were professionally involved in promoting one or other of the three indicators showed little confidence in their own efforts. Most notably, Robert Worcester timidly predicted an overall majority of 101 seats when his own polling organisation, MORI, had consistently been suggesting a double-digit Labour percentage swing from the Conservatives — which on any calculations would have given an overall majority nearer 200 than 100.
Railings and Thrasher were equally unwilling to be guided by the findings of their 'local government' method. This indicated that the Conservatives would receive 31 per cent of the total votes cast and Labour 44 per cent — an underestimation in each case of just 0.4 per cent compared with the outcome. But Railings and Thrasher as individuals made strong downward adjustments, telling Reuters that Labour's overall majority in seats would be either 81 (Railings) or 65 (Thrasher).
I was also a member of the Reuters panel and forecast a hung parliament. In so doing, I of course rendered a signal service to many other members, including Worcester, Railings and Thrasher, whose forecasts thus seemed somewhat less awful. But my starting-point had been simply to dismiss the three indicators. Instead, in a series of articles in The Spectator and the Sunday Telegraph, I relied on what. as I thought, 'history' had to teach us. I argued that a 'meltdown' by a major party was highly improbable in the absence of an organic split such as took place in 1905 or 1981, or of an acute national crisis as in 1931. I pointed out that 'since the second world war 13 general elections [before 1997] saw an average swing of around 2 per cent; the largest swing being 4.7 per cent to the Conservatives in 1970 and the largest swing to Labour being 12 per cent in 1964'. Hence I was essentially uninfluenced by opinion polls, parliamentary by-elections and local government elections, which were all pointing to something radically different.
It is right to add that this approach had served me well in the 1992 general election. For then 'history' was also my guide — with 1846, 1905-06 and 1931 much in my mind. Hence from 1989 onwards, almost alone among pundits, I predicted in the Sunday Telegraph that the still potent consequences of the SDP defection of 29 Labour MPs would suffice to see off Labour and, incidentally, that after another defeat John Smith would replace Neil Kinnock.
The failure of my 'method' in 1997 has, however, persuaded me that I must never again rely on history to trump the three indicators. So does this mean that in 2001 I am predicting a massive Labour victory?
By no means. For the three indicators on this occasion are simply not in harmony. The opinion polls, to be sure, are pointing to another Labour landslide. But it is a different story with the other indicators. The two most recent by-elections, held at Preston and at West Bromwich West on 23 November last, saw swings to the Conservatives from Labour of nearly 9 per cent — probably enough, if repeated nationally in the general election, to deprive Labour of its overall majority. The contrast with the Wirral South by-election, held shortly before the 1997 general election, could not be more striking. There the swing was 17.2 per cent in the other direction — just 2 per cent more than the swing in the same seat in the general election. As for recent localgovernment results, Railings and Thrasher on their website state that these also represent a 9 per cent swing to the Conservatives since the last general election. But they do not apparently expect this swing to be reflected in an early general election. Instead, in February 2001 their working calculation was that there would be only a 4 per cent swing to the Conservatives – leaving Labour with an overall majority of 80. So, as in 1997, they are second-guessing what their own 'method' is telling them. Last time, of course, it was they, and not their 'method', who were wrong.
Given these conflicting signals, which of the three indicators should we favour? My preference over the years has tended to be for the two that are based on what real voters do in real ballot-boxes. For example, in June 1970 I was unsurprised by Edward Heath's general-election victory, which reflected more closely the local government results of the previous month than the opinion polls that had evidently beguiled Harold Wilson into asking for a dissolution. Yet the situation today is very different from that in 1970. For the gap between the opinion polls and what real voters have recently been doing is massively greater and hence must induce some doubts on the part of every thoughtful pundit.
On the one hand, those who incline to rely on real voters have to acknowledge that the turnouts have recently been at such historically low levels that they may have lost any utility. On the other_ hand, those who favour opinion polls need to reflect on what happened last autumn, when at the height of the tanker-drivers crisis Labour's massive lead vanished overnight. And Labour's recovery was almost as rapid. There is, of course, no precedent for volatility on this scale. This may indicate that Labour's lead in the opinion polls is based on extremely shallow levels of commitment and hence on willingness to vote. And it may mean that only eve-of-poll opinion polls are worthy of any attention.
With the three indicators in such unique disarray, I therefore feel justified in relying for my forecast. as in 1997, on the lessons of history. True, the ones I invoked last time let me down. But a new one has taken my fancy, which I shall only abandon if it is invalidated in two successive general elections. It is that the British have become fundamentally volatile in their approach to politics. This transformation has come in two stages. First, in 1962 the Conservatives lost the Orpington by-election to the Liberals with a 25 per cent swing. Nothing remotely comparable had been seen in the postwar era. But after Orpington huge swings in by-elections became commonplace, as did great variations in turnout. By contrast, in the nine general elections after 1962 the electors showed much steadiness and, ultimately, tribal loyalty to their favoured political parties.
In the tenth, however, in 1997, all that changed, with a swing to Labour three times larger than anything that had been experienced in the postwar era. Yet the economy was in much better shape than in 1992 when the swing had been modest indeed. I conclude therefore that, where general elections are concerned, we are now in the equivalent of the post-Orpington era.
So I am accordingly confident that Labour will suffer a massive collapse in support at some point. And maybe it will actually happen in 2001. If so, however, we should not assume that the Conservatives will gain voters on the same scale that Labour loses them. In 1997 the Conservatives lost 4.46 million votes but Labour had a net gain of only 1.95 million — with 2.33 million fewer votes cast in 1997 than in 1992. My expectation for 2001 is that we shall see more of the same. •The largest gain, in short, will be in abstentions, mainly as a result of some two million votes being lost by Labour. In this scenario the Conservatives would gain many seats even if their own total vote remained much the same as in 1997.
Such gains might seem of comparatively little importance if, as I expect, Labour nevertheless for the first time in its history comfortably retains office for a full second term, with, say, an overall majority of 35 seats, thus giving the illusory impression that the dragon of volatility has been slain. But a massive swing from Labour to abstention would be a triumph for volatility all the same and, more importantly, would prepare the way for something even more spectacular on the next occasion.
Dr David Carlton is Lecturer in the Department of Politics and International Studies at the University of Warwick.
Previous page
