Earlier this week Nick Timothy and Tim Montgomerie had a short spat on Twitter about the reasons for the Conservative party’s failure to win a majority in June’s general election. Nick Timothy tweeted that:
‘We had a small majority and a divided cabinet, party and country’
Let that be your narrative. We had a small majority and a divided cabinet, party and country. We messed up the election…
— Nick Timothy (@NickJTimothy) September 26, 2017
When other things are equal, people are generally less likely to vote for parties that are seen as divided. If the Conservatives had been seen as divided in the run-up to the election, that might have hampered their chances of doing well. In fact, the Conservatives were seen as one of the most united parties. Whatever the reasons for the party’s failure to win seats, perceptions of division cannot be one of them.
I base this claim on data from the British Election Study (BES). For over thirty years, the BES has asked survey respondents whether they see the main parties as divided or united. The current edition of the BES – which has been running since substantially in advance of the 2015 election – is no different. In each ‘wave’ of the survey, voters are asked whether they see each party as very united, fairly united, neither united nor divided, fairly divided, or very divided. The graph below shows, for the campaign wave which ran from the 5th May to the eve of the election, the proportion of respondents who saw each of the four main UK-wide parties (Conservatives, Labour, the Liberal and Ukip) as very or fairly united, very or fairly divided, or neither united nor divided.
The Conservatives top the table: 61 per cent of respondents during the campaign saw them as very or fairly united, and only 23 per cent as very or fairly divided, for a net score of +38 per cent. The Liberal Democrats follow not very far behind (+32 per cent), with Labour bringing up the rear with a net score which was firmly negative (-65 per cent).
These figures reflect averages from a very large number of respondents (there were over 34,000 respondents involved in the campaign wave). Could this average conceal big differences which emerged only over the course of the campaign?
Daily figures for voter perceptions of how united each party was are shown in the second graph. Although the Conservatives came to be seen as less united over the course of the campaign, and Labour more united, these differences were not huge: they started with a net score of 40 per cent, and finished the campaign with a net score of 35 per cent. Far more impressive were the Labour gains, but then Labour started from a very low base.
I’ve described these changes as small: they are certainly small compared to the gap in perceptions before and after the campaign. The final graph shows responses to the same question plotted above, but this time asked of the post-campaign wave, fielded between the 9th and 23rd June. From a net score of almost plus forty, the party has fallen to a net score of -44. Only Ukip – then, as now, leaderless – was seen as more divided. The only party that was seen as more united than divided were the Liberal Democrats.
I’ve suggested that the Conservatives were seen as much more united than divided during the course of the campaign, and that this can’t account for their poor performance (relative to initial expectations). It is possible, of course, that parties can be divided in ways that aren’t visible to voters, and that these can have knock on consequences for electoral performance. Party divisions may make it more difficult to get canvassers out and about into the places where they are most needed. Still, if the figures from the BES post-election wave are broadly correct, the party may look back fondly on a period when its divisions were still by and large invisible to the public.