Several things stand out:
- The turnout was actually higher in 2014 than in 2012 by 2,800* votes.
- The Labour vote was only 500* votes down on 2012.
- There was no Liberal Democrat candidate in 2014 but in 2012 there had been 10,000* Lib Dem votes. By that time the 2010 general election Lib Dem surge (“Cleggmania”) was in the distant past so these were probably long-term Lib Dem supporters.
- The English Democrats, a right-wing party of no national significance but big in South Yorkshire, lost 14,500* votes, the Conservatives lost 2,500* votes from 2012 levels.
- UKIP increased its votes by about 30,000*.
But it’s not obvious what was happening underneath. Probably some erstwhile English Democrats and Conservatives moved to UKIP and the slight increase in the turnout was mostly to their benefit as well - about 20,000* votes in all. Where did the other 10,000* come from? At a guess, many came from Labour but their departure was compensated for on this occasion by the arrival of former Lib Dems.
Which raises an interesting question – what would have happened if there had been a Lib Dem PCC candidate? As a guess, something like this, which assumes (middle column) that 5,000 votes go back to the Lib Dems from Labour and 2,500 from both the Conservatives and UKIP. The numbers don’t matter too much because the key point is that almost certainly Labour would not have had more than 50% of the vote in the first round:
In these circumstances the second vote system used for PCC elections would have come into effect. Where the first preference vote had been for Conservative, Lib Dem or English Democrat and the second preference was for Labour or UKIP (with the two largest vote shares), the votes would be transferred.
There would then have been a second count of the Labour and UKIP votes. A similar argument to that above can be used to estimate what might have happened, assuming that all the voters had expressed second preferences, which they almost certainly didn’t**. It is worth considering what might have been the best case for UKIP. They might have picked up all the Conservative and all the English Democrat votes as second preferences. If they had also acquired all the Lib Dem second votes they would have won, but more probably, as above, at least 50% of the Lib Dems would probably have gone to Labour and 2,500* might have been ‘wasted’ on the Conservatives or not expressed any second preference at all. Even if 2,500* Lib Dems had voted UKIP as second choice, the second round would have given Labour its majority. It’s therefore not surprising that in the real ballot UKIP, even when the Labour first round margin over 50% was wafer-thin, didn’t ask for a recount.
But the margin of 29 wouldn’t have existed if there had been a Lib Dem candidate in the first round. I have no idea why this was the case, but, if nothing else, this PCC by-election offers an insight into how useful an informal Labour Lib Dem pact could be in normal parliamentary elections, even though these are “first-past-the–post”. If the Lib Dems were to soft-pedal their campaigning in Northern constituencies and concentrate their resources on holding onto their existing seats, particularly in SW England where, if Labour were to do the same in return, there would almost certainly be benefits to both. It would reduce UKIP’s impact on Labour in the North and the Conservative threat to the Lib Dems in the South West.
* Rounded figures.
** For example, in the 2012 Avon & Somerset PCC election second round, only 56% of the votes cast for the eliminated candidates indicated a second preference for one or other of the two candidates remaining.