This is the accessible text file for GAO report number GAO-05-699R entitled 'Preliminary Report on the Effects of COPS Funds on the Decline in Crime during the 1990s' which was released on July 7, 2005. This text file was formatted by the U.S. Government Accountability Office (GAO) to be accessible to users with visual impairments, as part of a longer term project to improve GAO products' accessibility. Every attempt has been made to maintain the structural and data integrity of the original printed product. Accessibility features, such as text descriptions of tables, consecutively numbered footnotes placed at the end of the file, and the text of agency comment letters, are provided but may not exactly duplicate the presentation or format of the printed version. The portable document format (PDF) file is an exact electronic replica of the printed version. We welcome your feedback. Please E-mail your comments regarding the contents or accessibility features of this document to Webmaster@gao.gov. This is a work of the U.S. government and is not subject to copyright protection in the United States. It may be reproduced and distributed in its entirety without further permission from GAO. Because this work may contain copyrighted images or other material, permission from the copyright holder may be necessary if you wish to reproduce this material separately. United States Government Accountability Office: Washington, DC 20548: May XX, 2005: The Honorable F. James Sensenbrenner: Chairman: Committee on the Judiciary: House of Representatives: Subject: Preliminary Report on the Effects of COPS Funds on the Decline in Crime during the 1990s: Dear Mr. Chairman: Established under the Violent Crime Control and Law Enforcement Act of 1994 (Pub. L. 103-322), which authorized appropriations of $8.8 billion for it, the Community Oriented Policing Services (COPS) program mission was to advance the practice of community policing as an effective strategy in communities' efforts to improve public safety and had as a goal providing for 100,000 new police officers. VCCLEA--the largest federal crime bill in the history of the country--was enacted during a period of increasing crime, particularly serious violent crimes, such as murder, rape, aggravated assault, and robbery.[Footnote 1] For example, between 1983 and 1992--the year before the national decline in crime began--the number of serious violent crimes known to the police increased from about 1.3 million to over 1.9 million (or about 50 percent), and the serious violent crime rate per 100,000 population increased from 538 to 758 (or by 39 percent). In 1994, the COPS Office began making grants to local law enforcement agencies, and by 2001, it had made about 23,000 grants totaling $7.3 billion. Of these obligated amounts, by the end of 2001, local law enforcement agencies had expended (or drawn down) about $5 billion. The COPS grant expenditures amounted to about 1 percent of total local law enforcement expenditures between 1994 and 2001. At the same time that local agencies were drawing down COPS funds, serious crimes declined. For example, between 1994 and 2001, the number of violent crimes declined from about 1.9 million to about 1.4 million (or by 23 percent), and the violent crime rate per 100,000 population declined from 714 to 504 (or by 29 percent). Given the comparatively large expenditure of COPS funds for local law enforcement and the correlation between these expenditures and the decline in crime, the questions of whether, and if any, how much, the COPS grants contributed to the decline in crime merits attention. Yet, this issue has received only modest attention. One recent study concluded that the COPS grants contributed to reductions in crime in the 1990s. However, we previously reviewed the study and reported that its methodological limitations were such that the study's results should be viewed as inconclusive.[Footnote 2] Moreover, neither the study we reviewed nor another study of effects of COPS grant funds on crime attempted to isolate the ways in which COPS funds could affect crime.[Footnote 3] For example, the studies did not examine if COPS- funded police officers were associated with reductions in crime. In response to our review of the prior study, you asked us to undertake our own independent evaluation of the impact of COPS grants on the decline in crime that occurred during the 1990s. This correspondence reports partial findings regarding three inter-related questions about the extent to which, if any, that COPS grants affected the decline in crime in the 1990s: (1) How were the COPS grant funds distributed among local law enforcement agencies, and to what extent did the distribution of funds correspond to the distribution of violent crime? (2) Did COPS grant funds lead to changes in the types of policing tactics that are associated with crime prevention? (3) Did COPS grant funds lead to increases in the number of sworn police officers, and if so, what was the impact of these COPS-funded officers on the decline in crime during the1990s? Our full report--due for release in the fall 2005--will address our research objectives in greater. To address our reporting objectives, we created and analyzed a unique database consisting of observations on over 13,000 local law enforcement agencies covering the years from 1990 to 2001. For each agency, we compiled data on COPS and other federal law enforcement grant obligations and expenditures, crime rates, and sworn officers. We also compiled data on factors that the literature suggests are related to changes in crime, including local economic conditions--such as employment rates and per capita income--and demographic variables--such as the percent of the population aged 15 to 24, and the racial and gender composition of the population. We also added to the database information from two surveys of nationally representative samples of police departments that reported on the types of policing tactics that they implemented. Prior to developing our database, we assessed the reliability of each data source and in preparing this report, we used only the data that we found to be sufficiently reliable for the purposes of our report. We analyzed the data on COPS and other federal grant obligation and expenditure amounts to describe how many agencies received COPS grants and the types and amounts of grants and expenditures. We compared the amounts of COPS grants with characteristics of each agency's community, such as their population size and crime rates. To address the extent to which COPS grants led to changes in policing tactics, we analyzed data from the two surveys to assess changes in policing tactics between agencies that received COPS grants and those that did not receive them.[Footnote 4] To assess the effects of COPS funds on officers and crime, we developed and estimated so-called fixed-effects regression models of these relationships. Our regressions estimated the effects of COPS and other federal law enforcement grants on officers and crime, respectively, while controlling for differences among agencies in their pre-COPS program trends in crime rates and sworn officers, differences over time in socio-economic factors that could affect both the number of police officers and crime rates, such as unemployment and per-capita income, and changes in population composition. To help to isolate the direction of causality between officers and crime, we used a statistical instrument, and to control for unmeasured sources of variation, we used agency, state, and year fixed effects. Because of the complexity of the statistical models that we used to estimate the effects of COPS grants on crime, we reviewed our approach and methods with a panel of expert researchers. The panel consisted of criminologists, economists, statisticians, and practitioners, and it was convened for us by the National Research Council. See enclosure I for additional details on our data and methods. We conducted our work between November 2003 and May 2005 in accordance with generally accepted government auditing standards. Background: The COPS Office distributed grants in some 36 different program categories. The largest grant program category was the COPS hiring grants, which required agencies to hire new officers and at the same time to indicate the types of community policing strategies that they intended to implement with the officers that they were to hire with these grants. The hiring grants paid a maximum of $25,000 per officer per year (or at most 75 percent of an officer's salary) and generally required that local agencies provide a match to cover the remaining salary and benefits. Agencies were also required to retain COPS-funded officers for at least one year after the end of the grant. COPS grants were also intended to encourage changes in policing practices. Agencies were asked to report the types of tactics that they planned to implement with their COPS hiring grants. Problem-solving that pertained to increased enforcement activities, place-oriented tactics that focused on addressing crime problems in specific buildings, neighborhoods or other places, and collaborating with community residents by increasing officer contact with citizens and improving citizen feedback were among types of tactics adopted. In addition to COPS hiring grants, there were several other major categories of COPS grants programs. Making Officer Redeployment Effective (or MORE) grants were used to purchase equipment and hire civilians, with the goal of expanding the amount of time current law enforcement officers can spend on community policing.[Footnote 5] Some COPS grant programs provided funds for innovations in policing. For example, the Distressed Neighborhoods Pilot Project grants provided funds to communities with high levels of crime and/or economic distress to hire officers and implement a variety of strategies to improve public safety, and the Methamphetamine Initiative provided funds to state and local agencies to support a variety of enforcement, intervention, and prevention efforts to combat the methamphetamine problem. Finally, there was a variety of miscellaneous COPS grant programs. For example, the Regional Community Policing Initiative provided funds for training officers and representatives of communities and local governments, and COPS in Schools provided funds to law enforcement agencies to hire and train school resource officers to help prevent school violence and improve school and student safety. Each year COPS was required to distribute half of the grant funds to agencies with populations exceeding 150,000 and half of the grant funds to agencies with populations of 150,000 or less.[Footnote 6] Assessing whether COPS grants contributed to the decline in crime in the 1990s is complicated by many factors. Nationwide, the decline in crime began in 1993 before the COPS program made its first grants. Further, as crime declined during the 1990s, the amount of COPS grant funds increased. The general decline in crime makes it difficult to isolate the effect of COPS funds on crime. COPS grants were distributed in ways that make rigorous evaluations of their causal impacts difficult to determine. For example, the majority of police agencies received at least one COPS grant, as we estimated that 75 percent of police agencies received some type of COPS grant in our study period and among larger agencies (i.e., those serving populations of 100,000 or more persons) over 90 percent received at least one COPS grant. This distribution of COPS grant recipient agencies limits the number of agencies that could be used as comparison agencies against which to assess the effects of COPS grant funds on the decline in crime. Further, within the COPS' program allowable limits on amounts of funds that agencies could request, agencies generally chose the amount of grant funds they wanted and they generally received the amount of COPS funds that they requested. This introduces potential selection bias into the analysis, in that agencies may have both self- selected themselves on whether to participate in the COPS program and on the extent to which they would participate. To assess the effects of the COPS grant, under conditions of selection bias, it is necessary to isolate the effects of the grants from an agency's underlying capacity to address crime problems in its community and its choice to participate in the COPS program. To address these issues, we relied on methodological developments in research on crime that aim to disentangle the effects of programs such as COPS from other factors that could affect crime rates at the same time. For example, statistical models based on the use of a panel of data--or repeated observations on the same units, such as police agencies, over several time periods--allow for partial identification of the effects of COPS funds on crime by taking into account the variation between agencies over time in the amounts of expenditures and their relationships to crime rates. These methods also allow for the introduction of controls for pre-existing differences between units (agencies) and controls for differences over time that can help to identify causal relationships. In addition, crime research has adopted the use of statistical instruments as a method for identifying the causal direction of factors that could be determined simultaneously, such as the relationship between police officers and crime rates. Finally, by explicitly identifying mechanisms through which a program can have its effects--such as increases in officers attributable to COPS funds and their effects on crime--the program's model can be tested to rule out spurious correlations between inputs (such as COPS funds) and outcomes (such as crime). Results: Of the $7.32 billion in total COPS grant funds obligated between 1994 and 2001, the majority was obligated in the form of hiring grants, which accounted for about $4.69 billion, or 64 percent of all obligations. About 82 percent of agencies received at least one COPS grant, of any type, during this period, and about 75 percent received at least one hiring grant. By 2001, about 70 percent of the obligated amounts had been drawn down (or spent) by the agencies that received the grants.[Footnote 7] In the sample of COPS grantee agencies for which we obtained population and crime data and could make comparisons,[Footnote 8] we found that the COPS Office distributed grants according to statutory requirements--in that about half of the monies went to agencies covering populations of more than 150,000 and half went to agencies with populations of less than 150,000. In the aggregate, COPS grant funds were also distributed in proportion to the total volume of index crimes, but they were not distributed in proportion to the volume of violent crimes in the agencies that fell into these two population categories. For example, agencies serving populations of less than 150,000 persons received about 53 percent of COPS funding but they accounted for about 39 percent of violent crimes reported to the police. Agencies serving the smallest populations (such as places with populations of fewer than 25,000 persons) received an even larger share of COPS grant funds than the share of violent crimes that they contributed to the national total. Our analysis of changes in policing tactics shows that COPS grant funds were associated with increases in police agencies' adoption of tactics that evaluations have shown to be effective in reducing crime. Our comparisons of the pre-COPS grant program (i.e., 1993) levels of four types of policing tactics with their levels during the COPS program (i.e., in 1997) in the Policing Strategies survey sample of police agencies showed firstly that among all police agencies there were increases in the adoption of problem-solving, place-oriented, community- collaboration, and crime-analysis policing tactics. Specifically, among all agencies in the sample, between 1993 and 1997, the level of use of these tactics increased by about 32 percent--as measured by our summative index of policing tactics. However, our results also showed secondly that agencies that received COPS grant funds between 1993 and 1997 increased their adoption of problem-solving and place-oriented tactics more than did the agencies that did not receive COPS grant funds during these years. For example, agencies that received COPS grant funds between 1993 and 1997 increased their level of problem- solving tactics by 45 percent--as measured by our summative index of tactics--while those that did not receive grants increased them by 32 percent. In addition, our analysis of changes in tactics during the COPS program in the National Evaluation of COPS survey data from 1996 to 2000, also showed that COPS grantee agencies had larger increases in their use of problem-solving, place-oriented, and community- collaboration tactics than did the agencies that had not received a COPS grant by 2000. Agencies that received a COPS grant before 1996 had the largest increases in place-oriented and community- collaboration tactics, while agencies that received a COPS grant between 1996 and 2000 had larger increases in these tactics than did the agencies that had not received a COPS grant by 2000. Our fixed-effects regression models of the effects of the variation in the timing and amount of COPS hiring grant expenditures on the levels of sworn officers over the years from 1990 to 2001 showed that COPS hiring grant expenditures were associated with increases in the net number of sworn police officers per capita. We obtained these results after controlling for agency-level differences in the timing and amount of other COPS grant expenditures and other federal law enforcement grant expenditures; after controlling for annual changes in local economic and demographic conditions in the county in which an agency was located; and after controlling for changes in state-level factors that could affect the level of sworn officers. Specifically, from our analyses of data for police agencies serving populations of 10,000 or more persons--which covered about 75 percent of the total population in the U.S.--we estimated that each $25,000 in COPS hiring grant expenditures per year was associated with a net increase in the stock of sworn officers of about six-tenths of a sworn officer. When we assessed the effects of COPS grant expenditures in specific years on the level of sworn officers in these years, we found that compared to the baseline, pre-COPS program year of 1993, COPS expenditures in 1998 through 2000--three years in which COPS expenditures were at or near their peak amounts of about $815 million per year--were responsible for an estimated increase in the number of sworn officers per capita of about 3 percent above the levels that would have been expected without the funds. Upon projecting the results from our analysis of our sample of agencies to the entire U.S. population, we estimated that in the years from 1998 to 2000, COPS grant funds paid for about 18,000 officers in each of these years. Using COPS hiring grants as a statistical instrument to isolate the causal direction of the relationship between sworn officers and crime over the years from 1990 to 2001 in our sample of agencies serving populations of 10,000 or more persons, we found that COPS-funded increases in sworn officers per capita were associated with declines in the index crime rate and declines in the rates of murder, aggravated assault, robbery, burglary, larceny, and motor vehicle theft. From our regression results we calculated that a 1 percentage point increase in the level of COPS-funded sworn officers per capita was associated with about a 0.4 percent reduction in the index crime rate and about a 1 percent reduction in the violent crime rate. Among types of violent crimes, the estimated effects of changes in officers on crime varied; for example, a 1 percent increase in the level of officers was associated with a 2 percent reduction in robbery rates but a 0.5 percent reduction in the rate of aggravated assaults. These effects held after we controlled for the effects of other federal grant program funds received by agencies, local socio-economic and demographic changes that could affect crime, and state-level factors--such as increases in incarceration, changes in sentencing practice, and state- level changes in other programs such as welfare--that could also affect crime. In our analysis, the total effect of COPS grant expenditures on crime rates depended upon the level of COPS grant expenditures in a given year, and level of expenditures varied from year to year. For the years from 1998 through 2000--when COPS grant expenditures were at their peak levels--and among the agencies in our sample, we estimated that COPS grant expenditures were associated with about a 1 percent annual reduction in the index crime rate from its 1993 level, and about a 3 percent annual reduction in the violent crime rate from its 1993 level. When we projected these results from the analysis of our sample to the entire U.S. population, we estimated the annual reduction in crimes attributable to COPS funds nationwide. For 1998--when COPS expenditures amounted to about $820 million or about 1.5 percent of all local law enforcement expenditures--we estimated that COPS grant expenditures were associated with a reduction of an about 212,000 index crimes and 68,000 violent crimes from their levels in 1993. These crimes due to COPS grant expenditures amounted to about 8 percent of the decline in index crimes between 1993 and 1998 and about 12 percent of the decline in violent crimes over this period. During the years 1999 and 2000--when COPS expenditures also amounted to about $815 million or 1.5 percent of all local law enforcement expenditures--and crime continued to decline, we calculated that the COPS-funded reductions in crimes accounted for about 5 percent of the total reduction in index crimes and about 9 percent of the total reduction in violent crimes from their 1993 levels. Agency Comments: TO COME: If you or your staff have any questions concerning this report, please contact Laurie Ekstand at (202) or by e-m ail at Ekstrandl@gao.gov or William J. Sabol at (202) 512-3464 or Sabolwj@gao.gov. Key contributors to this report were: Sincerely yours, Signed by: Laurie Ekstrand, Director: Homeland Security and Justice Issues: Signed by: Nancy Kingsbury: Managing Director, Applied Research and Methods: Enclosures - I: [End of section] Enclosure I: Elements of the Database and Methods Used to Assess Effects of COPS Funds on Sworn Officers and Crime Rates: We constructed a database that consisted of up to 12 years of data (covering the period from 1990 to 2001) for each of 13,133 law enforcement agencies. These agencies represented about 75 percent of all local law enforcement agencies that reported data to the UCR; they covered about 95 percent of the population in the United States and about 95 percent of all crimes reported in the UCR. We constructed our database in the form of a panel, in which we obtained repeated measures on key variables in each agency over time. The database contained information on federal grant amounts, crime, officers, and socio-economic and demographic factors associated with crime. The types of information contained in our database were: * Grant obligation amounts to and annual amounts expended by each recipient of a COPS grant; * Annual amounts of other federal local law enforcement grants expended by both agencies that received COPS funds and those that did not; * Annual data on the number of index crimes and each category of index crimes, along with annual data on the number of such crimes per 100,000 population; * Annual observations on the number of sworn police officers and the number of officers per 10,000 population; * Annual data on economic and demographic factors that are related to crime, such as employment, per-capita income, relative size of persons in "crime-prone" age group of 15 to 24, the gender and racial composition of populations, and total population. We linked data from the various sources to each law enforcement agency contained in our sample of agencies. The sources of data used to compile the annual observations from 1990 to 2001 on local police departments included: * FBI Uniform Crime Reports (UCR)--Annual data files on the number of crimes and sworn officers reported by each agency to the UCR. The crime data are available separately for each Index Crime, which include the crimes of murder and non-negligent manslaughter, rape, robbery, aggravated assault, burglary, larceny, and motor vehicle theft. In addition, the data on sworn officers represent the number of officers in each agency on October 31 of each year. We used the originating agency code (or ORI) variable and census place codes to link crime and officer data to other data sources. * COPS Office--Annual data on each grant awarded and amounts obligated for each grant. The COPS data include the ORI code and other information that we used to link with other data sources. * Office of Justice Programs Financial Data--Annual data on the amount drawn down from each grant awarded by OJP. Because OJP and COPS share data, these OJP data also included COPS grant draw downs. The data on draw downs of funds best represent expenditures of federal grant monies in the years in which the dollars were spent. By contrast, amounts obligated in a given year may not be spent in that year. We used information about Census place (FIPS) codes and about OJP vendors to link data. * Bureau of Economic Analysis--County level data on per capita income and employment. We linked these data to agency-level data using census place codes. Local economic conditions within each county are applied to each agency within a county. * Census of Population Intercensal Estimates--Annual data for each county from 1990 to 1999 on population totals, and population breakdowns by gender, race, and age. We linked these data using census place code identifiers, and we extrapolated from underlying trends in each population category to obtain estimates for 2000 and 2001. In addition to the sources above that provided annual data on local policing agencies, we obtained and included in our database information from two separate surveys of nationally representative samples of local law enforcement agencies about the types of policing tactics that they implemented. Both surveys consisted of two waves of observations on the same police departments. The first survey, the National Survey of Community Policing Strategies (or "Policing Strategies Survey"), was conducted by the Police Foundation in 1993 (pre COPS), and Opinion Research Corporation (ORC) Macro International, Inc., and the Police Executive Research Forum (or PERF) in 1997 (during the COPS program). We used the Policing Strategies Survey data that contain information on 1,242 police agencies that responded to both waves of the survey, and of these, we were able to link the data from 1,236 agencies to our larger database on crime, officers, monies, and economic conditions. We used the Policing Strategies Survey data to compare changes in the types and levels of policing tactics that occurred during the COPS program with pre-COPS levels of tactics. The second survey, the National Evaluation of the COPS Program (or "National Evaluation of COPS"), was conducted by the Urban Institute, and it consisted of two waves of observations on a sample of 1,225 agencies in 1996 and again in 2000. We used the data from this survey to compare changes in policing occurring during the period of the COPS program. We were able to link the data from these two sources to our larger database using various identifiers. Prior to developing our database, we assessed the reliability of each data source and in preparing this report, we used only the data that we found to be sufficiently reliable for the purposes of our report. Methods Used in Our Analyses: We used various methods to analyze each of our reporting objectives. Methods to Analyze the Flow of COPS Funds: To assess the flow of COPS funds, we use the OJP Financial data to compute the amount of COPS funds obligated by COPS grants and the amount expended--or drawn down--by local police agencies during the period from 1994 to 2001. To describe the overall COPS funding trends by grant type, we analyzed the universe of agencies in the OJP data that received any federal law enforcement grant during the period 1990- 2001, regardless of whether or not the agency received a COPS grant during the period. For analyses of COPS funds by agency population sizes and for comparisons of funding levels with levels of violent and total index crime, we limited our analysis to the subsample of agencies whose crime and population data we were able to link to the OJP data. This resulted in a sample of 11,187 agencies--or 85 percent--of the 13,133 agencies in our crime sample. These agencies accounted for 86 percent of the reported index crimes in the U.S. (between 1990 and 2001). We capture the vast majority of COPS funds, but we exclude some COPS funds from our analyses either because we could not match them to agencies that reported data to the UCR or because the funds went to agencies that we excluded from our analyses.[Footnote 9] In analyzing the flow of COPS funds to local law enforcement agencies, our objectives were to assess whether the COPS grants were distributed according to statutory requirements and to assess the effect of these requirements on the distribution of COPS funds in relation to the volume of serious crimes occurring in the agencies that received COPS grants. Methods to Analyze Changes in Policing Practices and Tactics: To address whether COPS grants were associated with increases in the adoption of crime-preventing policing tactics, we analyzed the Policing Strategies and National Evaluation of COPS survey data. We compared differences in the levels of tactics between agencies that received COPS grants and those that did not receive grants during the two periods covered by the surveys. With the Policing Strategies Survey, we compared pre-and post-COPS program differences in the adoption of tactics (or from 1993 to 1997), and with the National Evaluation of COPS Survey, we compared changes in the adoption of tactics that occurred within the COPS program (or from 1996 to 2000). Each of the surveys reported data on the policing tactics used by agencies. Survey responses were obtained from knowledgeable officials within each agency, such as the police chief or the chief's designee. The number of items related to policing tactics differed between the two surveys. For the data in each survey, we classified items into tactics categories based on our assessment of the policing literature. We then assessed statistically the degree to which the items that we included in each of our four categories were correlated with each other and uncorrelated with items in the other categories. The Policing Strategies Survey data contained 30 items related to policing tactics. We combined eight tactics pertaining to increasing officer contact with citizens and improving citizen feedback into a community collaboration index. We used items on the crime analysis units within police departments to create our index of crime analysis. We combined seven tactics pertaining to increasing enforcement activity or place management in buildings, neighborhoods, or other specific places into an index of place-focused tactics. And we compiled the data on twelve items that reflected organizational efforts to reduce or interrupt recurring mechanisms that may encourage crime into a problem- solving tactics index. The classification of items from the Policing Strategies Survey into our four indexes of types of policing tactics is shown in table 2. Table 1: Policing Tactics and Strategy Items Utilized to Create Summative Indexes in the Policing Strategies Survey. Community collaboration: * Agency uses foot patrol as a specific assignment; * Agency uses foot patrol as a periodic expectation for officers assigned to cars; * Agency uses citizen surveys to determine community needs and priorities; * Agency uses citizen surveys to evaluate police service; * Patrol officers conduct surveys in area of assignment; * Patrol officer meet regularly with community groups; * Supervisors maintain regular contact with community leaders; * Agency uses citizens as volunteers within the police agency. Crime analysis: * Agency has a decentralized crime analysis unit/ function; * Agency has a centralized crime analysis unit/function; * Supervisors manage crime analysis for geographic area of responsibility. Place-oriented practices: * Agency designates some officers as "community" or "neighborhood" officers; * Agency uses building code enforcement as a means of helping remove crime; * Geographically based crime analysis made available to officers; * Command or decision-making responsibility tied to neighborhoods or beats; * Patrol officers enforce civil and code violations in area; * Fixed assignment of patrol officers to specific beats or areas; * Agency uses other regulatory codes to combat drugs and crime. Problem-solving practices: * Specific training provided officers for problem identification and resolution; * Training for citizens in problem identification or resolution; * Landlord/manager training programs for order maintenance; * Interagency involvement in problem identification and resolution; * Agency has revised procedures to deal with neighborhood problems; * Multidisciplinary teams to deal with special problems such as child; * Specialized problem solving unit; * Patrol officers work with citizens to identify and resolve area problems; * Organization has been redesigned to support problem solving efforts; * Line supervisors make final decision about which problems are to be addressed; * Line supervisors make final decision about how to handle most community problems; * Line supervisors make final decision about application of agency resources to solve problem in geographic area of responsibility. Source: Policing Strategies Survey, 1993 and 1997. Note: Each individual items is coded dichotomously (yes/no) to indicate whether an agency implemented the specific tactic. [End of table] The National Evaluation of COPS Survey contained data on 19 items relating to policing tactics. We classified these items into the same tactics categories as we did with the Policing Strategies Survey data. Many of the items in the National Evaluation of COPS Survey were worded in the same way as in the Policing Strategies Survey. Table 2: Policing Tactics and Strategy Items Utilized to Create Summative Indexes in the National Evaluation of COPS Survey. Community collaboration: * Regular community meetings to discuss crime; * Surveys of citizens to determine general community needs and satisfaction with your agency; * Citizen action/advisory councils in precincts or beats; * Officers analyze community resident's comments to identify recurring patterns of crime and disorder on their beats; * Considering neighborhood values in creating solutions or planning projects; * Varying styles of preventive patrol (e.g. bikes, walk and talk); * Joint projects with local businesses to reduce disorder or petty crime. Crime analysis: * Analyzing crime patterns using computerized geographic information systems; * Officers analyze and use crime data to identify recurring patterns of crime and disorder on their beats. Place-oriented practices: * Clean up/fix up projects with community residents; * Joint projects with community residents to reduce disorder such as loitering, public drinking; * Beat or patrol boundaries that coincide with neighborhood/community boundaries; * Alcohol, housing, or other code enforcement to combat crime and disorder. Problem-solving practices: * Designating certain patterns as "problems" or "projects" requiring non-traditional responses; * Analyzing problems with business or property owners, school principals, or property managers or occupants; * Analyzing problems with probation/parole officers or others who monitor offenders; * Using agency data to measure effects of responses to problems; * Documenting problems, projects, analyses, responses, failures, and successes in writing; * Team approach instead of chain of command for prevention, problem solving, and law enforcement. Source: National Evaluation of COPS Survey, 1996 and 2000. Note: Each individual items is coded dichotomously (yes/no) to indicate whether an agency implemented the specific tactic. [End of table] From each of the surveys, we developed summative indexes of the overall number tactics and the number of tactics within each of the 4 categories of policing tactics. We then computed and compared changes in the mean levels of tactics between the COPS grantee and non-grantee agencies. To control for differences between agencies and trends over time in the adoption of policing tactics that could account for differences in tactics, in our analysis of the Policing Strategies Survey data and the National Evaluation of COPS Survey data, we estimated regressions that controlled for social and economic characteristics of the places in which police agencies were located and pre-existing trends in officers and crimes. Methods to Estimate the Effects of COPS Funds on Officers and Crime: We adopted similar approaches to estimating the effects of COPS funds on officers and crime, as described below. Methods to Estimate the Effects of COPS Funds on Officers: To estimate the effect of COPS funds on officers, we used fixed-effects regression methods that permitted us to assess changes in the levels of sworn officers per 10,000 persons as a function of COPS funds, other federal funds, local economic conditions, and changes in the age, gender, and racial composition of local populations. The fixed effects regression models allow us to control for two sources of unmeasured variation (i.e., omitted variables): The pre-existing differences among the agencies in our sample that are constant within agencies over time and the differences within agencies over time in relation to the overall trends in variables. By adopting these models we are able to control for the effects of unmeasured variables that vary over time between agencies and that might be correlated with our dependent variables. We introduced fixed effects at the level of the local law enforcement agency. In addition, to control for state-level influences on officers that we were unable to observe directly--such as changes in state sentencing practices--we also introduced into our models state- by-year level fixed effects. Finally, to control for underlying trends in the pre-COPS grant period in sworn officers and crime, we estimated regressions that analyzed these trends, and then we classified each agency's trend within population size groups. This allowed us to compare agencies within size categories that had similar trends in officers and crime prior to the COPS program. Specifically, we separated the agencies in 4 groups based on the growth rate in both officers and in crime during 1990-1993, which is prior to the implementation of the COPS program. We constructed each combination of these groups, which produced 16 cells. These cells were then interacted with each year, and 4 population categories, for a total of 704 effects. In essence then, each agency is being compared with another agency that had a similar "trajectory" of crime and officers in the pre-COPS period.[Footnote 10] We analyzed the data for 5,199 police agencies with populations of 10,000 or more persons. We estimated several regressions of the effect of COPS funds on sworn officers, and we included as time-varying independent variables the per capita amounts of COPS hiring grants, COPS innovative grants, COPS MORE grants, COPS miscellaneous grants, Edward Byrne Memorial State and Local Law Enforcement Assistance Discretionary Grants, Local Law Enforcement Block Grants (LLEBG), the employment to population ratio, per capita income, the percent of the population that was male, the percent of the population that was nonwhite, and the percent of the population that was between the ages of 15 and 24. All of our economic and population measures were observed at the level of the county that contained the local law enforcement agency. Our regressions included the agency level fixed effects, year effects, effects for the pre-COPS trends in the growth of officers and crime, population weights (to allow us to estimate national-level effects), and state-by-year fixed effects. Methods to Estimate the Effects of COPS Funds on Crime through Officers: We estimated the impact of COPS funding on crime through these funds' effects on changes in officers. We made use of the fact that, unlike the other COPS grant types, hiring grants were earmarked specifically for the hiring of officers. Consequently, variation in the number of officers coming from hiring grant should be unrelated with other changes in police expenditures. In this sense, we used COPS hiring as an instrument to isolate the direction of causality between officers and crime rates. We then estimated population-weighted regressions of the impact of the flow of COPS dollars on crime rates, net of pre-existing trends in crime rates and growth of officers, and net of economic conditions, population change, and other amounts of funding for local law enforcement. We used the population weights to allow us to develop estimates for the nation as a whole. We estimated our regressions on the same sample of 5,199 agencies with populations of 10,000 or more that we used in our officers equations. We estimated crime equations separately by type of crime and population size group. Under the assumption that hiring grants can be used as an instrument, we used the results from the two regressions--officer rate on COPS funds and crime rate on COPS funds--to calculate the elasticity of crime with respect to officers (i.e., percent change in crime rates attributable to the percent change in officers). More specifically, we estimated the effect of COPS funds on crime using a reduced form equation that included measures of specific types of COPS grant funds (e.g., hiring, MORE, innovative, and miscellaneous grants) expended, measures of other federal grant funds expended, and the controls for socio-economic and demographic changes in the population. We included the 704 variables that controlled for pre-COPS trends in officers and crime, and we introduced agency level, year level, and state-by-year fixed effects. The state-by-year fixed effects allow us to control for unmeasured state-level sources of variation with crime, such as increases in state incarceration rates, changes in state sentencing practices, and changes in other state programs--such as welfare reform--that could affect crime rates. We estimated regressions separately for the index crime rate and by type of index crime. After obtaining the coefficients from our officers and crime regressions, and to obtain estimates of the effects of COPS funds on crime through officers, we then calculated the elasticity of crime with respect to officers in any given year. The elasticity of crime with respect to officers provides an estimate of the effect of a 1 percent change in the level of sworn officers per capita on the per capita crime rate. Using these elasticities, we then apportioned the amount of the reduction in crime that we could attribute to COPS funds. To project the effects of COPS funds on crime to the nation as a whole, we then weighted the estimates of COPS effects on crime in our sample up to the nation as whole. [End of section] FOOTNOTES [1] VCCLEA contained other provisions to address violent crime, such encouraging states to increase the use of incarceration for violent offenders through the Violent Offender Initiative and Truth-in- Sentencing grants, enhancing penalties for gang crimes, and expanding the number of Federal death penalty offenses. [2] The study that we reviewed was: Zhao, J. and Thurman, Q. A National Evaluation of the Effect of COPS Grants on Crime from 1994 to 1999 (December 2001). Our review of this study was reported in: GAO: Technical Assessment of Zhao and Thurman's 2001 Evaluation of the Effects of COPS Grants on Crime, GAO-03-867R (Washington, D.C.: June 13, 2003). [3] In addition to the Zhao and Thurman work cited above, David Muhlhausen has assessed the effects of COPS funds on crime rates. Using county-level data on crimes and funds, Muhlhausen found that other than the Innovative grant program, COPS grants were not associated with reductions in crime. See: Muhlhausen, D. Do Community Oriented Policing Services Grants Affect Violent Crime Rates (Washington, D.C.: Heritage Foundation, May 25, 2001). [4] The first survey was National Survey of Community Policing Strategies, and it was conducted by the Police Foundation in 1993 to provide information in what was occurring and what needed to occur in the development and implementation of community policing. In 1997, Opinion Research Corporation (ORC) Macro International, Inc., and the Police Executive Research Forum (PERF) conducted the National Survey of Community Policing Strategies Update, a longitudinal follow-up to the previous survey, using the same sample as the 1993 survey. The 1997 survey was designed to provide information on the most current practices and trends in community policing. In the remainder of this letter, we refer to the two waves of this longitudinal survey as the Policing Strategies survey. The second survey was the National Evaluation of the COPS Program survey, which was conducted by the Urban Institute between 1996 and 2000. It was a nationally-representative sample of law enforcement agencies that were contacted in 1996 and again in 2000. In the remainder of this letter, we refer to this second survey as the National Evaluation of COPS survey. [5] MORE grants in 1995 could also be used to pay for police officer overtime costs. [6] Of funds available in any fiscal year, up to 3 percent may be used for technical assistance or for evaluations or studies carried out or commissioned by the Attorney General. The requirement to allocate the funds by size of agency population applies to the remaining funds in any fiscal year (USC 42, chapter 46, subchapter XX, sec. 3793, (a) (11) (B). [7] Due to lags between the time when grants are obligated and all of the funds are expended, not all COPS obligated amounts during the years from 1994 to 2001 were drawn down or expended in that same period by the agencies that received the grants. [8] This sample consisted of 11,187 agencies, or 85 percent of the agencies that reported crime data to the UCR. This group of agencies accounted for 86 percent of the reported index crimes in the United states between 1990 and 2001. [9] The COPS Office received applications from law enforcement agencies that did not have originating agency (ORI) numbers. These numbers are used by the UCR to identify agencies that report crimes. The COPS Office assigned the agencies that did not have ORIs a "ZZ" ORI code. This code was based upon a number that the applicant agency reported to which the COPS Office appended a "ZZ" code. We were unable to match these "ZZ" agencies to agencies in the UCR. However, our analysis of these agencies suggests that they were mostly smaller agencies of recent origin as well as consortia, tribal and private research organizations. [10] This approach was first implemented by Evans, W. and Owens, E in Flypaper COPS (College Park, Md.: Univeristy of Maryland, April, 2005).