De-constructing the recovery
Is there anything fishy about Britain's recent GDP numbers? At least one economist thinks there might be. To be precise, he thinks there might be a problem with the figures for the construction sector, which has accounted for an unusually high share of Britain's recent growth.
I hesitate to get into this, having got embroiled in a similar debate about Britain's official output data this time last year (see my post Small difference: Britain's third-quarter GDP). You'll remember, at that time, many economists thought the economy was stronger than the official numbers implied. The numbers for the third quarter were considered especially dubious. Everyone was expecting the numbers to show positive growth - instead they showed the economy continuing to shrink.
At that time, Danny Gabay, of Fathom Consulting, was one of the few economists prepared to stand up for the ONS. The preliminary GDP data weren't perfect, he said - but they were better than anything else out there. Since then, the official estimate for growth in those three months has jumped around a bit, but the official story is still that the economy shrank by 0.3% in the third quarter of 2009. The first estimate was of a decline of 0.4%.
Perhaps the numbers will be revised up sharply in a year or so, as the critics suggest. The point is that Gabay is not usually one to knock the ONS. But the construction figures for the past six months have him a bit concerned. If he's right, that would suggest the recovery has not been quite as strong as we thought. Construction has accounted for 40% of the UK's economic growth in the past two quarters, even though it accounts for just 6% of the economy.
The figures show real construction output has been growing at an annualised rate of 27% since the beginning of April. A number of explanations have been offered for this: for example, companies had to catch up on the ground lost due to bad weather in the first quarter. Some have also pointed to a flurry of public construction projects in the months leading up to the election.
But we're not just talking catch-up. I'm not sure public investment can explain it either. After all, the figures show the construction sector making up all of the ground that it lost in the recession, in just those two quarters. After the last recession it took 11 years (see chart below).
Is that plausible? The man from Fathom is not so sure, for three reasons.
First, there's the surveys, like the PMI. These do suggest a sharp rebound this year, but they are still well below their long-term averages, and Gabay notes that the PMI has actually decelerated since April.
Second, and more curious, you can look at the labour market figures, to see whether this construction boom is leading to a similar jump in employment. The answer is not really. Instead the official numbers show each worker suddenly becoming enormously more productive.
Now, it's perfectly usual for productivity to rise sharply coming out of a recession, as output comes back faster than jobs. Manufacturing productivity is now 7.5% higher than a year ago (it fell sharply during the recession because, as I've discussed many times, output fell much further than employment). But the construction sector is streets ahead, with a 13% increase in output per head.
You have to wonder whether one of those figures - either the output, or the number of heads - might be slightly wrong. It's not beyond our wildest imaginings that construction companies might, shock, have people working for them that they don't tell the ONS about (or the tax man). But it's not clear why they would be doing this a lot more than they did, say, two years ago. As the chart below shows, these productivity figures are much stronger than in any recent boom.
Finally, there is the awkward fact that the ONS has just changed the way it measures construction output, with the introduction of a new, monthly, output survey. The new survey got its first outing in - wait for it - the second quarter of 2010, the same quarter in which construction output jumped by nearly 7%. The ONS clearly think the new series is reliable, or they wouldn't have introduced it. But, having not used it before, they can't say for the sure that it's not made a difference.
It's possible that Britain is in the middle of the biggest construction boom in recent memory. Given the lead-times on construction, you'd want it to lead the recovery rather than lag, and there are some enormous buildings going up in the City. But has the sector really make up all of the ground lost in the recession in just six months? I leave it for you to judge. I'd especially like to hear from people in the construction business, what things feel like to them.
In the meantime, we can take some comfort from this week's very strong manufacturing figures for October - output grew by 0.6%, the strongest figures since March. Happily, Britain's recovery is built on more than construction. It also seems to have some momentum - even if the middle of 2010 turns out to be slightly less marvellous than the ONS thought.