How Labour can still lose

A triumphant conference; enormous poll leads. Can anything go wrong? Yes, write Patrick Dunleavy and

London's political in-crowd long ago concluded that the general election of May 2001 will be a shoo-in for Tony Blair. And nearly all the political portents seem to point that way: they suggest that the Prime Minister could well be thinking not just of a second term but of that controversial third term, too.

Labour's opinion poll rating remains steady at about 50 per cent with the Tories still stuck at rock bottom, on 28-30 per cent. Though the SNP and Plaid Cymru may steal seats in Scotland and Wales, Labour's remarkable English hegemony looks firm. Further, low inflation and low unemployment look set to continue, while Gordon Brown's surplus may allow tax cuts as well as more investment in the NHS and education.

So the establishment, mesmerised by the scale of Blair's win in 1997, assumes that new Labour is on track for another sweeping victory - hence those handsome business donations to party funds.

Almost the only person who remains cautious, it seems, is Blair himself. And he is right to be so. If he is getting good advice, he will know full well that his astonishing 1997 electoral landslide was built on sand.

General elections in Britain are arbitrary, even perverse, political events. It is not simply that the votes cast in first-past-the-post elections for Westminster do not translate accurately into seats won in the House of Commons, and that the winning party invariably gets an "inflated" reward in terms of seats. It is also that variations in the trends in votes for the other parties can have major repercussions on the headline result between the two main parties.

Since 1945, both Labour and the Conservatives have always won a slightly higher percentage of parliamentary seats than of votes: on average, for every 1 per cent of votes, they have won just over 1.1 per cent of seats. The variations around this average have been small. In 1997, however, the equation changed dramatically: to Labour's huge advantage and the Tories' disadvantage. For every 1 per cent of votes Labour got 1.5 per cent of the seats, while the Tories got only 0.8 per cent.

In other words, Labour did far better (by about 50 seats) and the Tories far worse than would normally be expected, even under a first-past-the-post system that tends to inflate the parliamentary representation of the winning party.

But it wasn't only Labour that excelled in 1997. The Liberal Democrats enjoyed an annus mirabilis. Since they and their predecessors became serious contenders in the 1974 elections, the Liberal Democrats have averaged 0.14 per cent of seats for every 1 per cent of votes. In 1997, they still suffered from first-past-the-post, but not nearly so much. This time they got 0.47 of the seats for every 1 per cent of votes and so more than doubled their number of MPs from 22 to 46.

For Labour it was the biggest landslide in seats since 1945. Yet if the electoral system had operated as it had done in every other recent election, Labour's majority over all the other parties (including Northern Ireland MPs and Scottish and Welsh nationalists) would have been no more than about 80. In other words, neither Labour nor the Lib Dems could claim unprecedented levels of popularity. In 1997, Labour's share of the popular vote (44.3 per cent) was almost identical to its performance (44.1 per cent) in 1964, when Harold Wilson, the then Labour leader, got a majority of four seats. Moreover, a smaller portion of the electorate voted in 1997 (71 per cent) than in 1964 (77 per cent). As for the Liberal Democrats, their vote actually fell by one percentage point between 1992 and 1997.

What explains these unparalleled results? The answer is simple. First, it wasn't that new Labour was so popular, but that the Tories were remarkably unpopular. Labour and Liberal Democrat voters in effect formed an integrated bloc of 61 per cent of the electorate against the Tories on 31 per cent. Under first-past-the-post, both parties were bound to make huge gains in seats once their voters got their act together, voting tactically in each constituency for whoever was most likely to beat the Tory. In that sense, the Tories faced a "national coalition" rather as Labour did, in a more formal sense, in 1931.

Now, however, dislike of the Tories is fading from the political scene. The intense discontent of 1997 may well be only a distant memory by 2001. After all, governments are inevitably, unfairly even, the primary focus for discontents; and this is as true for the Liberal Democrat voters who did so much to secure Blair's 1997 triumph as for other voters.

Liberal Democrat discontents are likely on two fronts. First, they are the people who backed Paddy Ashdown's call for extra spending on education and health; they will ask, in 2001, whether there has been significant improvement in these services. Second, they rallied round new Labour's democratic programme, which now looks far less imposing than it did in 1997, with a false Freedom of Information Bill, waning prospects of electoral reform and apparent ministerial preference for an appointed second chamber. Moreover, it is doubtful that the two parties will be able to stand as closely together in 2001 as they did in 1997. They have drifted apart, not only on democratic questions, but also on public spending, social policies and the government's caution on Europe.

So Labour cannot expect people to vote tactically as they did in 1997 and therefore cannot expect the electoral system to work as it did then. In other words, Millbank ought at least to consider the scary prospect of writing off its 50-seat bonus, along with 13 Liberal Democrat seats, and treating its present overall majority as only 80 seats. That, remember, is based on the voting shares at the last election, when Labour got 44 per cent and the Tories 31 per cent.

Can Labour expect to maintain or improve that position? The opinion polls can only be conjectural, being so distant in most people's minds from the actual election. Almost every other signal - including the most recent by-elections in Wigan and Hamilton South - points towards the swinging pendulums of normal party rivalries, and steep falls both in Labour support and the willingness of its voters to go to the polls. In the Scottish Parliament elections in May 1999, Labour fell by 12 percentage points from how electors said they would vote for a devolved parliament when asked in 1997. In Wales, the drop was even steeper, at 19 points. In the European elections, Labour's share of the vote was down by 18 points, with a record-breaking 24 per cent low turnout. In local council elections in 1999 and 1998, Labour's vote share was down by 10 points, and turnout fell badly, to 29 per cent in both years.

New Labour did not mobilise voters in great numbers in 1997 and it does not enthuse them now. Party leaders may claim that low turnouts are a sign of a contented electorate, but they won't fool the people, even if they fool themselves. Rather, they should study two US political scientists, Adam Przeworski and John Sprague, who argue in Paper Stones, a history of electoral socialism in western Europe, that social democratic governments have nearly always struggled to consolidate their grasp on power. To win in the first place, they have to create a fragile coalition of working- and middle-class voters. In government they then tend to focus on keeping their middle-class support, with the result that working-class supporters either defect or fail to vote at the next election.

Blair's own centrist political style is likely to compound the turnout problem. Historically, in all democratic countries, people tend to vote in greater numbers when they feel that big issues are at stake. Blair's insistence on evading and downplaying big issues dulls people's fears and expectations. Moreover, progressive middle-class voters already resent the government's failure to do more to promote social equality, green policies and genuine democracy at Westminster. They may well look for alternatives. In 1999, voting for parties other than the three main contenders rose to 38 per cent of the poll in Scotland, 36 per cent in Wales and 28 per cent in the European elections, albeit all under PR systems.

As for "heartland" Labour people, the government may not be able to count on active voting, even if Gordon Brown delivers on economic fundamentals. To be sure, there is cautious redistribution towards families and towards the working poor (though not towards those on pensions or benefits). But new Labour may find it hard to satisfy its core supporters on most of the issues that really matter to them - work opportunities, schools, health, transport, the environment. Daily experience reveals the shallowness of what has been achieved so far.

There are signs that No 10 is starting to recognise all this. But those who believe that opinion polls never lie; that it is contented rather than disillusioned people who can't be bothered to vote; and that the 1997 landslide represents a permanent realignment that Labour can rely on into the next century are building castles in the sky. Our perverse electoral system once denied Labour a Commons majority even when it won more than half the votes cast (1951). It may be crueller to a party that cannot hope for more than about 44 per cent of the vote in a era of diminishing turnout.

The Tories don't look as if they can win in 2001. But Labour can still lose. Expect renewed talk of hung parliaments.

Patrick Dunleavy is professor of government, London School of Economics; Stuart Weir is director of Democratic Audit, Essex University