In a post at Quartz, University of Michigan economics professor Miles Kimball and University of Michigan undergraduate student Yichuan Wang write that they have crunched Reinhart and Rogoff's data and found "not even a shred of evidence" that high debt levels lead to slower economic growth.
And a new paper by University of Massachusetts professor Arindrajit Dube finds evidence that Reinhart and Rogoff had the relationship between growth and debt backwards: Slow growth appears to cause higher debt, if anything.The Huffington Post
Reinhart And Rogoff's Pro-Austerity Research Now Even More Thoroughly Debunked By Studies
Mark Gongloff
An undergrad? Ouch.
5 comments:
Just to be accurate, R&R’s research never claimed to have shown a CAUSAL relationship between debt and growth. I.e. their research only claimed to have found an ASSOCIATION. Though obviously, liars that they are, they’ve made every effort to imply or suggest that their research bolsters the “causal” argument.
At least the following BBC article says R&R’s research never claimed a causal relationship. See 3rd paragraph here:
http://www.bbc.co.uk/news/business-22466551
And in this article Brad DeLong says “Second, RRR present a correlation--not a causal mechanism…”. See:
http://www.project-syndicate.org/blog/risks-of-debt--the-real-flaw-in-reinhart-rogoff
This says nothing because virtually all modern mainstream economics is based on statistics, which is correlation not causation.
In science, causation is stated through laws. Where are the laws in economics and I don't mean things that are called "laws," like the law of supply and demand, which has not causal formulation in terms of independent and dependent variables.
This should not be surprising since there are no laws in other social sciences either, because they are not ergodic like the natural (physical) sciences, mostly that is. Remember how Einstein freaked out at the idea that QM is probabilistic rather than deterministic, implying that the physical universe is not deterministic?
What this implies is that science is a posteriori rather than a priori. A priori systems are imaginative, and the causality imputed from axioms is asserted about the model rather than proved concerning the actual events that the model putatively models, that is, explains and predicts. That's why prediction is used to test the generality of explanations, since many explanations of events are possible and not all are necessarily good predictors.
Just because a syntactically elegant explanation can be constructed doesn't imply that it is semantically true. Mainstream economists have made the mistake of becoming infatuated with their own models to the point of imputing to them characteristics they don't have.
Bad science.
He who oppresses the poor to make more for himself or who gives to the rich, will only come to poverty. Proverbs 22:16
If not in this life, then the next.
There are a variety of statistical methods available for in the social sciences, medicine etc. that help in distinguishing causal connections from accidental correlations, and that can be used to discern the direction of causation, as well as rule out common cause explanations. There is no difference in principle between trying to find out whether high public debt tends to suppress economic growth and trying to figure out whether high levels of certain neuro-inhibitors tend to suppress respiration.
The empirical confirmation and disconfirmation of causal hypotheses does not depend on introducing any a priori conjectures or assertions into the model. What is sought is not metaphysical certainty, merely higher or lower degrees of probability and conviction.
If professional economists didn't aggressively apply these techniques to the questions that Rogoff and Reinhart raised than it is probably either because (i) R&R didn't share their data, (ii) other economists were mostly too lazy or too busy with other things to bother, (iii) the community of economists is slavishly impressed by academic hierarchy and elite institutional affiliations, or (iv) the community of economists was on the whole so in love with the policy recommendations that R&R's work seemed to point toward that they didn't want to probe further in a way that might undermine their ideological preferences.
Dan I believe it is not possible to show causality statistically outside of carefully controlled experiments, which are generally difficult to construct and replicate in the social sciences. Such experiments would have little general application. Nothing on the order of R & R would qualify.
Economics applies regression analysis to pre-existing data, which usually requires making many assumptions. As Yeva and Randy pointed out, R & R makes many untenable assumptions about the data, e.g., that govt debt has the same relevance across different monetary system when there is ample evidence showing that this assumption is incorrect.
Post a Comment