The debate about the size of the output gap (the difference between what output is now and what it could be without generating accelerating inflation) is taking place in most countries as we struggle to emerge from the Great Recession. This post is about how policy makers handle that uncertainty. If you think you are pretty sure what the size of the output gap is, you need read no further.
Let me try and crystallise this uncertainty in a simple binary fashion. One possibility is that since the recession innovation and technical progress (lets call this ‘underlying productivity’) has not grown as fast as we might have expected based on pre-recession trends. We have two major pieces of evidence for this. The first is survey data where firms are asked how much spare capacity they have. If underlying productivity had continued at pre-recession pace, we would expect them to be reporting more spare capacity than they currently are. The second is the behaviour of inflation: if the output gap was large and negative we would expect inflation to be falling faster than it is. (A nice summary is provided by Greg Ip.)
The other possibility is that supply is somehow following demand. Now the concept that economists normally have of supply does not do this: we think of production functions with capital, labour and technologically determined productivity. There may be some endogeneity here: less demand leads to less investment, which reduces capital a bit, and if innovation has to be embodied in new capital this can slow technical progress. Perhaps in recessions where wages are low, people want to work less. But the evidence from previous business cycles is that this endogeneity is not large enough to make supply largely follow demand. In normal business cycle downturns, firms report that they have plenty of spare capacity and inflation falls.
Although the traditional ‘production function’ notion of supply has served us reasonably well in the past, today it gives us a serious puzzle - why has underlying productivity slowed so much following this particular recession? Now the extent of this puzzle varies from country to country. The productivity puzzle is particularly acute in the UK, and much less in the US [For an analysis of the US position, see here (HT Econbrowser).] . In the UK, if actual productivity really does tell us how underlying productivity has behaved since the recession (as the survey evidence and perhaps inflation suggests), it would be a slowdown that is almost unprecedented in UK history. (For some international examples of previous finance led recessions, see Mark Thoma here.)
Keeping things simple, we can think of two policy responses. The ‘pessimistic’ response is to assume that underlying productivity growth really has slowed, while an ‘optimistic’ response would hope that once demand gets going, supply will follow. Although this issue is normally discussed in the context of monetary policy, long term fiscal plans also depend (in a rational world) on estimates of long term output (i.e. on supply) rather than short term movements in demand. In the UK there is a target for the future cyclically adjusted budget deficit. Last year we saw the OBR revise down its estimate for underlying supply, and future fiscal plans were tightened as a result. They could well do the same again quite soon (see this recent study from the Social Market Foundation).
Which response we choose (or more realistically, lean towards) depends in part on which view about supply we think is more probable. But it should also depend on the size of the mistake that will be made if this belief is wrong. The mistake we make if we adopt the optimistic view and go for expansionary policy but it turns out supply has indeed fallen is straightforward: we get inflation without any additional growth. The mistake we make if we are pessimistic and pursue a more contractionary policy is more unusual. If supply follows demand, and yet we adjust demand to match supply, we run the danger of self-fulfilling pessimism. In monetary policy we focus on inflation rather than the output gap, and because inflation is sticky, the output gap steadily falls away without any above trend growth, and inflation stays near target. With fiscal policy we cut future spending to match lower expectations about supply, but expectations about future government spending influence spending decisions today, reducing demand and therefore supply.
So how costly is each mistake? Higher inflation without higher growth is not good, but fortunately we will see it when it happens, and we have a well tested tool with which to deal with it - raising interest rates. So we will get a year or two of above target inflation. The mistake on the other side is potentially much larger, for two reasons. One I have talked about before: at the zero lower bound, it is more difficult to raise demand once we realise the mistake. The other is the more worrying. It may take some time before we see the mistake. At some point we will realise that we could have had GDP a good few percentages points higher than we did for a good few years, without a significant increase in inflation. But it might take some time for that realisation to occur. (For example, an unexpected positive demand shock that does not raise inflation.) That could lead to a cumulated percentage output loss in double figures. These are lost resources that we will never get back. There is an even worse outcome, which is that although supply might be elastic to demand in the short term, it becomes less so as demand stays low. In this case the cost in terms of lost GDP might continue forever.
In these circumstances, it seems to me that policy should lean towards being optimistic about supply, even if we are somewhat sceptical that it is really following demand - because the costs of self-fulfilling output gap pessimism could be so high.
This has some similarities with the position policymakers found themselves in during 2010. Although the chances that markets would suddenly stop buying government debt in countries outside the Eurozone seemed small even then, the scale of the damage if they had done so was thought sufficiently large that it justified switching to austerity. I have always thought that this reaction was quite understandable, and my main quarrel has been that positions were not changed when it became clear during 2011 that this risk was negligible. If policymakers were highly risk averse then, a similar attitude should make them very worried about self-fulfilling output gap pessimism today.