Thursday, October 15, 2015

GDP growth is not exogenous

Ken Rogoff in the Financial Times argues that the world economy is suffering from a debt hangover rather than deficient demand. The argument and the evidence are partly there: financial crises tend to be more persistent. However, there is still an open question whether this is the fundamental reason why growth has been so anemic and whether other potential reasons (deficient demand, secular stagnation,…) matter as much or even more.

In the article, Rogoff dismisses calls for policies to stimulate demand as the wrong actions to deal with debt, the ultimate cause of the crisis. As he argues, given that government expenditures have kept expanding (he uses the number for France at 57% of GDP) it is hard to argue in favor of more spending.

But there is a perspective that is missing in that logic. The ratio of debt or government spending to GDP depends on GDP and GDP growth cannot be considered as exogenous. Assuming that the path of GDP is independent of the cyclical stance of the economy does not sound reasonable but, unfortunately, it is the way most economists think about a crisis. A crisis is seen as a temporary deviation of output but the trend is assumed to be driven by something else (innovation, structural reforms,..). But that logic runs contrary to evidence on the way investment and even R&D expenditures behave during a crisis. If growth is interrupted during a crisis output will never return to its trend. The level of GDP depends on its history, what economists call hysteresis. In that world reducing the depth of a crisis or shortening the recovery period has enormous benefits because it affects long-term GDP.

[To be fair to economists, we are all aware of the persistent dynamics of GDP, but at the theoretical level we tend to explain it with models where the stochastic nature of the trend is responsible for the crisis itself rather than assuming that other factors caused the crisis and the trend reacted to them.]

In a recent paper Olivier Blanchard, Eugenio Cerutti and Larry Summers show that persistence and long-term effects on GDP is a feature of any crisis, regardless of the cause. Even crisis that were initiated by tight monetary policy leave permanent effects on trend GDP. Their paper concludes that under this scenario, monetary and fiscal policy need to be more aggressive given the permanent costs of recessions.

Using the same logic, in an ongoing project with Larry Summers we have explored the extent to which fiscal policy consolidations can be responsible for the persistence and permanent effects on GDP during the Great Recession. Our empirical evidence very much supports this hypothesis: countries that implemented the largest fiscal consolidating have seen a large permanent decrease in GDP. [And this is true taking into account the possibility of reverse causality (i.e. governments that believed that the trend was falling the most could have applied stronger contractionary policy).]

While we recognize that there is always uncertainty when estimating this type of macroeconomic dynamics using one particular historical episode, the size of the effects that we find are large enough so that they cannot be easily ignored as a valid hypothesis. In fact, using our estimates we calibrate the model of a recent paper by Larry Summers and Brad DeLong to show that fiscal contractions in Europe were very likely self-defeating. In other words, the resulting (permanent) fall in GDP led to a increase in debt to GDP ratios as opposed to a decline, which was the original objective of the fiscal consolidation.

The evidence from both of these paper strongly suggests that policy advice cannot ignore this possibility, that crises and monetary and fiscal actions can have permanent effects on GDP. Once we look at the world through this lens what might sound like obvious and solid policy advice can end up producing the opposite outcome of what was desired.

Antonio Fatás