Review of Murder at the Margin, by Marshall Jevons. Glen Ridge, New Jersey: Thomas Horton and Daughters, 1978, 168 pages.

Public Choice, 1979, Vol. 34, No. 2 (1979), pp. 233-236

In Murder at the Margin, Marshall Jevons tries to do for economics what G. K. Chesterton did for Catholicism in the Father Brown stories and Harry Kemelman is currently doing for Judaism in Friday the Rabbi Slept Late and other days of the week. A detective mystery may seem a strange vehicle for explaining and defending a world view, religious or otherwise,· but it is, after all, a microcosm of the application of human wisdom to the understanding of human behavior. Any system of thought that claims to explain the what and why of everything should have an easy time deducing who put the poison in the soup.

This is especially true of economics, which, thanks to the work of such 'economic imperialists' as Gary Becker and Gordon Tullock, no longer defines itself as the study of the allocation of limited resources to diverse ends, but as the science of understanding human behavior on the assumption that individuals rationally pursue ends, that, in Von Mises' words, 'human action is purposive behavior'. Criminal behavior in particular has been analyzed in this way by Becker, Landes, and others. One might hope that an economic detective story could not only introduce economics, a notoriously unpalatable subject, to noneconomists in an entertaining fashion, but at the same time make plausible its claim to be the new master science, the organizing principle around which any scientific understanding of human behavior must be built.

This is a difficult task, and unfortunately Murder at the Margin does not accomplish it. It fails at two levels. Considered simply as a detective story it is not a very good one; although the writing is often entertaining, the characters are superficial and the author relies heavily on telling instead of showing. The denouement depends on an economic theorem which few

readers will either know in advance or correctly understand after the author has explained it; it thus violates one of the central rules of the genre, that a sufficiently clever reader should be able to figure out the solution for him self. This is even worse for a mystery whose detective is an economist than it would be if he were a chemist or engineer using some technical trick to identify the murderer; one of the beauties of economics is that understanding it depends more on clear thinking and less on rote knowledge than with almost any other science.

As a mystery, the book is neither much better nor much worse than one would expect for a first book by an unknown author. As an explanation of economics, it is much worse than one would expect from an author with two such distinguished names. The hero, Professor Spearman, is a Nobel Prize-winning economist; the author goes out of his way to tell us how brilliant he is, and how fascinated with his own discipline. Spearman is continually noting and commenting on the economic behavior of those around him; the author uses these comments to give the reader brief lessons in economics. It is undoubtedly true that some great economists find it interesting to apply economic analysis to the world immediately around them, but one of the things that distinguishes great economists from other people is that the resulting observations are nontrivial ones. Professor Spearman's applications of economics are, with only a few exceptions, observations that when prices fall demand increases. Perhaps this is all the economics which mystery readers can absorb, although I doubt it, but as a portrait of an economist it is about as satisfactory as if a great mathematician was shown continually adding up columns of numbers while pointing out with delight that the sum was the same whatever order you added them in.

Of Professor Spearman's two nontrivial applications of economics one is the crucial deduction by which the murder is solved; since the laws that govern the reviewing of detective stories forbid giving away the solution, I will simply say that I believe the theorem the author applies is inapplicable to the particular problem to which he applies it. The other exception involves a 'Prisoner's Dilemma'. Professor Spearman describes this as involving a minimax solution, and says 'the safest strategy is for him to plead guilty, figuring his partner will do the same and against the chance that his partner might even plead innocent'. This, of course, is wrong; minimax or saddle point is the solution concept appropriate to two-person fixed-sum games, of which Prisoner's Dilemma is not one, and one of the things that makes the Prisoner's Dilemma seem so paradoxical is that the decision of one prisoner to confess does not depend on his gambling that his partner will also confess; as long as the two men's actions are independent, confession dominates refusing to confess whatever the other prisoner chooses to do.

More serious than such technical mistakes is the methodological error that underlies all of Professor Spearman's detective efforts — his misapplication of the rationality axiom. Rationality involves two assumptions. The first is that individuals choose the best possible means to their ends. Taken by itself, this has no predictive content; whatever the individual does can be explained by assuming that doing that particular thing - buying a good at a higher price than necessary, for example - is itself the end sought. One must therefore add a second assumption limiting ends. While hard to define exactly, this amounts to assuming that utility functions are reasonably simple - at least simple enough so that if all of the obvious ends one might seek are best served by buying a good at the lowest possible price, one's utility function will not have been specially tailored for the perverse purpose of refuting that conclusion.

What makes this assumption useful is not that it is always true. There may well be individuals - those, for instance, who have just spent several hours arguing with an economist - who receive positive utility, at least for a while, by acting 'irrationally'. What makes the assumption useful is that it describes fairly accurately the predictable element of human behavior. People do, to a considerable extent, act rationally for reasonably simple goals; to the extent that they do so an economist can predict their behavior. This point is made very clearly by Alfred Marshall in Chapter 2 of his Principles of Economics. Where, as in much of economic reasoning, all that matters is what most people do, or what people do on average, or what the predictable element in people's action is, the rationality axiom may lead to conclusions in which we have a good deal of confidence; even where it is important to know what a single person will do it may give us at least a best guess. But the idea that economic theory requires everyone to act in a rational, predictable, calculable way all of the time is not economics but a parody of it, and one that has often been used to ridicule economics by those who neither understand nor believe in it.

Marshall Jevons has apparently not read Alfred Marshall. Professor Spearman's detective deductions depend on the suspects behaving with perfect rationality and having no innocent motives for their actions save those immediately apparent. To pick an example almost at random, Pro fessor Spearman temporarily identifies a Harvard colleague as the murderer on the grounds that, on the night of the murder, he drank less than usual - even though the drinks, which he normally had to pay for, were on that occasion free. No connection is made between the murder (which occurred earlier in the day) and the suspect's sober violation of the law of demand. A little while later an innocent explanation for the behavior is discovered and Professor Spearman promptly concludes that the colleague (who is known to have disapproved of the victim and approved of the murder) must be innocent. Similarly, he concludes that the victim's widow must be innocent on the grounds that she could get more money by divorcing him and collecting alimony than by killing him and collecting social security survivor's benefits. It apparently does not occur to him that the woman, who dislikes her husband and is independently wealthy, might have relevant arguments in her utility function other than income.

Having said so much about what is wrong with Murder at the Margin, it may seem paradoxical to express my hope that the author will try again. The book is entertaining, and it succeeds in teaching some bits and pieces of elementary economics along the way; its faults stem from the difficulty of thinking up economic explanations which are both ingenious and correct, with the result that those which the author does come up with are mostly either trivial or wrong. This is the same difficulty that confronts the economist in his ordinary work. The advantage Marshall Jevons has over the rest of us is that he can invent his own facts and situations so as to make them amenable to analysis.

If he does publish a second book (and rumor has it that he intends to) there are two important changes I hope he will make. The first is to recognize that economic arguments rarely yield more than probabilities, and accordingly to make the solution less dependent on the idea that a single 'irrational' act is little short of proof positive of guilt. The second is to draw his collection of 'economic observations of a great economist' from life instead of from the opening pages of an elementary text. There are, after all, a number of great or at least near-great economists available to be studied, and those who are no longer living have mostly written books. It should be possible from those sources to compile some really ingenious applications of economics to the real world for Professor Spearman's use.

David Friedman

Center for Study of Public Choice

Virginia Polytechnic Institute and State University