Since the 1960s, conventional wisdom had it that elections are won or lost, if not in the TV studio alone, then in a combination of the studio, the newspaper office and the party HQ. Then in the 1990s Labour, taking its lead from the Liberal Democrats, discovered the potency of local campaigning in marginal seats. But this campaigning was centrally directed and, often, centrally staffed.
Now a Lancaster University team suggests a different view. David Denver, Gordon Hands and Iain MacAllister analysed the past three general elections in detail; looked at previous research that asked voters about the local campaigning they had experienced; and asked the main parties' agents in every UK constituency for numbers of campaign workers involved, leaf-lets delivered, and so on. They conclude that "the more intense the local campaign, the more likely people are to vote; the harder individual parties campaign, the more likely people are to vote for them".
The team found that, in 1992, in constituencies where there was a strong local campaign, turnout increased on average by 2.9 per cent. Where there was a weak local campaign, it fell by 0.5 per cent. So campaigning made a difference in turnout of 3.4 per cent. It was even higher in 1997 (4.4 per cent), and higher still in 2001 (4.6 per cent). This may help explain why turnout fell dramatically in 2001: as the parties concentrated more resources on fewer seats, it seems, the enthusiasm of voters in seats that were not targeted went into decline. The authors predict that, as parties continue to focus on "key seats", overall turnout will continue to fall.
Yet only the most benign (and probably unsuccessful) parties see their mission as increasing turnout: parties are about winning. None the less, the research suggests that local campaigning not only increases turnout but also increases prospects for success for those parties that do it.
Paradoxically, the harder the Tories campaigned locally in 1992 and 1997, the worse they did - but this apart, all the evidence from Labour and the Lib Dems (and the Tories in 2001) indicates a strong link between local campaigning and electoral success. Where Labour barely campaigned in 1992, its share of the vote was 1.5 per cent below average; where it campaigned strongly its share was 2.4 per cent above, giving a net benefit of 3.9 per cent. The Lib Dems gained an even more impressive 5.9 per cent from local campaigning. Boundary changes make such calculations for 1997 impossible; but the researchers calculate that in 2001, a strong local campaign was worth an extra 0.8 per cent of the vote to the Tories, 2 per cent to Labour and 3.2 per cent to the Lib Dems.
Such gains are marginal when there is a national landslide on the scale of 1997 or 2001. But, as the researchers point out, parties that fail to maintain the strength and vibrancy of their local organisations are likely, in some circumstances, "to suffer a possibly severe electoral penalty".
So new Labour's enthusiasm to turn itself from a party of members into one of supporters may be a long-term mistake. No matter how extensive a bank of centrally directed telephone canvassers a party establishes and how many tonnes of direct mail it despatches, "campaigns will always require volunteers on the ground to . . . knock on the doors of voters on polling day", say the researchers.
The research is published in the journal Political Studies, available from Blackwell Publishing (01865 791 100)