Recession and recovery, Krugman and Mankiw, evil and wonkish
There are a few things I want to comment on today, but I'll start with a quick analysis of the latest Krugman-Mankiw debate.
The Obama administration projects that when the recession ends, growth will be faster than the long-term average, as the economy catches up with its permanent trend. This is called the trend-stationary assumption - Paul Krugman supports it and Greg Mankiw disagrees, citing an alternative hypothesis called unit-root.
Though I often disagree with Mankiw, this time I fear he is right. Both parties are citing different analyses of historical data. In economics, data is always easy to argue about, especially macroeconomic time series data of which there are rarely enough to make unambiguous inferences. Neither are going into detail about their answer in terms of an underlying theoretical model. Of course, models are also easy to argue about, but in this case it seems relatively easy to at least expose the implicit assumptions, if not to know which are right.
The question can be posed as follows:
- Assume that trend growth is 2.5%
- Then assume that we have an 18-month recession, during which growth averages -1%
- By the end of the recession, GDP is 7% lower than it would otherwise have been
- Growth then resumes
- Does it catch up with the trend again - as in chart 1 below - or not - as in chart 2?
Chart 1: Actual catches up with trend
Chart 2: Actual does not catch up with trend
Krugman's and the CEA's argument is that if it does not catch up with trend, employment will never return to normal levels. But this makes a big assumption about the capacity of the economy.
Does the underlying capacity of the economy continue to grow at 2.5% even during a recession?
I would love to think that it does, but I don't really believe it. In some models, capacity growth comes from growth in the economy's stock of investment; in which case it seems unlikely to be growing right now, with net corporate borrowing falling and investment low. In others, it comes from innovation and knowledge spillovers; and if innovation is reduced in the recession then it won't continue.
In this case, the graph looks like the following, and underlying capacity never reaches the same level as in the first two scenarios:
Chart 3: Trend is hurt by the recession, and actual catches up to it
If innovation is mainly responsible for growth, will we keep innovating in a recession? Maybe innovation will actually increase, as unemployed entrepreneurs or programmers try new things. But if you like this model, you have to accept that productive innovation may actually have fallen in the last few years, when we had high utilisation of resources and effort was diverted into speculation on property and finance. In which case underlying capacity has not been growing, and the recession may simply be the result of overheating demand reaching its limit, and unemployment returning to its natural rate.
In this case, demand may have overshot capacity (financed either by foreign loans, or by running down the capital stock - disinvestment), and the graph looks like this:
Chart 4: Trend was hurt before the recession, and actual overshoots, undershoots and then converges on it
Krugman's assertion that capacity keeps on rising might be correct - but that probably depends on one of the following conditions:
- The recession is short enough not to significantly affect innovation and investment
- Growth depends on factors that are not (negatively) affected by recessions
- Underlying capacity growth will accelerate beyond trend as the recession ends
The second depends on your model. There is a literature of growth theory on which I am not an expert, but most of the modern work aims to model endogenous growth, arising from factors like intellectual property investment that arise within the economic system. That given, it seems likely that these would be affected by a recession. But you could argue that the effect of a 7% recession is only a 7% reduction in intellectual investment; thus capacity growth falls only from 2.5% to 2.33%, hardly a crippling fall. On the other hand, this type of investment might be proportional to corporate profits instead of revenue, in which case a recession has a disproportionate impact on it.
(Incidentally, one of the main exponents of new growth theory is Paul Romer, who is not married or related to Christina Romer, who as head of the CEA produced the administration's trend-stationary growth forecasts.)
The third assumes that capacity can grow at well over trend for a few years in a row; but my own feeling is that this growth is asymmetric. It can certainly fall below trend but it's harder for it to rise above. But I am hesitant to write off this option completely without evidence.
After all this speculation, either view could be right - and indeed, the truth is probably in between. It seems likely that much, but not all, of underlying capacity growth is dependent on current economic performance. Thus, a recession will be a brake on capacity growth, and the rebound will not be quite as strong as Krugman hopes - but it will indeed be faster than long-term average growth.
So I hope Mankiw's unit root hypothesis is wrong, but I fear Krugman's trend stationary alternative is also incorrect. We'd like Mankiw to be more wrong than Krugman in this case. I won't stir the pot by suggesting what the long-term trend looks like on that question.
Update: Greg Mankiw responds:
The theory behind the unit root empirical results remains murky. There are several mechanisms that might be working together. Innovation is one, as you suggest. Another is the impact on workers, who lose experience and job skills in downturns.