Thursday, August 11, 2011

Has the US economic policy been Keynesian for centuries?

Suppose the abstract of a paper starts with "It is demonstrated that the US economy has on the long-term in reality been governed by the Keynesian approach to economics independent of the current official economical policy." My first reaction is that of puzzlement, as I would not have thought as the US being particularly keen on Keynesian policy, except for the recent years (which are not considered in the quoted study). But again, data may speak differently from policy intentions, so let us dig deeper.

A. (Agung?) Johansen and Ingve Simonsen come to this stunning conclusion by looking at the correlation between (nominal?) (federal?) public debt and the Dow Jones Industrial Average. One can first question whether public debt is a good indicator of Keynesian policy. Public deficits or even public expenses would be better. And does the DJIA represent the US economy? It is certainly not an indicator of current activity, but rather of expected present value of future profits from a particular class of firms.

Whatever. Let us go with that. The analysis is done by computing over the 1791-2000 sample a sliding correlation between these two indicators over a five-year window. Surprise, the correlation is zero most of the time, except during some wars when it is strongly positive (and strongly negative during the second war with the Seminole Indians). From this they conclude the Keynesian policy was mostly pursued during wars. Now let us take a step back: the authors show that there is by their definition no Keynesian policy during peacetime. But during wartime, the government is credited with a policy geared towards expansion of the DJIA. They, one may ask, if this is the government overwhelming policy, as the authors seem to believe, why did the US wait so long to get into the two World Wars when the opportunity was there? I cannot make sense of all this.

10 comments:

Vilfredo said...

It is just another foray of physicists into Economics. What did you expect beyond some senseless data mining?

Kansan said...

EL, you are going for the low-hanging fruit here.

Anonymous said...

Why do you even bother discussing such junk?

Economic Logician said...

Because some people believe it, and if it stays unchallenged, more will believe.

In private email, one of the authors dismissed my comments with the claim that Economics is not a science, while Physics is. So there.

Anders Johansen said...

No, I don't bother. You havn't understood a thing. We do not speak
about the effective economic policy in peace time, but look at the
correlation of New Deal and the Cold War. The public dept and public
spending is the same difference. The huge influx of publice money
strengthen the private sector enormously and hence influence the
economy in peace time. The hottest thing in decades has been the
internet created by Pentagon, i.e., tax money. The html language was
created by CERN, i.e., tax money.

Last: Economy is not a science since it's axiomatic as mathematics,
but Physics is. That's why there's no Nobel Prize in economy, but
there is in Physics. The pseudo-Nobel Prize was created by swedish
banks in the 50'ties. The Nobel family is now reconsedering the use of
their family name for this price.

Kansan said...

Wow, and this guy calls himself a scientist? This is now becoming interesting!

ivansml said...

Hillarious paper. Paragraphs of cluelessness followed by lame data analysis which doesn't even support the conclusion. If anyone wanted to write an econophysics paper parody, it would probably look like this.

Some specifics, in addition to what EL said: no, EMH doesn't imply that only monetary policy is "allowed". Joint hypothesis problem doesn't make finance a religion. Economics is not axiomatic and nowadays, only a minority of economists do pure theory. Increase in public debt is not necessarily a sign of keynesian policy (e.g. GWB tax cuts). Correlation doesn't imply causality. And so on...

Anonymous said...

The authors are clearly crackpots.

Anonymous said...

Physicists are smart but they do not know Economics. For the past 30+ years, there is no academic employment opportunity (no job market at all) for the top Physics graduates. Since the late 1980s, Wall Street started hiring the brightest Physics PhDs for their strength in Analytic Thinking, Applied Mathematics, and Programming skills. Their success in Wall Street lead many Physicists think that they can do Economics without knowing Economics. Their works are mostly garbage.

On the other hand, many top Financial Economists came out of Physics. They switched field in the Graduate School and earned their PhD in Economics or Finance (sometimes a 2nd PhD after Physics) and made meaningful contribution in Finance/Economics. Bob Merton, Steve Ross, George Constantenides, Jiang Wang, Andrew Lo, John Cochrane etc are a small sample of ex-Physicists I can name right away.

There are also a very very small number of highly talented people who are self-taught in Economics but originally trained in Physics and Applied Mathematics. Fischer Black is well-known within and outside of academia. Personally, I was fortunate to take a Fixed Income Asset Pricing PhD course from Akash Bandyopadhyay, a PhD in Physics, at Chicago. It was one of the best courses I have taken in the grad school. His depth of knowledge in Economics/Finance used to amaze us.

Physics and Economics both study Dynamical Systems. Physicists can reproduce a system in its original state and re-run an experiment as many times as necessary. So their measurements are very precise and the models are rigid. Economists deal with human behavior which changes over-time and it is nearly impossible to precisely estimate the dynamics from one time-series of data. So mutually contradicting models can sit next to each other peacefully in Economics, which is not acceptable in Physics or Chemistry. As far as I understand, that is the only major philosophical difference.

Anonymous said...

The Dow Jones Industrial Average was created in 1896. How do they have data going back to 1791?