In London, it’s horses for courses

Two steps forward, one step back. It looks as if enthusiasts for constitutional reform will be able to enjoy a double celebration next week – but, if they are honest, they will also admit that one feature, the voting system for choosing London’s mayor, will need to be redesigned.

The reasons to celebrate are, first, that voters who use proportional representation seem to like it and, second, that more and more people are adopting a “horses for courses” approach to their choice of party, rather than blindly following the same party at all times – or merely using local votes to punish the failings of national governments.

Evidence that experience of PR is helping the cause of reform comes from new research conducted by the academic group Crest, the Centre for Research into Elections and Social Trends. Its latest study, published last week, demolishes the assertions of the status quo camp that voters would find PR ballot papers hard to use and that the experience would make them yearn for the simple familiarity of first-past-the-post.

Crest’s research finds that only one in ten voters in last year’s devolution elections found the ballot papers hard to fill in. In Scotland, a two-to-one majority thinks that PR should continue to be used to elect the parliament, and a clear majority favours extending PR to local government. However, views about extending PR to the House of Commons are more mixed, with Scottish and Welsh electors evenly divided between wanting a “fair” voting system and one that produces a winner.

Second, continuing support for Ken Livingstone in London’s mayoral election shows how readily voters are abandon ing their normal party allegiance if the conditions are right.

PR helps this process by boosting the prospects of smaller parties, but voters can go walkabout without it. In 1997, the general election was fought on the same day as elections to England’s county councils. Both were traditional first-past-the-post contests. Yet they produced some striking contrasts. In Cambridge, votes for county councillors divided almost evenly between Labour (21,000) and Liberal Democrats (18,000). But the very same voters overwhelmingly backed Labour (by 27,000 to fewer than 9,000) when choosing their MP. In Pendle, three times as many people voted Lib Dem for the county council as for parliament.

These were extreme examples, but on average, according to Colin Rallings and Michael Thrasher of Plymouth University, Labour parliamentary candidates performed four points better, Tories two points better and Lib Dems seven points worse than their council counterparts. The last time that a general election and local elections were held on the same day was in 1979. Rallings and Thrasher conclude: “Many more electors cast votes for different parties at the 1997 general and local elections than had done so at the comparable contests in 1979.”

Such a conclusion is no doubt maddening for party strategists, who prefer simple loyalty to complex judgement. But it is surely good for democracy. At the end of March 1982, the Tories looked as if they were heading for terrible results in the local elections five weeks later. Instead, the party triumphed across the country.

What happened? Did dozens of Tory councils suddenly buck up their ideas? Did Labour authorities simultaneously descend into corruption and incompetence? No, neither thing happened. Instead, Argentinian forces invaded the Falklands at the beginning of April, and British voters rallied round Margaret Thatcher and the task force sent to recover the islands. Local elections became little more than a giant, nationwide opinion poll.

Thatcher well understood that local elections were not really local at all. Her attacks on the powers of local government may be deplored on the grounds of democratic principle, but she could at least argue that voters did not really respect the local character of municipal power, so why should she?

This started to change in the 1990s – remember the way Wandsworth and Westminster resisted the anti-Tory tide in 1990 and 1994 by promising to keep local taxes down. And voters have rightly punished unsuccessful Labour councils, from Hackney and Islington to Sheffield and Liverpool. The effect has been to make it increasingly difficult to translate local election results into national trends.

The one reform that plainly isn’t doing its job is the supplementary vote being used to elect London’s mayor. This was to allow Londoners two mayoral votes – a first choice and a second choice. On the night of 4 May, first-choice votes will be counted. If no candidate wins an outright majority, then the top two candidates survive into a second count. The second preferences of supporters of the eliminated candidates will then be added to the tallies of these two finalists. Whoever then has the bigger total vote will win.

The aim is to ensure that the mayor enjoys the popular support of an outright majority of London voters. But this ambition may well be thwarted. The reason is that there is no clear, undisputed rival to Ken Livingstone. Voters who are determined to block him may end up casting both their votes for eliminated candidates.

Future elections for mayor – of any city – should be conducted by the Alternative Vote, under which voters have as many preferences as there are candidates, and place them all in – 1, 2, 3 and so on.

There is a strong case for electing MPs by AV as well – as I first argued in the New Statesman 20 years ago. But that is another story.