Everything is in Blair's favour

With an election announcement expected within days, Ivor Crewe argues that all Labour's fears about

"It's in the bag, Tony!" exclaimed the Sun, the day after the Budget, simultaneously announcing its endorsement of Labour and pronouncing Labour's victory as certain. With uncharacteristic modesty, the paper left it unclear whether Labour owed its unstoppable election triumph to the Sun's support or vice versa. But unstoppable, assumes the Sun and just about everybody else, it is.

The only exceptions are William Hague, who has to keep his spirits up, and leading Labour lights, who conjure up nightmares of misleading polls, mass abstention in the Labour heartlands, rural revolts and Son of Fuel Tax protests. "Remember 1970", warned Ken Livingstone on the BBC's Question Time. Nineteen seventy? Tony Blair was an oblivious 17-year-old at that election, but Labour veterans still shudder at the memory. In June 1970, all but one of the final pre-election polls pointed to a secure Labour win; but a late swing propelled the plodding Edward Heath into No 10.

The Sun is right and Livingstone is wrong. There is no parallel between 1970 and now. Harold Wilson's 1966-70 Labour government presided over devaluation, austerity budgets, a humiliating retreat on trade union reform and the loss of 12 by-elections to the Conservatives. Blair's 1997-2001 government has presided over the strongest economy since the mid-1950s and held every by-election it has defended. Wilson sprang his snap election when Labour had only pulled ahead in the polls the previous month, having trailed a long way behind for the preceding three years. Blair can call an election in the knowledge that Labour has enjoyed double-digit leads in the polls for more than eight years, except for six weeks after last autumn's fuel tax revolt.

It is true that the polls are almost certainly exaggerating the true Labour lead. The final polls overestimated the actual Labour vote by an average of 2.7 per cent in 1987, 4.2 per cent in 1992 and, despite the pollsters' elaborate adjustments, 3.6 per cent in 1997. They underestimated the actual Conservative vote by 1.1 per cent, 4.5 per cent and 0.9 per cent, respectively. The various contributory factors - sampling bias, "secret" Tories hidden among the "don't knows" and "refusers", differential abstention - will probably recur. Differential abstention may erode the Labour vote even more this time. But even if inflation of the Labour lead is at the higher end of the historical range - say, 8 per cent - Labour remains at least as far ahead of the Conservatives as in 1997.

And that is the pessimistic analysis. It ignores two compelling reasons for believing that Labour's true position is even stronger. First, the foundations underpinning Labour's poll lead are much sturdier than they were just before the 1997 election (see table). Labour's advantage over the Conservatives in the voters' eyes, whether in terms of leadership, the economy, or the main issues, is decidedly greater now than four years ago. In February 1997, Blair was preferred to Major as PM by an 11-point margin; that has now stretched to 25 points over Hague. Then, Labour was nine points ahead as the party of economic competence; under the most popular Chancellor since the war that lead has widened to 19 points. In every policy area (except, notably, health, where disillusion has set in) the voters have the same faith in Labour as in the Conservatives - or more - as just before the last election.

Secondly, while the polls may be overstating Labour's vote advantage, perhaps by more than in 1997, they are understating Labour's seat advantage, and again by more than they did in 1997. There was already a strong pro-Labour bias in the electoral system in 1997. Labour MPs were elected by a much smaller vote, on average, than Conservative MPs because they typically represent smaller constituencies with a lower turnout. Such was the bias that if, against all expectations, the Conservatives completely close the gap in the coming election, and the two parties tie at 38 per cent apiece, Labour would be still be left with 80 more seats than the Conservatives and a workable overall majority of 17. But the pro-Labour bias will probably be even more pronounced this time. Constituency boundaries have not been changed, but continuing migration will have transferred Labour voters from safe Labour seats in the urban north to marginals in the suburban and small-town south.

Moreover, any national swing back to the Conservatives is likely to vary in three ways that help Labour. The first is the pattern of differential abstention. Current polls and past elections indicate that it will be most pronounced in safe Labour seats and least marked in the marginals. In 1997, the turnout fell least in Labour's target seats, where both main parties concentrated their electioneering. This time, campaign resources will be even more ruthlessly focused on the marginals.

Labour will also be helped by the regional distribution of its marginals. The regions with the largest pro-Labour swings since the 1997 election, according to MORI (and corroborated by local by-elections), are the prosperous new-economy areas below the Humber-Severn line - the East Midlands, East Anglia, south-east, London and south-west (see table). These regions account for 48 per cent of British constituencies, but 65 per cent of Labour- Conservative marginals. If the Labour vote haemorrhages disproportionately in its heartland regions, the consequential loss of seats would be relatively small.

The third reason is the "new-incumbent" factor, first identified by David Butler and Michael Steed in the 1979 election. They noticed that the swing against the Labour MPs first elected at the previous election was consistently smaller than the regional average. The personal vote built up through constituency service and personal contacts by the previous Conservative MPs had disappeared, and been replaced by the personal vote garnered by the fresh Labour incumbent. In marginal seats, where MPs have the strongest incentive to nurse their constituents, the new-incumbent factor is typically worth between 1,500 and 2,000 votes.

Thanks to the Labour landslide of 1997, there will be more new incumbents defending their seats than at any time since 1950. Of the 57 Labour-Conservative marginals with a majority of less than 10 per cent, 44 were won against a Conservative MP seeking re-election - typically an MP who had had three terms or more to build up a personal vote. Labour's class of 1997, with little to occupy it at Westminster, has had plenty of time to dig in locally. Labour should perform better in its marginals than elsewhere. Most new Lib Dem MPs should beat Tory challengers for the same reason.

It's in the bag - and the bag will need to be bigger than Labour's doubters think.

Ivor Crewe is vice-chancellor and professor of government at the University of Essex