Since the start of the recession, the output per worker in the U.K. has fallen by 3 per cent. This is extraordinary. In a recent speech, Ben Broadbent, a member of the Bank of England's Monetary Policy Committee, noted that if the pre-crisis relationship between output and employment still held, the number of jobs would have fallen by 8 per cent since mid-2007. Yet employment has risen since then. It has also apparently risen since early 2010, despite stagnant output and falling public sector employment.
Is productivity falling, instead of rising, as one would expect?
One explanation could be a big shift to part-time work. In a recent article, Joe Grice, chief economist of the Office for National Statistics, showed that the number of hours worked had fallen more than the number in employment since early 2008 - but only by 2 per cent. Thus, even output per hour has fallen.
Another possibility could be that output is increasingly under-recorded or the number of workers increasingly over-recorded. On this Mr. Grice stated that "nothing has come to light that would lead to major concern about the reliability of either the GDP or labour market statistics". Indeed, it is quite hard to imagine what could produce such a growing under-recording of GDP or over-recording of employment.
A further reason for believing that the collapse in productivity growth is real is that much the same has happened to Germany, France and Italy. Indeed, after a surge in 2009 and 2010, productivity growth has collapsed in the U.S., too.
If it is true, why has it happened? I can imagine three plausible explanations: cyclical labour hoarding; substitution of labour for capital; and huge obstacles to the efficient deployment of capital.
Labour hoarding is surely part of the story. Corporate finances are strong, which makes the hoarding affordable - and indeed, desirable in the medium term. Moreover, real wages have been falling, making it cheaper to retain workers. Yet, though this must be part of the story, it cannot be all of it. It does not explain why employment has recently been rising. Moreover, as the stagnation continues, hoarding must make less sense. Yet employment stays robust.
Falling real wages also justify substituting labour for capital. I suspect this is part of what has happened. Yet it is hard to believe it explains much of a shortfall of 12 per cent in output per worker, relative to the pre-crisis trend.
Mr. Broadbent suggests a third cause: misallocation of capital due to a defective financial system. The decline in the rate of overall investment is not big enough to explain the productivity puzzle. But he notes that prices and profitability have diverged across sectors to a greater extent than normal. Yet this has not triggered a large reallocation of capital to high-return activities. As a result, "some firms are kept in business . . . despite making relatively low returns. Others, able to expand but unable to obtain the finance to do so, are forced to substitute labour for capital".
The problem with this argument is not in direction, but in degree. If one is to accept the argument, one must believe that so many high-return projects have been languishing - unfinanced - that productivity growth has halted. I do not buy it.
Labour hoarding, substitution of labour for capital and failings in finance must explain parts of the puzzle. But we do not understand what has happened.
The most important question, however, is always how to react to the extent of the uncertainty. I suggest three reasonable responses.
First, it would be wrong to assume that the capacity has been lost forever. This could be a dangerous, self-fulfilling prophecy. As Martin Beck of Capital Economics notes, historical experience suggests that economies may make up lost output.
Second, policies need to tackle both supply and demand. That is why efforts to improve the operation of the financial system make sense. I have proposed that the government should insure the tail risk on bank lending to small and medium enterprises. Similarly, the case for increased public investment in productive infrastructure is overwhelming today, particularly given low interest rates.
Finally, the country needs an ambitious macroeconomic target, but one that also caps inflation. As my colleague, Chris Giles, notes, the obvious one - in an environment of such uncertainty - is for nominal GDP. In the second quarter of this year, nominal GDP was more than 10 per cent below its pre-crisis trend, despite the big inflation overshoots. If the Bank were told to make up some of the lost ground, together with subsequent growth at up to 5 per cent a year, inflation might overshoot. But there might also be a surprising buoyancy of output and a recovery of at least some of the lost ground on productivity. It would surely be better to try and fail than to be sure of failing, by not trying.