Paul Ormerod: How sticky is unemployment?

How sticky is unemployment? Will it take three years to fall? The views expressed by the new Bank of England Governor, Mark Carney, on interest rates and unemployment remain a hot topic. Interest rates will not be raised until unemployment falls below 7 per cent, a process he thinks will take three years.

The perception which many people have of unemployment is indeed that it is rather sticky. Once it has risen, it takes a long time to come back down. Certainly, from a spatial dimension, the general view is correct. The average rate of unemployment in, say, the North of England or Wales has been higher than that of the South East for many years.

The same phenomenon is observed within each of the regions of the UK. Local areas where unemployment becomes high relative to the rest of the region then seem to be stuck with their unenviable position for many years.

Comparing unemployment rates at a local level over long periods of time is a fraught process, not least because local authorities are frequently reorganised, renamed and their boundaries altered. But the Office for National Statistics does have a consistent set of data over the 1990-2010 period.

So in 1990, for example, Liverpool had the highest unemployment rate in the North West. Twenty years later, it was still in top spot in this particular table. The same was true of Middlesborough in the North East, Hull in Yorkshire and Great Yarmouth in East Anglia.

Elsewhere, the dubious honour of having the region's highest unemployment rate did change over time, but not by much. In the West Midlands, Birmingham and Wolverhampton simply swapped the number one and two positions between 1990 and 2010. Even in the South East the rankings were pretty static. Hastings had the seventh highest rate in 1990s and was top in 2010. The Isle of Wight fell from first to third.

But when we look at unemployment over time, a completely different picture emerges.

Unemployment is now 7.8 per cent, and the Governor thinks it might come down to 7 per cent in three years' time, a difference, obviously, of 0.8 percentage points. Of course, it might not fall at all, but the evidence for real momentum in the economy is growing rapidly.

Once unemployment does start to come down, what does history tell us? Going right back to the late 19th century, there are 40 separate instances of unemployment falling by 0.8 per cent. In 32 of these, it took just a single year to accomplish. In another seven, it took two years. The only instance, in well over a century, of unemployment taking three years to be reduced by 0.8 percentage points is 1984-87.

Of course, the economy is not governed by physical laws which are fixed for all time, and history can be misleading. All we can say is that it would be very unusual, but not entirely unprecedented, for a 0.8 per cent drop in unemployment to take three years. A betting person would have to say that the odds were against.