sabato 24 luglio 2010

Cowen's "Risk and business cycle": a stocktaking.

Cowen's critique is the most thorough critical analysis of ABCT. Compared with Wagner's, Yeager's, Leijonhufvud's, Caplan's and Tullock's is probably the most interesting, and surely the most complete.

Here are his key points (there are other critiques, but I prefer these because they appear more relevant):

1. Why do consumption and investment comove during the boom and bust?
2. Why are interest rates so effective in causing malinvestment?
3. Why do entrepreneurs commit non-random, but biased, errors during the boom?
4. Why do central banks play a key role in miscoordination, whereas the market may be inherently unstable on its own terms?
5. Why does finance play no role in Austrian theorizing?

Point 1 requires a theory of overproduction, such as the one in Time and Money by Garrison. Machlup wrote something about capital consumption, and Hayek sometimes talked about the topic. Mises talked about overconsumption without defining the concept and justifying its possibility. I think the critique can be answered, but a convincing answer has not been found yet. Hayek went the wrong direction claiming that consumption falls during the boom through forced savings, but this is utterly incredible.

I don't get point 2. It appears that interest rate movements in neoclassical models do not have large effects... but this is due to the fact that money is useless in those models. However, the effectiveness of monetary policy in creating malinvestment relies on the potency of its transmission channels, and I don't know if a proper answer has been found to this objection.

Point 3 is about ratex. The standard Austrian answer is that ratex don't hold, but the only justification is Carilli and Dempter's prisoner dilemma argument. More can be done to analyse the issue: the result will surely focus on coordination problems (O'Driscoll's book is key), prisoner's dilemmas (Carilli & Dempster), and the role of money in coordination (Mises). Although I agree with this line of argument, Hayek's attempt to give a thorough microfoundation to the issue failed because of excessive complexity (he never wrote part two of The pure theory because of this).

I think that neoclassical economics don't get the foundation of the argument: a direct attack to Lucas's "signal extraction" view of money is required. Money is not (only) a signal, it is purchasing power: those who get it have higher spending power, and draw real resources away from others. This is an externality: externalities beget prisoner's dilemmas to be effectively counteracted, and this creates coordination problems. Translated into "neoclassical" words the argument may become more appealing.

Point 4 is based on a misperception of the Austrian literature: central banks are not required for intertemporal disequilibrium, they just enhance them. I don't know if banks can learn to avoid boom/bust cycles if left alone, but I'm sure that if they are shielded from the consequences of their own errors they will never learn, and cycles will be longer, deeper and more harmful.

Point 5 is a mystery. Except for sparse notes by Mises and Hayek on credit money, a book by Machlup whose focus was not business cycles, a paper by Antony Mueller (2001, QJAE), some notes by Chester Phillips and Strigl... I found no thorough analysis of finance in Austrian cycle theory. There is too much focus on money and not enough on credit: I'm convinced that financial accelerator models should be included in ABCT to make it a more complete theory. Cowen does something to solve this problem. Money is important in ABCT becasue it distorts the credit creation process, and monetary theory is important because it must explain how money affects market coordination through the credit market. Otherwise, ABCT would be a strand of monetarism. In my view, it is much closer to credit channel theories, such as Stiglitz, Bernanke, Geltler, Irving Fisher...

Despite, relative to other critics, Cowen has a fairly good understanding of ABCT, he sometimes wastes time on irrelevant arguments which make sense only within a homogeneous-goods theoretical framework. For instance, he says that entrepreneurs may react to an increase in loanable funds and interpret it as a fall in the demand for loans (i.e., a crisis). The underlying argument is that the fall in the demand for loans gives information about a fall in real returns and bad conditions. The problem is: who cares? I can make profits by increasing my investments, the world is too complex to interpret the result as either a variation of savings, of demand for loans, or an increase in fiduciary credit. The entrepreneur only knows that he can make an easy profit, or that competition will force him to act in that way. Cowen sometimes falls prey to macroeconomic habits of mind: the problem with macro is that it doesn't exist.

Interpretation of price signals to obtain information about the state of the economy is only possible in very simple frameworks (like those analyzed by Lucas, Stiglitz, Cowen) in which reality is so stylized that every signal can mean only one thing. In the real world entrepreneurs don't recreate the structure of the economy through monetary calculation, they only see if there are profits or losses in behaving in a certain way. Bayesian inference is not powerful when there are thousands of possible causes of the same effect.

A point in Cowen's arguments that strike me as absurd is when he claims that savings are stable during the business cycle. So, consumption is stable, savings are stable, GDP varies, but, national income identities assure us that GDP = Consumption + Savings + Taxes + Imports. Where's the variance in GDP? Probably Cowen is confusing personal savings and overall savings: I don't have data (I should check on Fred), but probably savings have the same variance of investments (very high), otherwise cycles could be only caused by large shifts in taxes and imports...

His pars construens is much worse than the critical part, however. He tries to build an efficient investment frontier along which the economy moves due to monetary policy, and higher growth is traded off with higher instability. In my opinion, it is the INEFFICIENCY of risk taking during booms that is the key aspect of business cycle theory, the economy trades off risk for nothing. Real growth just hides the unsustainability of malinvestment and misintermediation by soothing the inflationary impact of unsustainable policies. Booms during periods of technological progress last longer and are more dangerous.

Cowen's arguments are a good starting point to:
  1. Translate ABCT in comprehensible terms
  2. Add a financial dimension to ABCT
  3. Highlighting the critical points of ABCT
  4. Comparing ABCT with present-day theorizing
  5. Focusing on the role of risk
Although I haven't like the pars construens, I recognize that the pars destruens and the issues which are discussed are very important. Austrians' reaction to Cowen's book has not been constructive.

PS I forgot to mention that using Carilli and Dempster's paper to claim that Cowen is wrong implies a temporal error: that paper came four years later than Cowen's book. I'm a terrible philologist.

2 commenti:

  1. Book 4 of TPTofC is in fac the "part 2" of that book you fasely say wasn't written.

  2. It's about 60 pages (3 chapters), more a draft than a book.