Safe Seat Blues
… A Brief and Hazardous Analysis of UK Seats Contested Under First Past the Post and the Alternative Vote Systems
Looking at two very different UK seats under both First Past the Post and Alternative Vote indicates that the incumbents of safe seats may be in for a fight come the next election and parties and voters will have a range of new opportunties if the voting reform referendum yields a Yes result.
This commentary is based on the LSE 2nd preference analysis performed just after the last general election. I have assumed that 3rd and subsequent preferences split the same way as 2nd preferences. Thus votes allocated to a UKIP candidate from a Green Candidate are allocated as if the original vote had been for UKIP. This is a heroic assumption and the analysis and commentary should be treated with some caution. I put them up as a talking point rather conclusive evidence of anything. I am using data on the allocation of second preferences gathered during the 2010 British Electoral Survey and published in this article http://www.oxfordjournals.org/our_journals/parlij/gsq042.pdf I have looked at two seats I saw mentioned in a recent discussion between Pro (#Yes2AV) and Anti (#No2AV) commentators on Twitter. Mid Bedfordshire 2010 Results Conservative Nadine Dorries wins the seat with 52.49% of the vote under either First Past the Post or Alternative Vote. The swing from 2010 for Ms Dorries to lose her seat under First Past the Post is a hefty 13.8%. This is a safe conservative seat. Under AV a 9.6% swing from Conservative to Liberal Democrat is required for the Liberal Democrats to win Mid Bedfordshire. Even the most optimistic of Liberal Democrat activists or strategists would not put Mid Bedfordshire high on a list of target seats. The swing for UKIP to win the seat is 23.7%. Under AV this sort of swing from Conservative to UKIP (or to a conservative minded independent) might see the Liberal Democrats sneak through the middle to capture the seat on 5th preferences. A lot depends on how 3rd to 5th preferences are distributed. As I note in my introduction I have made a simple but perhaps overly simplistic assumption. The 2010 result is a bit of high water mark for the Conservative vote in this constituency. The three prior general elections have seen the Conservatives poll in the high 40’s. It’s difficult to tell if the 2010 result represents a return of Mid Bedfordshire folk to their Conservative hearts after the razzle dazzle of New Labour had worn off or if the 1997 to 2005 elections results are about par for the course. I note that the Labour vote was down slightly more in Mid Bedfordshire than it was across the country as a whole. In order to get a feel for how this seat might look in 2015 I have run the same analysis on the 2005 results. In 2005 Ms Dorries won the seat with 46.3% of the vote, with the Liberal Democrats and Labour splitting the remainder of the vote almost evenly, slightly to the Liberal Democrats advantage. 2005 Results In 2005 Mid Bedfordshire is a safe Conservative seat, under First Past the Post. Under AV the seat is decided on 5th preferences. Total votes for the Conservative Candidate were 23,345 (46.30%). Total votes for the Liberal Democrat and Labour Candidates were 23,341 (46.29%). Suddenly things begin to look a little more interesting. Unsettling perhaps if you are the incumbent. Under AV a 4.5% swing from Conservative to Liberal Democrat leads to an Orange win on 5th preferences. The Tories lead all the way until the last round when the Labour vote splits 66% to the Liberal Democrats with most of the rest of Labour second preferences being cast for candidates already eliminated. If the Liberal Democrats can persuade some Labour voters to switch their second preference from, say UKIP, the swing required is slightly less. If Labour voters split their second preferences 71% Liberal Democrat and 4% UKIP instead of the 66% and 9% predicted by the BES data then the swing from Conservative to Liberal Democrat required is 3.9%. This holds true if second preferences are captured from the Greens too. I think that a 4.5% swing is an achievable swing. It would put Mid Bedfordshire about a third of the way down the Liberal Democrats list of target seats. It should be noted that the current target seats probably become more marginal and therefore a 4.5% swing might not be a relatively attractive target if all constituencies are elected under AV. Plymouth Sutton and Devonport Plymouth Sutton and Devonport saw Conservative candidate Oliver Colville win the seat with 15,050 votes (34.29%) over Labour’s Linda Gilroy on 13,901 (31.67%). The Liberal Democrat candidate, Judy Evans, polled third a few thousand votes behind Ms Gilroy. A swing of 1.3% from Conservative to Labour would see Labour win the seat. A 7% swing from Conservative to Liberal Democrat would see the Liberal Democrats come from 3rd place to 1st place and win the seat over a 2nd placed Labour candidate. Under AV Mr Colville still wins the seat by 19,877 votes to 19,234 on 6th preferences. A swing of 2.7% in 2nd preference allocation from Liberal Democrat voters from Conservative to Labour would also see Labour win the seat. This is a Conservative / Labour Marginal seat that could easily become a three way marginal. Under AV the swing required from Conservative to Labour falls from 1.3% to 0.9%. This is fewer than 400 votes. Under AV Plymouth Sutton and Devonport is a slightly easier gain for Labour. Of interest to me is that the swing Labour to Liberal Democrat of 4% would have an impact on the result. Under First Past the Post the distribution of votes between the losing parties is of little interest. Under AV I notice that a 3.1% swing from Labour to Liberal Democrat would put the Liberal Democrats into second place approaching the last round and they would win the seat. There were 4,114 votes not cast for the one of the 3 top placed candidates, the margin between the first ranked candidate and the third ranked candidate is 4,221. Plymouth Sutton and Devonport becomes a three way marginal with the distribution of preferences from minor parties crucial to determining who progresses to the last round. Conclusions If it is safe to draw any conclusions from analysis of these two seats I would draw the following. Previously safe seats become marginal. Seats with what under First Past the Post would have unassailable majorities become winnable on single figure swings. However, I note that the Mid Bedfordshire AV winning margin of 4 votes still requires a substantial swing to unseat the incumbent. Marginal seats can become three way marginals and the preference voting system reduces the need for tactical voting. What impact this has on campaigning, consensus building and cross-bench co-operation is something that that political parties and voters will have to create themselves. I would suggest that personal qualities, individual voting record and individual positions to the right or left of parties will begin to have some influence on election outcomes. In closely fought seats the results become finer and harder to predict as they will turn on the distribution of later preferences. Whilst tactical voting is not necessarily eliminated it is moved from 1st preferences to later preferences.