Free Lunch: Vote tallies and seat counts

Δημοσιεύθηκε: 28 Απριλίου 2015 - 11:05

First-past-the-post arithmetics

The advantage of the UK electoral system in a two-party world is exactly the same as its weakness in a multi-party world: small differences in vote share are amplified into large swings in parliamentary seats. That makes for large majorities and hence ease of governing when the electoral competition is largely between two parties. But when small parties are big enough (or concentrated enough) that their vote share, too, can produce big swings in the seat count, it makes for unpredictable elections and no single-party majority.

The long-term implication is that if the smaller parties are here to stay, so are hung parliaments. Better get used to coalition building. The short-term headache, for election junkies at least, is to make sense of what may happen on May 7. The permutations are so many that it is tempting to give up trying to map the link between poll results and the composition of the next government.

Here are the possibilities in the FT's latest forecast, based on a model from Election Forecast:

But very helpfully, along comes Simon Wren-Lewis with a clever trick to make this more manageable. He observes that the polling uncertainty around the small parties' projected seat count turns out to be smaller than for Labour and the Conservatives.

That means, on current polls, that the most significant uncertainty is how the roughly 550 seats the two main parties are likely to get will be divided between them. Here are the latest projected party vote counts - the Conservatives and Labour have 449 in total:

Moreover, it is quite clear how the smaller parties, with the exception of the Lib Dems, align themselves: they straightforwardly fall into a rightwing and a leftwing block. That allows Wren-Lewis to map the possible outcomes on to the single dimension of the Labour-Conservative seat split.

It's worth reading his entire post, but the essence is as follows. If the Conservatives get more than 310 seats (240 or less for Labour), the coalition continues or the Conservatives govern alone. If they get fewer than 285 and Labour more than 265, Labour governs through some form of understanding with the Scottish National party. Wren-Lewis discusses the case in between, where the Lib Dems decide, in a detailed follow-up post.

A Nobel on executive pay

When Jean Tirole won the Nobel memorial prize for economics last year, the awards committee cited his work on market power and regulation. Less well known is his criticism of the bonus culture. In an FT blog post, Andrew Smithers - himself a scourge of distorted executive incentives - brings attention to Tirole's analysis (with Roland Benabou) of how performance-related pay may destroy efficiency within companies and whole sectors. Competition makes the problem worse in their model.

When companies use performance-based bonuses to compete for the most talented workers, they end up tilting remuneration too much towards observable outcomes and away from those that are hard to measure but nonetheless matter - such as long-term risk. In such situations, a bonus cap can make all companies (and the economy at large) better off by helping focus workers' efforts on the right things.