Gaming a Tory leadership election

It was a lot easier when he stood.
It was a lot easier when he stood.
A Facebook discussion I was in the other day ended up talking about the mechanics of Tory leadership elections, and it prompted a few thoughts. Just to be clear, these are all about electoral strategy for candidates in that putative election, not about their policies or personalities except in as much as they might influence their strategy.

A leadership election is a two-stage process. In the first round, MPs nominate candidates and then a series of eliminative ballots are held. The candidate with the least votes in each ballot drops out until only two candidates remain. Those two then go to a ballot of the party membership which decides the victor. If only two candidates are nominated, the process jumps straight to the membership ballot, if only one candidate is nominated (as happened with Michael Howard in 2003) they’re elected unopposed. Another important point to note is that there’s no provision for candidates to enter the race after the initial close of nominations – despite media speculation, the rules don’t allow for a stalking horse election.

Even without stalking horses, there’s still plenty of scope for strategy within the initial stage of the process. Candidates are not only concerned about getting themselves into the membership ballot but also who they’ll face while they’re there. This can be seen in the final MP ballot of the 2001 election where several of Iain Duncan Smith’s supporters reportedly backed Ken Clarke in an effort to ensure that it was Clarke, and not Michael Portillo, who Duncan Smith would face in the membership ballot. (It was perhaps a foretaste of his abilities as a leader that the scheme came close to a horrendous backfire as enough of them switched to Clarke that he only beat Portillo by a single vote)

The interesting effect of this system is that while they can’t end up with someone supported by a small group of MPs become their leader, it is possible to become leader if you can get a third plus one of the Conservative MPs to support you. With current numbers, that’s 111 MPs. If you can rely on that many supporting you, there’s no way that you can be stopped from getting into the membership ballot. Every vote short of that target makes it easier for your opponents to co-ordinate their strategy and block you.

This presents us with an interesting situation if we have a candidate who only has limited popularity with the MPs, but is popular with the membership. Assuming that candidate can persuade around a third of the MPs to back them, the other challengers face three options: they can try and coordinate their voters to exclude the other candidate from the membership ballot; they can fight it out between them for the remaining two-thirds of the electorate and see who does best; or they can agree to rally behind one candidate. The latter option would be accepting that the candidate with membership support would be on the membership ballot, but would ensure that his rival is seen as the clear choice of the MPs in the hope members would react positively to a candidate with clear Parliamentary support.

To illustrate this, assume a contest has got down to the final three candidates: A, B and C. A and B both believe that C is more popular with the membership than they are, so would prefer them not to face the membership. Both A and B would also prefer the other to C given the chance, and think they would have the chance to beat them in the membership ballot. Their best course of action depends on how popular they think C is amongst the MPs. If they think C has the support of less than a hundred MPs, it makes sense for them to keep competing with each other as both are still likely to beat C and make it to the membership ballot. If C is more popular, but still short of 111 MPs, then there is an incentive for them to co-ordinate their voters so that both of them still get more than C. If, however, they’re sure that C will get 111 or more MPs supporting them, then the incentive becomes to pick one of A or B to give them a resounding victory in the final MP ballot and go to the members as the clear choice of the Parliamentary party, in the hope that will help them beat C.

Where this gets interesting is that these courses of action give C an incentive to make their support look smaller than it is. If we assume there have been more than three candidates, and there have been other MP ballots before, it’s in C’s interest to get enough support to make it through to the final three and no more. The further A and B believe C is from having 111 MPs backing them, the less incentive there is for A and B to co-ordinate to stop C. C thus has an incentive to hide their real number of supporters until the final round in order to create their best scenario for winning: getting themselves on the membership ballot without a strong ‘unity’ candidate against them.

In other words, when the next Tory leadership election comes around, expect there to be lots of shenanigans and behind-the-scenes manoeuvring where the actual vote tallies may not reflect the real support candidates have. It’ll be fun to watch, if you can forget that whoever emerges from it all will likely be leading the country afterwards.

Is aiming for a brokered convention the only way Republicans can stop Trump?

The 1952 Democratic Convention - the last major brokered convention
The 1952 Democratic Convention – the last major brokered convention
If there’s one thing that gets US political geeks fired up, it’s the prospect of there being a brokered convention to choose a party’s Presidential nominee. Every four years, professional and amateur pundits look over the field of candidates and proclaim that there’s no way someone is going to get 50% of the delegates and seal the nomination, leading to a brokered convention where the nominee only emerges after a series of ballots. Every four years, these predictions are then scattered to the winds as each party manages to find a nominee who can do just that, and the pundits go back to watching their West Wing series 6 DVDs to get their brokered convention fix.

Hopes were high that this year’s massive Republican field would finally lead to a brokered convention. Surely having so many candidates with so many varied appeals to the electorate would prevent any front runner emerging, leaving the final decision to be made in Cleveland in July where headline writers were waiting to unleash multiple variations on ‘the mistake on the lake‘. Normal service, however, appeared to have resumed with Donald Trump rising from the field and winning a series of primary victories including taking the most states in yesterday’s Super Tuesday primaries.

Unlike his predecessors who emerged from the field to claim a hotly contested nomination (Mitt Romney and John McCain are the most recent examples), the rise of Trump has not been welcomed by the Republican leadership, with various parts of the Republican elite looking to find a candidate to stop him. While there are three remaining anti-Trump candidates in the race – Ted Cruz, Marco Rubio and John Kasich – they face a collective action problem where all would agree that a single non-Trump candidate would be best for the party, but all believe that the other two should drop out and back them.

This problem is compounded by the structure of the Republican primaries. Up to now, delegate allocation in them has been principally proportional according to the share of the vote received, so while Trump has won the most contests, he doesn’t hold a commanding lead in total delegates. (Electoral Vote currently give Trump 332 of the 1237 needed for the nomination while Cruz has 221, Rubio 122, Kasich 27 and Ben Carson 8) From March 15th, however, most states allocate delegates by a winner-takes-all method. As repeated British elections have shown, getting around 40% of the vote when your opponents are split is a very good way to take advantage of a winner-takes-all system. If no single ‘Stop Trump’ candidate emerges by then, the conventional wisdom goes, Trump will be able to mop up swathes of delegates by exploiting his rivals’ division and secure himself the nomination.

There is perhaps a way for the Republicans to avoid a Trump nomination from this position, but it would require them to embrace the idea of a brokered convention occurring. If none of Cruz, Rubio or Kasich can be persuaded to withdraw, they could instead agree a non-aggression and non-competition pact instead. Previous results and polling show that each of them has the potential to defeat Trump in different states, so the best option for them is to tacitly endorse Stop Trump voting. For instance, just as Cruz was able to defeat Trump in his home state of Texas, Rubio can beat him in Florida and Kasich in Ohio, if the other two aren’t competing for their votes there. In states where the delegate allocation is winner-takes-all, it makes sense for them to identify which of the three is the best placed to defeat Trump and leave the field clear for them, effectively dividing the remaining states between the three of them. It won’t give any of them the nomination before the convention – there doesn’t appear to be a strategy for that for any of them – but it keeps it from Trump.

Sure, there’s a big downside (for the Republicans, at least) in the party not having a nominee until the end of July, giving Hillary Clinton the opportunity to effectively campaign without opposition once she seals the Democratic nomination. It feels to me, though, that the choice for the Republicans is now that they either accept Trump as their candidate or plan for a brokered convention as the only way to stop him. A single Stop Trump candidate seems unlikely to emerge through the primary process, but combined action from his rivals can at least keep the nomination away from him even if they individually won’t benefit from doing so.

Then again, the Republican Party might just spend the next few months continuing to tear itself apart to the amusement of all those outside it, further confirming my theory that the corporate interest most benefiting from this year’s election is Big Popcorn.

Early thoughts on the Canadian election results

canadaresultWell, I got the number of Conservative seats almost right – right on close to 100, just wrong on which side of it they’d fall. For the other two, though, I seem to have got it quite wrong: underestimating the size of the Liberal surge and underestimating how far the NDP would fall. But anyway here’s some early thoughts:

There does appear to have been a decent amount of strategic voting (see here for a view from inside Canada) against the Conservatives. They actually outperformed the final polling projections in terms of the percentage of the vote they got, but underperformed in terms of the seats they won.

Liberals and NDP appear to have benefited from this in different ways. The Liberals have swept up a huge number of seats, gaining from both Conservatives and NDP, with NDP switchers giving them more of the former than they’d expect. The NDP appear to have limited their losses thanks to Liberals switching in seats they weren’t going to win. In some seats, the Liberals have steamrollered the NDP from second place, or jumped them to go from third to first to take a seat from the Conservatives, but when the Liberals were out of the running in an NDP-held seat, their voters seem to have kept a few NDP MPs in place where the Conservatives were the leading opposition.

The size of the Liberal victory is worth pointing out too, giving from how far down they’ve come in a single election. They’ve increased their number of seats fivefold (they’d have won 36 in 2011 on the new boundaries, and won 184 this time) and moved straight from third place into majority government. Yes, they’re an historic major party in Canada and 2011 was a frakishly bad result for them, so it’s not quite a shock insurgency, but I’m still struggling to think of another party that has made such gains in a single election. Then again, the volatility of Canadian electors and their willingness to shift dramatically during election campaigns is already a bit of an outlier, so perhaps this is to be expected given their political culture?

One interesting area of comparison between Canada and the UK could be the contrasting election experience of Justin Trudeau and Ed Miliband. Both were subjected to sustained criticism of their credibility before the election (Conservatives portrayed Trudeau as ‘just not ready’) but Trudeau appears to have turned that completely around, while Miliband was never able to. Was it just a case of Conservatives making expectations so low that Trudeau was able to easily surpass them, or was there something else there?

Finally, I’m sure the Trudeau name helped Justin, but I want to see polling to see just how important it was compared to the ‘not Michael Ignatieff’ factor. That, I think, could be a crucial distinction.

Canada and strategic voting

poll-tracker-seat-projectionsI’ve written here a couple of times about the upcoming Canadian general election, and now it’s finally here. Canadians are voting today for the 338 members of their new Parliament after a rather long election campaign, though unlike ours there’s been plenty of variation in the polls during the campaign (with the same level of doubt as to whether that will reflect the result). The Liberals have kept, and possibly extended, their lead over the last week of the campaign, while the Conservatives have wavered up and down and the NDP continued to slip.

(My finger in the air prediction? Liberals just over 150 seats, Conservatives just above 100 and NDP about 75. I reserve the right to proclaim my genius tomorrow if this is right, or completely ignore it if not.)

Beyond the result itself, what I’ll be most interested to see is if strategic voting (what we in the UK call tactical voting) has played much of a part in the results. From looking at the projected results for each riding from ThreeHundredEight, it’s clear that there are a lot of seats that are currently projected to be won by the Conservatives that could switch if enough voters switch to whichever of the Liberals and the NDP are best placed to defeat them, though the question is whether voters mutual antipathy to continued Conservative government is enough to make them switch, especially when the NDP and Liberals have been attacking each other quite heavily during the campaign. It’ll be an interesting test of what effect party signalling can have on the potential for voters switching, especially when there has been an organised movement to try and encourage it.

Having just written a dissertation in which it factors quite heavily, I’m obviously quite interested in the theory and practice of strategic/tactical voting, but I’m also interested in the names we use to describe it. On this side of the Atlantic, we use tactical voting, but in Canada (and the American literature on the subject), strategic voting is the term of choice. While the two terms are used interchangeably, I’m wondering two things: first, if there are potentially different phenomena at work that can be described using the two different terms, and secondly, if people react differently if asked to vote ‘strategically’ rather than ‘tactically’.

To take the second point first, as it could be just a personal foible, do people see the contrast between ‘strategic’ and ‘tactical’ as noteworthy? If you’re the sort of person who might use your vote for different parties in order to achieve an end would you react better to exhortations to vote strategically, or vote tactically? It doesn’t necessarily affect the votes themselves, but I think it’s interesting to consider the way in which the arguments are framed.

On the first point, however, I think there is possibly a case for seeing two different processes at work. These are still rough definitions, but perhaps we can best see tactical voting as the voting patterns in a single contest, while strategic voting (or perhaps strategic coordination) is the process that provides the framework for tactical voting. For instance, if a large number of voters in a particular constituency decide that they don’t like their MP and vote for whoever can defeat him, without the competing opposition parties or any formal structures encouraging it, there’s a high level of tactical voting going on, but little strategic voting/coordination taking place. Alternatively in a situation where parties were working together and encouraging vote switching, but little actually took place in the ground, there’d be a high level of strategic voting/coordination taking place, but little or no tactical voting. The question is whether that has to be a wider strategy in place for tactical voting to have any more than a limited local effect, or how much are voters capable of voting to achieve instrumental ends without coordinated strategic input?

Canada might provide some interesting data for that question in this election – is voter antipathy to Harper and the Conservatives sufficient on its own to motivate local tactical switching on a large scale, or will the ongoing Liberal-NDP rivalry (and lack of coordination) mute the effects of it? Hopefully, there’ll be some sort of answer amidst tomorrow’s results.