CRIME AND PUNISHMENT

Thinking about Thinking About Crime James Q. Wilson

think of only three possibilities: (1) he did not read my book; (2) he read my book, but in his zeal to get on with the task of using some caricature of me as a weapon with which to advance his cause, he forgot what he read; (3) he read and remembered the book, but chose to misrepresent it. I believe that the second is the correct hypothesis, but I have no direct evidence with which to reject the first or the third possibilities. A major example is his statement that I omit demographic variables other than the age structure of the population in accounting for crime increases, and that specifically I neglect the high unemployment rates among young males, especially black teenagers. At the very beginning of my book I cite the appalling figures about the rise of teenage unemployment, especially among blacks, during the 1960s. I return to the subject in a brief case study of Washington, D.C., summarize statistical analyses on the relationship between unemployment and crime, and conclude with an argument in favor of simultaneously increasing the costs of crime and the benefits of legitimate occupations. Incredibly, Curtis asserts that "Wilson fails to discuss these matters."

Theoretical Structure

W

r hatever the status of labeling theory as applied to crime,

it is indisputably a powerful force in dealing with theories of crime. No one who writes on this subject, it seems, can escape being labeled, and of course the "radical" criminologists eagerly and proudly label themselves. Thus I am, in Isidore Silver's view, a "classicist," or perhaps a "new classicist," though one who has taints of outright positivism. And in Lynn A. Curtis's sly imputation, I am part of a "reactionary and selfish conservatism" and a "new criminologist." Labels have some value, of course; they are a shorthand way of handling in brief compass complex arguments and useful techniques for remembering things that might be asked on a final examination. But for serious intellectual discourse they are a shriveling and misleading force, reducing complex arguments to superficial formulas and stripping important assertions of all sense of context and qualification. I tried to write a book about crime that would dispense with labeling entirely, but my reviewers will not let me get away with it. They want persons and ideas stuck into neat pigeonholes. Small wonder that sensible statesmen have concluded that crime is too important to be left to criminologists.

Demographic Variables The most extreme and perverse example of the strategy of argument by slogan that I have encountered is Lynn A. Curtis's criticism of views he imputes to me. As I read and reread his manuscript, my bewilderment and despair grew. Unless there is another James Q. Wilson who has also written a book about crime, what needs explaining is not crime, and not my views on crime, but Curtis's reading habits. I can 10

It is no small debating point. Central to any argument, such as mine, about the importance of deterrence in a crimereduction strategy is the following theoretical structure which I make explicit but which Curtis ignores. To the extent that potential offenders are rational and reasonably well informed, they will choose among consumption, legitimate employment, and illegitimate employment based on the expected net value (or net costs) of each activity. This is more or less the model Isaac Ehrlich uses to estimate the effect on aggregate crime rates of certain criminal justice and demographic factors, and it is even more the model that Ann Witte uses to estimate the effect on individual crime rates of the same factors. Nobody currently working in this field has, to my knowledge, argued that changing the costs of crime will affect crime rates without regard to changes in the costs and benefits of alternatives to crime. Everybody asserts, instead, that if the costs or benefits of property crime shift relative to the costs and benefits of legitimate occupations, then at the margin there will be changes in crime rates. The best available aggregate measure of the value of alternatives to crime is either the unemployment rate or the labor force participation rate for young males. I can only conclude from Curtis's misleading and irrelevant remarks on this subject that he simply has not thought the matter through very carefully or is unfamiliar with the technical literature. Curtis does raise one set of issues concerning deterrence that is of consequence, though he does not put the matter accurately. He argues that deterrence is of little significance because many criminals are "irrational," because the National Council on Crime and Delinquency has said that we know nothing about deterrence, and because it is possible (Continued on page 19)

SOC IErY

them" (Radzinowicz again), then distributive and retributive justice are not strangers but handmaidens. If the quest for both forms of justice is to continue within the framework of the liberal democratic state, then the ideas of the new wisdoms are more than limited; they are both irrelevant and profoundly retrogressive.El Isidore Silver is professor of constitutional law and history at John Jay College of Criminal Justice, City University of New York.

Wilson

(Continuedfrom page 10) that rising crime rates have deterred the use of imprisonment owing to the system being swamped. The first argument is wrong, the second irrelevant, and the third important but inconclusive. Homocides, rapes, and assaults are to Curtis "largely irrational and unplanned." There is ample evidence that many of these offenses are indeed unplanned and affected by the emotional state of the participants. There is no evidence that any but a tiny portion are literally irrational, in the sense that the offenders are insane, incapable of judging the consequences of their acts, or unaware of the immediate risks they run. Offenders may be more or less calculating, but only at the extremes of that continuum can they be said to be perfectly rational or wholly irrational. As long as they are distributed along that continuum, then some perceptible change in risk or benefit will, at the margin, modify their behavior. The problem for analyzing the deterrent effects of variables affecting altercative situations is not that all or most participations are irrational, but that the salience of socially controllable sanctions (arrest, imprisonment, fines) is very low because they are remote (quarreling persons rarely testify against one another in court) and infrequent, compared to the personally controllable sanctions at the disposal of the participants. It is a well-established psychological principle that the effect of a given reenforcement (reward or penalty) on behavior is, among other things, dependent on the ratio of that reenforcement to all other reenforcements operating on a person at the moment. Because of this tenet almost no study of deterrence finds a significant relationship between the probability of imprisonment and the rate of assaultive crime--not because such crime is irrational, but because the probability of imprisonment is not the sanction dominating that form of behavior.

Property Crime In any event, my book is not about assaultive crime, but about predatory property crime. With respect to these property offenses, such as robbery, Curtis argues that the "less sophisticated and unprofessional robbers" act impulsively and perhaps randomly, implying that even robbers are irrational. (Curtis does not actually say what inference he draws from this " f a c t . " ) March/April 1977

I cannot believe that he is asserting that robbers are indifferent to the money obtained by the robbery. Certainly John Conklin's study of robbers comes to quite the opposite conclusion, and it is the best monograph available on the subject. To be sure, many robbers behave casually or impulsively; so do most persons shopping in supermarkets, buying clothes, or leafing through this magazine. No one seriously supposes, however, that if the price of the goods being impulsively purchased went up sharply, the amount of these goods consumed would not (depending on their price elasticity and the availability of substitutes) go down. It happens every day. I cannot imagine why we would believe that robbers are any less (or more) rational in their robbing than they are in their purchases of clothing, CB radios, or beer, impulsive as all these things may be. In any event, we need not speculate: almost every study that has been done (there are now about forty) shows that, others things being equal, the robbery rate declines with increases in the cost of robbery (measured as the probability of arrest or the probability of punishment.)

Deterrence The second argument is no argument at all. The National Council on Crime and Delinquency (NCCD), which asserted that for every study showing deterrence to operate there are "at least two that did not," is an organization that has performed some important and useful services; but one must not forget that it is essentially a special interest group with a favored policy that it tries to sell to public officials. That policy is, roughly, deprisonization; and to that end Milton Rector and the NCCD have regularly urged programs aimed at reducing the certainty and severity of criminal sanctions. It is hardly surprising that the NCCD should believe that deterrence is unproved. Its claim, absent any detailed evidence, is to be given no more and no less attention than the rival claims of the National Rifle Association or People Against Handguns. I know of no serious scholar who believes that there are a large number of competent studies that disprove deterrence. I do know of many scholars who are unconvinced by studies that claim to prove deterrence, and this situation leads to Curtis's third argument. He is quite correct to say that Alfred Blumstein (and others) have pointed to methodological problems in many deterrence studies. I am a member of Blumstein's National Academy of Sciences panel, and we have been struggling with this issue for the better part of a year. Unfortunately, Curtis does not explain what the methodological problem is, how it might be overcome, and what assumptions one must make before one can accept or reject it. Briefly, there are a number of problems---errors in measuring key variables (such as the crime rate) and the existence of a mutual or simultaneous relationship between crime and punishment. The latter is the most important difficulty, because no amount of data or statistical manipula19

tion will eliminate it. If the rate at which sanctions such as imprisonment are imposed is affected by the rate at which crime occurs--as it might be if high crime rates overload the court system--then the observed negative correlation between crime rates and sanction rates cannot be interpreted as proving that deterrence is operating. One cannot choose, in short, between believing that sanctions deter crime or that crime deters the imposition of sanctions. There are three ways to cope with that problem. One is to assume that this simultaneous relationship exists and then to find some "identification restrictions" that permit one to assess the direction of causality in the equations. This approach is difficult because it is hard to think of a factor that will influence crime rates but that will not also influence sanction rates. Another way is to assume, on the basis of reasonable arguments and observations, that the relationship is not simultaneous--that sanctions are imposed at rates unaffected by how much crime exists. In my own research on the police 1 find little evidence that the police are swamped by crime and hence unable to arrest as many people as formerly; and I find considerable evidence that the higher the arrest rate, the lower the robbery rate. A third way is to use not cross-sectional studies of crime statistics, but actual experiments designed to reduce crime (as New York City did in its subway system) or to use data on individual criminals and to see the effects on their behavior that result from being punished. The matter is complicated, and the panel on which Blumstein and I serve has not yet reported. It may well turn out that aggregate studies of imprisonment rates are not the best way to measure deterrence effects, but that result hardly exhausts the available research strategies. Nor do the methodological difficulties of those studies, even if insuperable, disprove the existence of deterrence. Curtis should bear in mind, however, what he is throwing out if he rejects these statistical studies of deterrence. The very equations used to discover the apparent deterrent effect of imprisonment are also those used to show the meliorative effect of social programs such as increased employment and income redistribution. For example, Ehrlich finds that crime rates decrease with decreases in unemployment and increases in imprisonment. To the extent that methodological problems--measurement errors or simultaneity--prevent drawing inferences about deterrence from these equations, they also prevent drawing inferences about the value of changing social factors as well. Passionate Disagreement The remainder of Curtis's criticisms are, in general, untrue, irrelevant, or ad hominem. There is no systematic evidence that prisons are "schools of crime," if by that phrase is meant that, other things being equal, released prisoners are more likely to commit crimes than matched offenders who have not been sent to prison. I do not assert that social scientists believe that there is a simple, straightforward relationship between poverty, racism, and crime; indeed, I devote a whole chapter to showing that social 20

scientists have not generally argued this premise. Apparently this is another chapter Curtis either did not read or did not remember. Social scientists have some very complex views on crime, just as I do; simple labels are inadequate in either case. Lester Thurow would smile to be told that he has "thoughtful plans" that would reduce economic inequality or that economic inequality contributes substantially to crime. I do not doubt that Thurow favors having less inequality; but Curtis's simplistic description of his views suggests that here, as elsewhere, Curtis is in over his head. By this time the attentive reader will have surmised that I do not agree with Curtis, and might suspect that I bring some passion to the subject. The reasons for the disagreement I have stated; the passion I now explain. It is not that I am absolutely confident of the correctness of my analysis; Isidore Silver, to whom I shall turn in a moment, is nearer the mark when he suggests that on crucial points I am equivocal or uncertain. Whatever heat there is in my response to Curtis comes in part from the systematic misrepresentation of my views and in part from his tendency to use abusive, ad hominem slogans in lieu of reasoned arguments. I am told that I favor "benign neglect," whatever that phrase is. It is apparently something bad, but I doubt that Curtis knows what the phrase meant when it was used or even cares; it is a convenient stick. He can only "assume" that I do not believe in genetic differences between races. I say practically nothing about race in the book, and I certainly say or imply nothing about genetic differences among races. If Curtis is wondering what I meant about possible "biological" factors that contribute to crime, I would refer him to the writings of his mentor, Marvin Wolfgang, who believes--as I dtr---that though some intriguing correlations have been found between various neurophysiological states and violence, and though the subject deserves further inquiry, no clear causal patterns have yet been established. Finally, please spare me lectures about how whites cannot understand life in the ghetto. If one were to write that a black could not possibly understand whites, the writer would be promptly branded a racist, and properly so. Indeed, my interest in crime began when I lived in the midst of a black ghetto for five years. During that time I was never the victim of crime, but I saw and knew of countless blacks who were frequent victims; and I know something of the terror they experienced because of what America has allowed to occur on those streets. There are problems with white-collar crime in our society, but I suspect that black women fearful of going out in the street to cash their Social Security checks wish that Curtis would stop changing the subject to corporate malpractice whenever they complain about the crime that is their daily companion. Silver's Issues Isidore Silver's comments on me and other recent authors are a refreshing contrast to those of Curtis, not because Silver agrees with me----on crucial matters he does not--but because he raises the right questions and does so civilly though SOCI ETY

firmly. Silver essentially questions whether we really know very much about crime at all, and certainly whether we know enough to be very confident about policy recommendations. We do not know a lot, and our policy views ought to be quite tentative. Indeed, my own effort to be tentative and somewhat nonspecific has led Silver and others to find ambiguity in my views and to suppose that I mean something that I may not in fact mean. To the extent that false interpretations can fairly be placed in my arguments, I am chiefly to blame. I welcome this opportunity to clarify a few of them. First, I do not subscribe to the view that penal policy should be based on a prediction of future dangerousness. Ernest van den Haag and I have argued about this point before, and we shall no doubt argue again. The current and likely future state of the predictive arts is so imperfect as to make the imposition of sanctions based on predictions a profound injustice. My criticisms of parole are in part based on this skepticism about prediction, and for this and other reasons I would prefer to see the discretionary authority of parole boards sharply curtailed or eliminated altogether. Second, Silver is guessing when he suggests that l must favor imprisoning first offenders. Most first offenses are too trivial to warrant so severe a sanction; even the police are reluctant to send a first-time loser to prison, especially if he is a juvenile. There are, however, some first offenses for which I would expect imprisonmentimurder and rape are two, an offense committed with a firearm is another, the significant abuse of public trust is yet another. Third, I do not "shrug off" middle- or upper-class crime; instead, I devote a whole chapter, the second, to trying to explain why predatory street crime is important and why it ought to be possible to write a book about it without having first to dwell on the sins of the affluent. Silver may disagree with my reasons for writing about street crime, but he ought to state the reasons rather than shrug off my arguments-or write his own book about corporate crime. Finally, Silver wonders whether I am really and fully wedded to the notion that criminal policy ought to be based exclusively or primarily on deterrence. I am not. A strictly deterrent view leads directly to some miserable philosophical problems, most of which have been amply developed by the many critics of utilitarianism. Furthermore, I am not convinced, for reasons stated in my reply to Curtis, that we shall ever know enough about how deterrence works to be able to bet confidently that that policy alone will make a significant and acceptable difference in crime rates. Were I given the task of designing penal sanctions, I would begin as a retributivist--i.e., as one who sees the first test to be that of justice. I would try to propose penalties that seemed morally suitable for crimes and circumstances of various kinds. Silver and I might disagree about some of these penalties, though I am willing to guess that, locked in a room for a day or so, we would find that we disagree on relatively few. But in justifying that schedule of penalties, based in the first instance on a concept of "just deserts," I would try to estimate the gains to society that might result from the deterrent or incapacitative effects of those penalties. Such March/April 1977

facts and estimates would help society decide whether it agreed with those penalties and whether it was prepared to spend much or little to see them institutionalized.

Tolerating Crime These are subsidiary issues, however. Essentially Silver is asking whether a liberal democratic state may not have to tolerate a great deal of crime in order to remain liberal and democratic, unless it is prepared for profound (but specified) social and economic changes. If I thought that curtailing our liberties were the price of less crime, I would endure the c r i m e - - I can always flee from it (to the distant suburbs or to small towns), but I doubt that I could anywhere regain any lost liberties. The system designed by the framers of the Constitution has made it possible to retain and even enlarge our liberties during the midst of a prolonged and pervasive crime wave. Few other nations would have been so disposed--martial law, suspensions of the writ of habeas corpus, and arrests on suspicion are weapons readily available in almost every other nation in the world. But I do not agree that we should do nothing about crime. What Silver describes as "retrogressive" and "irrelevant" are, insofar as my proposals go, nothing of the kind. Is it retrogressive to imprison or otherwise punish persons who have raped, robbed, beaten, and stolen? Is it not more retrogressive to allow them to produce more victims or to encourage the victims to take the law into their own hands? To label the punishment of the guilty as retrogressive is to label justice itself as retrogressive. The " n e w classicism" is neither new nor classic, but a reaffirmation of some elemental standards of fairness appropilate to a society that is neither vengeful nor prepared to require of the family that it once again assume the responsibilities it had in the eighteenth century but has since surrendered to the secular state. If there are profound socioeconomic reforms that can achieve this end more simply, Silver ought to tell us what they are.F-I James Q. Wilson is Henry Lee Shattuck Professor of Government at Harvard University.

van den Haag (Continued from page 11) succeed in deterring all but the undeterrables. (But why do we have so many more irrational rapists and murderers than the Japanese, the English, or the Italians? Is it just a coincidence that the actual probability and the actual severity of our punishments is lower?)

Free Will and Deterrence Let me turn to Silver. He too attacks beliefs that neither Wilson nor I hold. We are accused of believing that "every man i s . . . capable of---and does--calculate his conduct," and that "free will runs rampant." (Elsewhere Silver mangles a quote from Punishing Criminals rejecting that view. He does not realize that his quotation contradicts his 21