Cornish and clarke 1986 the reasoning criminal

Who Can Help Me Criminal Theory Graduate Course?

DATE DOWNLOADED: Thu Feb 11 00:04:57 2021 SOURCE: Content Downloaded from HeinOnline

Citations:

Bluebook 21st ed. Derek B. Cornish & Ronald V. Clarke, Understanding Crime Displacement: An Application of Rational Choice Theory, 25 CRIMINOLOGY 933 (1987).

ALWD 6th ed. Cornish, D. B.; Clarke, R. V., Understanding crime displacement: An application of rational choice theory, 25(4) Criminology 933 (1987).

APA 7th ed. Cornish, D. B., & Clarke, R. V. (1987). Understanding crime displacement: An application of rational choice theory. Criminology, 25(4), 933-948.

Chicago 17th ed. Derek B. Cornish; Ronald V. Clarke, "Understanding Crime Displacement: An Application of Rational Choice Theory," Criminology 25, no. 4 (November 1987): 933-948

McGill Guide 9th ed. Derek B Cornish & Ronald V Clarke, "Understanding Crime Displacement: An Application of Rational Choice Theory" (1987) 25:4 Criminology 933.

AGLC 4th ed. Derek B Cornish and Ronald V Clarke, 'Understanding Crime Displacement: An Application of Rational Choice Theory' (1987) 25(4) Criminology 933.

MLA 8th ed. Cornish, Derek B., and Ronald V. Clarke. "Understanding Crime Displacement: An Application of Rational Choice Theory." Criminology, vol. 25, no. 4, November 1987, p. 933-948. HeinOnline.

OSCOLA 4th ed. Derek B Cornish and Ronald V Clarke, 'Understanding Crime Displacement: An Application of Rational Choice Theory' (1987) 25 Criminology 933

Provided by: University of Memphis Cecil C. Humphreys School of Law

-- Your use of this HeinOnline PDF indicates your acceptance of HeinOnline's Terms and Conditions of the license agreement available at

https://heinonline.org/HOL/License -- The search text of this PDF is generated from uncorrected OCR text. -- To obtain permission to use this article beyond the scope of your license, please use:

Copyright Information

https://heinonline.org/HOL/Page?handle=hein.journals/crim25&collection=journals&id=945&startid=&endid=960
https://heinonline.org/HOL/License
https://www.copyright.com/ccc/basicSearch.do?operation=go&searchType=0&lastSearch=simple&all=on&titleOrStdNo=1745-9125
UNDERSTANDING CRIME DISPLACEMENT: AN APPLICATION OF RATIONAL CHOICE THEORY

DEREK B. CORNISH London School of Economics

RONALD V. CLARKE Rutgers, State University of New Jersey

It has been claimed that the rational choice perspective, which sees criminal behavior as the outcome of decisions and choices made by the offender, can provide a usefulframework for analyzing crime control poli- cies. By developing the concept of "choice-structuring properties," which refers to the constellation of opportunities, costs, and benefits attaching to particular kinds of crime, this paper attempts to develop rational choice theory in order to improve analysis of crime displacement--a concept fre- quently invoked by the critics of opportunity-reducing measures of crime prevention.

The model of the offender as a decision maker underlies much criminologi- cal work recently undertaken by psychologists, economists, and sociologists of deviance (Clarke and Cornish, 1985; Cornish and Clarke, 1986a). This "rational choice" perspective on crime assumes that offenders seek to benefit themselves by their criminal behavior; that this involves the making of deci- sions and choices, however rudimentary on occasions these choices might be; and that these processes, constrained as they are by time, the offender's cogni- tive abilities, and by the availability of relevant information, exhibit limited rather than normative rationality. Our own formulation of rational choice theory was founded on the additional premise that the decision processes and the factors taken into account are likely to vary greatly at the different stages of decision making and among different crimes. For this reason, we drew attention to the needs both to be crime-specific when analyzing criminal choices and to treat decisions relating to the various stages of criminal involvement in particular crimes (initial involvement, continuation, desis- tance) separately from those, such as target selection, relating to the criminal event itself (Clarke and Cornish, 1985; Cornish and Clarke, 1986a).

A RATIONAL CHOICE PERSPECTIVE ON CRIME DISPLACEMENT

Our intention in developing an emphasis upon criminal decision making was to provide a general framework for thinking about the prevention and

CRIMINOLOGY VOLUME 25 NUMBER 4 1987

CORNISH AND CLARKE

deterrence of crime, but our particular interest in rational choice theory arose out of work on "situational" crime prevention-a range of preventive meas- ures, including defensible space architecture, target-hardening, and neighbor- hood watch, designed to reduce the opportunities for, and increase the risks of, committing specific kinds of crime (Clarke, 1983). Despite evidence of its utility, critics have seized upon one apparent weakness of the approach: that preventive measures which increase the difficulties of a particular crime will merely result in criminal activity being "displaced"-for example, to other targets, times, places, or types of crime (Reppetto, 1976; Gabor, 1981). Cru- cial to this objection is the belief that, to the offender, many if not most crimes are functionally equivalent-a view that derives from the traditional hydraulic view of offending as the product of enduring criminal drives or dispositions (Cornish and Clarke, 1986b).

Crucial to the viability of situational approaches, on the other hand, is the contrasting view that displacement is far from inevitable and occurs only under particular conditions. Rational choice theory assumes that offenders respond selectively to characteristics of particular offenses-in particular, to their opportunities, costs, and benefits-in deciding whether or not to dis- place their attentions elsewhere. Indeed, since the existence of criminal dis- positions is questioned, so too is the corresponding notion of criminal "energies" which have to be displaced into alternative actions. If frustrated from committing a particular crime, the offender is not compelled to seek out another crime nor even a noncriminal solution. He may simply desist from any further action at all, rationalizing his loss of income (for example) in various ways: "It was good while it lasted"; "I would have ended up getting caught"; and so on. Such an analysis is consistent with the available empiri- cal research, which is indicative of the contingent nature of displacement. For example, the fitting of steering column locks to all cars in West Germany in 1960 brought about a 60% reduction in car thefts, whereas their introduc- tion only to new cars in Great Britain displaced theft to the older, unpro- tected vehicles (Mayhew, Clarke, Sturman, and Hough, 1976). Again, while a variety of security measures dramatically reduced airliner hijackings in the early 1970s (Wilkinson, 1977), a police "crackdown" on subway robberies in New York City displaced robberies to the street (Chaiken, Lawless, and Ste- venson, 1974).

Research of this kind, however, which merely analyzes crime patterns, is likely to yield only limited information about displacement. This is because, just as reductions in target crimes brought about by situational measures may be modest and difficult to detect, especially when crime as a whole is rising, so, too, evidence of displacement may lie concealed within the same overall crime statistics. Moreover, such research on its own fails to provide an ade- quate explanation for the occurrence or absence of displacement, although reasons may sometimes be inferred. Given these problems, additional ways of

934

RATIONAL CHOICE THEORY

investigating displacement are needed and, in particular, studies which focus upon the offender's own explanations for his decisions and choices.

THE CONCEPT OF CHOICE-STRUCTURING PROPERTIES

A more promising approach to the study of displacement is suggested by rational choice theory's emphasis upon the need to adopt a crime-specific focus when attempting to explain or prevent criminal behavior. Rather than assuming that potential offenders are fueled by a general disposition to offend which makes them relatively indifferent to the nature of the offense they com- mit, the rational choice perspective asserts that specific crimes are chosen and committed for specific reasons. Decisions to offend, in other words, are influ- enced by the characteristics of both offenses and offenders, and are the prod- uct of interactions between the two. Thus, the final decision to become involved in a particular crime is the outcome of an appraisal process which (however cursory) evaluates the relative merits of a range of potential courses of action, comprising all those thought likely in the offender's view to achieve his or her current objective (for example, for money, sex, or excitement).

It follows that an understanding of the factors which the offender takes into account when performing this rudimentary cost-benefit analysis is neces- sary. These factors relate both to offense and offender characteristics but, for the present, can be usefully viewed as those properties of offenses (such as type and amount of payoff, perceived risk, skills needed, and so on) which are perceived by the offender as being especially salient to his or her goals, motives, experience, abilities, expertise, and preferences. Such properties pro- vide a basis for selecting among alternative courses of action and, hence, effectively structure the offender's choice. The characteristics of offenses which render them differentially attractive to particular individuals or sub- groups (or to the same individuals and groups at different times) have there- fore been termed choice-structuring properties. It follows that the readiness with which the offender will be prepared to substitute one offense for another will depend upon the extent to which alternative offenses share characteristics which the offender considers salient to his or her goals and abilities. A recog- nition of the contingent, crime-specific nature of criminal decision making therefore has important implications for an understanding of displacement.

In the absence of information from offenders, some a priori selection of properties thought likely to be salient to offender decision making has to be made. For illustrative purposes, this is attempted later in the paper in rela- tion to two broad groups of offenses-those of theft involving cash and of illegal substance abuse (Table 1). The concept of choice-structuring proper- ties was first employed, however, in the attempt to clarify policy issues relat- ing to gambling and suicide (Cornish and Clarke, in press).

CORNISH AND CLARKE

In the case of gambling (and following the work of Weinstein and Deitch, 1974), choice-structuring properties such as number and location of gambling outlets, frequency of events on which bets can be made, time elapsing before payment of winnings, range of odds and stakes, degree of personal involve- ment, skills needed or perceived, and "nerve" required, were employed to identify forms of gambling more or less designed to encourage high degrees of involvement and to attract the participation of particularly susceptible indi- viduals (Cornish, 1978). In Britain, the widespread provision of "betting shops" in prime urban locations enables off-course gambling to take place throughout the afternoon. These premises offer a vast range of simple and complex betting strategies, a feeling of personal involvement and challenge fostered by the exercise of handicapping skills, and an atmosphere of "action" encouraged by the rapidity of events and payouts, presence of other gamblers laying bets and collecting winnings, and the use of live television commentary from the course-a combination of properties which provides an environment designed to encourage continuous gambling. The contrast with the choice- structuring properties of lotteries is significant: lotteries are held relatively infrequently, involve lengthy periods between staking and payout, offer the minimum of personal involvement, little scope for social interaction or the exercise of skill (real or perceived), a limited range of odds and bets, and very long odds against winning. The prime attraction to their adherents, there- fore, is the possibility they offer of a big "windfall" for very little initial out- lay. "Numbers," on the other hand, while ostensibly rather similar to the lottery, offers a wider variety of staking levels and odds, a larger number of events and swifter turnaround, greater perceived scope for the invocation of personal luck, and more social interaction-features which go some way to explaining why attempts to promote lotteries as legal alternatives to the num- bers racket have proved unsuccessful (Kaplan and Maher, 1970).

It is in examples like these, where activities are examined in some detail, that the value of choice-structuring properties in clarifying the unique con- stellations of motives, opportunities, rewards, and costs offered by different forms of gambling becomes evident. Attention to these parameters also sug- gests a means of controlling participation in potentially dangerous forms of gambling through regulation of these properties. Indeed, this strategy appears to guide the efforts of regulatory bodies and legislators when moni- toring and controlling certain forms of gambling such as betting and gaming, and those of promoters when trying to increase rates of participation and encourage escalation of involvement into more profitable forms. Manipula- tion of the choice-structuring properties of bingo, for example, in order to shorten the duration of individual games, the development of "linked bingo" to enable larger prizes to be offered, and the introduction, as "interval games," of gaming machines-whose choice-structuring properties, especially when deliberately manipulated by casino promotors (Hess and Diller, 1969),

RATIONAL CHOICE THEORY

tend to encourage continuous gambling-all provide graphic examples of these strategies (Cornish, 1978).

In the case of suicide, properties of the various methods such as the degree of prior planning necessary, the courage required, likely pain, distastefulness of method, extent of disfigurement, time taken to die when conscious, scope for second thoughts, and chances of intervention, were used to explain why, when deprived of more acceptable methods, people do not always turn to other means of killing themselves. Domestic gas, for example, used to have particular advantages as a method of suicide: it was painless, very widely available, required little preparation, was highly lethal (death could take place in less than half an hour), was not bloody, and did not disfigure. These features help to explain how the detoxification of domestic gas-a method that had formerly accounted for over 50% of all suicides (Kreitman, 1976; Kreitman and Platt, 1984; Clarke and Mayhew, in press)-brought about a 35% decline in the national rate of suicide in Britain during the 1960s. Some population subgroups such as the elderly and the less mobile may have found these advantages particularly compelling; there is evidence, for example, that suicidal women are more attracted by self-poisoning and more repulsed by violent and bloody methods (Marks, 1977). Since the needs and circum- stances of particular subgroups may make certain methods uniquely attrac- tive, then, it seems likely that reducing opportunities to use particular methods need not simply result in displacement to others, but can bring about genuine gains in the prevention of suicide deaths. Thus, an apparently obvi- ous alternative to gassing, such as overdosing, which might appear to offer many of the same advantages, may nevertheless be subject to disadvantages which limit its viability as a substitute; for example, access to the most lethal drugs may require the cooperation of a doctor, or long-term planning and the faking of relevant symptoms, in order to build up sufficient quantities, while the range of more accessible nonprescription drugs may be either less lethal or, in the case of other alternatives such as domestic poisons, more painful to ingest (Clarke and Mayhew, in press).

CHOICE-STRUCTURING PROPERTIES OF CRIMES

Identifying an activity's unique blend of choice-structuring properties emphasizes its distinctive features and this, in turn, facilitates the making of comparisons between different activities. But, because crimes are such a het- erogeneous group of behaviors, it is not immediately clear on what basis to group crimes for comparison. One possible criterion is suggested by the aim of the exercise, which is to enable the conditions under which displacement is more or less likely to occur to be specified. Since few would expect displace- ment across behaviors engaged in for widely differing purposes, the goals of offending could provide the primary criterion for selecting the crimes to be

CORNISH AND CLARKE

compared. Thus, crimes whose main purpose appears to be to obtain money might be analyzed together, while those whose goal is sexual outlet would need to be separately analyzed. Some a priori determination-later refined by empirical research-of the purposes being served by particular offenses will therefore need to be made before they are grouped together in order to ana- lyze their choice-structuring properties. Although it may be the case that many crimes serve a mixture of goals, one of these will usually be dominant. This will provide the appropriate criterion for analysis, the remaining subsidi- ary purposes taking on the role of further choice-structuring properties for the particular offenses being compared.

For the sake of simplicity, it has so far been assumed that the individual chooses only from among criminal alternatives when seeking to achieve his goals. Given the wide range of noncriminal alternatives also available to the offender, however, confining comparisons of choice-structuring properties to those among criminal means alone may seem unduly restrictive. A crime such as drunken driving, for example, whose purpose is very specific and tem- porary (that is, the need to get home after drinking) and in relation to which

alternative crimes are few or none, illustrates the point that for some crimes most, if not all, of the alternative means being compared will be noncriminal. In addition, displacement will usually be directed in such cases to legal behav- iors: more likely alternatives to drunken driving may be to call a cab, use public transportation, or walk, rather than to persuade an equally drunk companion to drive the car instead. Notwithstanding this example, it seems intuitively more likely that criminal behavior will usually be contemplated only after legitimate means have been foreclosed or rejected. Drunken driv- ing, it could be argued, is a special case since one of the effects of alcohol may be to short-circuit this usual sequence. Under these circumstances, the capac- ity of the otherwise law-abiding citizen to consider the long-range conse- quences of his actions may be temporarily impaired, and this may lead him to entertain criminal actions much sooner (Campbell and Gibbs, 1986: 126, 177). If criminal means are usually only considered at a later stage, this may suggest that they do in fact have something in common with each other and that these features provide some justification for limiting comparisons to crimes alone. But, while this meets the above objections, it also opens the door again to the very dispositional explanations of offending that the rational choice perspective was designed to challenge, since it suggests explanations in terms of offender characteristics, such as the tendency to select means which offer immediate gratification of needs, regardless of the consequences for others. Consequently, the preference at this stage is to defend confining com- parison to crimes alone, not because criminal behavior is inherently different from other behaviors, but on pragmatic grounds alone: it is the possibility of displacement to other crimes which constitutes the major problem for crime- control policy.

RATIONAL CHOICE THEORY

Before embarking on a more detailed discussion of their application to the problem of crime displacement, it may be useful to provide hypothetical lists of the choice-structuring properties of two quite different offense groupings: those designed to yield cash (for example, burglary, theft with or without contact, shoplifting, mugging, bank robbery, fraud, tax evasion, and auto theft); and those concerned with the ingestion of illegal substances (such as marijuana, opiates, LSD, cocaine, "crack," amphetamines, barbiturates, and volatile substances).

As can be seen from Table 1, while specifying the dominant purpose and confining comparisons to criminal means takes one some way toward the goal of drawing up lists of choice-structuring properties, the resulting groupings of offenses will usually be rather broad. While it may be tempting to try for somewhat narrower arrays of offenses, such as those sharing a common modus operandi, this may be unhelpful when estimating the likelihood of dis- placement since it may result in the omission of important choice-structuring properties. In turn, their omission may make it difficult to explain, for exam- ple, why burglars who prey on distant affluent suburbs would never consider breaking into apartments in their neighborhood; why the shoplifter might be reluctant to contemplate mugging; or why the computer fraudster might give up crime entirely if it became too difficult to continue his frauds. In the course of his investigation of robbers' decision making, for example, Feeney (1986) notes the surprising fact that many of them thought burglary too unpredictable and risky.

Similar considerations apply to offenses of illegal substance abuse. An analysis of their choice-structuring properties indicates that different sub- stances provide different experiences, and this-together with considerations of availability, cost, risk, expertise required, and social context of usage- suggests that displacement and escalation among substances may be more limited than is usually thought. Information from opiate abusers, for exam- ple, suggests that a desire to join a specific drug culture of users may be an important determinant of initial involvement (Bennett, 1986); an alternative culture such as that represented by teenage glue sniffing may be seen to offer rather different, and less attractive, experiences in terms of social cachet, excitement, and alternative life-style. In addition, the specific psychological effects of the drugs themselves may restrict substitutability: today's energetic, acquisitive "yuppie" cocaine user may typically be of similar social back- ground to the 1960s cannabis-using hippie, but the effects of cocaine may be more in tune with modern life-styles and aspirations than those produced by cannabis.

Choice-structuring properties may also highlight similarities between apparently different behaviors. For example, crimes such as burglary on a public housing project, in a middle-class suburb, or in a wealthy enclave may,

CORNISH AND CLARKE

Table 1. Choice-Structuring Properties of Two Offense Groupings

Theft Involving Cash

Availability (numbers of targets; accessibility) Awareness of method (e.g., pickpocketing vs. insurance fraud) Likely cash yield per crime Expertise needed Planning necessary (pickpocketing vs. bank robbery) Resources required (transport; equipment) Solo vs. associates required Time required to commit Cool nerves required (bank robbery vs. computer fraud) Risks of apprehension Severity of punishment Physical danger Instrumental violence required Confrontation with victim (mugging vs. burglary) Identifiable victim Social cachet (safecracking vs. mugging) "Fencing" necessary Moral evaluation

Illegal Substance Abuse

Availability (glue from hardware stores vs. prescription drug) Awareness (special knowledge of doctors or pharmacists) Social cachet (cocaine vs. heroin) Solitary vs. social Knowledge/skills required to administer (heroin vs. marijuana) Technical equipment required (heroin) Dangerousness of substance (crack vs. marijuana) Primary method of administration (injecting vs. smoking) Different forms substance can take Nature of psychological effects Number, type, and severity of side effects Dependency Length/intensity of "high" per dose Financial costs Legal penalties Detectability Interference with everyday tasks Moral evaluation

940

RATIONAL CHOICE THEORY

for some offenders, have fewer attractive properties in common than appar- ently different offenses, such as burglary or mugging, committed in their own neighborhoods. While the latter offenses may involve different skills or risks, these may be counterbalanced by the advantages of offending within familiar territory. For these reasons, again, the most appropriate level of analysis for choice-structuring properties would seem to be at the most general level con- sistent with the likelihood of displacement.

Since the lists in Table 1 derive from a rational choice perspective on offending, they both concentrate upon the opportunities, costs, and benefits of the various alternatives being compared. Though no particular attempt has been made to reconcile differences between the two lists, some categories of choice-structuring properties (especially the more generally applicable ones such as "availability") are common to both, while others inevitably reflect unique features of each offense grouping. The properties listed are not neces- sarily those taken into account by the offender, who may not be fully aware either of the range of properties involved or of the part they play in his deci- sions. Rather, the properties listed have been selected on a priori grounds as being of most relevance to the task of comparing offenses and, hence, of estab- lishing the likely limits of displacement within each offense grouping. Thus, there is likely to be more displacement between particular theft offenses where they share similar profiles of choice-structuring properties-for exam- ple, where the likely cash yield per crime is comparable, where similar skills and resources are required, and where the physical risks are the same. In contrast, where the profiles differ, this may clarify why displacement is unlikely to occur. Lastly, some choice-structuring properties may have a more pivotal role to play in decisions concerning displacement. It is generally accepted, for example, that some offenders will not contemplate crimes which involve the use of violence.

Little is known at present about offender decision making, and because of this the above lists may need modifying in the light of empirical research. But even at this stage such lists should provide a useful tool for those involved in crime prevention. By directing attention to those features of crimes which make them attractive to particular groups of offenders, such an approach will make it easier for policy makers to anticipate the direction and amount of any displacement to other forms of crime. In the past, for exam- ple, uncritical and often hidden assumptions that illegal substances are equivalent in their attractiveness and effects may have had damaging effects upon policy formation through their tendency to encourage preoccupation with the inevitability of displacement and escalation. Careful attention to choice-structuring properties of different activities, however, will enable the accuracy of assessments to be improved about the likely costs and benefits of undertaking new crime prevention initiatives in relation to specific forms of crimes. The lists will also alert policy makers to action that needs to be taken

CORNISH AND CLARKE

in order to forestall criminal displacement or even to facilitate displacement to noncriminal alternatives. Finally, lists of choice-structuring properties should assist in the evaluation of crime prevention initiatives by helping to orient the search for displacement.

CHOICE-STRUCTURING PROPERTIES AND OFFENDER PERCEPTIONS

The choice-structuring properties in Table 1 attempt to provide a compre- hensive list of the salient ways in which crimes with similar goals differ from each other. Although policy makers require such comprehensive information in order to think constructively about displacement, it should not be assumed that offenders will utilize the data in a similar way. As mentioned above, they may lack information about the full range of offenses that could satisfy their goals, they may be unaware of the extent to which available opportunities have structured their choices, they may be ignorant of all the costs and bene- fits of the different offenses, and they may assign particular importance to certain choice-structuring properties (such as eschewing the use of violence, or restricting selection of victims to those of particular socioeconomic or eth- nic groups), which then come to exert a disproportionate influence upon involvement and displacement decisions. Moreover, in practice, offenders may not always take account of the full range of properties. For example, the choice-structuring properties listed in Table 1 are mainly relevant to an indi- vidual's initial decision whether or not to get involved in a particular crime. They may have rather less application to more immediate decisions relating to the commission of a particular offense (or what may be termed the criminal "event"), although a similar comparison process-albeit using a different and more restricted range of properties-undoubtedly takes place when potential targets or victims are being compared. The present lists would become more salient again when, having committed the offense, the offender had to decide whether to continue with a particular form of crime or to desist. Last, as a result of the experience of committing the offense in question, further choice- structuring properties may become apparent to the offender and existing ones may assume a different value. Thus, the degree of steady nerves required may only become apparent once a mugging has been attempted.

As well as exemplifying one of the major premises of the rational choice perspective-that the offender's decision-making processes will tend to dis- play limited rather than normative rationality-the above points also illus- trate the dynamic nature of criminal decision making. Thus far it might well appear that a rather passive role has been assigned to choice-structuring properties in that it has been implied that offenders' needs lead them to search out suitable criminal opportunities in their environments. But, as the term implies, choice-structuring properties may often play a more active role in

RATIONAL CHOICE THEORY

generating offending. Some of the opportunities may offer a constellation of properties sufficiently attractive to provide a temptation to crime, as is often argued to be the case with petty offenses such as shoplifting. These points underline the three-fold distinction made by Maguire (1980) and by Bennett and Wright (1984) among offenders who seize, search for, or create opportu- nities. It is also clear that, as well as specifying features of behaviors (kinds of gambling, methods of suicide, types of crime), choice-structuring properties implicitly specify salient characteristics of the actor, such as his or her needs, preferences, personal characteristics, and perceptions. In other words, the term "choice-structuring property" is a relational concept designed to pro- vide an analytic tool for increasing an understanding of the interaction between person variables and arrays of behaviors-in the case of crime, to specify more closely offenders as well as the offenses they commit. Thus, where crime displacement occurs, a knowledge of the choice-structuring properties which the offenses share may permit more accurate identification of the subgroups of offenders involved; and this may well prove a more fruit- ful way of investigating the interface between offense and offender-and, in particular, issues relating to specialization and generalization (Cornish and Clarke, in press)-than the more static and rigid offender typologies of tradi- tional criminology.

Greater knowledge about all these matters would undoubtedly improve policy makers' ability to predict the likelihood and direction of displacement. But, as well as requiring more information about the way offenders perceive and utilize the choice-structuring properties of crimes, more needs to be known about the criminal opportunity structure within which the offender operates if a complete picture of the determinants of displacement is to be given (Cook, 1986b). First, at a macro level, more ecological research is required in order to explore the changes in opportunities and, hence, in crime rates, brought about by changes in routine activities, life-styles, and commer- cial practices. As has been indicated above, the detoxification of domestic gas in Britain brought about a substantial decline in the national suicide rate dur- ing the 1960s. In the same way, participation in gambling rises whenever new facilities are created (Cornish, 1978). In relation to crime, Wilkins (1964) showed how rises in the rates of auto theft in Britain parallel the increased rates of new car registrations, and Cohen and Felson (1979) showed how increases in burglary in the United States reflected the rise in "stealable" property and in numbers of women working outside the home. More recently, Tremblay's (1986) research on credit card bank frauds has indicated how the introduction of new commercial marketing strategies may also some- times have unforeseen consequences. Thus, a move by certain Canadian banks to extend facilities for check cashing to nonregular customers able to guarantee the transaction by means of a credit card, offered existing credit

CORNISH AND CLARKE

card thieves a novel and lucrative way of preying directly upon the banking system itself instead of upon retailers alone.

The processes through which these changes in opportunities at the macro level take place also require elucidation. The escalation in deaths from car exhaust fumes in Britain from the beginning of the 1970s, for example, sug- gests that learning may have an important role to play in determining changes in suicide rates over time, as people gradually come to identify a novel and attractive method of suicide (Clarke and Lester, 1987). At the micro level, Tremblay's work provides some hints about the circumstances under which, for one particular form of crime, such diffusion of innovation might come about. Previous experience in committing similar forms of crime may sensitize offenders to new variations on their favorite themes; member- ship of criminal knowledge networks may speed up the diffusion of informa- tion among specialists, while the media may spread such information more widely among the noncriminal population. The more dramatic the event- such as hijacking, bank robbery, rape, murder, or suicide-the more vivid, detailed, and widespread the coverage, and the more often, at the time of the event and subsequently at committal and trial, the details are repeated. Under these circumstances, the likelihood of "copycat" offenses may be fur- ther enhanced.

RATIONAL CHOICE THEORY AND CRIME-CONTROL POLICY