Counting the unemployed

What with the huge deficit, I thought it’d be nice to save the government some money. The money I can save it is whatever it’s planning to spend on this:

The government has launched an investigation into the gap between the number of people out of work and those who are claiming unemployment benefit. It comes after data last month showed the jobless rate, under International Labour Organisation (ILO) rules, rose to 7.6% in the three months to May. At the same time, the rate of people claiming unemployment benefit in June was just 4.8%.

The Labour Force Survey interviews a large number of people to find out how many are looking for work. This ILO-backed method has consistently produced higher unemployment numbers than the claimant count of those receiving Jobseeker’s Allowance. Hardly surprising: it’s easy to say ‘yes’ in response to a survey, but harder to jump through the hoops required to qualify for JSA.

So, has the gap between the two sets of numbers changed over time? And in particular, has it changed very recently?

Yes and yes.

The graph below compares like with like: rolling three-month averages of the LFS and claimant count figures as a percentage of the working-age population. I’ve also included (measured on the right-hand scale) a comparison, dividing the claimant count by the LFS number to show what proportion of people who want work are actually signing on; the higher this proportion, the closer the two measures.

(Using calculations based on ONS data.)

As you’d expect, there’s an extremely strong correlation between the two measures of unemployment: they go up and down pretty much in tandem. But you may not have expected the almost equally strong correlation with the proportion of unemployed people who are signing on.

Certainly, some of the fall in this proportion that took place from 1994 to 2007 will have been due to various changes that tightened the eligibility criteria for JSA. And some of it will have been due to the relative erosion of the benefit’s value, making it less worth the effort of the bureaucracy for unemployed people who have some other means of temporary support.

But there has to be something else going on: how else to explain the rises in this proportion during the early 1990s and current recessions? (This recent change, please note, is the exact opposite of what the innumerate commentators have been fretting about. The problem that the government is investigating does not exist.)

I’d guess that when unemployment is high, people who are looking for work realise that it’s going to take longer to find it, and so they’re likelier to need the financial support that signing on can bring. But when the economy is strong and jobs plentiful, work is easier to find and so signing on for a little while is less important to people who aren’t in immediate danger of destitution.