Tim asks an intelligent and important question:

Which comes first? The trust that allows a high-tax high-benefit society? Or the high-tax high-benefit society that engenders the trust?

It's certainly true that equality, trust and economic success go together, but as Tim says, disentangling cause and effect are tricky.

This, I suspect, goes to show just how damned difficult the social sciences are.

Think of it this way. There are (at least) two mechanisms here.

1. The causal link between inequality and trust isn't necessarily linear. As Eric Uslaner has stressed, a big cause of distrust is segregation; people distrust people they have little contact with. This implies that small changes in income equality - the sort that don't change social relations between rich and poor - might not affect trust. But bigger changes, which lead to the rich living in gated "communities" and the poor in ghettos would reduce distrust. This means there might also be long time lags between rising inequality and distrust.

2.Trust can be cyclical. When trust is high, people have an incentive to exploit others' trust - by being conmen or welfare scroungers, say - and their bad behaviour leads to increasing distrust.But when distrust is very low, people have fewer opportunities to behave badly - because no-one trusts them with their money - and this leads to more trustworthy behaviour and then to greater levels of trust.

We can see the current popularity of tougher rules for welfare claimants as an example of mechanism 1; the poor, confined to housing estates, are segregated from others, which causes those others to fear they are "scrounging". It's possible that the continued segregation - exacerbated by housing benefit cuts - will strengthen this mechanism.

But eventually mechanism 2 might take over. As life on welfare becomes (even) less pleasant, we'll have (even) fewer "scroungers", and this might lead the median voter to become more supportive fo higher benefits. He might eventually figure: "I'm hearing less about 'scroungers' and welfare benefits seem mean. We should raise them."

In this way, cycles of trust can generate cycles of welfare benefits. We're on the downward part of the cycle now, but we might not remain so. The history of the UK welfare state is more one of fluctuating levels of provision than ever-increases or ever-decreases.

Herein, though, lies a problem. Both these mechanisms contain time lags of unknowable length. Also, we cannot see in advance the relative power of these mechanisms. This implies that we cannot know when the cycle will turn.

For me, this just confirms Jon Elster's wisdom. The social sciences are not a body of laws and deep parameters, but rather a ragbag of mechanisms, some obscure. We might be able to identify which mechanisms worked after the fact. But we cannot use them to make forecasts.