Monday, February 14, 2011

Has The Slump In Construction Been The Same As In Other Sectors?

Paul Krugman claims that the numbers shows that the increase in unemployment in the construction hasn't been greater than in other sectors. He does so by comparing the ratio of the number of unemployed in 2010 with the number of unemployed in 2007 for different sectors.

But this methodology is misleading for several reasons. First of all, if you double unemployment that is 3% initially the increase will be far smaller than if unemployment was 8% initially. In the first case, the more relevant increase in percentage points will be only 3 percentage point and in the second case the increase will be 8 percentage points. And the increase in the construction unemployment rate has been 13.2 percentage points, compared to 5 percentage points for the overall unemployment rate and just 2.8 percentage points for the sector education and health service.

And secondly, because some unemployed finds jobs in other sectors (thereby in many cases crowding out existing job seekers in that area and thereby increasing unemployment in that sector) or drop out of the work force. And with construction unemployment being higher than other sectors for a long time they are more likely to do so than workers in other sectors. Evidence of this can be found in employment statistics. While the number of formally unemployed that was previously construction workers rose 1.1 million, construction employment fell by 2 million (a 26.6% decline).

By contrast, while the number of formally unemployed that was previously working in the "Education and health services" rose by 668,000, the number of people employed in that sector rose by 1.2 million (a 6.5% increase).

In a related statistical note, the big decrease in construction employment means that the increase in the number of unemployed underestimate the increase in the unemployment rate.

When one sector has a 20.6% unemployment rate and another has a 5.8% unemployment rate and when even more astonishingly one sector see a 26.6% decline in employment and another see a 6.5% increase, there is clearly a strong element of structural change.