Although the complex cycle of local elections can be a source of much confusion, there are occasions when it can be turned to advantage. Monitoring a party's progress or decline over not just one year but rather a series of years serves to operate as a belt-and-braces technique and strengthens any conclusions about the ebb and flow of popular support that we might wish to make.
The method works as follows. For each by-election, there is always a previous year when the ward was last contested at the annual May elections. By-elections held during 2002, therefore, will have an equivalent May election in at least one of the four years between 1999 and 2002. If a by-election occurs in a shire district with whole council elections, for example, that means that the point of comparison is 1999 when the seat was last fought. By contrast, a by-election in a metropolitan borough with annual elections could be compared either with the result in 2000 if the by-election took place before May or with that in 2002 if it was fought after May.
by-election and the relevant preceding May election
and where there have been no boundary changes in the intervening period. Of these, 40 were fought originally
in 1999, six in 2000, 15 in 2001 and 25 in 2002. Armed with this knowledge, we can calculate the average change for each party between this year's by-elections and each of the last four rounds of May elections.
The analysis is further strengthened if it can be related to the 'national equivalent vote' which then prevailed. This is an estimate of vote share for the three main parties had the elections been fought across the country as a whole. In May 2000, for example, the Conservatives were comfortably out in front with 38% of the national equivalent vote, with Labour on 30% and the Liberal Democrats on 26%. In 2002, by contrast, it is estimated that the Conservatives would have obtained 34%, only just ahead of Labour on 33% and with the Liberal Democrats back on 25%.
The final piece of the jigsaw involves taking the average change in each party's share of the vote and adding it to or subtracting it from that year's national equivalent vote.
What can this method tell us about the state of the three main parties, just six months away from the crucial test of the 2003 May elections?
For the Conservatives the news is, as anticipated, somewhat gloomy. During 2002 the party has failed to make significant by-election gains, even incurring some notable losses, as November's results confirm. The four comparisons show the party ranging between 32% and 34% of the national vote, fractionally better than in recent opinion polls but a long way short of resurrecting the party's chances of unseating Labour. The party is well down on the highs of 2000 (see table) but, more significantly, its support has hardly changed compared with 1999. Given that the majority of contests next year were previously fought then, the prospects of the Conservatives making dramatic gains do not look good.
Before the government becomes too excited by this news, however, it must acknowledge its own difficulties. While three of the four comparisons show Labour with between 31% and 33% of popular support, the fourth, comparing 2002 with 1999, suggests the party can count on attracting fewer than three in 10 voters. That appears to imply that by-election results in the shires - the location of most of the seats which have not been contested since 1999 - have been particularly poor for the party. Labour too must face next year's local elections with some trepidation.
Which leaves the Liberal Democrats. The party's by-election performance shows its vote share on the rise, whichever year provides the point of comparison. More significant still, the party's 2002 by-election share is on average four percentage points higher than in the same seats in 1999. Indeed, the party has gained more seats from Labour in the shires than have the Conservatives. If this pattern continues, then a poor Labour performance in 2003 is likely to mean good news for the Liberal Democrats, supporting the party's view that it, rather than the Conservatives, is now regarded as the alternative party of choice by many former Labour voters.
Nor should we forget the political fringe. Disillusionment with all the political parties is likely
to have played some part in the BNP's victory in Blackburn last month, and there is evidence of a growing willingness by voters to look outside the mainstream. In May 1999, 10% of votes were cast for independent and 'other' party candidates. It will be little surprise if that figure increases again.
Labour may excuse the loss of seats next year as natural for the party of government at the mid-term. It will be much more difficult for the Conservatives, however, if their progress remains sluggish. But that is precisely what by-election results are suggesting will happen. The circumstances look ripe for yet another crisis and inevitable demands for a change of leadership.
Colin Rallings and Michael Thrasher
Directors, LGC Election Centre, University of Plymouth
By-election results November 2002
Council Ward Result Majority Turnout%
07 November 2002
Aberdeen City CouncilNewhills LD gain from Lab50.9 over Con 43.0
Bedfordshire CC North East BedfordshireCon held 47.3 over LD 23.5
Epping Forest DCWaltham Abbey North EastLD gain from Con17.1 over Con25.7
Epping Forest DCWaltham Abbey South WestLD gain from Con22.6 over Con18.3
Lewisham LBC Downham LD gain from Lab8.9 over Lab 26.9
Stratford On Avon DCStratford AlvestonLD gain from Con15.9 over Con35.3
Stroud DC Dursley Con held 2.6 over LD 35.8
08 November 2002
Council Earlston Gordon & DistrictCon gain from LD24.5 over LD65.8
14 November 2002
Bracknell Forest BCHanworth Con held 3.9 over Lab 22.0
Flintshire CC Leeswood Lab gain from Ind23.5 over IND 48.9
Kingston upon Hull
City Council Marfleet Lab held 17.8 over LD 20.2
Penwith DC St. Buryan Con held 34.8 over LD 46.2
Salisbury DC Fisherton & Bemerton VillageLab held14.8 over Con 27.4
Wiltshire CC Salisbury St. Paul Lab held 16.9 over Con 24.3
Taunton Deane DCRuishton & CreechLD held 4.1 over Con 44.0
21 November 2002
Darwen BC Mill Hill BNP gain from LD0.9 over Lab 39.4
Bolton MBC Daubhill Lab held 5.0 over LD 21.0
Dacorum DC Woodhall Lab gain from Con2.5 over Con 23.0
Sutton LBC Carshalton Central LD held 7.2 over Con 29.9
28 November 2002
Durham CC Burnopfield & DiptonInd gain from Lab 1.4 over Lab 22.0
Hertfordshire CC Meriden Tudor LD gain from Lab21.5 over Lab 27.7
Mole Valley DCBookham South LD held 13.8 over Con 40.7
Mole Valley DCDorking NorthLD gain from Con 20.9 over Con 40.2
Trafford MBCDavyhulme East Con held 15.2 over Lab 50.0