By Lewis Baston and Seth Thévoz
Part 1 of the analysis set out the 2015 general election trends in changes of votes across the 57 seats the Liberal Democrats were defending.
The Local Election Dimension
As well as the general election, there were also local elections across much of England. While these were also a disaster for the Liberal Democrats, the contours of the disaster were somewhat different.
Local elections taking place on the same day as a general election allows a finer-grained analysis of the trends; the differences between local and parliamentary voting patterns throw up information about the strength of personal votes and incumbency and the number of voters who make different choices at each level. They also illuminate the distribution of votes for parties within parliamentary constituencies, although this paper concentrates on the overall differences between local and general voting in whole constituencies.
Of the 57 seats that elected Liberal Democrat MPs in 2010 and were being defended in 2015, there were comparable local elections in 29 of them (plus the target seat of Watford). The 11 Scottish seats, the seven in London, three in Wales and three in Cornwall are excluded because they had no local elections, and nor did Berwick-upon-Tweed, Cheltenham and Chippenham. Westmorland & Lonsdale did have partial local elections, but too much of the constituency (including the main town of Kendal) was not up for election this year, and it has therefore been excluded. Compiling local election results by constituency is a contentious exercise – the cautious reader is directed to the appendix to this section – and in making the comparison, it is also necessary to put figures in vote share terms, rather than in absolute number of votes.
There were eight 2010 Liberal Democrat seats out of the 29 with comparable local elections where the result of the local and national elections was different, namely:
- Norfolk North, where Norman Lamb won despite the Conservatives winning in the local elections;
- Eastbourne, Thornbury & Yate, Bath, Yeovil, Portsmouth South and most dramatically Eastleigh where the Conservatives won the parliamentary seat and the Lib Dems ‘won’ the local elections in the equivalent area. In Eastleigh, there was a 16.5% Conservative lead in the General Election voting and a 5% Lib Dem lead in the local elections.
- Bristol West was won by Labour in the General Election, and the Greens in the local elections. The Lib Dems were third in both contests.
Among the eight constituencies where the Liberal Democrats did better (relative to their main competitor party) in the general election:
- There were four seats that were unambiguously Conservative-facing (Norfolk North, Wells, Solihull, Lewes);
- Two where the Conservatives were second in 2010 but where Labour were the principal competitors in 2015 (Leeds North West, Cambridge);
- Two where the competition was unambiguously between Labour and Lib Dems (Burnley, Bradford East).
Among the 21 constituencies where the Liberal Democrats did better in the local elections and worse in the general election there were:
- Three unambiguously Labour-facing seats (Manchester Withington, Redcar, Birmingham Yardley);
- Two Labour-facing seats with a significant Green presence (Norwich South, Bristol West)
- One seat where the Conservatives were second in 2010 but where Labour were the principal competitors in 2015 (Sheffield Hallam);
- Fifteen constituencies which are Lib Dem v Conservative contests, of which at most two had serious UKIP presences (Torbay, Southport, Devon North, Colchester, Eastbourne, Thornbury & Yate, Cheadle, Hazel Grove, Yeovil, Portsmouth South, Somerton & Frome, Taunton Deane, Eastleigh, Dorset Mid & North Poole).
The constituency results demonstrate a wide range of patterns. The very worst results – relatively and absolutely – in the general election were in a number of Con/Lib Dem marginals in southern England: Watford, Mid Dorset & North Poole, Eastleigh, Taunton Deane, Somerton & Frome, Portsmouth South, and Yeovil. These made up the cases that fit the general description, but where the Lib Dem MP managed to keep pace or exceed the local election vote all the more remarkable – Wells, Torbay and Lewes in particular, plus Norman Lamb’s singular success in Norfolk North.
In some of the constituencies, the dominant net effect was simply that the Conservatives ran way ahead of their local election vote and the Lib Dems way behind – Eastleigh is the clearest case of this.
This does indicate that there was a ‘soft’ Lib Dem vote that was prepared to support the party’s local election candidates but voted Conservative nationally – there has always been an element of this in some constituencies including Eastleigh, but it seemed particularly large in 2015. This does suggest there was something – in some constituencies anyway – to the explanation offered by some Lib Dem strategists that fear ended up propelling these people into voting Tory. However, short of committing to a permanent alliance with the Tories, it is difficult to imagine what more the Lib Dems could have done to reassure these voters since 2010! Winning back this group will depend on them losing faith in the Conservatives as offering stability and competence, and on the alternative (be it Labour or a more complex arrangement) being detoxified.
While the same pattern applied to some extent in Yeovil, there was probably more churn going on. UKIP stood a particularly incomplete slate in the local elections, meaning that they were bound to do better in the general election at the expense of the other parties – probably all three other parties in varying proportions. But there was still a net transfer from Lib Dem local to Conservative general going on under the churn as well as protest voters splitting their support between Lib Dem local and UKIP general.
Torbay was different again – while there may have been a LG-LD GE-Con vote in there, the dominant pattern looked like a tactical rally around the two main parties in the parliamentary contest – local government voters for Labour and Green supporting Adrian Sanders and – probably – many UKIP local voters going Conservative nationally.
There was a puzzling pattern in Bath that probably reflected a large number of cross-currents between local and national; but it was clear that even if one assumes most of the Independent vote in the local elections was right wing, there was a net transfer from Green to Conservative and UKIP for the general election, presumably via a LG-Green GE-LD vote and a LG-LD GE-Con vote.
The presence of a significant bit of the Conservative GE vote that split their support with the Lib Dems at local level is interesting, but there is a risk of drawing the wrong conclusions – that it was small or non-existent in Torbay and huge in Yeovil suggests that the Lib Dems positioning further to the right may not be the solution. It may well reflect that the Conservative general election vote in 2015 was highly conditional and uncertain – and possibly one that was pushed in the Tory direction by the repeated support of many Lib Dems for the broad claims made for Conservative government in terms of economic management and stability.
In Sheffield Hallam Nick Clegg did better than his local running mates, but Labour’s Oliver Coppard exceeded the local Labour score by even more. The Conservative and UKIP shares were, unusually, lower in the General Election, probably reflecting tactical voting for Clegg, but the Green local vote rallied more strongly than usual around Labour in the General Election. Comparing for a moment with Cambridge, another seat where Labour displaced the Tories as the challenge party:
Julian Huppert seems to have been able to attract significant support – nearly enough to win – from people who voted for other parties in the local elections – probably in his case a considerable number of Green supporters as well as some from Labour and the Conservatives.
Birmingham Yardley was an unusual Labour-facing analogue to the Eastleigh pattern; more usually in the Labour-facing seats the local vote seemed to collapse alongside the general election vote. The local candidates in Yardley were able to outscore the parliamentary vote by a clear margin; it does seem that there was a significant proportion of the LG-LD vote that could not support the party’s national stance and voted Labour (or another party) rather than do so.
While there is more analysis to do on all of this, even a cursory glance suggests some interesting patterns. The tables below show the aggregate results for each category of seat.
A further stage of the analysis will involve comparing local election performance in 2011-14 with local and general performance in 2015. It was true in some seats that the Lib Dems remained ahead in local election votes throughout the parliament but lost in 2015 (Hazel Grove and Birmingham Yardley being the most striking cases). Particularly in the Conservative-facing seats, traditional Liberal Democrat campaigning methods managed to keep the party viable in 30-40 per cent turnout local elections but failed when the electorate was swollen to general election levels.
Looking forward, the local elections suggest some possibilities for further targeting and some seats that should be taken off the boards. There seems some logic in regarding seats in which the Lib Dems won the local elections as being the most likely targets for regaining – particularly Bath, where there was no incumbency factor, and Eastbourne; there are at least pluralities of voters in the six seats concerned that were prepared to put a cross in the Lib Dem box even in 2015. The least encouraging results from the point of view of future Lib Dem gains are those where the local government vote was weaker than the parliamentary vote and the 2015 parliamentary vote benefited from an incumbency effect which will not be present next time.
Conclusion
By far the largest tranche of seats lost was due to the loss of ‘protest votes’ – something the party could ill afford to do without. At least 20 seats turned Conservative, not because of great transfers from the Lib Dems to the Tories (or even Labour), but because of major net transfers of support from the Lib Dems to a combination of both UKIP and the Greens (although this included some cross-voting). In Scotland, the party suffered heavily from the rise of the SNP juggernaut (as did Labour), but its vote actually held up surprisingly well in previously-held seats – even though it also collapsed elsewhere in Scotland – and it remains competitive in half of the 10 seats it lost to the SNP. In most Labour-facing seats, there was a noticeable Labour recovery (which was not found in many Lab-Con marginal outside London); and combined with the Lib Dem collapse, this proved devastating in every Labour-facing Lib Dem seat except Leeds North West and Sheffield Hallam, both of which had previously had Labour in third place. Where the party survived, it survived (in part) due to Conservative underperformance in the eight held seats, in two cases due to a UKIP intervention; but mostly, local factors and chance helped the seats to buck the national trend.
There is some evidence to support the existence of an ‘incumbency factor’ for popular local MPs, and comparing local and national election results shows just how large it could be in some seats; however, its extent was by no means uniform. All seven previously-held seats for which data exists where a new PPC was standing showed no GE ‘bounce’, and a minus (or zero) score over council election vote share – but exceptions in Scotland like Gordon flew in the face of this. Furthermore, at least six incumbent Lib Dem MPs seeking re-election showed negative ‘bounces’ over local election results: John Hemming in Birmingham Yardley, Mark Hunter in Cheadle, David Laws in Yeovil, Mike Thornton in Eastleigh, Steve Webb in Thornbury and Yate; and last but not least, Simon Wright in Norwich South – who had the worst result of any incumbent Lib Dem MP and came fourth. The incumbency effect should therefore not be overstated as a factor. Even in four of the six seats where the ‘incumbency effect’ was largest (Cambridge, Devon North, Lewes and Torbay), it was still not enough for the party to cling on to the parliamentary seat.
If a party polls 7.9% nationally, it can expect to win anything between fifteen seats and zero. In this light, holding onto eight seats can be seen as a triumph in light of the party’s dismal national showing, and indeed, the question should not be why the Lib Dems won so few seats, but how they managed to hold on to so many given the low support they attracted nationally; by contrast, losing nearly two-thirds of its once-sizable electoral coalition, which was consistently ±3% from 20% in the six previous general elections in 1987-2010, must surely rank as one of the great cock-ups of British electoral history.
If Lib Dems are to take back any or many of these seats, they will need a strong, persuasive, positive, distinctive anti-Establishment positioning; as the party’s centrist stance – characterised by its “Stronger economy, fairer society” slogan – did little to draw or retain its pre-existing coalition of voters, and the evidence overwhelmingly points to it having actively repelled many of them.
Lewis Baston is a psephologist and writer on politics, history and elections and past Director of Research at the Electoral Reform Society and research fellow at Democratic Audit. He has written several books of electoral geography as well as the biography of Reggie Maudling and a short book about scandals and corruption.
Dr Seth Thévoz is a political, cultural and social historian of Britain from 1800 on, and specialises in social networks, and detecting corruption among legislators. He is a member of the Federal Finance and Administration Committee, and in his spare time he is Honorary Librarian of the National Liberal Club.
An appendix to this article can be found here.