In my last post I noted that the US was abnormal in comparison to the other G7 countries. In all the other countries GDP fell by more than employment, so labour productivity fell. That's what normally happens in a recession. Okun's Law says so. But US GDP fell by less than employment, so labour productivity increased. The US broke Okun's Law. The US is abnormal.
What caused the US to be abnormal? There are any number of ways in which the US is different from the other G7 countries. It's hard to tell which one of those differences caused the US to break Okun's Law. So let's increase the sample size.
Spain is even more abnormal than the US. Ireland is a bit abnormal too. (Thanks to commenter David Fitzsimmons for the suggestion, and thanks very much to Stephen for the graphs.) What do those three countries have in common that caused all three to be abnormal? My guess: a big drop in construction employment.
Here is Stephen's graph for the G7 countries. It's the GDP/employment ratio. (You can see the separate graphs for GDP and employment on my previous post.)
US productivity increased. All the other G7 countries obeyed Okun's Law. Their productivity fell during the recession, and only began to recover as their employment began to recover. The US is abnormal. The US broke Okun's Law.
Now lets add Spain and Ireland to the mix (but take out Japan and Italy just to make space). Here are Stephen's graphs of: GDP; employment; and the GDP/employment ratio.
Ireland started out normal, with its GPD dropping faster than employment. Then as the recession deepened in 2009, it turned abnormal. GDP continued to fall, but employment fell faster. Productivity increased back to what it was before the recession.
All three countries broke Okun's Law. All three are abnormal.
What's going on?
Here's my guess. It's because all three countries had a big fall in construction. But what's so special about construction? My guess: construction workers have general human capital, not firm-specific human capital. Labour hoarding is what causes productivity to fall in a recession. And labour hoarding only makes sense when workers have firm-specific human capital.
Okun's Law is a purely empirical law. It says that, when a country goes into a recession, the percentage fall in GDP is typically two or three times as large as the percentage fall in employment. So labour productivity is pro-cyclical.
Okun's Law has always been a theoretical puzzle for Keynesian and Monetarist economists, or anyone who believes that recessions are caused by a fall in aggregate demand. If there are diminishing returns to labour, we ought to see productivity rise when employment falls. And if it's the lower-skilled workers who lose jobs disproportionately in a recession, which is what normally happens, that's even more reason why we would expect to see average productivity rise. But Okun's Law says it falls. How come?
The simplest plausible explanation has always been labour hoarding. When demand falls, and firms cut output because they can't sell as much, they don't need as many workers. But if they lay off the old workers, they might not still be around when demand returns to normal and the firm wants to hire them back. And it might be very costly to hire new workers instead. So you keep them on the payroll, even if they are not producing much.
The labour hoarding hypothesis makes sense: if firms believe the drop in demand is temporary; and if hiring new workers is costly. If both those conditions hold, keeping unneeded workers on the payroll will increase costs in the short run, but may reduce costs in the long run.
Why would hiring new workers be costly? There are costs of advertising and interviewing new workers. And the costs of having a new worker on the payroll who turns out to be unsatisfactory. But the biggest cost may be firm-specific training.
There's general human capital, which is useful at many firms, and firm-specific human capital, which is only useful at one firm.
If your workers have firm-specific human capital, and you lose them, and need to hire new workers when demand returns, it will be costly to train the new workers for the particular skills they will need in your firm. So you won't want to lose your existing workers, and you will keep them on the payroll even if you don't need them right now.
But if your workers have general human capital, it doesn't matter if you lose them. You can lay off all your bricklayers in a recession, and just hire new bricklayers when demand returns.
My guess is that the construction trades require a lot of general human capital and very little firm-specific human capital. Two building sites can swap bricklayers easily, with minimal re-training. "Welcome to your new job. The Porta-Potty is over there. Start laying bricks here". Am I right?
Or, it might be that firms think that construction work won't be coming back for a long time, because of the overhang of unsold houses. So it doesn't make sense to keep the workers on the payroll producing nothing for years. Is that plausible?
In particular, what happened to construction in Ireland in early 2009, when Ireland started to break Okun's Law? If my guess is right, then something big happened.
[Off-topic slightly: notice how the countries that had the worst recessions tended to be the ones where productivity increased? That doesn't look good for the theory that says recessions are caused by a negative productivity shock.]