Evidence is converging to show that casinogambling causes significant increases in crime. Takenaltogether, casinos impose crime and other costs –paid for by society, including those who do not gamble– that exceed their benefits and represent substantialburdens on nearby populations. Because casinogambling fails a cost-benefit test, policymakers shouldgive serious consideration to options that includeimposing taxes equal to the costs casinos impose,restricting casino expansion, or banning casinogambling altogether.Crime is affected by multiple factors includingpopulation density, the number of males and femalesin different age ranges, percent of each age group thatis white, percent of each age group that is black, percapita personal income, unemployment rates, percapita retirement compensation, per capita incomemaintenance payments, and “shall issue” laws (givingcitizens the right to carry concealed firearms uponrequest—believed to reduce certain crimes). Hence,connecting any single cause such as casinos to crimeis controversial. Only by careful sifting of a large bodyof data can the effect of casinos be separated fromother causes to establish a connection. The gamblingindustry naturally has resisted research findings thatlink casinos to more crime.How do researchers conclude that casinos causecrime and measure the size of the connection? Thereare two ways—the first is through the study ofproblem and pathological gamblers and the second isthrough statistical analysis of crime numbers.Connecting casinos to crimePathological gambling is a recognized impulsecontrol disorder of the Diagnostic and StatisticalManual (DSM-IV) of the American PsychiatricAssociation. Pathological gamblers (often referred toas “addicted” or “compulsive” gamblers) are identifiedby a number of characteristics including repeatedfailures to resist the urge to gamble, loss of controlover their gambling, personal lives and employment,reliance on others to relieve a desperate financialsituation caused by gambling, and the committing ofillegal acts to finance gambling. Problem gamblershave similar problems, but to a lesser degree.It appears that a significant proportion of thepopulation is susceptible to problem or pathologicalgambling. The latent propensity becomes overt whenthe opportunity to gambleis provided and sufficienttime has elapsed for theproblem to manifest.Pathological gamblers aregenerally found toconstitute one or twopercent of the populationand problem gamblersare another two to threepercent in areas wherecasino gambling is available. One study of gamblers intreatment found that 62 percent committed illegal actsas a result of their gambling. Eighty percent hadcommitted civil offenses and 23 percent were chargedwith criminal offenses, according to a 1990 MarylandDepartment of Health and Mental Hygiene survey. Asimilar survey of nearly 400 members of Gambler’sAnonymous showed that 57 percent admitted stealingto finance their gambling. Moreover, the amounts arenot small. On average they stole $135,000, and totaltheft was over $30 million, according to the testimonyof Henry Lesieur from the Institute of Problem Gam-bling before the National Gambling Impact StudyCommission, Atlantic City, New Jersey, January 22,1998. The National Gambling Impact StudyCommission’s final report issued in June 1999 reportedthat among those who did not gamble (had notgambled in the past year) only 7 percent had ever beenFor more ongambling, seeLegal Gambling inIllinois: A Primerat the IGPA web site:www.igpa.uillinois.edu
——————————————————————————–
Page 2
Policy Forum is an occasionalpublication that presents discussions andresults of recent research on importantpolicy issues. Manuscripts by universityand non-university authors are welcome.Scott Koeneman, Managing EditorIGPAUniversity of Illinois1007 W. Nevada St.Urbana, IL 618012 Casino gambling causes crimeincarcerated. In contrast, more than three times thisnumber (21.4 percent) of individuals who had beenpathological gamblers at any point during their lifetimehad been incarcerated.By tallying up the crimes of pathological andproblem gamblers and the associated costs to societysuch as police, apprehension, adjudication, andincarceration costs, the average crime costs to societyof an additional pathological or problem gambler(some studies lump the two groups together) can bedetermined. Recent research using this methodologyfound that an average problem gambler costs society$10,112 per year. Crime costs constituted $4,225, or 42percent of these costs.Combining crime costs with studies of the preva-lence of pathological and problem gamblers providescrime cost figures for society as a whole. Using thenumbers just reported implies annual crime costs peradult capita of $57. This number can be compared tothe crime costs found by the second method forrelating casinos to crime.Connections in crime statisticsA second way to determine the effect of casinos oncrime is to look directly at aggregate crime statistics.The advantage is that the method is direct, and—because it looks at more than just the crimes commit-ted by problem and pathological gamblers—it is moreinclusive. The disadvantage is that it may be difficult todistinguish the share of crime related to casinos fromthe mass of other crime that occurs all the time.Moreover, the period of major casino expansion in theUnited States, 1991 to 1997, coincides with a period ofsecular decline in overall crime rates. It would betempting, therefore, to observe that crime fell after aparticular casino was introduced and from this con-clude that the casino reduced crime. Such a conclusionwould be false if crime would have fallen even furtherwithout the casino. Finally, because the effects ofcasinos might differ in different areas, a large samplecould be needed to reliably pinpoint the truth.In research conducted at the University of Illinoisand the University of Georgia and with these factors inmind, Professor David Mustard, Cynthia Hunt Dilleyand I examined crime statistics for all 3,165 counties inthe United States for twenty years beginning with year1977. This period covers the period of introduction ofcasinos in all counties with the exception of Nevada.The number of offenses for the 7 FBI Index I offenses(robbery, aggravated assault, rape, murder, larceny,burglary, and auto theft) was obtained from theFederal Bureau of Investigation’s Uniform CrimeReport County-Level Data. We obtained U.S. CensusBureau data to control fordemographic, income,and other variables thataffect crime as describedabove. In all, 54 variableswere used to explainobserved crime ratesacross counties and time.We included twelvevariables to identify eachyear from four yearsbefore the opening of thefirst casino in a county toseven years after itopened. These variablesserve two purposes: first,to distinguish the effectsof casinos from changes that preceded their opening(for example, a trend toward lawlessness conceivablycould lead to the opening of a casino instead of thereverse); and second, to sort out the timing of thoseeffects (an effect on crime could take several years todevelop). To find the dates for the first casino openingwe contacted state gaming authorities in every state,called casinos to find opening date or date of first ClassIII gambling (in many cases casinos began as bingohalls and switched at a later date), and used casinointernet website information to check our data. Thefinal list was verified against the annually producedExecutive’s Guide to North American Casinos.What did the data show? If property crime ratesare indexed so that 1982 rates equal 100, then thecrime rate in 1991 was 99.7 in non-casino counties(counties that had no casinos during the sampleperiod) and 100.3 in casino counties (counties that hada casino by the end of the sample period)—hardly anydifference at all. However, looking at the same statisticjust 5 years later—after casinos had begun operation inthe majority of the casino counties—the indexes stoodat 82.1 for non-casino counties and 93.7 for casinocounties. The raw data suggests, therefore, that 12.4percent of the crime observed in casino countieswould not be there if casinos were absent. A similarpicture emerges for violent crimes.The problem with using the raw data for infer-ences, however, is that direct comparisons do not take“The raw datasuggests… that12.4 percent ofthe crimeobserved incasino countieswould not bethere if casinoswere absent. ”
——————————————————————————–
Page 3
Policy Forum 3into account other factors that cause crime. Forexample, it is well known that crime rates in areas ofhigh population density tend to be higher. What ifbetween 1991 and 1996 casino counties experienced asignificant increase in their population density? Thensome or all of the increased crime might be due to thechange in population density. This is why we collectedso many other variables and applied regressionprocedures to them—to separate the changes in crimerates due to other factors from those due to casinos.After adjusting for all of the other factors aninteresting picture begins to emerge, both in terms ofthe share of crime in casino counties due to casinosand in terms of the pattern that the changes take overtime. The data indicated that compared to non-casinocounties there was no discernable difference in crimein casino counties in the four years before casinosopened that could be attributed to the opening ofcasinos. (We did not expect to find any connection, sothis finding was anticipated.) For the first three yearsafter the casino began operation, there also was nosignificant impact on crime rates. After the third year,however, crime rates began to rise in casino countiescompared to those without casinos. By 1996, casinosaccounted for 10.3 percent of the observed violentcrime and 7.7 percent of the observed property crimein casino counties. Estimates of the share of crimeattributable to casinos in 1996 for individual crimesranged from 3 to 30 percent. Auto theft was thehighest, followed by robbery at 20 percent. (In addi-tion to stealing an auto, auto theft includes taking partsof cars such as expensive sound equipment as well asthings from or out of a car.) The values for the rest ofthe offenses were between 3 and 10 percent.Criminologists in the late 1980s and early 1990sestimated the cost per victimization of different typesof crime. Applying these costs to the implied numberof offenses for each crime due to casinos and dividingby the adult population of casino counties in 1996produced an annual cost for casino-induced crime of$63 per adult capita. This figure is remarkably close tothe $57 per adult capital crime cost estimated throughthe study of problem and pathological gamblers.ImplicationsCritics of casino gambling point to a number ofsocial costs. In addition to the direct governmentalcosts of regulating casinos and providing socialservices occasioned by gambling, these include thecosts of bankruptcy, illness, suicide, harm to families,lost economic output, and crime, among others.Research to pinpoint the size of many of these costs isstill in its beginning stages. This paper has describedresearch directed to determining the costs to societyof just seven Index I crimes tracked by the FBI:Larceny, burglary, auto theft, robbery, aggravatedassault, rape, and murder. In areas with casinos theevidence points to costs of $63 per adult per year, butother studies that provided information on all of thesocial costs of casinos suggest that the total is over$100 per adult annually. Estimates implying costs of$135, $150 and more are common.The social benefitsof casinos are theincrease in profits andtaxes from casinos(casino profits and taxesless lost profits and taxesof other businesses dueto casinos) plus theconvenience value toconsumers of havingcasinos nearby com-pared to having to travelgreater distances togamble. Research on thebenefits suggests theyare no larger than $40per adult annually. Thuscasino gambling fails acost-benefit test by asubstantial margin interms of Index I crimesalone.It is an open question whether casino gamblingcan be offered in a way that allows citizens who couldgamble without harm to do so while at the same timepreventing the creation of problem and pathologicalgamblers and the social costs already discussed. Ifcasino gambling cannot be offered in ways that causeit to pass a cost-benefit test, then banning it (as wasdone until recently) is preferable on economic terms.In light of the evidence, what can a responsiblelegislator do? One option is to tax casinos by anamount equal to the costs that they impose on society.As we have shown, a conservative estimate of thesecosts is about $100 annually per nearby resident. Sincecasinos typically take in revenues of around $200 peradult each year from nearby residents, such atax would represent 50 percent of casino revenues. Ifsocial costs were ultimately determined to be higher,required taxes would also be higher. Withtaxes imposed at the appropriate level, some casinoswould go out of business. Only those casinos thatcould pass a cost-benefit test by compensating societyfor the damage they do would operate. Dr. Earl Grinols is a professor in the Department ofEconomics, University of Illinois Urbana-Champaignand an affiliate of the Institute of Government andPublic Affairs, University of Illinois.“If casinogamblingcannot beoffered in waysthat cause it topass a cost-benefit test,then banning it(as was doneuntil recently) ispreferable oneconomicterms. ”
——————————————————————————–
Page 4
The Institute of Government and Public Affairsserves society by helping to improve public policiesand the performance of government.It does this by disseminating re-search on public policy issues and thepublic decision-making process, andfacilitating the application of suchresearch to the issues and problemsconfronting decision makers andothers who address public issues.This is done through basic andapplied problem-solving research, the communicationof research results to other researchers throughscholarly publications, and the application of researchresults through public service and continuing educa-tion programs that help practitioners understand andaddress the issues they face.IGPA is distinct among premierpolicy research units because of itsdedication to public service. Ourwork not only advances knowledgebut also provides practical assistanceto government and public policymakers. Such assistance takes theform of direct consultation, member-ship on boards and commissions,formally contracted research projects, training forpublic servants, information development and dissemi-nation, testimony before legislative bodies, surveys,program analyses, and analyses for the public media.Policy ForumInstitute of Government and Public AffairsUniversity of Illinois1007 W.Nevada St.Urbana, Illinois 61801Printed on recycled paper.4 Casino gambling cause crimeVisit the IGPA Web Site: www.igpa.uillinois.edu
