I’ve started the week £50 better off thanks to a canny spread bet I had placed on Labour’s likely performance at the next general election. A few months ago, anticipating a “Brown bounce” once the chancellor moved into Number 10, I placed a bet “buying” Labour seats at 283, which meant that for every seat over that number that Labour gained at the next election, I would win my stake of £7 (Mike Smithson over at politicalbetting.com has a rather more eloquent explanation).

The good thing about spread markets is that you don’t need to wait for the outcome of the event in question to cash in—you can profit simply by taking advantage of the movement of the market. This is why I was able to make money from the Brown bounce without having any particular view on how Labour is likely to perform at the ballot box. Once the effect of the bounce had become clear in published opinion polls, the markets started to move in Labour’s favour, and I was able to “sell” my bet at 290, making myself a tidy profit of £49 (290-283=7, multiplied by my £7 stake).

Of course, had I turned out to be wrong about the Brown bounce and had Labour’s poll ratings continued to decline, I would have lost £7 for every seat under 283 I sold the bet at (though due to loss aversion I would probably have held off selling until the election itself, with possibly catastrophic results). This is one reason why spread betting is not for the faint-hearted—as I discovered when I made a disastrous move into cricket spreads during England’s test series against Pakistan last year. But I have done well out of political betting—in addition to my Brown bounty, I made a very tidy packet betting on Blair’s departure date after the 2005 election, when the reduction in Labour’s majority seemed to persuade punters that the PM would be out in little over a year—which suggests that a combination of knowledge, canniness and timing can be profitable in betting markets.

The more interesting question is whether betting markets are of any value in making predictions, in a “wisdom of crowds” sense. A gaggle of prominent American economists and political scientists recently published a statement calling for restrictions on prediction markets to be loosened, arguing that greater use of markets that yield payments for correct predictions would “substantially improve decision-making in the private and public sectors.” They point to the Iowa Electronic Market, which since 1988 has apparently predicted the results of US presidential elections better than opinion polls 75 per cent of the time.