Monday, November 22, 2010

Economists, Handwaving, and the Culture of Economic Modeling

One of the most typical form of hand-waving you will see by some economists is when they insist that one policy is more efficient than others is the assertion that policy X should be adopted, even though it results in a distribution of income that disadvantages a large group of people who are less affluent on average to benefit a smaller group of people who are much more affluent on average. Many times, economists will point out that policy X is optimal, because you could theoretically tax a percentage of the increased profit flowing to winners in the more affluent group and transfer those resources to the losers in the less affluent group and the overall result would be more income for everyone. A good example is this blog post by University of Chicago economist Gary Becker, where he argues that the Russian policy of adopting an export ban on wheat that would effectively result in higher real incomes for those who do not own farms is a mistake, since it prevents farmers from enjoying windfall profits due to unexpectedly high prices for wheat on international markets. The argument goes that "Russia would be better off if it allowed farmers access to world prices, and it could give income transfers to urban families that offset at least some of the resulting higher cost to them of wheat and other foods."

Unfortunately, this sort of thinking, which could be taken directly from an Economics 101 textbook and does not consider any of the particularities of the Russian situation as being important, is very inadequate in terms of determining the best policy going forward. One wishes that economists like Becker would remember the old economic saying that "there is no such thing as a free lunch". An elementary error that Becker makes is his basic failure to account for the costs of redistributing income (or the costs of failing to do so.) What Becker fails to consider is that it not "free" to redistribute income.

First of all, those who earn extra income from accessing higher prices on international markets are very likely to fiercely resist redistribution on a political level. Those who oppose such redistribution are generally likely to succeed. Politically, it is much easier to adopt policies that result in more favorable distributions of income (even if there are some economic inefficiencies) than it is to redistribute income after it has been earned.

One might ask whether it should matter if there is a successful redistribution. Why should it matter whether the gains are distributed to a only a few? Isn't a $1,000,000 allocated to one person as good as $1,000 dollars allocated to a thousand people? Shouldn't Russia be indifferent between these two outcomes? And the answer is, of course not, due to the diminishing marginal utility of a dollar. That is, one person is not likely to derive as much utility out of a million dollars as a thousand people would out of a thousand dollars. Clearly, dollars that are spent on basic needs provide the most utility; only when one has enough quality food to eat and adequate shelter and adequate access to the other basic necessities, does one look to derive pleasure from luxury goods. One can much more easily do without a luxury good than one can do without a necessity. So, to the extent that the political failure to redistribute income would have the effect of depriving people of basic necessities, there would be a significant loss to Russian society.

Second, even if it is politically possible to redistribute income, government programs to redistribute income are potentially vulnerable to corruption. That is, there will be a temptation by those in charge of such redistribution to redirect wealth to themselves or to favored groups. Such corruption may not defeat a program to redistribute gains entirely, but their cost may in fact exceed the costs of the export ban that Becker is criticizing.

Third and finally, even if a program of redistribution is undertaken with minimal corruption, there is still administrative overhead involved with establishing a bureaucracy to oversee the redistribution, determining eligibility for redistribution, and prosecuting instances of fraud by those who falsely claim eligibility for benefits or attempt to claim benefits more than once. Perhaps such costs can be minimized by carefully structuring the government bureaucracy in charge of redistribution, but the challenges of doing so, both politically and logistically, should not be discounted. Also, newly established bureaucracies can sometimes be very difficult to disband politically when they have served their purpose.

Maybe Becker's conclusion that the Russian export ban on wheat is bad policy is right. But that would be more a matter of sheer luck than sound analysis. The universe of policy possibilities facing can be divided into three categories.

(1) Impose an export ban on wheat. This has enforcement costs as well as the increased income for farmers that is foregone by accessing international markets.

(2) Allow the export of wheat, but do not make any attempt to redistribute any of the windfall profits accruing to farmers due to unexpectedly high prices in international markets. This option has significant costs due to the diminishing marginal utility of a dollar and the fact that food is a basic necessity rather than a luxury item.

(3) Allow the export of wheat, but set up a bureaucracy to redistribute a portion of the windfall profits accruing to farmers. This option has significant costs in terms administrative expense, possible corruption, the establishment of new bureaucracy, and the political resistance of new bureaucracies to being disbanded when they have served their purpose.

So, what Becker has failed to do is try to compare these options and estimate the costs of the various choices. Implicitly, Becker waives his hand and says that the losers from the decision to allow the export of wheat can be more than compensated by the gains of the winners and everyone will be better off. But he fails to consider the costs of such transfer programs and whether it is even politically feasible. Basically, Becker's analysis proceeds as if the cost of transfer would be zero, when this is far from likely to be the case.

Finally, it should be noted that to actually do this analysis correctly, one would have to be very knowledgeable on the particulars of the Russian situation. In this case at least, an analysis based on nothing more than the abstract principles gleaned from an Economics 101 textbook is woefully incomplete. You have to know something about the political situation in Russia to determine whether a transfer of income to compensate the losers is even politically feasible and you have to know something about the corruption situation to estimate the likely costs of administering such a program. It may very well be that a temporary export ban is in fact the best policy for maximizing the welfare of the Russian people, when one does a full analysis that takes into account the costs of alternative policies. It is impossible to know for sure without doing the full analysis.

What is amazing to me though, is that professional economists like Gary Becker could make such elementary errors. If I were to guess about the origins of this mistake, I would chalk it to what I will call a "culture of economic modeling" where economists get so comfortable making simplifying assumptions for their models, that they make elementary analytical errors when it comes to policy analysis. Here, Gary Becker is doing nothing more than assuming that the costs of a transfer to compensate the losers are zero or de minimus. It is very natural for economists in particular to make such assumptions, since they are always making simplifying assumptions when it comes to creating economic models. But, when you consider the complications of the real world, in this case the simplifying assumption seems to be extremely dubious. At the very least, serious analysis of the political feasibility and the costs of doing transfers is worthy of serious analysis before assuming that such costs are insignificant.  I think this goes to show that it is very important for economists to carefully consider the applicability of their simplifying assumptions when doing real world policy analysis. I think the habits of mind that are useful for creating abstract academic economic models with no urgent need for real world applicability may need to be carefully adjusted for real world policy analysis.

No comments:

Post a Comment