My questions are for Professor Jansen.
I want to start by saying that I am no particular great fan of the first past the post system. In fact, I am the author of a motion adopted in the last Parliament to change the way we elect our Speaker from a first past the post system. It's actually a process in which you drop people off the ballot one at a time to a preferential ballot.
Having said that, first past the post is not the best of systems I can imagine, but it's also not the worst. I would define the worst system as that which has a predicable outcome in terms of which parties are winners or losers, that effectively allows—albeit not in every election into the future—the next election to be predictably affected, at least in part to be rigged, by choosing a particular system. I don't think that that particular sin can be laid at the feet of any STV model or of MMP, but I do think that the alternative vote system, single-member preferential votes, may have that problem.
I'm turning to a paper that you co-authored with Peter McCormick that was published on November 30 of last year in which you point out something that others have pointed to as well, that had the 2015 election been conducted using the alternative vote and everybody had voted the same way they actually voted, with the same preferences and the same second preferences, the Liberals would have gone from winning 184 seats to winning 205 seats. Interestingly, in the 2011 election, in which we know the Liberals got less than half the votes under the current system than they did in 2015, the Liberals also would have benefited.
I don't know if you've done any further research into prior elections—2008 and 2006—to see whether it is a consistent pattern or not. Let me ask that question as a starting point.