BRACE yourself to hear the phrase ‘data dependent’ — a lot — over the next month. With this week’s release of the US Federal Reserve minutes reinforcing expectations of a rate rise, perhaps as soon as September, Fed officials are at pains to stress that any final decision will depend on the macroeconomic data; and thus be ‘data dependent’.
But as the Fed keeps tossing around those ‘D’ words, a certain irony hangs in the air. In a corner of the data world — productivity — the picture looks so baffling it is hard to see how anyone could depend on those numbers.
And while this data fog has not yet attracted much public debate, the mystery could complicate the Fed’s policy challenge — not least because it makes it hard to tell how fast the economy can grow before it needs more rate hikes.
To understand the issue, take a look at some numbers assembled by Alan Blinder, an economics professor at Princeton and a former vice-chairman of the Fed. Mr Blinder calculates that in the period between 1995 and 2010, productivity in the US economy grew on average by 2.6pc each year.
Central banking is an art, not a science, especially when the time comes to change course
That meant the potential trend growth rate in the US economy, or the speed at which the economy could grow sustainably ignoring capacity issues, was roughly 3pc. (The trend rate is usually calculated by adding population growth and productivity increases.) However, since 2010, overall average productivity increases have tumbled to just 0.65pc; indeed, over the past year private sector labour productivity, excluding farming, has grown by a paltry 0.3pc according to the quarterly data series.
This is peculiar on many levels. For one thing, the trend flies in the face of the popular belief that the western world is undergoing a technology boom. After all, the internet revolution has given us mobile phones with the computing power of 1970s rockets, while analysts such as the Oxford Martin school are now predicting that digitalisation has become so powerful that it will replace half of all US jobs in the next couple of decades.
In what adds to the sense of mystery, the Fed itself seems to be using a very different understanding of productivity. Recent projections from Fed policy board members, for example, imply they think potential ‘trend growth’ in the economy is around 2.15pc. That implies productivity growth of around 1.75pc.
While this week’s minutes revealed that the Fed has recently ‘trimmed projected rates of productivity gains and potential output growth’, it is far from clear how big this ‘trim’ is — or why the Fed’s numbers have recently been so at odds with statistics. “The Fed is clueless about the trend rate of growth of productivity,” Mr Blinder concludes. “Just like everyone else.”
One explanation may be that statisticians cannot keep pace with technological trends; they might be missing part of the output generated by, say, smartphones. Another possible problem is that the last decade’s credit boom made finance seem more productive than it really was before 2007, which makes the post-crisis data look far worse.
A further factor is that corporate investment has recently been weak, while there is a growing skills gap in America’s workforce. This would normally be expected to lower productivity.
Then there is another, even more intriguing related issue: a technology time lag. As Mr Blinder notes, if you look at the productivity figures in a wider context, there are two other notable points. First, America is not the only country where productivity has tumbled; this is seen in places such as the UK too. Secondly, this is not the first time such a swing has occurred: back in the 1970s and 1980s there was another slump after an earlier boom. This might be just a coincidence. But what is striking about the 1970s was that it was also a period of dramatic technological change; the onset of the computing era.
And while logic might suggest that innovation should boost productivity, the problem with new technology, as economists such as Andrew McAfee of MIT point out, is that it takes time for companies to harness. So, just as it took a couple of decades before computers raised US productivity trends, it might take time before the economy is truly boosted by today’s smartphones.
If this theory turns out to be correct (as I suspect it is), it suggests that eventually those productivity numbers should rise sharply. The problem, however, is that it could take several years. Until then, better keep watching the productivity numbers — if nothing else, because they show that central banking is an art, not a science, especially when the time comes to change course.
Published in Dawn, Economic & Business, August 24th, 2015
On a mobile phone? Get the Dawn Mobile App: Apple Store | Google Play
Dear visitor, the comments section is undergoing an overhaul and will return soon.