Guest writer Dr Stephen Fisher, of Oxford University, on why winning the most votes doesn't always mean winning the most seats
The race between Labour and UKIP for the largest share of the vote in Britain at this week’s European Parliament elections seems too close to call. Polls have varied in which party they place ahead. An average of YouGov European vote intention polls for the month to 16th May (weighting recent polls more highly) has the two parties neck-and-neck on 27%. Adjustment for self-reported chances of actually voting would see UKIP in the lead.
With such a close competition, who wins on votes and who wins on seats might be different.
The UK has a form of proportional representation (PR) for European Parliament elections. But it is regional not national.
The use of regional rather than national PR can help small parties with concentrated support. Plaid Cymru won a seat in 2009 but it did not achieve the 1% of the vote UK-wide that would be the threshold for representation if a national system of PR had been in operation.
Typically, though, regional PR advantages big parties over small ones. Both the Greens and BNP would have won more seats in 2009 if national PR had been used.
The effective threshold is close to 10% in most regions. The South East, as the largest region, has 10 seats and so 7% of the vote might be sufficient to win a seat. Meanwhile a party needs 20% in the North East to win one of the only three seats up for grabs there.
Since it matters where votes are won as well as how many, the winning party on votes may not win the most seats.
This has happened in general elections in Britain (1951 and February 1974), the USA (2012) and in Australia (on many occasions). It is less likely with regional PR than first-past-the-post (FTPT). But it is not just a theoretical possibility for this year’s euro elections.
Using YouGov data on European vote intention over the last month it is possible to get large enough sample sizes in each region to estimate the pattern of change in the share of the vote across regions.
UKIP look like they will again make much more limited gains in Scotland and Wales than in England. Apart from London, UKIP look set to gain more votes in the eastern regions of England (North East, Yorkshire & Humber, East of England, and South East) than in west. Change in the Conservative share is somewhat the mirror image, with the party holding up relatively well in Scotland, Wales and London and falling most heavily in the eastern regions of England. This would seem to confirm broader polling evidence that UKIP take more votes from the Conservatives than any other party, especially in European elections.
Labour’s recovery since 2009 is strongest in the north of England, and this is reflected by correspondingly heavy Liberal Democrat losses in those three regions. Again this in line with the pattern of party switching we have seen in the polls. Labour’s recovery since 2010 has been driven primarily former Liberal Democrat voters. The regional pattern of this switch will be of concern to both parties. Labour need to make greater progress in the south to win a general election and the Liberal Democrats risk losing all EU representation from outside the south.
Most polls have suggested the Greens are at risk of losing their two seats in the South East and London, but very recent polls suggest they might do better than in 2009. Unless the polls are seriously wrong, the BNP vote has all but collapsed. They will surely lose their two seats.
While the estimated regional variation in the change in vote share since 2009 confirms broader developments in party competition with lessons for next year’s general election, in truth the implications for the European Parliament are likely to be more limited. By comparison with a uniform regional change projection, regional differences only affect the outcome of just three seats.
The large thresholds discussed above mean that deviations from uniform regional swing need to be massive to affect seat outcomes.
The north/south differences mean the Liberal Democrats gain one in London but lose one in Yorkshire & the Humber, compared with uniform swing. This affects where they are represented without any net effect on their total number of MEPs.
The overall net effect of the regional differences is Labour to lose two seats to UKIP. If this pattern is replicated on the night a marginal vote share win for Labour could be accompanied by a two-seat lead on seats for UKIP.
Acknowledgements:I am very grateful to Joe Twyman and others from YouGov for providing the individual-level polling data on which the analysis is based.
Stephen Fisher is the University Lecturer in Political Sociology and the Fellow and Tutor in Politics at Trinity College, Oxford University. His research focuses on electoral behaviour and social attitudes.
Main image: Getty
