Draft of Letter to Krugman

I’ve been working on this for a couple of days.   If anyone is reading… comments?

Dear Dr. Krugman,

I write to encourage you to add the following argument to your arsenal.

The neo-classicists argue that the market-clearing price is always the optimal price, but this is wrong.  As we know from game theory, locally rational decisions can produce collectively sub-optimal outcomes.  A depression is tragedy of the commons.

What is the externality that is lost in the depression?  Information.  Rational transactions are like scientific experiments.  They are informative both in their success and in their failure.  We aren’t normally aware of it, but when two parties transact there is an important benefit to third parties — it tells them something about the private beliefs of the transacting parties.  The transaction serves the market by calling attention to its basis and its outcomes.  This allows a third-party observer to form expectations about what the outcome “should be” based on his/her own information, which in turn gives the observer the ability to learn from the transaction’s outcomes.    Thus, the rational transactions of two parties creates the opportunity for third parties to gain information, improving their future decisions.

The market allocates capital efficiently given a level of information, but it does not allocate capital efficiently to produce information.  In a depression, there are fewer transactions and so there is less information gained.  Decisions by actors to hold their money are rational, but the market as a whole loses out.  The neo-classicists are correct in saying that for the government, or any other third party, to “force” investments is to spend capital on experiments that are likely to fail.  But the failure of investments provides information.  Neo-classicists would also be correct in saying that arbitrarily chosen, irrational transactions do not provide as much information as rational ones.  This is true, but slightly biased experimentation reveals more information than no experimentation.

Government investment is needed, and the emphasis should be on informativeness.

Sincerely,

Drew Margolin