On reasonable projections

Alice has nicely set out how an election resulting from a non-confidence vote in a minority Parliament can generally only be expected when at least one party misreads its political prospects. (Of course, an election precipitated by the government alone such as the one in 2008 requires no such miscalculation.)

But in case anybody is wondering how it's possible for all parties to conclude they have something to gain from an election when that seems impossible on its face, take a look at the latest Nanos poll:

Nationally, the popularity of the Conservative Party declined by a single percentage point, to 39 per cent, with both the Liberals (28 per cent) and NDP (20 per cent) up one point from the month before. These numbers are well within the margin of error (3 per cent) and suggest little or no change in support for the three national parties.

In other words, all three of the national parties in Parliament are slightly above their 2008 vote share numbers, giving each the opportunity to reasonably foresee circumstances where it could gain seats. And the Bloc is less than a percentage point down from its 2008 vote share in Quebec, which combined with more vote splitting among the federalist parties will provide plausible scenarios where it could add seats as well.

Of course, the parties should also be able to recognize the risks involved in reading the opportunities and dangers incorrectly. (And of course nobody is about to make decisions based on a single commercial poll.) But particularly when the message for ages has been that nobody can expect to gain much out of an election, recent trends seem to offer everybody at least some reason for hope of improving their current standing - and that may well be enough to result in an election.