Wednesday, October 9, 2013

Wait a (second) moment...

Marco Buti and Pier Carlo Padoan reply to the criticism that I had raised in an earlier blog post to their previous article on how to strengthen the European economy recovery. I can see that their argument is now more balanced where fiscal policy consolidation is mentioned as a potential factor to explain the dismal recovery. We still probably disagree on how much of a factor it was and whether there were alternatives to the policies that Euro countries implemented but at a minimum it is good to see that it is not included as a possible factor. On the issue of reforms it is hard to disagree with them on the need for further reforms in Europe, the real debate is whether these reforms will pay off fast enough and if they do not, what is the role for traditional demand policies (monetary and fiscal).

But there is something else in their article where my reading is quite different from theirs: the role of policy uncertainty. Before I express my views let me state that one of my most cited research papers is about the role of fiscal policy volatility in reducing economic growth (here is an example of my work in this area), so I am very open to the idea that volatility in policy can be detrimental to growth. But I have always been surprised that uncertainty and volatility are some times used to refer to episodes where the possibility of a bad scenario is increasing and this is not quite the same as an increase in uncertainty. Let me explain.

When we talk about volatility we are referring to an increase in the variance (which is a "second order moment" in statistics, that's the origin of my title) while we keep the mean constant (the mean is a "first order moment"). So increases in uncertainty or volatility only apply to circumstances where on average we expect a similar outcome but now we have a higher probability of both a better and a worse outcome. What we have seen in Europe during the crisis is very different. Quoting from Buti and Padoan:

"The unprecedented increase in tail risks in 2011 and first half of 2012, when the survival of the Eurozone was widely questioned, qualifies as such an uncertainty shock."

This is not (just) an uncertainty shock. The mean is also changing. The average scenario ahead looks much worse now that the Eurozone might collapse. Strictly speaking it might be that the variance has also increased but the change in the mean is probably more relevant than the change in the variance. The fact that a Eurozone collapse is now possible means that we face a much worse scenario ahead for the Euro countries (regardless of how uncertain we are about that scenario). And why was the Eurozone about to collapse? Because we are in the middle of a really bad crisis. And what is making the crisis so bad? Many things but one of them is the inappropriate policy mix (fiscal and monetary). So it is really uncertainty? No, it is simply a measure of how low expectations are getting and expectations are endogenous to current outcomes. There are, of course, statistical methods to try to separate each of these factors and establish a proper measure of uncertainty and a true causal relationship to growth but my reading of the academic literature is that this is not what we are doing and we are still dealing with correlations and comovements in variables without a clear understanding of the truly exogenous variation in policy uncertainty. All the bad news are included in what is being called an uncertainty shock.

Antonio Fatás