By contrast, though - as Alistair Darling admitted in his Pre-Budget report - GDP has been worse than expected. Back in March, the consensus was that GDP would fall 3.1 per cent this year. It now looks like being 4.5 per cent down on 2008.
As a matter of arithmetic, then, the big surprise of this recession is that productivity - GDP per worker - has slumped. In the year to Q3, the number of workforce jobs fell by 2.1 per cent, whilst GDP fell 5.1 per cent, so productivity is down 3 per cent. My chart shows that productivity has fallen more sharply in this recession than it did in the early 1990s one, and more protractedly than it did in the 80s recession*.
This is especially puzzling because one thing suggests the opposite should have been the case. A big story of this recession has been that firms have been unsure about the availability of credit, and therefore anxious to cut borrowing and raise cash. This is why inventories were slashed. But this should have caused firms to cut labour costs savagely as well. But in fact they haven't, in aggregate.
What's going on? Here are four possibilities:
1. "Employment has been underpinned by public sector jobs." True, public sector employment rose by 290,000 in the year to Q3. However, private sector employment fell by only 3.1 per cent in the year to Q3. This is less than the drop in private sector GDP. So we still have a puzzle.
2. "Workers have priced themselves into jobs." But the evidence for this is mixed. True, real wages fell in this recession whereas they didn't in the 1990s one. But they also fell in the 80s recession - and that didn't stop unemployment soaring. And there's little evidence that the demand for labour is so price-elastic as to explain much of the surprising strength of labour demand. After all, the lesson of the introduction of the National Minimum Wage is that labour demand isn't very price-responsive.
3. "Technology is different now." In the 1980s, a lot of labour was unskilled. Employers could therefore sack workers in the knowledge that they could be easily replaced if or when demand recovered. Today, though, workers are more skilled. Sacking these is dangerous; you might not be able to re-hire workers of similar calibre if you need to. So employers have had more incentive to hoard labour.
4. "There's been an adverse technology shock." Economists typically think of recessions as being caused by falling demand. There is, though, a minority view that recessions are instead due to falls in supply potential; it just becomes harder to produce things. A fall in productivity might be evidence that this has happened. Many mainstream economists are rude about this theory: how can it be, they ask, that the entire economy can simply forget how best to make things? But productivity shocks needn't be aggregate ones. Xavier Gabaix of Stern School of Business in New York has recently showed that difficulties in just a handful of firms can generate aggregate fluctuations. And research at the New York Fed has found that swings in the efficiency of investment - which can show up as falls in labour productivity - can be sources of recession.
There is a big difference between theories 3 and 4. Theory 3 implies that when demand picks up, productivity growth will soar, as workers who are under-employed in their jobs become busier. If this happens, then - unless wages rise as well - unit wage costs will fall either profit margins will expand or inflation fall. Either would be great for equities.
If, however, theory 4 is right, the recovery in output might be slow and firms will find it much harder to expand margins. Which is not so good for equities.
For what it's worth, my hunch favours 3. But I'm not 100 per cent confident.
* Today's figures don't explicitly report Q3 productivity. I've estimated it by dividing gross value added at basic prices by the number of workforce jobs. My measure of wages is not average earnings - this data began only in 1990 - but rather aggregate compensation of employees, divided by the number of jobs, and then deflated by the RPI excluding mortgages.
