The musical chairs model says that shocks to nominal GDP—or total spending in the economy—drive unemployment. When nominal GDP falls, there is no longer enough spending to sustain the same number of jobs unless wages fall. Because wages are slow to adjust, unemployment rises instead. In Mr Sumner’s words:

If you stop the music and pull a couple of chairs away (lower NGDP), a few of the contestants will be sitting on the floor (unemployment).

To illustrate this, Sumner uses American data to show a close correlation between unemployment on the one hand, and the ratio of the hourly wage to GDP per person on the other (which I’ll call the “musical chairs ratio”).  I’ve reproduced his chart below:

I was interested in how this model performs using British data, and I’ve plotted the result below.* There is certainly a strong correlation between the two series, most notably in 2008 when unemployment spiked. But it is striking for how long the musical chairs ratio declines while unemployment stays stubbornly high. By the time unemployment at last dips below 8% in 2013, the musical chairs ratio is nearly back at its pre-crisis level. There were plenty of chairs, yet many workers remained on the floor.

Why did unemployment remain stubbornly high? One explanation could be a negative supply-side shock: commodity prices increased sharply in 2011, which might have caused additional structural unemployment. Another could be the swathes of public sector redundancies overseen by the coalition government. If public sector workers find it hard to move to the private sector, this might interfere with the predictions of the model.

*”Britmouse” did a similar thing two years ago. Since then, the national accounts have been revised.