On second choices

Alice's post on second-choice preferences in the 2008 election is well worth a read. But the most valuable part of the information collected through the Canada Election Study figures to lie in its ability to present some background to calculations along the lines of Malcolm's guest post on the retention rate required to make a non-compete agreement the least bit useful.

For those who aren't familiar, Malcolm concluded that based on the 2008 federal election results, we wouldn't see a non-compete agreement result in a majority government (or in any party exceeding the Cons' seat total) absent an overall net retention rate of 70% in an agreement involving the Libs, NDP and Greens. So let's look at a simplified and perhaps unduly generous net retention rate for types of vote shifts based on the CES data, calculated as follows:

((votes to non-compete partner as second choice) - (votes to Cons)) / total votes for original party

Using that formula, here are the actual net retention rates for various types of votes in the 2008 election:

Lib to NDP: 13%
Lib to Green: -10%
Lib to (NDP + Green): 28%
NDP to Lib: 17%
NDP to Green: -2%
NDP to (Lib + Green): 36%
Green to Lib: 19%
Green to NDP: 13%
Green to (Lib + NDP): 46%

Now, it's worth highlighting that the above doesn't factor in any question as to the likelihood of the vote switching as intended in a non-compete scenario. And the numbers for the three-party scenario are almost certainly on the high side, since they assume that a voter whose first two choices are both part of the non-compete will vote for the other partner as a third choice in the absence of any data to that effect.

But even granting every favourable assumption to a three-party non-compete agreement, its application to the 2008 vote totals would result in...effectively the same situation we have now, with the Cons holding the most seats of any party, and any alternative requiring the Bloc's support to win power.

So Alice's numbers look to offer yet another indication that the more important work to be done involves shifting baseline vote preferences, as well as laying the groundwork for the argument that the confidence of the House of Commons trumps one party's claim to the most seats - not pretending that a non-compete agreement will have any meaningful effect.