The point of an electoral college is to ensure a leader is acceptable to both MPs and party members, though Eaton highlights the potential tensions were a candidate not ahead in both sections, winning the leadership with Parliamentary support against a plurality of party members, is the favourite choice of members but not MPs, or indeed surges home with a large trade union lead while being narrowly behind in the other two sections.
To better display the idiosyncracies of the Labour system, here are some key figures:
- The vote of one MP is worth nearly 608 party members and 12,915 affiliated members.
- The vote of one party member is worth 21 affiliated members.
- An MP's vote is worth 0.12 per cent of the total electorate, a party member's vote is worth 0.0002 per cent and an affiliated member's vote is worth 0.00000943 per cent.
That seeks to capture the relative value of the ballot papers which go out. But the values of the different sections are different when they are counted, which may be what counts in the end. Each section is still one-third of the college, however many votes are cast or remain valid in later rounds. So lower turnout or more drop-outs in any section makes individual votes in that section more valuable.
In practice, this means that party member votes will probably be worth between four or five times those of affiliated members, rather than 20 times.
The question is, firstly, what turhout is in returning votes and, ultimately, what we could call "final round turnout": the proportion of votes which express a preference between the top two candidates, and so are still being counted in the decisive count.
MPs and MEPs will turn out at 100% or thereabouts. (It is unlikely that an acting leader or ex-leader will not cast a ballot to remain neutral even under anonymity; and no MPs plan to publicly abstain, as Neil Kinnock crucially did in the 1981 deputy leadership). 99%+ should express a preference between the last two candidates, excepting the possibility that a couple of Campaign Group MPs are entirely neutral between Milibands. (But perhaps not if a surprise candidate makes the run-off, and some supporters of an eliminated Miliband have not given 2nd/3rd preferences because they didn't think that possible. I have heard that there were a handful of MP "exhausted" ballots in the final 2007 deputy contest, with enough voting power in theory to switch the result from Harriet Harman to Alan Johnson in 2007).
Party membership turnout may well be around 70%. There is no fixed total yet, since new members can join until September 8th and get a vote. Membership turnout was 69.1% in the CLP section 1994 leadership contest, with 172356 valid votes cast in a membership which was then higher, at around 250,000. (Turnout was 77% in the Tory contest in 2005, and 64.1% in the last LibDem contest in 2007).
YouGov found that 9% of Abbott, Balls and Burnham supporters said "neither" at the end of July, when asked which Miliband they would prefer, with Ed Miliband 51-33 ahead of his brother, and 6% "not sure" at that point.
If we assume that a third of votes go to candidates placed 3rd, 4th and 5th, and rate of "exhausted" ballots were to approach 10% of these votes (or 3.3% of all votes cast), this would reduce "final round turnout" by around 2.5% of the total eligible membership, for example from 69% to 66.66%. This is probably most likely among a minority of Abbott voters, and the "exhausts" proportion could end up lower. (To the extent this occurs, it also means first preferences count a little more than at first glance).
Affiliated turnout is still anybody's guess. It was 19.1% in the 1994 election - around 770,000 votes of just over 4 million ballot papers issued, and 8% in the 2007 deputy leadership. The ratio was 4.5 votes to 1 between the affiliate and members' section. (Turnout also differs a lot within this section: Fabian turnout in 2007 was 50%, six times above the trade union average, in the deputy leadership; socialist societies may again find their turnout rates similar to party members' turnout; not union turnout).
Again, there will be some exhausted ballots and some think this could be a little higher in this section. If this were again to apply to 3-5% of ballots cast, it would still only reduce overall "final round turnout" among eligible voters by under 1%, say from 19% to 18% for example.