Police and Crime Commissioner elections – where the 18.5% turnout figure came from

Chris Game

Getting exciting, isn’t it? Just 78 days and 21 hours (at the time of typing) till polling stations open for the Police and Crime Commissioner (PCC) elections on 15th November. At least, that’s what Birmingham City Council newsroom’s dedicated website says –   It’s been running for nearly three weeks now and, given the dearth of information emanating from the Government, is well worth a visit.

It particularly is if you happen to be planning to set your alarm for polling day. There’s a competing countdown clock on the Get Out and Vote! website, set up to boost the participation of British Muslims in our national life, but it’s set 12 hours behind the Council’s, which, even allowing for the extra hour at the end of British Summer Time, seems odd. Unless it’s a tactic aimed at generating a last-minute voting surge and repeating the queuing embarrassment caused at several polling stations at the General Election. 

If so, I fear it’s seriously misconceived.  Voting, let alone queuing, seems likely to be at a premium. A couple of weeks ago, we heard an embarrassed, and embarrassing, Nick Herbert, Minister for Policing and Criminal Justice, repeatedly refuse to tell the BBC Today programme’s Evan Davis whether a turnout as low as 15% would be acceptable for this radical and controversial innovation. Evidently it would – any turnout at all, in the Minister’s view, representing greater democratic legitimacy than the present system of appointed police authorities. 

Davis’ 15% seemed to be plucked from the proverbial thin air, but we now have something apparently much more authoritative. The Electoral Reform Society (ERS) have done some sums and are asserting that the PCC elections “are set to have the lowest voter turnout of modern times – projected at 18.5%”. Brilliant – not ‘under 20%’, or 19%, or ‘around 18%’, but an eye-catchingly precise 18.5%. And, judging from the frequency with which the projection has been quoted, it’s worked – even though none of the mentions I’ve seen either explain or question just how the categorical claim was arrived at. I’m trusting, dear readers, that you may be a tad more curious.

In fact, the methodology is disarmingly simple: you think of a baseline number, then subtract stuff from it. The baseline figure chosen by the ERS is 34%, on the grounds that “recent local election turnouts are in this region”.  Surprisingly, considering how fundamental it is to the whole exercise, there is no further justification, yet it is certainly questionable. 

One difficulty is that the ‘region’ in which recent local turnouts have fallen is actually rather large. This year was calamitous – a 32% turnout in the English local elections taken as a whole.  It was also, however, the lowest overall percentage for 12 years.

Last year’s picture was significantly different. The overall average turnout across English authorities was around 43%, comprising all metropolitan boroughs (38%), and most of the unitaries (41%) and shire districts (44%). Birmingham and Coventry, 28.4% and 27% this year, both managed 37% in 2011, and these disparities between the two years were not exceptional.

I’m not suggesting that 2011 was more typical than 2012. Part of the reason it was ‘good’ was that it was the year in our four-year electoral cycle when the ‘all-out’ district and unitary councils are elected, and they consistently produce higher turnouts than those electing their members one-third at a time. All the English councils voting in 2012 elect by thirds, have elections in three years out of four, and, perhaps not surprisingly, have relatively lower turnouts.

All I suggest, then, is that 2012 was not typical, at least of the past decade. Yet, in choosing 34% as a baseline, ERS have picked a figure that, while 2% higher than 2012, is at least 2% and generally around 4% lower than any other aggregate figure in the past 10 years.

The remainder of the ERS projection involves estimating the percentage drop in turnout likely to be caused by three additional turnout variables, and subtracting these estimates from the 34% baseline.

First of the three is the fact that the PCC elections will take place in cheerless November, rather than what in most years ought to be the lustier month of May. Voter turnout in council by-elections has been shown to be statistically related to the number of hours polling stations are open in daylight, and therefore to sunset times. Studying over 4,000 by-elections held between 1983 and 1999, Professors Rallings and Thrasher of Plymouth University’s Local Elections Centre found a 6.6% average difference between turnouts in May by-elections (38.1%) and those in November (31.5%). Call it 6%, and the 34% drops down to 28%.

Secondly, there is the Government’s refusal to allocate state funding for mailshots, as in parliamentary elections, in which information about each candidate is posted out to voters. At up to £35 million it would be too expensive, say ministers. Instead, there will be an information pack from the Electoral Commission to all households, explaining about the elections, and a single national website, giving details of all candidates that will be posted free to those electors motivated to request them.


There is parliamentary election evidence that turnout can be boosted by up to a third when candidates receive mailings both from sitting MPs and their main challengers. There are no free mailings in local elections, so it could be argued that this factor has already been allowed for in choosing a local turnout figure for a baseline. The ERS, however, think it needs to be further adjusted, by a rather arbitrary 5.5% – so we’re down now to 22.5%.


Finally, there’s the absence of party political broadcasts. Derided as they often justifiably are, PPBs have been shown to be at least as effective as local campaigns in getting a party’s less committed supporters to drag their indolent butts along to the polling station. There will be no such mobiliser this time, but the effect is hard even to begin to estimate, so let’s just say a further 4% off the baseline. And so, ladies and gentlemen, we’re left with a projected turnout of 18.5%. It’s not rocket science, hardly even political science, but could you do any better?




Chris is a Visiting Lecturer at INLOGOV interested in the politics of local government; local elections, electoral reform and other electoral behaviour; party politics; political  leadership and management; member-officer relations; central-local relations; use of consumer and opinion research in local government; the modernisation agenda and the implementation of executive local government.