Economic Method, Theory and Policy: Krugman Edition

In a recent attack on causal-realist economists [or the "Austrianish Horde" as he called it], Paul Krugman again reverts back to using a baby-sitting co-op as a model for the economy. His doing so provides an excellent example of what I tell my students the first day of my Foundations of Economics course. I explain that too often people wish to begin by answering policy questions without making sure their theory is correct. This is a disastrous mistake, because one's policy conclusions are directly related to one's understanding of economic theory. It is our economic theory that leads us to conclude that particular economic policies will result in particular economic consequences. For example, increasing the money supply by a billion dollars a day will result in its purchasing power being lower than it would be without the monetary inflation.

Krugman's recent blog post is a case in point. Krugman is well-known to be a Keynesian fiscal and monetary activist. He pushes this policy agenda in large part because of the way he conceives of the economy. He thinks we can model the economy as a baby-sitting cop-op that uses baby-sitting tickets to purchase sitter services. In doing so, Krugman uses what Rothbard called a "false mechanical analogy" of model building.  Rothbard criticizes the use of such model building because "the 'models' of economic and political theory are simply a few equations and concepts which, at very best, could only approximate a few of the numerous relations in the economy or society." As such, they often abstract so far from reality as to be irrelevant at best or destructive at worst.

Krugman's baby-sitting co-op model is an example of the latter. In this model the output is baby sitting and the "money" is the tickets. Krugman used this model way back in 1998 in the first edition of his Return of Depression Economics, by which he meant return of Keynesian economics.

What I said about the model in my review of the book, still applies today:

Krugman’s main analytical model is a quintessential example of his strengths, such as they are, and weaknesses. While attempting to explain the workings of the economy in simple terms that the general population can readily understand, he hitches his analytical wagon to an article using a baby-sitting co-op in 1970s Georgetown as a model for the macroeconomy. As a result, Krugman makes fundamental errors regarding how the economy works. In an attempt to efficiently ration baby-sitting services among the members, the baby-sitting co-op issued coupons. Each member family paid a baby-sitting ticket whenever they used the co-op and received a ticket whenever they baby-sat for one of the other members of the co-op. Purposely leaving out the details, Krugman tells the reader that members of the co-op suddenly increased their demand to hold baby-sitting tickets. Consequently, there was not enough aggregate baby-sitting demand for the services of those members in the co-op who were looking to baby-sit in order to increase their ticket incomes. In other words, Krugman explains, the baby-sitting co-op went into a recession.

The attraction of this model is its seductive simplicity. Krugman is often quite good at taking issues and problems the general reader finds unmanageably complex and explaining them in ways simple to understand. While this is, of course, a virtue, it is only a virtue if his explanation accurately reflects reality. The chief responsibility of the economist is to get the analysis straight. For this, the
baby-sitting model will not do.

The fundamental error of this model is that it only has one good: baby-sitting. This leads the reader to think of economic output as if all goods produced in the economy are homogenous units making up one aggregate. The lesson of the model is that if a recession occurs, it must be that there is not enough demand for this homogenous output. In the baby-sitting model, the problem is that members
demanded to hold too many tickets. For the economy as a whole, as Krugman views it, “a recession is normally a matter of the public as a whole trying to accumulate cash” (p. 11). People decrease spending in order to increase their cash balances. The supply of goods not demanded sits idle, and unemployment results.

In reality, of course, the plethora of goods bought and sold in the world economy are heterogeneous. What causes a recession is not too little “aggregate demand” or “too much capacity” due to overinvestment. Recessions are a product of malinvestment, resulting from government intervention in credit markets. If the government increases the money supply through credit expansion by artificially lowering interest rates, an incentive is created for entrepreneurs to invest in too much production of some higher order goods and not enough production of other lower order goods. It is not that too much investment is occurring in every sector of the economy, rather, investment that is occurring is being directed toward producing the wrong things, from the point of view of the people who make up society.

A very telling characteristic of Krugman’s analysis is that he argues that recessions will persist until aggregate demand picks up due to monetary inflation. Again Krugman, alluding to both the baby-sitting co-op and the economy, states that the recession “can normally be cured simply by issuing more coupons” (p. 11). The immediate question that should come to mind is why the surplus was not eliminated by a fall in the price of baby-sitting. We do not know. Krugman does not even bring it up! The model assumes that the prices are fixed at a one-ticket-to-onenight-of-baby-sitting ratio. This lulls the reader, and it seems Krugman himself, into forgetting that prices will adjust downward to eliminate any surplus due to a drop in demand. It is curious, to say the least, that in a book with the word economics in the title, the author does not get around to discussing even the mere possibility of a price decrease in the face of a surplus until page 155, that is, until the reader has read 92 percent of the text.