Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
7 Lawful Policing L earning whether the police have been successful in reducing crime and fear of crime is important to our normative evaluation of the police, as well as our scientific interests in understanding their behav- ior. There is an equally important set of criteria by which citizens make judgments about the police: their fairness and restraint. Concern about these issues arises primarily because the police are authorized to use force and authority in carrying out their mission. When the authority of the state is used, those who are subjected to it have a right to understand its use and to query whether it has been brought against them fairly and justly. We char- acterize this concern as the lawfulness of the police, as a companion to the concern about their crime control effectiveness. Lawfulness is closely linked to the concept of legitimacy, or the perceptions of citizens regarding police fairness. Legitimacy is the subject of Chapter 8. The same tools that have enabled researchers to examine what police do, why they do it, and their crime control effectiveness can also be turned to the study of whether the police are fair and legitimate. Some of the earli- est and most important pieces of police research were animated by concerns about fairness and lawfulness. Empirical methods were developed to dis- cover the extent to which the police had discretion in how they behaved and how they used the discretion available to them. The achievement of those studies was to show that the world of policing was shot through with dis- cretion, and that the choices the police made about how to use their discre- tion were only imperfectly guided by the legal standards that they ought to have been following. Concerns about whether the police behave fairly, whether they are racially biased, whether they are corrupt, whether they are 252
LAWFUL POLICING 253 brutal, whether they are aggressive or trigger-happy have continued to exist and to be investigated using the methods of the social sciences. Our review of the social science evidence on police lawfulness focuses especially on research that examines police compliance with the U.S. Con- stitution, state and other laws, and with the policies of their departments. Concern about police use of force and the intrusive exercise of their author- ity leads political institutions to develop rules that bring police into compli- ance with such standards. The rules were adopted over time to promote the legitimacy of the police, as a rule-bound institution engaged in the even- handed pursuit of justice and as the protector of individual liberty. We summarize research on police compliance with rules and standards for in- terrogations, search and seizure, the use of excessive and deadly force, and corruption. Courts, legislatures, mayors, and city councils impose standards for conduct on police; the empirical question driving much of this work has been the extent to which they are followed--that extent serves as one mea- sure of the legitimacy of police in a free society. POLICE COMPLIANCE WITH THE LAW AND THE CONSTITUTION The legitimacy of police activity is closely tied to police compliance with legal standards. The goal of ensuring lawfulness in police conduct finds expression in rules guarding individuals' rights to liberty and prop- erty, freedom from unreasonable intrusions, and fair and equal treatment. The Fourteenth Amendment to the United States Constitution forbids dep- rivations of life, liberty, or property without due process of law. The "rule of law [means]...that the citizen should be free from arbitrary power" (LaFave, 1965:64). Through their power to enforce laws--or not--police are invested with a great deal of control over the liberty interests of indi- viduals. Concern about these issues has led to sustained scholarly interest in both the question of how the police use their discretionary power and how best to control it. In recent decades, concern about the exertion of arbitrary police power, as well as racial bias in the exercise of police power, has driven important developments in constitutional law. In Brown v. Mississippi (297 U.S. 278, 1935), the Supreme Court invalidated the defendant's conviction by a Mis- sissippi court because the conviction rested almost entirely on a confession extracted through torture. In overturning Brown's conviction, the Supreme Court articulated a constitutional standard of fundamental fairness for the evaluation of police practices--with regard to practices pertaining to inter- rogation--to be applied to states via the Fourteenth Amendment. Brown v. Mississippi marked the first time the Supreme Court was willing to regulate police practices at the state level, and it set an important limitation on how
254 FAIRNESS AND EFFECTIVENESS IN POLICING the police can use their force and authority to generate evidence to be used at trial. The role of race in determining how Brown was treated also dem- onstrates how closely the articulation of constitutional criminal procedural rights in the United States has been tied to this country's history of racial discrimination (Klarman, 2000). The constitutional rules that prescribe the manner in which police are allowed to legitimately carry out their tasks have their genesis not only in the Court's concerns about how much state authority is appropriate to use in the pursuit of law enforcement objectives, but also in the Court's continuing concern with the fairness of policing--in particular, with its efforts to guard against racial discrimination. Like Brown, later criminal procedure decisions--a muscular body of interconnected doctrines that govern the everyday machinery of law en- forcement--also were influenced by the Court's interest both in restraining the state and rooting out institutional racism and other forms of bias (Kahan and Meares, 1998). In the law, rules of criminal procedure provide the state-level link between constitutional principles and the daily actions of participants in the criminal justice system. Historically, law enforcement was a key instrument of racial repression, in both the North and the South, before the civil rights revolution of the 1960s. Modern criminal procedure reflects the Supreme Court's contribution to eradicating it. The Court, be- ginning in the 1960s, erected a dense network of rules to delimit the permis- sible bounds of discretionary law enforcement authority. These decisions, while reflecting a concern for the effects of racism, make up a more general regime of criminal procedure that tells individual police officers when and how they can interact with criminal suspects on the street;1 minutely regu- late the nature of police interrogations of suspects in custody;2 and spell out in detail the procedures that police must follow before they conduct searches or engage in related forms of surveillance (LaFave, 1996). These rules con- strain, in both substance and form, the authority of police to maintain pub- lic order.3 Although rarely couched as such, the unmistakable premise of these doctrines was the assumption that the historically empowered groups in communities could not be trusted to police their own police because of the distorting influence of racism. 1See, e.g., California v. Hodari D., 499 U.S. 621, 638-39 (1991) (describing interlocking maze of doctrines governing "consensual encounters," "investigative stops," and "seizures of person"). 2See, e.g., Davis v. United States, 512 U.S. 452, 458-62 (1992) (describing multiple "pro- phylactic" rules designed to conform interrogation to right to counsel). 3See, e.g., Papachristou v. City of Jacksonville, 405 U.S. 156 (1972); Kolender v. Lawson, 461 U.S. 352 (1983).
LAWFUL POLICING 255 Of course, the articulation of rules does not automatically lead to legiti- mate policing. The rules of criminal procedure are not self-enforcing, nor are the laws regulating the honesty and ethical behavior of police officers. Even assuming that rules consistent with legitimate policing have been ap- proved by the legislature or adopted by a court, it is necessary for individual officers to actually follow them. The next sections of this chapter review research on the extent to which police follow the rules that govern their behavior in four key areas: interrogations, use of deadly force, searches and seizures, and police corruption. Police Interrogations The Supreme Court began extensive regulation of local police practices in a case involving a clearly unlawful interrogation and racial bias. In a series of cases between 1936 and 1964, the Supreme Court reviewed the particular facts of cases in order to determine whether a confession was voluntary and in accord with the due process clause of the Fourteenth Amendment.4 The difficulties inherent to administering a fact-based vol- untariness test led the Court to develop the more easily administered, and perhaps more effective, method of controlling interrogation techniques in Miranda v. Arizona (372 U.S. 436, 1966). Miranda holds that any confession obtained during custodial interroga- tion is compelled in violation of the right against self-incrimination unless a police officer has first given specific warnings to the person being ques- tioned and unless that person specifically waives those rights. In the famous "Miranda warning," police are required to advise persons they question (1) that they have the right to remain silent, (2) that police are entitled to use any statement they make against them, (3) that they have a right to counsel, and (4) that counsel will be provided if they cannot afford one. Prior to the Miranda decision, the operating standard for confessions or otherwise in- criminating statements elicited from a suspect in custody was "voluntari- ness." Only the Federal Bureau of Investigation (FBI) operated under spe- cific instructions to issue a warning informing suspects in custody of their rights. Miranda is no doubt the best-known criminal procedure decision in the United States. The positive obligation handed to law enforcement as the result of the decision provoked an immediate reaction, reflecting society's division over the police effectiveness issue (Dalton, 1984:62-63). There were complaints that Miranda requirements would undercut the ability of police to catch and convict criminals, and it was predicted that issuing Miranda 4See, e.g., Fikes v. Alabama, 352 U.S. 191 (1957).
256 FAIRNESS AND EFFECTIVENESS IN POLICING warnings would unnecessarily hamstring police in their pursuit of crimi- nals. Since that time, empirical studies have found that Miranda require- ments only minimally curtailed the investigative function of police (See- burger and Wettick, 1967; Wald et al., 1967; Zeitz et al., 1969; Black and Reiss, 1967a, 1967b; Milner, 1971a; Schaefer, 1971). In a recent opinion in another case involving police powers, the chief justice of the Supreme Court acknowledged the status of Miranda, stating that the ruling "has become embedded in routine police practice to the point where the warnings have become part of our national culture."5 The question is, do police regularly adhere to its strictures? The bulk of the studies of Miranda occurred shortly after the decision.6 Research on the ruling is of two types. A set of effect studies examine confession rates before and after Miranda was decided, to see whether the warnings affected the number of confessions obtained by police. The second stream of research involves the direct observation of police behavior. One of the earliest empirical assessments of police compliance comes from the research project directed by Donald Black and Albert Reiss that was already under way when the Supreme Court announced its decision (Black and Reiss, 1967a, 1967b). Their findings belied the police conten- tion that interrogation was necessary in obtaining evidence against a sus- pect. Especially in felonies, Black and Reiss noted that there typically was alternate evidence against a suspect that police could rely on, including physical evidence and other witness testimony (1967b:53). They also found, following the Miranda decision, that when police officers initially encoun- tered and arrested suspects, the required warning was rarely given (Black and Reiss, 1967a:102-109). The lack of compliance observed by Black and Reiss may have stemmed from the fact that the decision had been handed down only recently. Subse- quent studies found that police by and large did issue the Miranda warning, albeit in an often perfunctory and superficial fashion (Wasby, 1970; Leiken, 1971; Baum, 1979). Summaries of Miranda effect studies suggest that the warning as it is actually delivered may have resulted in a reduction in confessions of be- 5Dickerson v. United States, 120 S. Ct.2326, 2336 (2000). 6See Neil A. Milner, The Court and Local Law Enforcement: The Impact of Miranda (1971); David W. Neubauer, Criminal Justice in Middle America (1974); John Griffiths and Richard Ayres, Faculty Note, A Postscript to the Miranda Project, Interrogation of Draft Protesters, 77 Yale L.J. 395 (1967); Lawrence S. Leiken, Police Interrogation in Colorado: The Imple- mentation of Miranda, 47 Denv. L.J. 1 (1970); Richard Medalie et al., Custodial Police Inter- rogation in Our Nation's Capital: The Attempt to Implement Miranda, 66 Mich. L. Rev. 1347 (1968); Neil A. Milner, Comparative Analysis of Patterns of Compliance.
LAWFUL POLICING 257 tween 4 and 16 percent.7 Milner (1971a, 1971b), examining the use of Miranda in four Wisconsin cities, attributed some intercity variation he found to differences in police training and access to other sources of legal information across departments. Miranda studies based on interviews with police or offenders or on official records cannot tell us directly whether police follow the Miranda requirements; for this, the best evidence is direct observation. Leo (1998) has compiled the best data on the topic. In an observational study of police detectives, Leo found that detectives provided Miranda warnings in all the cases in which they were legally required to do so, or in approximately 96 percent of the cases. In the remaining cases, the Miranda warning was not legally required because the suspect was not technically "in custody." Sev- eral other empirical studies also conclude that police overwhelmingly fol- low the guidelines set forth in Miranda.8 Leo also found that police seek to present the Miranda warnings in ways that encourage individuals to waive their constitutional rights. The only social, legal, or case-specific variable that was a significant predictor of the suspect's likelihood to invoke Miranda rights was the existence of a criminal record. After Miranda rights were invoked, 96 percent of suspects were returned to jail; in the remaining 4 percent of cases, detectives contin- ued to question the suspects after informing them that information pro- vided could not be used in a court of law against them. Cassell and Hayman (1998) also observed the implementation of Miranda. They concluded that "evidence suggests that police have adjusted to Miranda by shifting to non-custodial `interviews' to skirt Miranda's re- 7Law & Soc'y Rev. 119 (1970); David W. Neubauer, Confessions in Prairie City: Some Causes and Effects, 65 J. Crim. L. & Criminology 103 (1974); Cyril D. Robinson, Police and Prosecutor Practices and Attitudes Relating to Interrogation as Revealed by Pre- and Post- Miranda Questionnaires: A Construct of Police Capacity to Comply, 3 Duke L.J. 425 (1968); Roger C. Schaefer, Patrolman Perspectives on Miranda, 1971 Law & Soc. Ord. 81; Richard Seeburger and R. Stanton Wettick, Jr., Miranda in Pittsburgh --A Statistical Study, 29 U. Pitt. L. Rev. 1 (1967); Otis Stephens et al., Law Enforcement and the Supreme Court: Police Per- ceptions of the Miranda Requirements, 29 Tenn. L. Rev. 407 (1972); Michael Wald et al., Interrogations in New Haven: The Impact of Miranda, 76 Yale L.J. 1519 (1967); James W. Witt, Non-Coercive Interrogation and the Administration of Criminal Justice: The Impact of Miranda on Police Effectuality, 64 J. Crim. L & Criminology 320 (1973); Evelle J. Younger, Interrogation of Criminal Defendants--Some Views on Miranda v. Arizona, 35 Fordham L. Rev. 255 (1966); Evelle J. Younger, Results of a Survey Conducted in the District Attorney's Office of Los Angeles County Regarding the Effect of the Miranda Decision upon the Pros- ecution of Felony Cases, Am. Crim. L. Q. 32 (1966). 8See Tracey L. Meares and Bernard E. Harcourt, Transparent Adjudication and Social Sci- ence Research in Constitutional Criminal Procedure, 90 J. Crim. L. & Criminology 733, 769, tbl. 11.
258 FAIRNESS AND EFFECTIVENESS IN POLICING quirements" (p. 228). Although much of their evidence for the change in behavior is anecdotal, Cassell and Hayman do indicate that 70 percent of their observed interviews were custodial, while 30 percent were noncusto- dial. Whether this represents a change from pre-Miranda figures is unclear. Cassell and Hayman further argue that police were less successful in non- custodial interviews. The fact that police have begun to use them more despite their lower rate of success suggests to the authors that, "contrary to the view of some defenders of Miranda, . . . interrogating police officers believe the Miranda rules are harmful to their efforts" (p. 229-230). An- other way of interpreting Cassell and Hayman's finding is that Miranda actually does provide individuals protection against state power. Research in this area is dated, perhaps reflecting earlier findings that police were by and large complying with the letter of the law, and that it did not appear to have much effect on their effectiveness. Other custodial rights issues have not been investigated at all. Because the most compelling re- search in this area involves the direct observation of police behavior, in- cluding both by patrol officers and detectives, it is laborious and expensive to conduct. The resulting data typically represent only one jurisdiction, making it difficult to infer the effects of departmental policy and other man- agement-level factors on the implementation of legal rules. Excessive and Lethal Force The potential for using force underlies many of the functions exercised by the police. In fact, one definition of the police is that it is the body lawfully authorized to exercise lethal force against citizens of the state, hold- ing a virtual monopoly over this power. As a price for holding this mo- nopoly, and because its use is inevitable, there are standards for the applica- tion of force by police. When force is appropriate to use, and how much and what kind of force may be used, are all defined both by constitutional rules and by statutes that create a liability for both police departments and individual police officers who misuse the force and authority entrusted to them. In addition, police administrators usually strive for economy in the use of force; good police work, in this view, is policing that employs only the force that is required in a particular situation, and not more. A key concern to policing in the 21st century is the proper use of force, since the real or perceived misuse of force can thoroughly and quickly undermine police legitimacy. In the United States, use of lethal force has been a major source of conflict between minority groups and the police since the inception of the institution (Walker, 1977). Numerous studies have demonstrated that blacks are shot and killed by the police in numbers that are vastly dispro- portionate to the number of whites who are shot and killed. Between 1950
LAWFUL POLICING 259 and 1960, one study found that blacks were killed by Chicago police at a rate of 16 per 100,000, compared with a rate of 2 per 100,000 for whites (Robin, 1963). From October 1966 to October 1974, the Memphis Police Department recorded approximately 225 instances of firearm discharges to attempt to stop fleeing felon suspects; 114 of those shot by police were suspected of nonviolent property crimes. Of the 114 shot, 96 were black. In a study of fleeing felons in Memphis between 1969 and 1974, one researcher found that police officers shot and killed 13 blacks and 1 white person (Fyfe, 1982). Many see such disparities in the exercise of force lying at the core of challenges to the legitimacy of policing in the United States in the 21st century. The constitutional rule adopted by the Court to circumscribe the use of deadly force by police officers is a product of another case, Tennessee v. Garner. In 1985, the Supreme Court overturned that state's permissive flee- ing felon rule, which allowed police officers to use "all the means necessary to effect an arrest," even of an unarmed fleeing felon, after notice of inten- tion to arrest was given. Scholars have found that guidelines contained in Garner led to a few changes in state law. However, by 1990 only 4 of the approximately 30 states whose laws fell outside the Garner boundaries had changed their laws to bring them into compliance with Court guidelines.9 Most of the remainder continued to legally authorize shootings of nonvio- lent fleeing felons (Fyfe and Walker, 1990). There were more changes in the administrative policies of departments governing police use of force (Skol- nick and Fyfe, 1993). These rules in turn have been evaluated and appear to have had an effect on police shootings. Tennenbaum (1994) concluded that Garner reduced fatal police shootings in the United States by about 60 per year. This is a difficult research area. There is no central repository of data on police use of force, and the data collected nationally by the FBI on the use of lethal force are deeply flawed. They are voluntarily submitted and they do not include injuries that fall short of death. The data that are sub- mitted are known to be inaccurate, based on comparisons between them and other data sources (Fyfe, 2002). Gaining access to local records on this potentially explosive and politically sensitive topic is difficult, and agencies have much to lose by cooperating with researchers. As a result of these difficult circumstances, virtually every study of police use of force has been based on the records of one or a small number of local police departments. 9See, e.g., Paul G. Cassell and Bret S. Hayman, Police Interrogation in the 1990s: An Em- pirical Study of the Effects of Miranda, in The Miranda Debate: Law, Justice, and Policing 222, 231 (Richard A. Leo and George C. Thomas III eds., 1998) ("Our data thus support the emerging consensus that police play by the Miranda rules.").
260 FAIRNESS AND EFFECTIVENESS IN POLICING Furthermore, it makes a difference how the data are obtained. On one hand, studies conducted in agencies that voluntarily open their records to researchers appear to represent agencies that are most confident of their professionalism. In general, these studies show that racial minorities are subjects of police force at higher rates than for whites, but that shootings by police are fairly racially balanced (see, for example, Fyfe, 1981, and Blumberg, 1981). On the other hand, studies of agencies that are forced to open their records because of suits alleging use of excessive force, or Free- dom of Information Act suits by media organizations, tend to find more racial disparity in the use of force, a great deal of disparity in the use of lethal force, and a higher rate of shootings of racial minorities that appear to be questionable (Fyfe, 2002). For example, Meyer (1980) found in Los Angeles that blacks were more often unarmed when they were shot, and Fyfe (1982) found that blacks in Memphis were more often shot in circum- stances that were not as threatening to the officer. Official case files typi- cally are the source of data on these incidents, and these documents present a version of events in which every incentive exists for the organization to present a favorable version of events, and much of the information is based on paperwork completed by the officer who used the force. One of the few firm conclusions that can be drawn from this research is that rates of police use of force and lethal force are highly variable. Fyfe (2002) analyzed the results of a project conducted by the Washington Post that assembled data on fatal police shootings in 51 large municipal and county police and sheriff's departments during 1990-2000. Fatal shootings rates for county police departments varied by a factor of 14, while for city departments the ratio of shootings from top to bottom was 8 to 1; among sheriff's departments, it was almost 6 to 1. In a seven-city study by Milton et al. (1977), the top to bottom ratio was also 6 to 1. Another general and important conclusion from these analyses is that most police use of force is nonfatal. In one 6-agency study, only 18 percent of adult custody arrests involved the use of physical force or threats of physical force, and most of the force was confined to threats, the use of restraints, weaponless tactics, and control holds (Garner and Maxwell, 1999). A final conclusion is that, as noted above, there is usually racial disparity in the use of nonlethal force, and often considerable racial disparity in the use of lethal force. Despite problems with the data, there is evidence of the positive effects of legal and administrative efforts to control police use of force. In a study of the use of force by the New York police department, Fyfe (1979) found that a policy change by that agency led to a precipitous drop in shootings by officers there. He also found that New York City police rarely shot un- armed people. Sparger and Giacopassi (1992) conducted a follow-up study in Memphis, the jurisdiction in which the Garner decision originated; they
LAWFUL POLICING 261 found a dramatic reduction in racial disparities in police shootings in the post-Garner period. In a study in Philadelphia, Waegel (1984) used data collected from media reports by a nongovernmental organization, the Police Project of the Philadelphia Public Interest Law Center, to examine the effect of a 1973 state statute restricting use of force by police officers. He concluded that officers did not comply with the statute, and that compliance did not in- crease with time (p. 128): "With but one exception (1977), in every year the most frequent category of lethal force incidents involved an unarmed per- son who was fleeing after the commission of a felony (1970-1972) or an unarmed person who had not committed a violent felony (1974-1978)." However, in a later analysis, Skolnick and Fyfe (1993) concluded that ad- ministrative changes imposed by the city's mayor Frank Rizzo worked against the intent of the statute. In response to the state legislation, he re- scinded the city's long-standing and more restrictive deadly force policy, thus effectively expanding the potential role for deadly force in the city. Later, when a new mayor reestablished the more restrictive deadly force policy, fatal shootings declined in Philadelphia by 67 percent in the follow- ing year (Fyfe, 1981). Skolnick and Fyfe (1993) were also critical of the use of media reports as a source of data for police shootings, finding that one- third were not reported in the press. (For more on this, see also Fyfe, 1988 and 1987). The committee has deliberated on the problems inherent with data col- lection on police use of lethal force; we remain convinced that the obliga- tion to collect such information outweighs its inherent difficulties. As we have already noted, an effort to collect these data has been undertaken by the International Association of the Chiefs of Police (IACP). Like the development of the Uniform Crime Reports in the early part of the 20th century (Cross, 1917), implementing a voluntary data collection initiative on the use of force by police agencies is a challenging endeavor filled with complexity and the potential for error and subversion. While operationalizing concepts can be difficult in any large national study, the definitional problems in this area are readily apparent. The IACP defines police use of force as "the amount of force required by police to compel compliance by an unwilling subject" (Henriquez, 1999:20). This definition excludes excessive force, since excessive implies that the amount of force exceeds that required to compel compliance. If anything, it sounds like a definition for "reasonable" force, not use of force more generally. It also imposes a normative standard, forcing localities to define the amount of force required to compel compliance and then count those instances in which that level of force was used. A better definition would rely on the officers' overt forceful behaviors, whether justified or not, rather than on
262 FAIRNESS AND EFFECTIVENESS IN POLICING divergent local interpretations of how much force is necessary to compel compliance. Furthermore, the anonymous nature of the data collection means that there is no way to check the reliability or validity of the data. The voluntary nature of agency reporting also has several undesirable con- sequences, most notably that the resulting collection of data cannot be treated as a representative sample of police in the United States. Therefore the data cannot be used to draw inferences about police more generally. Furthermore, since there is no mechanism for ensuring consistent reporting by agencies, the data also cannot be used to calculate trends in police use of force (Henriquez, 1999). Many of the problems with the IACP data will probably iron them- selves out as police agencies continue to participate in the initiative and develop standards for reporting. Even if the data collection becomes rou- tine, there is still one fundamental problem: the anonymous reporting sys- tem means that data from this study cannot be merged with data from other sources, such as the Uniform Crime Reports, the census, or other data sources. This limitation means that the range of potential explanatory vari- ables will be limited; thus the data are not well suited for explaining differ- ences among police organizations. The development of measures useful for assessing interagency varia- tions in use of force is still in its infancy. Although isolated studies have collected national data on police use of force, the only systematic effort to record police use of force nationally is still several years away from realiz- ing its potential as a data source on policing. The committee recommends legislation requiring police agencies to file annual reports to the public on the number of persons shot at, wounded, and killed by police officers in the line of duty. As long as the only comprehensive national data set on police use of force continues to operate in a voluntary and anonymous fashion, it will fail to capture the data that are, arguably, the most important of all. Seizures and Searches The Fourth Amendment states that "the right of the people to be secure in their persons, houses, papers and effects, against unreasonable searches, shall not be violated, and no Warrants shall issue, but upon probable cause, supported by Oath or affirmation, and particularly describing the place to be searched, and the person or things to be seized." Searching citizens and their property is a basic law enforcement tool for solving crimes and build- ing criminal cases against defendants. The framers of the U.S. Constitution recognized the state's legitimate interests to engage in searches, but they were so concerned about the risks of abuse that they enshrined their pro- scription of "unreasonable" searches and seizures in the Fourth Amend- ment, the first of several to address concerns about due process of law.
LAWFUL POLICING 263 Although Americans have long been interested in the contribution their police make to reducing crime, they also expect the police to operate by the rules of law (Roberts and Stalans, 1997:Ch. 7). Since 1961, the Supreme Court has used the Fourth Amendment to establish rules regulating every- day police search and seizure practices. In Mapp v. Ohio (367 U.S. 643, 1961), the Supreme Court applied an exclusionary rule, which already gov- erned federal prosecutions, to the states as well. The rule excluded the fruits of unlawful searches from consideration in court. The Court did so in order to deter unlawful police conduct and secure the guarantee of the Fourth Amendment against unreasonable searches and seizures, since there ap- peared to be no feasible alternative to controlling such behavior. The larg- est body of social science research on police compliance focuses on the effects of the exclusionary rule. Immediately following Mapp, researchers began a series of studies con- cluding that, at least in some areas of the country, the exclusionary rule may have caused police officers to adhere more closely to the requirements of the Fourth Amendment (Goldstein, J Crim Law, 1967; Amsterdam, Minn Law Rev, 1974). Skolnick's (1966) study of Eastville and Westville police is one early example. Skolnick employed direct observation to identify com- mon police search practices. The study notes that officers frequently skirted constitutional standards. It identifies an officer outlook that motivated and justified such actions: a focus on discovering and controlling crime and applying conventional morality to standards of police behavior in dealing with suspects. In this light, technicalities of constitutional law are viewed as unreasonable obstacles to reasonable police goals, and any normative pres- sure to conform to law is substantially mitigated. Other studies were mixed: some found that the exclusionary rule deterred misconduct (Canon, 1974; Wasby, 1976), some found little or no effect (Oaks, 1970), and some were inconclusive.10 A new wave of studies followed the Supreme Court's embrace of a cost/ benefit approach to the exclusionary rule in U.S. v. Calandra.11 These at- tempted to measure and balance the costs (excluded evidence) and the ben- efits (deterrence of police misconduct) of the exclusionary rule. Many focused exclusively on the social costs of the exclusionary rule ("lost con- victions"), and they overwhelmingly concluded that those costs were mar- ginal.12 10See Brief amicus curiae of the Florida Chapter of the National Bar Association in support of the Respondent-Appellee, Tennessee v. Garner, No. 83-1035, 3 (1984). 11Note, Effect of Mapp v. Ohio on Police Search-and-Seizure Practices in Narcotics Cases, 4 Colum. J. Law & Soc. Probs. 87 (1968). 12414 U.S. 338 (1974).
264 FAIRNESS AND EFFECTIVENESS IN POLICING More recently, researchers have returned to examinations of actual po- lice practices. Studies based on field observations provide the most reliable effect data (Cannon, 1991) but are relatively rare in the exclusionary rule literature. Older studies sponsored by the American Bar Foundation con- cluded that police officers knew the relevant constitutional standards and did their best to work around them when searching suspects.13 In addition to being dated, these studies suffer from a reliance on interviews with the police rather than observations of behavior in the field. In 1986, Orfield (1987) studied Chicago narcotics officers to determine how the exclusionary rule has influenced their behavior. Orfield conducted structured interviews with 26 narcotics officers. He found that the exclu- sionary rule had changed police, prosecutorial, and judicial procedures on the institutional level. At the individual level, Orfield found that the officers he interviewed generally adhered to the exclusionary rule and that the rule deterred constitutional violations--at least if the goal was to prosecute the offender, as opposed to confiscate contraband. LaFave (1965) and LaFave and Remington (1965) reported earlier that restrictions on searches and seizures were ineffective when their purpose was instead to disrupt illicit networks or assert police authority on the street. Other studies based on interviews with police officers have been used to estimate the frequency of illegal police procedures (Canon, 1991; Krantz, 1979; Milner, 1971a; Wasby, 1976). This method has the advantage of offering interviewed officers an opportunity to speak confidentially about their work, which could encourage them to be frank regarding their prac- tices. However, police surveys and interviews suffer from a variety of other limitations. Some officers may be unaware that certain search practices in which they engage are illegal (or legal), thus producing unreliable estimates of the rate of illegal search practice. In addition, officer recall of search practice may be faulty. Studies that give officers hypothetical scenarios to respond to may not reflect what they actually do on the job. In a study of officers in the early 1990s, Gould and Mastrofski (2001) conducted field observations of 115 searches conducted by 44 lead offic- ers.14 They conclude that 29 percent of the searches were conducted uncon- stitutionally. This extrapolates to a rate of 6-7 illegal searches per 100 resi- dents over a year's time. However, they report that only three violations were rated so egregious as to "shock the conscience" (p. 36). Importantly, 13See, e.g., Comptroller of the United States, Impact of the Exclusionary Rule on Federal Criminal Prosecutions (1979); Peter Nardulli, The Societal Cost of the Exclusionary Rule: An Empirical Assessment, 3 Am. B. Found. Res. J. 585 (1983). 14American Bar Foundation, Survey of the Administration of Justice (1957); L.P. Tiffany, D.M. McIntyre, and D.L. Rotenberg, Detection of Crime: Stopping and Questioning, Search and Seizure, Encouragement and Entrapment (1967).
LAWFUL POLICING 265 this research includes data from searches that do not lead to arrest or a suppression motion. In fact, Gould and Mastrofski found that "only 7 per- cent of the suspects who were arrested or cited experienced an unconstitu- tional search, while 43 percent of those who were not arrested or cited experienced a violation of their constitutional right" (p. 36). Clearly, stud- ies that limit their focus to lost convictions (which require first an arrest and then a suppression motion) can miss a significant number of unconstitu- tional searches. Gould and Mastrofski also found that "when constitutional violations occur, they are more likely to involve the frisk of a suspect's outer clothing than a full search of his person or possessions" (p. 38). Fur- thermore, the violations were generally not exceptionally invasive; the au- thors estimate that only two or three of the searches would "rise to the level of egregiousness required for civil liability" (p. 38). Finally, the data indi- cated that "illegal searches were highly concentrated in a few officers....[S]ix officers (14 percent of the officers in the sample) accounted for 22, or two- thirds, of the illegal searches" (p. 53). The design of the study did not en- able them to determine the eventual fate of the 7 percent of cases in which they observed unconstitutional sanctioning. Gould and Mastrofski had limited data on the background of the ob- served officers, so their analysis focused mostly on the features of the sus- pects and the situation. They found that the most powerful predictor affect- ing the probability of an unconstitutional search was whether the officers were explicitly looking for drugs and whether the search was a pat-down as opposed to a more intrusive search. They concluded that this pattern of effects suggested that the city's "war on drugs contributed to the rate of illegal searches. Notably, the study found no evidence that blacks were more likely than whites to receive an unconstitutional search. Interestingly, al- though a small number of officers accounted for a disproportionately large share of the unconstitutional searches, these officers did not conform to a "rotten apple" portrait. Instead these officers were strongly committed to community policing and to treating citizens with respect. Their searches, both legal and illegal, were performed in a respectful, even solicitous man- ner. The researchers found that the rate of illegal searches appeared to be high, but most police transgressions were relatively minor. Given that so few of the illegal searches were officially recorded, the exclusionary rule would prove irrelevant to preventing or correcting most violations. The Supreme Court's U.S. v. Leon decision in 1984 permitted a "good faith" exception to the exclusionary rule for faulty warrant searches. One post-Leon study tracked hundreds of cases in seven cities throughout the legal process, from the beginning of the search warrant request to disposi- tion and disposition of appeal (Sutton, 1986). The research drew on official records, firsthand observation of warrant reviews, and interviews with po- lice and court decision makers. The study found that the search warrant
266 FAIRNESS AND EFFECTIVENESS IN POLICING process was rarely used, the process was perfunctory (with magistrates serv- ing more as allies than independent judges committed to maintaining due process standards), and it was often subverted by circumventing the war- rant requirement, ignoring it, or meeting it through fabrication or falsifica- tion of evidence. Another group of researchers did a pre-post study of search warrant practices in the same seven cities and found that they remain un- changed by Leon, thus adding to the body of research that finds little or no effect from appellate court rulings on police practices (Uchida and Bynum, 1991). More recently, the New York Office of the Attorney General conducted a study of police stop and frisk practices in New York City.15 The study analyzed 175,000 forms completed by police officers over a one-year pe- riod. Researchers compared patterns of stops with census data, crime statis- tics, and demographic information, yielding a quantitative portrait of stop and frisk practices.16 In analyzing this and other studies like it, it is important to note that the Supreme Court legitimated many stop and frisk practices in Terry v. Ohio.17 In Terry, the Court interpreted the Fourth Amendment by balancing indi- vidual interests in liberty (privacy, property, and autonomy) against soci- etal interests in safety (which include, of course, the stopped and searched person's interests). The Court departed from its then long-time insistence that police searches and seizures be justified by the level of particularized suspicion known as probable cause.18 The Court instead determined, focus- ing on the reasonableness clause of the Fourth Amendment, that a protec- tive pat-down of a stopped person's clothing that fell short of a full-blown search need be justified only by reasonable, articulable suspicion that crimi- nal activity was afoot and that the person with whom the officer was deal- ing may be armed and dangerous.19 15See Jon B. Gould and Stephen D. Mastrofski, Suspect Searches: Using Constitutional Stan- dards to Assess Police Behavior (unpublished manuscript on file with author). 16Civil Rights Bureau, Off. Of the Att'y Gen., The New York City Police Department's "Stop & Frisk" Practice: A Report from the Office of the Attorney General (December 1, 1999). 17The forms that provide the fodder for this analysis are known as UF-250s. According to the New York Police Department's Patrol Guide, a police officer who stops, questions, and frisks an individual on the basis of reasonable suspicion must complete a UF-250. In situations in which probable cause exists before stops are conducted, no form is completed. In situations that fall outside these four contexts, a police officer may fill out a form if he or she desires to do so. The pool of forms analyzed in the study contained about three quarters mandated reports and the rest voluntary. Id. at 91. The forms covered stops that occurred in 1998 and the first three months in 1999. 18392 U.S. 1 (1968). 19Terry, 392 U.S. at 27.
LAWFUL POLICING 267 The study of the New York Attorney General's Office used the form submitted by police officers themselves to identify stops that clearly meet the constitutional standard of reasonable suspicion according to Terry and its progeny and those that did not. The data may be biased because they were drawn from police reports. Examining them, Fagan and Davies found that the facts as stated did not meet the reasonable suspicion standard of Terry in 14 percent of cases (Fagan and Davies, 2000). The New York report also examined the disparate effect of stop and frisk practices. When classified by race, the percentage of wrongful stops-- those that did not articulate reasonable suspicion--were about the same for blacks, Hispanics, and whites: 16, 14, and 17 percent, respectively. These differences were not statistically significant. The New York report used decennial census estimates to calculate the rate at which persons of different races were stopped, and concluded that the rate for blacks was much higher than that for whites. The report also used race-specific arrest totals to cal- culate crime-based stop rates (stops per arrest), and found that by this mea- sure blacks were stopped at twice the rate of whites. However, Chapter 8 considers the methodological problems involved in using population counts and other measures of at-risk populations as a basis for examining patterns of racial profiling, concluding that these must be considered with caution. In grappling with the issue of police compliance with constitutional or legal standards, it is crucial to understand the nature of the evidence, and stop and search procedures are no exception. Studies based on official records have limitations when they are used for estimating the frequency with which police violate constitutional protections regarding search and seizure. First, records that document searches can also be used to evaluate their performance, and the officers whose performance is to be evaluated are the very persons responsible for completing the documents. They are well aware of these downstream uses and may understandably craft their reports to cast their own performance in the most favorable light, perhaps by leaving out uncomplimentary material, being vague, or even falsifying records. Second, official records, whether police- or court-generated, are substantially incomplete. Many searches go undocumented by police be- cause no arrest was made. A systematic field study of the police in the early 1990s reported that 39 percent of searches without warrant resulted in an arrest or citation that would produce an official document of the search, and only 9 percent of unconstitutional searchers were so documented (Gould and Mastrofski, 2001). Even when searches are reported, a thor- ough judicial review of the legality of that search depends on defense counsel's filing a suppression motion. Given the cursory review afforded to most criminal cases by defense counsel and the ubiquity of quick plea bar- gains in routine criminal cases, court documents constitute an unreliable means of assessing the constitutionality of police practices.
268 FAIRNESS AND EFFECTIVENESS IN POLICING Police Corruption Corruption is another way that police deviate from the law. Corruption is usually characterized as the abuse of authority for personal gain (Barker and Carter, 1986:3; Goldstein, 1975:5; Klockars et al., 2000:1; Lersch, 2000; Sherman, 1974:5).20 It has also been described as "profit-motivated misconduct." Police corruption can take a variety of forms: gratuities (free meals, small gifts); bribes (not to enforce the law or enforce it selectively, to provide information about police activities and investigations, to provide false testimony, to secure a promotion or desired assignment within the department); and theft, burglary, and other misapplications of police power that benefit the officer (shakedowns of drug dealers, accepting payment from private citizens for harassing and arresting persons when there is no legal justification) (Walker, 1999:245-248). There is debate about whether some of these practices, such as the acceptance of gratuities, should be con- sidered corruption.21 There is also debate about whether off-duty offenses by officers, which could range from drug abuse and burglary to domestic violence, should be considered. As a result, studies vary somewhat in how corruption is defined and measured. We draw on a broad range of these studies in our review, which covers how much corruption there is, its causes, what is known about how to control it, integrity as an alternative measure, and corruption's effect on the legitimacy of police. Measuring Corruption It is difficult to measure police corruption reliably, because those who have the most knowledge about the corrupt acts are themselves usually implicated in those practices and have a vested interest in keeping these practices undetected (Klockars, 1999:208). Those who bribe police officers are quite unlikely to report it to authorities that might seek to put a stop to 20Terry, 392 U.S. at 30. Justice Warren's majority opinion said very little about stops that accompany or are antecedent to frisks, but Justice Harlan in concurrence emphasized that the frisk in the case depended on the reasonableness of the accompanying stop, and that the right to frisk must be automatic "if the reason for the stop is, as here, an articulable suspicion of a crime of violence." Id. at 33. 21Some analysts consider that corruption should also include abuses of authority that do not render a personal benefit but are in fact undertaken to further what the actor perceives as larger organizational or societal objectives that are widely embraced, such as reducing crime and delivering justice (Muir, 1977:271). Such forms of "noble cause" corruption as giving false testimony "to win the war on crime" have been noted at least as far back as the 1931Wickersham Commission (Barker, 2002:12). While this is certainly worthy of study, we are guided here by Klockars' caveat not to cast the conceptual net regarding corruption too broadly, perhaps rendering it useless for purposes of understanding its causes.
LAWFUL POLICING 269 such practices. Officers understandably do not report their own corrupt practices and, as with many other occupations, are reluctant to report the corrupt practices of their colleagues. Some argue that the problem is espe- cially severe in policing because of a strongly developed informal code of not reporting their colleagues' misbehavior and even condoning it (Klockars et al., 2000:2; Stoddard, 1995). Some have suggested that the problem is not as severe now as it once was (Sherman, 1977). Citizens who know about corrupt practices are reluctant to report them to the authorities be- cause their own illicit behavior may be revealed, or because they fear retali- ation by the police or that the authorities will not believe their claims. And many citizens simply do not have knowledge of specific corrupt acts. Thus, the usual methods of collecting data on commonly measured legal offenses (official crime reports and surveys of victims or offenders) fail to capture most of the corrupt events. This problem is compounded by the considerable variation that exists in mechanisms used to uncover, expose, and punish corruption. Communi- ties in which police departments, other state and local government institu- tions, and the press are vigorous in pursuing corruption may well find it, while communities that are less energetic may not, regardless of how much is actually going on. That is, we lack a measure of corruption that is inde- pendent of the effectiveness of the policies and procedures we are trying to evaluate. We must acknowledge at the outset that the available scientific evidence on corruption is severely flawed as a result. Those who look for national statistics on police corruption will not find them (General Accounting Office, 1998:10). Some police departments publish annual statistics on the status of complaints against the police and their disposition, and on rare occasion these will indicate the nature of the misconduct. Usually, however, this information is not made public in any detailed, systematic fashion.22 The FBI's Public Corruption Unit maintains statistics on corruption cases in which it has jurisdiction. It reported that the number of officers convicted annually as a result of its investigations ranged between 83 and 150 during the 1993-1997 fiscal year period (Gen- eral Accounting Office, 1998:11). Most of what we know about police corruption is drawn from investi- gations that follow in the wake of scandals. Among the most widely known recent reports are those of the Knapp Commission report (City of New York, 1973), the Mollen Commission report (City of New York, 1994), and the Pennsylvania Crime Commission report (1974). 22One recent statistical analysis of police misconduct in New York City between 1975 and 1996 draws on information about a wide range of misconduct, including, but not limited to, corruption as we have defined it (Kane, 2002).
270 FAIRNESS AND EFFECTIVENESS IN POLICING The Knapp Commission's report on New York City followed a corrup- tion scandal that surfaced in 1970. The report created a typology of corrupt practice, distinguishing "grass-eaters," officers who accept gratuities and solicit small payments from legitimate and illegitimate enterprises, from "meat-eaters," those who aggressively and constantly exploit their position for substantial financial gain. The commission speculated that grass-eaters comprised the vast majority of corrupt officers and that meat-eaters were only a small portion of the entire force, but that was a measure of their faith rather than any data. The commission concluded that corruption was "an extensive, Department-wide phenomenon, indulged in to some degree by a sizable majority of those on the force and protected by a code of silence on the part of those who remained honest" (City of New York, 1973:61). The pervasiveness of the corruption uncovered by the Knapp Commis- sion stands in sharp contrast to the typical claim that corruption is limited to a few rotten apples (due to the moral deficiency of a few individuals) or at most, "rotten pockets" (a few small work groups in which corruption is widely practiced) (Sherman, 1974:7; Skogan, 1979). Although the rotten apple view is the usual conclusion of police leaders and boards of inquiry, independent researchers have often concurred with this assessment, such as with the River Cops in Miami in the late 1980s (Delattre, 1989); the New York City drug corruption scandal of the early 1990s that was investigated by the Mollen Commission (Barker, 2002); the Rampart scandal in Los Angeles; the Chicago drug scandal of the mid-1990s; and the Seattle inci- dent of a detective stealing money from the home of a man killed in a police shootout (Barker, 2002). An attempt at assessing the extent of corruption across cities was of- fered in the wake of the corruption scandals revealed by Knapp Commis- sion investigation. Sherman (1978) examined New York and three other cities that were attempting to reform their police departments. He acknowl- edged the impossibility of gathering valid and reliable data on corruption events in these cities, but he argued that it was possible to classify the degree to which corruption was organized in each city--how much cooperation among police occurred in corrupt practices. His cross-department study employed a variety of indicators of the degree of organizational corruption: the level of active cooperation among two or more officers (based on inves- tigative accounts); the level of passive cooperation among officers (e.g., fail- ure to report corrupt activities of other officers, the number of officers work- ing for a superior officer who accepts bribes for ordering subordinates not to enforce the law); and the level of citizen-police cooperation (e.g., the ratio of citizen-consenting types of corruption to the number of citizen- victimizing types of corruption). The study found that corruption was orga- nized to a high degree before the scandals surfaced, even though a high level
LAWFUL POLICING 271 of organizational effort at corruption control existed in three of the four departments. Most academic experts and observers of policing agree that most police officers are honest (General Accounting Office, 1998:11), but there is little evidence to evaluate this claim, or more important, to provide more precise estimates of the true nature and extent of corruption. Are police more likely to be corrupt than those in other occupations? There is no systematic evi- dence on this, although it seems likely that police corruption is more news- worthy than that in occupations not imbued with such authority and high expectations of rectitude. The public may regard such corruption as a greater threat to the social good and be more sensitized to it when it is publicized. Interestingly, however, the public tends to hold the police in fairly high esteem for honesty and integrity compared with other occupa- tions to which they are routinely compared (Gallagher et al., 2001:63). National surveys repeatedly show the police receiving fewer negative rat- ings about their honesty and ethical standards than about two-thirds of the other occupations listed, including judges, funeral directors, governors, jour- nalists, and lawyers (Bureau of Justice Statistics, 2001:114). In recent years, the level of negative ratings for lawyers has been about 30 percent higher than for police. Causes of Corruption The research literature is long on theory and short on evidence about what causes police corruption. The putative causes of corruption can be grouped into four categories: characteristics of individual officers, opportu- nities for corruption, characteristics of police organizations, and character- istics of the larger environment (Walker, 1999:250). Characteristics of Individual Officers. One set of explanations focuses on the personal character of officers (Muir, 1977; Sherman, 1974). Refusing to accept applicants for police positions who have demonstrated serious moral lapses (e.g., committed crimes or abused drugs) is one way that depart- ments act on the theory that officers themselves carry with them the seeds of their own success or failure regarding corruption. Studies do not provide a rigorous indication of how successful such selection strategies are in re- ducing corruption. However, in general, personality-based explanations are not popular among researchers. They note the poor performance of psycho- logical screening tests in predicting and explaining past and future perfor- mance (Burbeck and Furnham, 1985:65; Daley, 1980; Malouff and Schutte, 1980; Talley and Hinz, 1990), and that such explanations fail to account for varying situations that influence officer behavior, or to take into ac-
272 FAIRNESS AND EFFECTIVENESS IN POLICING count the effect of the police organization and environment (Dwyer et al., 1990). A perhaps more viable theory of corruption posits that officers who perceive a pathway for legitimate advancement in their career are less likely to pursue corrupt practices than those who perceive limited opportunities for career advancement (Sherman, 1974). Career opportunities serve as both an incentive for staying on track (to increase the prospects of success) and a deterrent to corruption (to avoid placing their career prospects at risk). Opportunities for Corruption. Opportunities to act corruptly arise from exposure to the temptations of corruption and the capacity to conceal it. High discretion, low visibility jobs with stable work groups in areas (geo- graphic or functional) with high levels of lucrative illicit trade provide the greatest opportunity for corruption. Although almost all police enjoy con- siderable discretion in their work, detectives enjoy a great deal of freedom about how they spend their work days, and they also work out of uniform, which reduces their visibility (General Accounting Office, 1998:11). When they operate in work groups with changing membership, they have the op- portunity to acclimate new members to the norms of the group and to assess their willingness to tolerate the level of corruption the group finds acceptable. Narcotics detective units are thought to experience especially high levels of opportunity for corruption because they deal with preventing traffic in a commodity that commands very high sums of money and pur- veyors who are highly motivated to bribe police to engage in nonenforce- ment or selective enforcement (General Accounting Office, 1998; Manning and Redlinger, 1977). Sherman (1978:9) concluded from his four-city study that management strategies to eliminate opportunities for corruption avail- able in the environment proved effective. This was done by reducing police responsibilities for enforcing certain laws that exposed them to corruption pressures (e.g., gambling), urging the public not to offer gifts to officers, and stricter enforcement of bribery laws against the public. Characteristics of Police Organizations. Police organizations are thought to exert a considerable influence over the level of corruption in the depart- ment, through both formal means (policies, rules, and procedures) and in- formal means (the subculture norms and practices of police workers) (Klockars, 2003; Sherman, 1974, 1977). Many formal aspects of the orga- nization are explicitly designed to limit corruption, by creating systems to define it (by establishing rules), detect it (internal affairs investigations of allegations, integrity tests), and punish it. Departments that had become more authoritarian and bureaucratized were found to have lower post-re- form levels of corruption (Sherman, 1977). Specific management strategies associated with success at controlling corruption were high levels of person- nel turnover; holding supervisors accountable for their subordinates' per-
LAWFUL POLICING 273 formance; closer surveillance of operations; and termination of policies that encouraged corruption (e.g., ending quotas for vice arrests, which encour- aged perjury about hard-to-get evidence). Sherman (1978) also found that covert and proactive techniques (integrity tests that provide officers the op- portunity to act corruptly) were more effective than standard investigations based only on formal complaints about past practices. The tolerance of the police work group for corrupt practices constitutes an informal organizational characteristic of considerable theoretical impor- tance for explaining high levels of corruption (Klockars, 2003; Pennsylva- nia Crime Commission, 1994; Sherman, 1974). On one hand, an accepting or tolerant attitude toward corruption among officers eases the transition to corrupt practices ("Everybody thinks it's ok"), and it minimizes the risk of disclosure by coworkers. On the other hand, when there is widespread intolerance of corruption among officers, those who are involved in it risk becoming an outcast and denied the protections of a loyal support group. Characteristics of the Environment. Researchers have looked to a wide variety of environmental influences on corruption that affect individual of- ficers and entire organizations. Perhaps the most obvious is the criminal law, which by defining what is criminal, regulates the demand for police corruption (Walker, 1999:250). By prohibiting the manufacture and sale of alcoholic beverages in the 1920s, the United States created a tremendously lucrative market for police corruption. Presumably, the pressure and re- wards for police corruption recede when popular products and services are legalized and regulated. While the criminal law is determined by relatively large political units (the states and the nation), other aspects of the environment relevant to corruption vary at the community level. Public standards about what con- stitutes intolerable corruption (and what does not) are believed to be an important influence on how local government generally and the police in particular are pressured and induced to behave (Goldstein, 1975; Klockars, 2003; Sherman, 1977). Communities in which a "public-regarding" ethos is dominant are hypothesized to have less corruption than those with a "private-regarding" ethos (Sherman, 1974:16). The former places highest priority on promoting the general community welfare; the latter focuses on promoting self-interest. Where public-regarding is the dominant perspec- tive, "good government" reformers have attempted to keep politicians from meddling in police operations, and this presumably has made it less likely for police corruption to flow from elected officials. One might also hypoth- esize that where political competition is perpetually close (that is, no single party or political group dominates local politics), there is less likely to be police corruption. This should be especially the case when the local pros- ecutor is of a different party from the municipal chief executive. Under
274 FAIRNESS AND EFFECTIVENESS IN POLICING these circumstances, the prosecutor is more likely to see political merit in uncovering and prosecuting police corruption. In a somewhat different vein, higher levels of culture conflict (based on race, ethnicity, religion, lifestyle) are thought to facilitate police corruption, because culture conflict leads to divergence on the rightness of some laws, and this in turn creates pressure for uneven enforcement of those laws (Sherman, 1974:17). Scandal itself is viewed as a powerful influence on the level of corrup- tion. The existence of scandal relies on news media vigorously investigating the possibility of police corruption (Sherman, 1974:30). Scandal appears to reduce the degree to which corruption is organized, at least in the short run (Sherman, 1978:12). However, whether the benefits of this publicity will endure depends heavily on the reform policies that are subsequently imple- mented and whether those policies are sustained or allowed to deteriorate over time. Recent scholarship in this field has returned to the individual level. There is a new focus on officer surveys, asking not about reports of their own offending, but rather measuring integrity via their attitudes, percep- tions, and self-reports of how they would respond to situations in which integrity is an issue (Klockars et al., 2000). The idea is that in departments of high integrity, officers will know and support the rules governing police misconduct; they will know the disciplinary threats associated with various rule violations; they will assess the department's disciplinary practice as fair; and they will be willing to report misconduct. This allows researchers to describe "in a fairly precise way, the characteristics of a police agency's culture that encourage its employees to resist or tolerate certain types of police misconduct" (Klockars et al., 2000:10). Klockars and colleagues (2000) surveyed officers in 30 police departments in the United States and found that overall a majority would not report a colleague who engaged in the least serious misbehaviors (e.g., accepting free meals and discounts), but that they would report someone who engaged in behaviors judged to be at intermediate or high levels of seriousness (e.g., accepting kickbacks from an auto repair shop for referrals, turning in a lost wallet while keeping the cash from that wallet). Their study also found that police departments varied considerably in the climate of integrity. Using integrity measurement as a means of indirectly addressing the corruption problem appears to show considerable promise. Surveys are rela- tively inexpensive and easy to administer; they are not nearly as threatening as the more intrusive forms of integrity testing that departments use to catch individual officers; and importantly, the method provides quantifi- able indicators that do not rely on officers' self-reporting their own or their colleagues' misdeeds. Nonetheless, there are limitations to this method. Per- haps most importantly, and as its advocates acknowledge, what officers indicate on an attitude survey about a series of scenarios does not necessar-
LAWFUL POLICING 275 ily predict with a high degree of accuracy how the officers would actually behave if actually confronted with those situations. The strongest test of this survey methodology would be to examine officers' actual behaviors. For example, do officers who score high on integrity actually report officer misconduct at a higher rate than those who score low? Do agencies that score high on integrity actually show less tolerance of discovered miscon- duct than those that score low? For these and other reasons, it is not safe to assume that organizations that score high on integrity will necessarily score low on corruption (Klockars et al., 2000:10). PROMOTING COMPLIANCE WITH LEGAL AND CONSTITUTIONAL RULES It is one thing to write down a set of rules that describe the proper use of police force and authority; it is quite another to have those rules reliably guide the day-to-day conduct of the police on the street. Indeed, if there is one important thing that decades of research on the police have taught, it is that there are many powerful factors influencing the conduct of police be- yond the laws and rules that are supposed to guide them. Nonetheless, it is important that society find effective means for guiding the conduct of the police, and a great deal of experience has accumulated with different ways of trying to do so. We present the evidence on this experience below. This section first considers research on the efficacy of criminal and civil actions against the police in controlling police misconduct. Both lethal and excessive force claims are governed by civil suits grounded either in state tort law or federal civil rights laws or both, and excessive force claims also are circumscribed by criminal law. The section next considers the effects of federal civil actions against police departments for "pattern or practice" claims that they promote unlawful behavior. These cases were authorized by the same Violent Crime Control and Law Enforcement Act that sup- ported the expansion of community policing. Administrative policies and procedures have also been adopted by policing agencies to address miscon- duct, including internal inspections and early warning personnel systems. This section also examines research on external accountability mechanisms, addressing what is known about controlling police corruption. Criminal and Civil Liability It is difficult to determine the effectiveness of criminal and civil liability in deterring police use of excessive force because data regarding the fre- quency and severity of excessive force incidents are weak. Among the myriad of obstacles to measuring excessive force is the fact that even official records regarding police abuse of force are rarely kept or are not available to the
276 FAIRNESS AND EFFECTIVENESS IN POLICING public (Human Rights Watch, 1998). In 1994, Congress mandated that the U.S. Department of Justice "acquire data about the use of excessive force" by police officers and "publish an annual summary of the data" (Bureau of Justice Statistics, 2001). However, even the Department of Justice has been unable to collect this information; in order to comply with the mandate, it has instead published statistics regarding the general use of force by police, acknowledging that "current indicators of excessive force...are all critically flawed." Notwithstanding this barrier, the data that are available indicate that criminal and civil liability for excessive force has little effect on police behavior. Administrative guidelines are an effective deterrent to the use of excessive force when they are consistently enforced by police management. Criminal Law Criminal prosecution represents one means of bringing to justice offic- ers who have broken the law, either through corruption or excessive use of force. At the local level, prosecutors may bring criminal charges against police officers for excessive force under general state laws regarding assault, battery, murder, etc. Some states also have specific statutes that make ex- cessive police force a separate crime. Federal and state grand juries also have the power to investigate and bring indictments against police officers for alleged criminal actions. Most experts regard criminal prosecution as an extremely limited ve- hicle for achieving meaningful accountability. Obtaining the conviction of officers is extremely difficult, as judges and juries exhibit a high level of deference to the testimony of police officers. Also, criminal prosecution of individual officers or even small cadres of officers does not address systemic organizational problems that contribute to officer misconduct (Vera Insti- tute, 1998). It appears that comprehensive data regarding rates of prosecution, con- viction, and sentencing are not compiled in most states. However, by all accounts, local criminal prosecution of officers for excessive force is ex- tremely rare (Human Rights Watch, 1998; Adams, 1996 in Geller &Toch). In Los Angeles, the district attorney's office prosecuted only 43 excessive force cases between 1981 and 1991, less than 0.25 percent of alleged acts of excessive force (Levinson, 1994). In 1998, a district attorney in Atlanta reported to Human Rights Watch that he remembered only three excessive force cases from the past five years.23 Researchers attribute this phenom- enon to two sources. The first is reluctance on the part of prosecutors to 23U.S. Commission on Civil Rights, supra note 8 at 131, 133.
LAWFUL POLICING 277 bring excessive force cases, due to lack of support by public officials, their need for good working relationships with the police, and the low probabil- ity of conviction as a result of evidentiary and witness credibility problems and jury identification with officers.24 In addition, there are very few refer- rals from police internal affairs units, which review excessive force com- plaints before deciding whether to refer them to the district attorney's of- fice.25 Between 1986 and 1990, 98 percent of excessive force complaints were not referred for prosecution by the Los Angeles Police Department's internal affairs unit.26 Excessive police force is a federal crime under 18 U.S.C. § 242, which makes it criminal for an individual acting under color of law to deprive any person of his or her civil rights. Investigations of such violations are con- ducted by the FBI and then referred to the Civil Rights Division of the Department of Justice. In order for a police officer to be convicted under section 242, the prosecutor must prove that he or she had "specific intent" to deprive an individual of his or her civil rights.27 Federal prosecution and conviction for excessive police force is also extremely rare. Less than 1 percent of cases reported to the Department of Justice lead to prosecution.28 In 1994, 96 percent of police misconduct cases recommended for prosecution by federal investigative agencies were not pursued by the Department of Justice.29 Of the official misconduct cases prosecuted, the rate of conviction is unusually low compared with other types of cases brought by the Civil Rights Division, for which there is al- most a 100 percent success rate. In 1994, 78 percent of federal police mis- conduct cases were successful; in 1996, only 64 percent were successful.30 One reason cited for the low conviction rate is the specific intent require- ment, which is considered a significant obstacle to conviction.31 Finally, even when police officers are convicted, they usually receive light sentences. Although section 242 allows a maximum sentence of 10 years in prison and a fine when bodily injury is inflicted, most convicted officers spend little or no time incarcerated.32 24Skolnick and Fyfe, supra note 25 at 205. 25Patton, supra note 29 at 801. 26Chevigny, supra note 32 at 101. 2742 U.S. 14141 (1994). 28Screws v. United States, 325 U.S. 91 (1945). 29Human Rights Watch, supra note 1 at 94. 30Marshall Miller, "Police Brutality," 17 Yale L. & Pol'y Rev. 149, 153 (1998). 31Human Rights Watch, supra note 1 at 99. 32See id. At 85; Cheh, supra note 5 at 247; United States Commission on Civil Rights, supra note 5 at 101.
278 FAIRNESS AND EFFECTIVENESS IN POLICING Given these statistics, it is not surprising that most researchers have concluded that criminal sanctions are not a significant deterrent to exces- sive force by police officers. The risk of prosecution and conviction is so small that officers are likely to believe (correctly) that they can use excessive force without criminal consequences.33 Civil Liability A second use of the courts involves civil suits by private individuals seeking damages for harm done by officers. In most states, victims of exces- sive police force can sue for damages under traditional theories of tort li- ability, such as assault or battery. The elements of proof and the legal stan- dards for determining civil liability vary from state to state.34 Injured persons may also bring civil suits against the police under federal law, spe- cifically 42 U.S.C. § 1983. Municipalities can also be held liable for a police officer's misconduct if it occurred pursuant to policy or custom.35 Suits grounded in federal civil rights laws seek to enforce constitutional stan- dards and have been acknowledged by the court to be important supple- ments to the exclusionary rule. On its face, section 1983 permits plaintiffs to recover not only mon- etary damages, but also equitable relief. However, the Supreme Court has severely restricted the availability of injunctions under section 1983, mak- ing it nearly impossible for plaintiffs to obtain injunctive relief from specific police practices or patterns of abuse.36 There is some evidence of a signifi- cant increase in civil litigation, at least in certain major departments over the years (Newell, Pollock and Tweedy, ICMA, 1992). A number of civil rights and civil liberties activists have embraced civil litigation as a reform strategy, assuming that rising damage awards will force public officials to undertake significant reforms in the police department, especially if they bring cases challenging existing police practices. Attempts to evaluate the effect of civil litigation have found that it has often not produced the de- sired reforms (McCoy, 1984; Yale Law Journal, 1979). Although civil suits alleging excessive force are difficult to win,37 the amount paid in police misconduct cases has grown steadily over the past 30 years,38 and many American cities are now paying massive amounts of 33Cheh, supra note 5 at 260 34Montell v. Department of Social Services, 436 U.S. 658 (1978) 35See Rizzo v. Goode, 423 U.S. 362 (1976); City of Los Angeles v. Lyons, 46 U.S. 95 (1983). 36See Rob Yale, "Searching for the Consequences of Police Brutality," 70 S. Cal. L. Rev. 1841 (1997). 37Jerome H. Skolnick and James J. Fyfe, Above the Law 202 (1993). 38The Christopher Commission, supra note 10 at 56.
LAWFUL POLICING 279 money in settlements with and judgments for victims of police brutality. Between 1986 and 1990, Los Angeles paid over $20 million in 300 suits against police officers alleging excessive use of force.39 Between 1994 and 1996, New York City juries awarded $70 million in claims alleging police misconduct, although some of these awards were overturned on appeal or were settled for lesser amounts to avoid appeal by the city.40 These amounts do not include the costs of litigation and negotiating settlement. Despite this financial burden, it is well recognized that civil suits offer little deterrence to excessive police force. With regard to individual police officers, there are two primary reasons for this lack of deterrent effect. First, legal fees, settlements, and judgments are paid for by the government. These awards are further limited by the immunity of local governments from pu- nitive damages. In most, if not all, jurisdictions, individual police officers are protected by law from having to pay legal fees and damages in civil suits.41 City attorneys are responsible for defending officers in misconduct cases; for the most part, the officer's only responsibility is to participate in depositions and trial testimony.42 Thus civil suits provide no financial in- centive to individual officers to cease using excessive force. Second, even successful excessive force suits do not lead to any sort of disciplinary action and have no effect on promotion in the vast majority of cases. In Los Ange- les, civil suits are not recorded in police personnel files, and they are not considered by internal affairs unit in investigating complaints.43 A 1992 investigation found that of 185 officers involved in 100 civil lawsuits in 22 states, only 8 were disciplined; 17 of the officers were actually promoted.44 The outcomes of civil suits are equally ineffective in influencing sys- temic change within police departments. Studies have shown that neither the Los Angeles nor the New York City police departments make institu- tional or policy changes in response to successful civil suits. The results of civil suits are not even monitored.45 One reason for this is that damages in these suits are not significant in relation to total police budgets and, more importantly, are not even paid out from the police budget but out of general city funds.46 More broadly, the minimal influence of civil suits on policy 39Matthew Purdy, "In New York, the Handcuffs are One-Size-Fits-All," New York Times, August 24, 1997, A1. 40See, e.g., Cal Gov't Code § 825 (West 1993). 41Alison L. Patton, "The Endless Cycle of Abuse: Why 42 U.S.C. § 1983 is Ineffective in Deterring Police Brutality," 44 Hastings L.J. 753, 768 (1993). 42Id. at 784. 43Human Rights Watch, supra note 1 at 82, citing Rochelle Sharpe, "How Cops Beat the Rap," Gannett News Service, March 1992. 44Paul Chevigny, Edge of the Knife 102 (1995). 45Id. 46U.S. Commission on Civil Rights, supra note 8 at 131, 133.
280 FAIRNESS AND EFFECTIVENESS IN POLICING and institutional change can be explained by the dominant viewpoint among public officials and police departments that civil damages are the cost of doing business rather than indicative of problematic patterns or practices.47 In Los Angeles, the cost of civil suits is considered "a reasonable price for the presumed deterrent effect of the [police] department's most violent re- sponses to lawbreaking."48 In an early study of this matter, LaFave collected data prior to Mapp v. Ohio's restrictions on search and seizure (LaFave, 1965). He identified three potential pre-Mapp sanctions for unconstitutional arrests and searches in Kansas, Michigan, and Wisconsin: tort suits, criminal penalties, and admin- istrative sanctions. Officers interviewed in the three relevant states expressed little concern about the threat of a tort action for false imprisonment fol- lowing an improper arrest. In all three states such suits were infrequent and typically brought only when there was an allegation of extreme physical force and violence. Similarly, researchers could not find one instance of criminal prosecution against a police officer for an unconstitutional arrest. Finally, LaFave notes that department discipline for arrest was infrequently used, especially if the person arrested was guilty. Although civil suits do not provide an effective deterrent to excessive force, there are rare instances in which a particular suit results in so much media attention that police departments are forced to respond.49 It is also possible that a particularly large judgment in a small city or town will lead to reform because the financial consequences are more severe when the judgment is paid out from a small budget.50 For the most part, however, the availability of recourse under state law and section 1983 to victims of ex- cessive force has had no influence on police behavior. Federal "Pattern and Practice" Actions In 1994, Congress enacted the Violent Crime Control and Law En- forcement Act, under which the federal government has the power to con- duct investigations and bring suit against any police department where there is "a pattern or practice of conduct by law enforcement officers...that de- prives persons of rights, privileges, or immunities secured or protected by 47Skolnick and Fyfe, supra note 25 at 205. 48Patton, supra note 29 at 801. 49Chevigny, supra note 32 at 101. 5042 U.S. 14141 (1994).
LAWFUL POLICING 281 the Constitution."51 As of October 2001, the Department of Justice was in the process of investigating 13 agencies for a pattern and practice of police abuse. Six additional investigations resulted in settlement, four through con- sent decrees and four through settlement agreements.52 The latest of these agreements, issued September 4, 2002, between the department and the police of Columbus, Ohio, came only after a civil complaint was filed. Generally, reforms introduced as the result of these agreements include implementing agreed-on best practices, including reporting systems and early warning systems. Provisions regarding the use of force are included in all of the settlement agreements. Each consent decree creates a role for a court-appointed monitor who oversees compliance with the agreement. The police departments involved are required, among other things, to improve officer training regarding the use of force, to provide detailed reports of each incident of force, and to initiate internal investigations whenever a criminal or civil suit alleging misconduct is brought against an officer. A 2002 consent decree with the Cincinnati Police Department specifically lim- its the use of chokeholds, chemical sprays, canines, beanbag shotguns, and 40 mm foam rounds.53 For the most part, these settlement agreements are too recent to deter- mine whether they have actually decreased the use of excessive force by police officers. However, the Vera Institute of Justice has been monitoring police activity in Pittsburgh, which entered the first of the settlement agree- ments with the Department of Justice in 1997. It reports that Pittsburgh police are largely in compliance with the consent decree.54 In light of this report, this approach to deterring excessive force appears promising, al- though its effectiveness in implementing permanent reform remains to be seen. 51U.S. Department of Justice, Civil Rights Division, Special Litigation Section, "Frequently Asked Questions," http://www.usdoj.gov/crt/split/faq.htm (June 19, 2002). The settlements covered the following agencies: Pittsburgh Bureau of Police, PA (consent decree, 1997); Steubenville Police Department, OH (consent decree, 1997); New Jersey State Police (consent decree, 1999); Los Angeles Police Department, CA (consent decree, 2001); District of Colum- bia Metropolitan Police Department (settlement agreement, 2001); Highland Park Police De- partment, IL (settlement agreement, 2001). As of April 12, 2002, the Department of Justice has also entered into a consent decree with Cincinnati Police Department, Ohio; see http:// www.usdoj.gov/crt/split (June 19, 2002). 52Memorandum of Agreement between the United States Department of Justice and the City of Cincinnati, Ohio and the Cincinnati Police Department, http://www.usdoj.gov/crt/split (June 21,2002). 53Vera Institute of Justice, "Pittsburgh's Experience with Police Monitoring," http:// www.vera.org/project/ project1_1.asp?section_id=2&project_id=13 (June 14, 2002). 54The Christopher Commission, supra note 10 at 26.
282 FAIRNESS AND EFFECTIVENESS IN POLICING Administrative Reform Research indicates that clear administrative guidelines regarding the use of force, coupled with consistently imposed sanctions for misconduct, reduces the incidence of excessive force. Not surprisingly, written policies are not in themselves an effective deterrent to police misconduct. Since 1979, the Los Angeles Police Department has restricted the use of force to that which "is reasonable and necessary" once "other reasonable alternatives have been exhausted."55 Clearly, this rule had little influence on the police officers involved in the beating of Rodney King in 1991. However, a writ- ten guideline can be sufficient in itself to encourage brutality. For example, the Los Angeles Police Department's official position for many years was that a chokehold is a nonlethal pain compliance technique, while virtually every other major police department considered the use of chokeholds to be a form of potentially lethal force. Perhaps as a consequence, the number of chokehold-related deaths in Los Angeles during the late 1970s was twice the combined total of the remaining 20 largest U.S. police departments.56 The effectiveness of administrative guidelines intended to deter exces- sive force depends on whether or not the performance measures and disci- plinary actions implemented by police management provide any incentive to abide by the rules. Most police departments measure performance in quantitative terms, for example, how many tickets were given or how many drugs dealers were arrested.57 As a result, police officers are not rewarded for the quality of their work. Skolnick and Fyfe compare this to measuring the performance of lawyers by the number of their cases without consider- ing whether the cases were won or lost.58 Furthermore, this approach en- courages officers to ignore guidelines when doing so helps the bottom line. The disincentives provided by performance measures can be exacerbated by the failure of police departments to take disciplinary action when it be- comes apparent that an officer has violated restrictions on the use of force. The Christopher Commission reported that there were a significant number of Los Angeles police officers who persistently ignored written guidelines regarding force, and, instead of being punished, they were often rewarded with positive evaluations and promotions.59 When disciplinary actions were taken, the punishments were inappropriately lenient.60 Independent com- 55Skolnick and Fyfe, supra note 25 at 42. 56Id. at 125. 57Id. 58Id. at iii-iv (foreword). 59Id. at 167. 60See Human Rights Watch, supra note 1 at 63-64.
LAWFUL POLICING 283 missions in New York City and Boston also found many shortcomings in the way that internal affairs divisions handled complaints.61 The influence of police leadership over individual officers, whether through performance and disciplinary measures or through general tone- setting, may be substantial. In a study of the Indianapolis and St. Petersburg police departments, Terrill found that the official position of police admin- istrators on the "style of policing practiced" was directly related to the amount of force used by officers. Police leadership in Indianapolis pro- moted a get-tough approach to policing, while St. Petersburg's approach was more toned down and emphasized a problem-solving model. As a re- sult, police officers in Indianapolis were more likely to use higher levels of force than those in St. Petersburg.62 Other researchers have also concluded that "whether through act or omission, the chief is the main architect of police officers' street behavior."63 It is not surprising, then, that there is wide consensus that police admin- istrators are significantly more effective than the courts in deterring the use of excessive force. The more difficult question is why so many police de- partments have failed to provide incentives to individual officers to abide by use of force rules. One explanation is that there is public and political pressure on police leadership to get tough on crime rather than deter the use of excessive force. Only when there is a particularly outrageous and well- publicized incident of police brutality does public attention turn to the issue of excessive force. Over the past 15 years, there have been a number of police beatings that received significant media attention and resulted in pub- lic and political demands for reform. Civil rights attorneys believe that this increased scrutiny of police force has led to change: "Fifteen years ago, police officers would beat people up all the time. Today, there are more limits on police.... People are also more sensitized to the issue and won't tolerate as much violence."64 Thus, to a certain extent, it appears that pub- lic and political pressure is the most effective deterrent to the use of exces- sive force. Policies and Procedures Internal accountability mechanisms have been developed at the initia- tive of police departments themselves, often in response to lawsuits or pres- 61William Terrill, Police Coercion 233 (2000). 62Skolnick and Fyfe, supra note 25 at 136. 63E.g., the 1989 beatings of alleged suspects in the Carol Stuart murder in Boston; the 1991 beating of Rodney King in Los Angeles; the 1997 beating and rape of Abner Louima in New York City. 64Patton, supra note 29 at 801 (quoting a civil rights attorney she interviewed).
284 FAIRNESS AND EFFECTIVENESS IN POLICING sure from citizen groups concerned about police misconduct. In some de- partments, internal mechanisms have been imposed by the federal courts as a part of the settlement of civil suits brought by the U.S. Department of Justice, in addition to suits by private litigants in Pittsburgh, Los Angeles, and Cincinnati. Finally, the constitutional decisions discussed above have had the effect of invalidating both state statutes and the policies and proce- dures of many police departments; many internal policies and procedures take the form they do today in order to conform with those decisions. Thus an important role of police departments' policies and procedural rules is to implement federal and state court rulings, along with state stat- utes and city ordinances. For example, the types of weapons that are made available to officers, the rules for their use, training in weapon use, report- ing requirements when they are employed, and the procedures for review- ing the appropriateness of their use, all make up part of a police depart- ment's use of force policy, which must be aligned with state and federal statutory and constitutional requirements. Most of the research on police compliance with administrative policies therefore examines the implemen- tation of constitutional guidelines in police agencies. It is at the administra- tive level that controlling police behavior becomes a "management prob- lem." However, to date there has been little research on the effectiveness of managerial and organizational strategies to motivate officer compliance with department policies and procedures. In contrast to the private sector, relatively little has been attempted, much less evaluated, to evaluate the effect of varying pay and perquisites, for example. More has been written on the impediments to managing police, including the effect of police unions, police culture, and the law and order politics of many cities. Administrative Rules The principal mechanism for building and maintaining accountability involves administrative rulemaking by police departments. This approach involves attempting to control the exercise of police officer discretion in critical incidents through (1) written policies that structure and confine dis- cretion, (2) the requirement that officers complete written reports about particular incidents, and (3) review by supervisory officers (Davis, 1975; Goldstein, 1977; Walker, 1993). One significant application of this approach on which there has been some research involves police officer use of lethal force (Fyfe, 1979; Geller and Scott, 1992). As we have noted, beginning in the early 1970s police departments replaced the permissive fleeing felon rule with the more restric- tive defense of life standard. Official policies also prohibited warning shots and other uses of firearms. Early studies found that administrative rules
LAWFUL POLICING 285 were successful in reducing the number of firearms discharges and the num- ber of citizens shot and killed (Fyfe, 1979; Sparger and Giacopassi, 1992). Research on administrative rulemaking has since been extended to other types of incidents involving the use of police authority. These include stud- ies of the use of physical force (Walker, Alpert, and Kenney, 2001), high- speed pursuits (Alpert, 1997; Alpert and Dunham, 1990), domestic vio- lence (Sherman 1992), the use of canines, and nonlethal technologies (e.g., pepper spray). There is limited evidence on the effectiveness of administra- tive controls in these areas, however. Controls over high-speed pursuits do appear to reduce the number of pursuits and consequently the number of injuries and fatalities (Alpert, 1997). Research on the use of physical force has been limited to investigating the prevalence of force incidents and the factors associated with its use (Bureau of Justice Statistics, 1999; Geller and Toch, 1995). The deterrent effect of mandatory arrest in domestic violence incidents has been investigated through a number of experiments, but there is limited evidence regarding officer compliance with domestic violence poli- cies in nonexperimental conditions (Sherman, 1992). Early warning systems represent a new internal accountability mecha- nism that has developed in recent years. In brief, an early warning system is a data-based management tool designed to identify officers who appear to have performance problems and to subject those officers to some ad- ministrative intervention, usually in the form of counseling or retraining (Walker, Alpert and Kenney, 2001). Early warning systems are an out- growth of the development of administrative rule-making since the required reports principally use data on use of force and other actions to analyze officer performance. Early warning systems have been growing rapidly and are recognized as one of the best practices in police accountability (Walker, Alpert, and Kenney, 2001). However, the evidence for the success of early warning systems is limited at this point. The only evaluation of early warning sys- tems found that they were effective in reducing the use of force and citizen complaints among officers subject to intervention, but the study itself was limited by a number of methodological problems, including questions about the official data that were used (Walker, Alpert, and Kenney, 2001). Never- theless, these systems embody the trend toward increased accountability of the police to the community--a trend that speaks directly to the importance of both perceived and objective legitimacy of the police. Internal Inspections Responsibility for enforcing internal standards of conduct lies with po- lice department internal affairs units, or what are increasingly called profes-
286 FAIRNESS AND EFFECTIVENESS IN POLICING sional standards units. Despite the fact that police management experts rec- ognize that internal affairs units play a critical role, very little research has been conducted on these units. Little is known about the organization, man- agement or staffing of these units. Nor is much known about the investiga- tive procedures used or patterns of discipline. There is a sizeable anecdotal literature on the failure of these units to maintain accountability and integ- rity (Kerner Commission, 1968; Christopher Commission, 1991; Mollen Commission, 1994), but none of these reports or investigations rises to the level of a scientific study. Police Data Collection The advent of sophisticated information systems and a growing aware- ness of the power of gathering data for internal accountability processes, has encouraged police administrators to routinely collect more data on po- lice operations and practices. Sometimes this is in response to external de- mands for closer police accountability. In the case of the debate over racial profiling, for example, the mandated collection of official data on traffic stops has emerged as the principal strategy on the part of civil rights groups for ending the practice, defined as a pattern of stopping black and Hispanic drivers on the basis of race or ethnicity rather than suspected criminal activ- ity (Harris, 1997, 1999, 2002; Police Executive Research Forum, 2000; Ramirez, McDevitt, and Farrell, 2000). As an accountability mechanism, traffic stop data collection operates on the same assumption as administrative rulemaking: that requiring offic- ers to complete official reports on critical decisions will simultaneously put them on notice that their actions are being scrutinized and create a system- atic data base that will permit determination of whether a pattern of illegal racial or ethnic discrimination exists. This is an area in which there is rapid policy change. A growing number of states have enacted laws requiring law enforcement agencies to collect data on traffic stops. In addition, a large and growing number of agencies have undertaken data collection voluntarily. Many of these efforts have yielded publicly available data sets on traffic enforcement patterns (San Jose Police Department, 2001; San Diego Police Departtment, 2000). A great deal of controversy surrounds the proper benchmark for interpreting traffic stop data (Bland et al., 2000; General Accounting Office, 2000; Har- ris, 1997, 1999, 2002). For the purposes of litigation, this question has been addressed by the use of observational studies of traffic patterns on interstate highways conducted in Maryland and New Jersey (Harris, 2002). In these cases, the data were interpreted as supporting the contention that race discrimination existed in those instances. More difficult is the question of determining the extent to which pat-
LAWFUL POLICING 287 terns in traffic enforcement that indicate racial or ethnic disparities repre- sent patterns of illegal discrimination: the challenge is to develop an appro- priate benchmark. Official reports to date generally find racial and ethnic disparities in traffic stops, searches, and arrests (San Jose Police Depart- ment, 2001). With respect to traffic stops, these studies use the resident population of the specific area in question as the benchmark. Walker (2001) argues that the resident population is not a valid benchmark because it does not represent the at-risk driving population, and that related benchmarks using estimates of the driving-age population by race and ethnicity are simi- larly flawed. Attempts to investigate possible discrimination in field stops of pedestrians with resident population data serving as the benchmark are similarly flawed because population data do not reflect either the overall pedestrian population or pedestrians who exhibit legitimate suspicious be- havior. Walker proposes using an early warning system approach in which the activity levels of groups of officers with comparable assignments serve as a benchmark for individual performance. In a sense, this would mean measuring police officers against other police officers, as opposed to mea- suring the citizens stopped by police against some satisfactory benchmark. With respect to searches and arrests of people stopped by police, the popu- lation of drivers or pedestrians stopped serves as the appropriate bench- mark, thereby facilitating the research effort. Other strategies for eliminating racial profiling have been advanced, including exhortation, training, in-car video cameras, and various adminis- trative controls, such as requiring officers to notify dispatchers about key features of stops (Fridell et al., 2001; Cohn, Lennon, and Wasserman, 1999). To date there are no studies investigating the effect of any of the proposed reforms designed to eliminate discrimination in traffic enforcement, al- though a General Accounting Office (2000) report on searches by the U.S. Customs Service suggests that administrative controls can simultaneously reduce racial disparities and increase the efficiency of enforcement efforts. Controlling Corruption With few exceptions, what is known about the effectiveness of various methods to control police corruption is based on the experiences of police leaders and the impressions and expertise of scholars. The U.S. General Accounting Office (1998:19-25) surveyed practitioners to identify strate- gies for preventing police corruption, especially that which is drug-related. They noted the importance of a commitment to integrity by top department leaders, holding managers accountable for corruption in their units, more rigorous hiring policies and more training for recruits, empowering inde- pendent auditors to oversee the handling of corruption allegations, and changing the police culture from protecting one's peers to protecting integ-
288 FAIRNESS AND EFFECTIVENESS IN POLICING rity. The study also reported the practitioners and experts' recommenda- tions on detecting corruption more effectively. The practitioners surveyed stressed the importance of placing officers suspected of corruption under surveillance, turning corrupt officers to testify against others involved in corrupt practices, interviewing arrested drug dealers to get information on police corruption, employing proactive investigations (including integrity testing) in areas that have high concentrations of drug-related corruption complaints, and developing an early warning system of factors and behav- ior patterns that may indicate a corruption problem. The above list is consistent with research-based recommendations, which have also included rewarding officers for good performance, pursu- ing federal prosecution, and working with the news media to mobilize pub- lic opinion to fight corruption and pressure elected officials to support ef- fective policies (Walker, 1999:254-63). We note that the systematic evidence supporting these claims is almost nonexistent, due to the extreme difficulty of obtaining valid and reliable measures of police corruption. EXTERNAL ACCOUNTABILITY MECHANISMS The tradition of local accountability of police is sometimes extended by the creation of external bodies charged with overseeing police operations. The logic behind this strategy seems apparent: what police do is important, and they are one of the most visible parts of local government. Where there is dissatisfaction with police internal accountability mechanisms, citizen groups have demanded the creation of external mechanisms as well. In re- cent decades citizen dissatisfaction, reflecting an apparent loss of perceived legitimacy, has been primarily centered in the black community and has traditionally been labeled "the police-community relations problem" (Kerner Commission, 1968). Citizen Oversight Bodies The principal demand for external accountability by civil rights groups has been for the creation of citizen oversight agencies, commonly termed civilian review boards (Walker, 2001). Specifically, citizen groups have de- manded that an external oversight agency be given responsibility for receiv- ing, investigating, and disposing of citizen complaints against agencies. This demand reflects the belief that police internal affairs units, in varying de- grees, discourage complaints, fail to investigate complaints thoroughly and fairly, and fail to discipline officers who are found to have committed mis- conduct. The police and their supporters generally deny that excessive force is a problem and argue that police departments are better equipped to in- vestigate complaints than are people who are not sworn officers.
LAWFUL POLICING 289 External citizen oversight agencies have been growing steadily since the late 1970s. By 2001 these were slightly more than 100 such agencies. Police departments in nearly 90 percent of the big cities are subject to some form of oversight regarding complaints. Citizen oversight also exists with respect to a number of medium-sized and small police departments, along with some county sheriff's departments. These figures are based on the broadest definition of external oversight and include any procedure in which there is some citizen input into the review process, however limited, by persons who are not sworn police officers. Virtually all of these agencies are autho- rized by local ordinance (Walker, 2001). Because of the bargaining that takes place in the legislative process that created them, citizen oversight agencies take a number of different forms. Walker (2001) classifies them by four different models of oversight: under this scheme, only Class I systems have original jurisdiction with regard to the investigation of complaints. Other institutional arrangements include boards that play an auditing or monitoring function with regard to police-run investigations. There is very limited evidence regarding the effectiveness of citizen over- sight agencies. As already noted, considerable anecdotal evidence exists re- garding the shortcomings of internal affairs units (Kerner, 1968; U.S. Com- mission on Civil Rights, 1981), but no comparative studies indicating that external citizen oversight agencies are more effective than internal affairs units. This kind of evaluation of internal accountability mechanisms poses a number of difficulties (Brereton, 2000; Walker, 2001). The published lit- erature generally fails to take into account the multiple goals of oversight agencies; these include, but are not limited to, conducting thorough and fair investigations of citizen complaints and building citizen confidence in the complaint process. Comparative studies of their effectiveness are also diffi- cult because local external and internal units or boards tend to handle very different kinds of cases and hear different types of allegations (Hudson, 1972, Walker, 2001). Blue Ribbon Commissions A traditional approach to pressing for greater accountability has been through the formation of authoritative commissions that investigate police problems and issue reports with recommendations for reform (Walker 1985). At the national level, the most important examples include the 1931 Wickersham Commission (National Commission on Law Observance and Enforcement, 1931), the President's Crime Commission (1967), the Na- tional Advisory Commission on Civil Disorders (the Kerner Commission) of 1968 and the American Bar Association Standards for Urban Police Func- tion (1973). There have also been numerous state and local commissions created in the wake of local legitimacy crises; the 1922 Cleveland Crime
290 FAIRNESS AND EFFECTIVENESS IN POLICING Survey initiated this trend. Other notable commissions include the Christo- pher Commission, formed in response to events associated with the Rodney King incident in Los Angeles, which published its findings in 1991; and both the Knapp Commission and the Mollen Commission, created in re- sponse to allegations of widespread corruption in the New York City Police Department. Blue ribbon commissions have played an important role in setting gen- eral standards for the police and in that respect guiding reform efforts across the country. In some cities, reform-minded police chiefs have been able to use the legitimacy crisis engendered by the investigations conducted by com- missions to effect needed changes in their departments (Murphy and Plate, 1977). The major weakness of this approach to reform is that blue ribbon commissions are temporary agencies that disband following publication of a final report. Consequently, there has been no mechanism to ensure imple- mentation of their recommendations (Walker 1985). For example, there has been controversy over the implementation of the Christopher Commis- sion report in Los Angeles (Christopher Commission, 1991; Bobb et al., 1996). Walker (2001) argues that the role of blue ribbon commissions has been superseded by the auditor model of citizen oversight, which involves a permanent body that has the capacity to oversee and report on the imple- mentation of recommended reforms (San Jose Independent Auditor, 1993- present; Special Counsel, Los Angeles County Sheriff's Department, 1993- present). Currently, New York City has the only permanent external corruption control agency in the nation, the New York City Commission to Combat Police Corruption. Not enough is known about the extent of police lawfulness, or police compliance with legal and other rules, nor can the mechanisms that best promote police lawfulness be identified. The committee recommends re- newed research on the lawfulness of police and a coordinated research em- phasis on the effectiveness of organizational mechanisms that foster police rectitude.