On my way back from the polling station, I got caught in a heavy shower. This reminded me that, from the perspective of instrumental rationality, voting is irrational: the cost of doing so (getting wet) far exceeds the benefit, the infinitesimally small chance that our single vote will make a difference.

However, millions of us today will ignore this thinking and turn out. This doesn't, I think, mean millions of us are so stupid we can't do basic cost-benefit calculations. It means instead that the instrumental perspective is a narrow one which ignores something important. But what?

One answer has been proposed by Andrew Gelman and Noah Kaplan (pdf). They say we have social preferences. When we vote, we don't just consider the benefit to ourselves but to others. And the tiny chance of a benefit to millions of people is big enough to outweigh the small cost of voting.

Another possibility - entirely consistent with that - is that we use what Robert Nozick has called (pdf) evidential expected utility (pdf). We think "if I vote then people who think like me are also likely to vote too which raises the chances of getting the candidate I want". Simon's right: collectively individually insignificant acts can be significant.

Both these theories are reasonable. But they run into a problem. It's hard to see how they can explain systematic differences in turnout. The young and poor are less likely to vote than the old and rich. Do they really have less social preferences? Are they less likely to use evidential expected utility - to be, in parlance of Newcomb's problem, one-boxers rather than two-boxers?

I doubt it. Instead, something else is happening. It's that there are many things we do not because they make sense in narrow cost-benefit terms, but because they symbolize the type of person we are. As Sam says, "voting isn't instrumental, aimed at affecting policy, it's expressive." It expresses an identity - who we are and (in my case and others) who we are not.

This, I think, contributes to explaining differential turnout. Older people were socialized into norms in which one did vote, often as an expression of class solidarity. With the decline of class-based voting, younger people have a weaker norm. And because none of our centrist managerialist parties try to speak for the worst-off, so many of the poor don't want to identify with them. As the IPPR's Matthew Lawrence says: "For many, democracy appears a game rigged in favour of the powerful and the well connected."

For me, this is another reason for voting Green. If the young and poor are to engage with mainstream politics - and it's moot whether the media-political establishment wants them to - then the parties must offer more than a shopping list for the median voter. They must embody a cause worth identifying with. And as Terry Eagleton says, the Greens are now the only party "unafraid of what George Bush Sr once called "the vision thing"". I have my doubts about that vision. But it represents a break from the managerialism of the main parties which has failed both to deliver good governance and to inspire the people. And I thus applaud it.