Earlier today, Tim Montgomerie, referring to a Times leader, tweeted:

(1) Record, unexpected jobs growth. (2) Massive reform of welfare incentives. Could the two be connected?

The answer could be: not much.

To see why, let's ask the question we must ask about all theories in the social sciences: what's the mechanism? Through what channel might "a tilting of the welfare system away from one that favours the dole towards one that favours work" reduce unemployment?

One possibility is that it encourages the unemployed to fill the jobs that are available rather than wait for the ideal job or lounge around the house.

If this were the case, we'd expect to see an inward shift of the Beveridge curve. We should see lower unemployment and fewer unfilled vacancies because the jobless would rather take the jobs that are available.

But this hasn't happened. My chart shows that since the Tories took office, the number of unfilled vacancies has risen. What's happened is a move along the Beveridge curve, rather than an inward shift in it. This corroborates micro-level evidence that the Help to Work scheme had only small effects in getting people into employment.

The big story since 2010 is not that the unemployed have been filling vacancies faster but rather that there's been an increase in the demand for labour.

Now, there are many reasons why the demand for labour has risen. But, in fairness to Tim, benefit reform might be one. Forcing the unemployed to look harder for work might have increased the effective supply of labour and so bid down wages which in turn has increased demand.

Personally, I suspect that whilst there's something in this, the power of this mechanism might be weak. For one thing, many of the unemployed want to work anyway, because for many of them joblessness is a source of great misery. For another, the minimum wage limits the extent to which wages can fall. And for a third, the price elasticity of demand for labour might not be very high, in part because the market is lumpy - it's probably less than unity (pdf) - which means that a given fall in wages won't create many jobs.

Very many things in the social sciences are true but not very much so. I suspect the idea that benefit reform has priced people into jobs is one such case.

But perhaps I'm wrong, and the reforms have worked as Tim thinks. Even if this is true, though, I'd keep quiet if I were he: does he really want to argue that the government is directly responsible for falling wages?