Tories could be trounced without Lib-Lab pacts: Ivor Crewe on the lessons that opposition parties could learn from last week's local election successes

The Conservatives' humiliation in Newbury and the council elections will intensify the debate on the Centre Left about a general election pact between Labour and the Liberal Democrats.

Advocates of a pact point to last Thursday's results. What produced the rout, they argue, was not simply the Government's depth of unpopularity but the unprecedented efficiency with which the anti-Conservative vote was distributed. The 1985 county elections offer a telling comparison. The Conservatives were almost as disliked then, managing only 36 per cent of the shire county vote compared with 35 per cent last Thursday. But in 1985 they won 1,316 seats, retained control in 10 out of the 47 counties and were the largest party in 15 others; this time, they won only 969 seats, lost control of all but one county and were the largest party in only eight others.

Two factors account for the difference. Firstly, both Labour and the Liberal Democrats advanced furthest in their own areas of strength - Labour in the urban North, the Liberal Democrats in the rural and small-town South. Secondly, both parties were helped by the mysterious failure of the other to stand in certain Conservative marginals. John Curtice of Strathclyde University estimates that on average Liberal Democrats did 11 percentage points better where Labour candidates stood down and Labour did 13 percentage points better where Liberal Democrat candidates stood down; in both cases the Conservative vote was almost unaffected. In the absence of an official pact, voters - and occasionally local parties - acted as if one existed.

The pro-pact lobby also cites recent polling evidence. In a Gallup poll in February fully 58 per cent claimed they would vote for a 'Labour/Liberal Democrat alliance led by Mr Smith with Mr Ashdown as his deputy' and in an ICM/Guardian poll last month 65 per cent said they would vote for a 'joint candidate' supported by Labour and the Liberal Democrats. On these figures Labour and the Liberal Democrats would crush the Conservatives at a general election, as on Thursday.

Labour and Lib-Dem traditionalists draw the opposite conclusion. Last week's results, they argue, demonstrate that their parties can trounce the Conservatives without a pact. Labour contested Newbury without sabotaging the Liberal Democrats and the Liberal Democrats contested most of Northamptonshire without preventing Labour from gaining control. If voters want to turn out Conservative administrations they will do so without party headquarters entering pacts.

Would a pact, in fact, work? The logic seems impeccable. The two parties would each agree to stand down in those marginal Conservative seats where in 1992 their own candidates came a poor third and the other party was the clear challenger. In reality, the idea is fraught with difficulty, indeed danger.

On closer inspection the electoral arithmetic barely adds up. In the absence of their own candidate, Liberal Democrats would not necessarily split in Labour's favour. True, they appeared to do so on Thursday - but only in Conservative-run counties. True, the ICM poll reports that 87 per cent would vote for a joint Lab-Lib candidate in a straight fight against a Conservative. But that figure reflects the current depth of Conservative unpopularity, which will have passed by the next election. In last year's general election Liberal Democrats' second preferences divided 34 per cent Con, 39 per cent Lab - too slender an advantage to Labour to help in any but the most marginal of seats. Paddy Ashdown cannot necessarily deliver.

In which case, argue some pact advocates, Labour should unilaterally withdraw from seats where only the Liberal Democrats have a good chance of ousting the sitting Tory. Such a bold move would not, of course, help to elect additional Labour MPs. But it would help to defeat additional Conservative MPs, thereby increasing Labour's chances of becoming the single largest party and forming a minority or coalition government. Labour would still get a return on the deal.

The return, however, would be very limited. In most safe Conservative seats where Labour came third, the Liberal Democrats were too far behind to have a serious chance of winning. In a few others, such as Falmouth & Camborne or Conwy, Labour did well enough to win next time. Suppose that three conditions were required for a Labour withdrawal from a Tory seat: that Labour came third in 1992 with under 15 per cent of the vote; that the Liberal Democrats came within 20 per cent of the Conservatives; and that the Labour vote was larger than the Conservative majority. The total number of such seats comes to 11. Given the political difficulties of a pact, it would hardly be worthwhile.

And the political difficulties would be enormous. Voters' mental maps of the party system would change. They would associate the Liberal Democrats with the Labour party and a Labour government. Some anti-Labour supporters of the Liberal Democrats would switch to the Conservatives. The net benefit to the Liberal Democrats might be zero, or even negative.

Moreover, while national party headquarters could instruct certain local parties not to field a candidate, it could not prevent local activists from supporting unofficial candidates, who might attract enough votes from defiant or confused voters to undermine the purpose of the pact. The inevitable wrangling between Walworth Road and local parties would be echoed among Labour MPs and be gleefully hyped up by the Conservative press as a 'Labour split'. The electoral damage inflicted by the rows might exceed the electoral advantage of a pact.

Nor can a pact exist in a political vacuum. The two parties would need to agree on a joint manifesto, act jointly at Westminster and agree in advance to form a coalition in the event of a hung parliament. The result: more rows, both within and between the parties. The attempt of the Liberals and SDP to create an alliance in the mid 1980s is not a happy precedent.

The solution for both parties is to denounce pacts in public while condoning them in private. They can nominate paper candidates in hopeless seats and encourage party workers to campaign in more favourable constituencies nearby. They can turn a blind eye to informal 'arrangements' at local elections. They could even tolerate a few formal deals, so long as they were initiated locally ('It's a decision for the local party'). That, after all, is what happened in the county elections. And it worked.

Ivor Crewe is Professor of Government at the University of Essex.