The NIESR forecasts (pdf) that unemployment will rise to almost 9% of the workforce later this year. Whilst I'm no fan of economic forecasts, there's a reason to fear it might be right. My chart shows it.

It shows that during the worst of the 2008-09 recession GDP fell far more than employment; although it was a severe recession for output, it was a relatively mild one for jobs.

We can (horribly!) roughly quantify this. If the relationship between GDP growth and employment growth had been the same since 2007Q4 as it was in the previous 20 years, the 4.3% drop in GDP since then would have caused a 3.3% fall in employment. But in fact, it fell only 1%. This means that employment is almost 700,000 higher than you'd expect, given what's happened to GDP.

Now, to some extent, this reflects a genuine slowdown in productivity, and to some extent labour hoarding. Insofar as it's the latter, we should expect two things. Some firms will respond to rising demand by increasing utilization of labour and unwinding this hoarding, rather than increasing hiring. And others, if they fail to get the recovery in demand they hope for, will trim their workforces.

On both counts, a rise in aggregate demand might not generate many net jobs.

There's a precedent for this. In the early 80s, GDP bottomed out in 1980Q4 and recovered thereafter, but it was not until early 1987 that employment returned to its 1980Q4 level.

Now, I'm not saying the jobs picture is that bad. But that episode is a warning - that even if we do get a nice recovery in output (which is a big if), jobs need not follow.