• Selected
• Entire List

Abstract:
We offer evidence that legalized abortion has contributed significantly to recent crime reductions. Crime began to fall roughly 18 years after abortion legalization. The 5 states that allowed abortion in 1970 experienced declines earlier than the rest of the nation, which legalized in 1973 with Roe v. Wade. States with high abortion rates in the 1970s and 1980s experienced greater crime reductions in the 1990s. In high abortion states, only arrests of those born after abortion legalization fall relative to low abortion states. Legalized abortion appears to account for as much as 50 percent of the recent drop in crime.

Abstract:
We offer evidence that legalized abortion has contributed significantly to recent crime reductions. Crime began to fall roughly 18 years after abortion legalization. The 5 states that allowed abortion in 1970 experienced declines earlier than the rest of the nation, which legalized in 1973 with Roe v. Wade. States with high abortion rates in the 1970s and 1980s experienced greater crime reductions in the 1990s. In high abortion states, only arrests of those born after abortion legalization fall relative to low abortion states. Legalized abortion appears to account for as much as 50 percent of the recent drop in crime.

Abstract:
Donohue and Levitt (2001) present a number of analyses that suggest a causal link between legalized abortion and reductions in crime almost two decades later when the cohorts exposed to legalized abortion reach their peak crime years. Joyce (2003) challenges that finding. In this paper, we demonstrate that Joyce's failure to uncover a negative relationship between abortion and crime is a direct consequence of his decision to focus exclusively on the six-year period 1985-90 without including adequate controls for the crack epidemic. We provide empirical evidence that crack hit the high-abortion early legalizing states harder and earlier. We then demonstrate that using precisely the same treatment and control groups as Joyce, but extending the data analysis to encompass the lifetime criminal experiences (as opposed to an arbitrary six-year window), the evidence strongly supports the hypothesis that legalized abortion reduces crime. We also show that our original results are robust to focusing on only the cohorts born immediately before or after Roe v. Wade. The data suggest that ease of access to abortion, rather than simply de jure legalization, is a critical determinant of the extent of the crime reduction.

Abstract:
Donohue and Levitt (2001) present a number of analyses that suggest a causal link between legalized abortion and reductions in crime almost two decades later when the cohorts exposed to legalized abortion reach their peak crime years. Joyce (2003) challenges that finding. In this paper, we demonstrate that Joyce's failure to uncover a negative relationship between abortion and crime is a direct consequence of his decision to focus exclusively on the six-year period 1985-90 without including adequate controls for the crack epidemic. We provide empirical evidence that crack hit the high-abortion early legalizing states harder and earlier. We then demonstrate that using precisely the same treatment and control groups as Joyce, but extending the data analysis to encompass the lifetime criminal experiences (as opposed to an arbitrary six-year window), the evidence strongly supports the hypothesis that legalized abortion reduces crime. We also show that our original results are robust to focusing on only the cohorts born immediately before or after Roe v. Wade. The data suggest that ease of access to abortion, rather than simply de jure legalization, is a critical determinant of the extent of the crime reduction.

Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.

Abstract:
Race has long been recognized as playing a critical role in policing. In spite of this awareness, there has been virtually no previous research attempting to quantitatively analyze the issue. In this paper, we examine the relationship between the racial composition of a city's police force and the racial patterns of arrests and crime. Increases in the number of minority police are associated with significant increases in arrests of whites, but have little impact on arrests of non-whites. Similarly, more white police increase the number of arrests of non-whites, but do not systematically affect the number of white arrests. The race of police officers has a less clear-cut impact on crime rates. It appears that own-race policing may be more effective in reducing property crime, but no systematic differences are observed for violent crime. These results are consistent either with own-race policing leading to fewer false arrests or greater deterrence. In either case, own-race policing appears more "efficient" in fighting property crime.

Abstract:
Race has long been recognized as playing a critical role in policing. In spite of this awareness, there has been virtually no previous research attempting to quantitatively analyze the issue. In this paper, we examine the relationship between the racial composition of a city's police force and the racial patterns of arrests and crime. Increases in the number of minority police are associated with significant increases in arrests of whites, but have little impact on arrests of non-whites. Similarly arrests of non-whites, but do not systematically affect the number of white arrests. The race of police officers has a less clear-cut impact on crime rates. It appears that own-race policing may be more effective in reducing property crime, but no systematic differences are observed for violent crime. These results are consistent either with own-race policing leading to fewer false arrests or greater deterrence. In either case, own-race policing appears more "efficient" in fighting property crime.

Abstract:
Measuring the relative likelihood of fatal crash involvement for different types of drivers would seem to require information on both the number of fatal crashes by driver type and the fraction of drivers on the road falling into each category. In this paper, however, we present a methodology for measuring fatal crash likelihood that relies solely on fatal crash data. The key to our identification strategy is the hidden richness inherent to two-car crashes. Crashes involving two drinking drivers are proportional to the square of the number of drinking drivers on the road; crashes with one drinking and one sober driver increase linearly in the number of drinking drivers. Imposing a limited set of assumptions (e.g. independence across crashes, equal mixing on the roads), we are able to estimate both the likelihood of causing a fatal crash and the fraction of drivers of each type on the road. Our estimates suggest that drivers with alcohol in their blood are at least eight times more likely to cause a fatal crash; legally drunk drivers pose a risk at least 15 times greater than sober drivers. Males, young drivers, and drivers with bad past driving records are all more dangerous, but the impact of these other factors is far less than that of alcohol.

Abstract:
Measuring the relative likelihood of fatal crash involvement for different types of drivers would seem to require information on both the number of fatal crashes by driver type and the fraction of drivers on the road falling into each category. In this paper, however, we present a methodology for measuring fatal crash likelihood that relies solely on fatal crash data. The key to our identification strategy is the hidden richness inherent to two-car crashes. Crashes involving two drinking drivers are proportional to the square of the number of drinking drivers on the road; crashes with one drinking and one sober driver increase linearly in the number of drinking drivers. Imposing a limited set of assumptions (e.g. independence across crashes, equal mixing on the roads), we are able to estimate both the likelihood of causing a fatal crash and the fraction of drivers of each type on the road. Our estimates suggest that drivers with alcohol in their blood are at least eight times more likely to cause a fatal crash; legally drunk drivers pose a risk at least 15 times greater than sober drivers. Males, young drivers, and drivers with bad past driving records are all more dangerous, but the impact of these other factors is far less than that of alcohol.

Abstract:
We analyze a unique data set detailing the financial activities of a drug-selling street gang on a monthly basis over a four-year period in the recent past. The data, originally compiled by the gang leader to aid in managing the organization, contain detailed information on both the sources of revenues (e.g. drug sales, extortion) and expenditrues (e.g. costs of drugs sold, weapons, tribute to the central gang organization, wages paid to various levels of the gang). Street-level drug dealing appears to be less lucrative than is generally though. We estimate the average wage in the organization to rise from roughly $6 per hour to $11 per hour over the time period studied. The distribution of wages, however, is extremely skewed. Gang leaders earn far more than they could in the legitimate sector, but the actual street-level dealers appear to earn less than the minimum wage throughout most of our sample, in spite of the substantial risks associated with such activities (the annual violent death rate in our sample is 0.07), There is some evidence consistent both with compensating differentials and efficiency wages. The markup on drugs suggests that the gang has substantial local market power. Gang wars appear to have an important strategic component: violence on another gang's turf shifts demand away from that area. The gang we observe responds to such attacks by pricing below marginal cost, suggesting either economic punishment for the rival gang or the presence of switching for users that makes market share maintenance valuable. We investigate a range of alternative methods for estimating the willingness of gang members to accept risks of death, all of which suggest that the implicit value that gang members place on their own lives is very low.

Abstract:
The market for sports gambling is structured very differently than the typical financial market. In sports betting, bookmakers announce a price, after which adjustments are small and infrequent. As a consequence, bookmakers do not play the traditional role of market makers whose primary function is to match buyers and sellers, but rather, take large positions with respect to the outcome of game. Using a unique data set that includes both prices and quantities of bets placed over the course of an NFL season, I demonstrate that this peculiar price-setting mechanism allows bookmakers to achieve substantially higher profits than would be possible if they played the role of the typical market maker. Bookmakers are more skilled at predicting the outcomes of games than bettors and are able to systematically exploit bettor biases by choosing prices that deviate from the market clearing price. While this strategy exposes the bookmaker to risk on any particular game, in aggregate the risk borne is minimal. Bookmaker profit maximization provides a simple explanation for heretofore puzzling deviations from market efficiency that were observed in past empirical work. I find little evidence that there exist bettors who are systematically able to beat the bookmaker, even given the distorted prices that bookmakers set. The results concerning whether aggregating across bettor preferences improves the ability to forecast outcomes are inconclusive.

Abstract:
It is typically difficult to differentiate empirically between deterrence and incapacitation since both are a function of expected punishment. In this paper we demonstrate that the introduction of sentence enhancements (i.e. increased punishments that are added on to prison sentences that would have been served anyway) provides a direct means of measuring deterrence. Because the criminal would have been sentenced to prison anyway, there is no additional incapacitation effect from the sentence enhancement in the short-run. Therefore, any immediate decrease in crime must be due to deterrence. We test the model using California's Proposition 8 which imposed sentence enhancements for a selected group of crimes. In the year following its passage, crimes covered by Proposition 8 fell by more than 10 percent relative to similar crimes not affected by the law, suggesting a large deterrent effect. Three years after the law comes into effect, eligible crimes have fallen roughly 20-40 percent compared to non-eligible crimes. This large deterrent effect suggests that sentence enhancements, and may be more cost-effective than is generally thought.

Abstract:
Differentiating empirically between deterrence and incapacitation is difficult since both are a function of expected punishment. In this paper we demonstrate that the introduction of sentence enhancements (i.e., increased punishments that are added on to prison sentences that would have been served anyway) provides a direct means of measuring deterrence. Because the criminal would have been sentenced to prison anyway, there is no additional incapacitation effect from the sentence enhancement in the short-run. Therefore, any immediate decrease in crime must be due to deterrence. We test the model using California's Proposition 8, which imposed sentence enhancements for a selected group of crimes. Using a differences-in-differences approach (comparing changes in eligible crimes relative to non-eligible crimes in California and the rest of the nation), Proposition 8 appears to reduce eligible crimes by four percent in the year following its passage and eight percent three years after passage. These immediate effects are consistent with deterrence. The impact of the law continues to increase five to seven years after its passage, suggesting that incapacitation may be important as well.

Abstract:
Agents are often better informed than the clients who hire them and may exploit this informational advantage. Real-estate agents, who know much more about the housing market than the typical homeowner, are one example. Because real estate agents receive only a small share of the incremental profit when a house sells for a higher value, there is an incentive for them to convince their clients to sell their houses too cheaply and too quickly. We test these predictions by comparing home sales in which real estate agents are hired by others to sell a home to instances in which a real estate agent sells his or her own home. In the former case, the agent has distorted incentives; in the latter case, the agent wants to pursue the first-best. Consistent with the theory, we find homes owned by real estate agents sell for about 3.7 percent more than other houses and stay on the market about 9.5 days longer, even after controlling for a wide range of housing characteristics. Situations in which the agent's informational advantage is larger lead to even greater distortions.

Abstract:
This article describes basic facts regarding the Black-White test score gap over the first four years of school. Black children enter school substantially behind their White counterparts in reading and math, but including a small number of covariates erases the gap. Over the first four years of school, however, Blacks lose substantial ground relative to other races; averaging 0.10 standard deviations per school year. By the end of third grade, there is a large Black-White test score gap that cannot be explained by observable characteristics. Blacks are falling behind in virtually all categories of skills tested, except the most basic. None of the explanations we examine, including systematic differences in school quality across races, convincingly explain the divergent academic trajectory of Black students.

Abstract:
This paper describes basic facts regarding the black-white test score gap over the first four years of school. Black children enter school substantially behind their white counterparts in reading and math, but including a small number of covariates erases the gap. Over the first four years of school, however, blacks lose substantial ground relative to other races; averaging .10 standard deviations per school year. By the end of third grade there is a large Black-White test score gap that cannot be explained by observable characteristics. Blacks are falling behind in virtually all categories of skills tested, except the most basic. None of the explanations we examine, including systematic differences in school quality across races, convincingly explain the divergent academic trajectory of Black students.

Abstract:
This paper reports on the results of a prospective implementation of methods for detecting teacher cheating. In Spring 2002, over 100 Chicago Public Schools elementary classrooms were selected for retesting based on the cheating detection algorithm. Classrooms prospectively identified as likely cheaters experienced large test score declines. In contrast, classes that had large test score gains on the original test, but were prospectively identified as being unlikely to have cheated, maintained their original gains. Randomly selected classrooms also maintained their gains. The cheating detection tools were thus demonstrated to be effective in distinguishing between classrooms that achieved large test-score gains as a consequence of cheating versus those whose gains were the result of outstanding teaching. In addition, the data generated by the implementation experiment highlight numerous ways in which the original cheating detection methods can be improved in the future.

Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.

Abstract:
Although the theoretical literature on corruption is well developed, empirical work in this area has lagged because it has proven difficult to isolate corrupt behavior in the data. In this paper, we look for evidence of corruption in an unlikely place: the highest echelons of Japanese sumo wrestling. This paper provides strong statistical evidence documenting match rigging in sumo wrestling. A non-linearity in the incentive structure of promotion leads to gains from trade between wrestlers on the margin for achieving a winning record and their opponents. We show that wrestlers win a disproportionate share of the matches when they are on the margin. Increased effort can not explain the findings. Winning on the bubble is more frequent when the two competitors have met often in the past. Success on the bubble tends to rise over the course of a wrestler's career, but declines in his last year, consistent with the game theoretic predictions. Wrestlers who are victorious when on the bubble lose more frequently than would be expected the next time they meet that opponent, suggesting that part of the payment for throwing a match is future payment in kind. Systematic differences across wrestling stables suggest that the stables play a role in facilitating the corruption. In times of increased media scrutiny, the match rigging disappears.

Abstract:
In the 1960's, Blacks and Whites chose relatively similar first names for their children. Over a short period of time in the early 1970's, that pattern changed dramatically with most Blacks (particularly those living in racially isolated neighborhoods) adopting increasingly distinctive names, but a subset of Blacks actually moving toward more assimilating names. The patterns in the data appear most consistent with a model in which the rise of the Black Power movement influenced how Blacks perceived their identities. Among Blacks born in the last two decades, names provide a strong signal of socio-economic status, which was not previously the case. We find, however, no negative causal impact of having a distinctively Black name on life outcomes. Although that result is seemingly in conflict with previous audit studies involving resumes, we argue that the two sets of findings can be reconciled.

Abstract:
Previous studies of the impact of changes in prisoner populations on crime rates have failed to adequately control for the simultaneity between those two variables. While increases in the number of prisoners are likely to reduce crime, rising crime rates also translate into larger prison populations. To break that simultaneity, this paper uses the status of prison overcrowding litigation in a state as an instrument for changes in the prison population. Overcrowding litigation is demonstrated to have a negative impact on prison populations, but is unlikely to be related to fluctuations in the crime rate, except through its effect on prison populations. Instrumenting results in estimates of the elasticity of crime with respect to the number of prisoners that are two to three times greater than previous studies. The results are robust across all of the crime categories examined. For each one-prisoner reduction induced by prison overcrowding litigation, the total number of crimes committed increases by approximately 15 per year. The social benefit from eliminating those 15 crimes is approximately $45,000; the annual per prisoner costs of incarceration are roughly $30,000.

Abstract:
Previous studies of the impact of changes in prison populations on crime rates have failed to control adequately for the simultaneity between those two variables. While increases in the number of prisoners are likely to reduce crime rising crime rates also translate into larger prison populations. To break that simultaneity this paper uses the status of prison overcrowding litigation in a state as an instrument for changes in the prison population. Overcrowding litigation is shown to have a negative impact on prison populations but is unlikely to be related to fluctuations in the crime rate except through its effect on prison populations. This methodology results in estimates of the elasticity of crime with respect to the numbers of prisoners that are two to three times greater than those of previous studies. The results are robust across all of the crime categories I examine. For each prisoner reduction induced by overcrowding litigation the total number of crimes committed increases by approximately 15 per year. The social benefit from eliminating those 15 crimes is approximately $45000; the annual per-prisoner costs of incarceration are roughly $30000.

Abstract:
This study presents an overview of modern field experiments and their usage in economics. Our discussion focuses on three distinct periods of field experimentation that have influenced the economics literature. The first might well be thought of as the dawn of "field" experimentation: the work of Neyman and Fisher, who laid the experimental foundation in the 1920s and 1930s by conceptualizing randomization as an instrument to achieve identification via experimentation with agricultural plots. The second, the large-scale social experiments conducted by government agencies in the mid-twentieth century, moved the exploration from plots of land to groups of individuals. More recently, the nature and range of field experiments has expanded, with a diverse set of controlled experiments being completed outside of the typical laboratory environment. With this growth, the number and types of questions that can be explored using field experiments has grown tremendously. After discussing these three distinct phases, we speculate on the future of field experimental methods, a future that we envision including a strong collaborative effort with outside parties, most importantly private entities.

Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.

Abstract:
School choice has become an increasingly prominent strategy for urban school districts seeking to enhance academic achievement. Evaluating the impact of such programs is complicated by the fact that a highly select sample of students takes advantage of these programs. To overcome this difficulty, we exploit randomized lotteries that determine high school admission in the Chicago Public Schools. Surprisingly, we find little evidence that attending sought after programs provides any benefit on a wide variety of traditional academic measures, including standardized test scores, attendance rates, course-taking, and credit accumulation. This is true despite the fact that those students who win the lotteries attend better high schools along a number of dimensions, including higher peer achievement levels, higher peer graduation rates, and lower levels of poverty. We do, however, uncover evidence that attendance at such schools may improve a subset of non-traditional outcome measures, such as self-reported disciplinary incidences and arrest rates.

Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.

Abstract:
Current education reform proposals involve improving educational outcomes through forms of market-based competition and expanded parental choice. In this paper, we explore the impact of choice through open enrollment within the Chicago Public Schools (CPS). Roughly half of the students within CPS opt out of their assigned high school to attend other neighborhood schools or special career academies and magnet schools. Access to school choice dramatically increases student sorting by ability relative to neighborhood assignment. Students who opt out are more likely to graduate than observationally similar students who remain at their assigned schools. However, with the exception of those attending career academies, the gains appear to be largely spurious driven by the fact that more motivated students are disproportionately likely to opt out. Students with easy geographical access to a range of schools other than career academies (who presumably have a greater degree of school choice) are no more likely to graduate on average than students in more isolated areas. We find no evidence that this finding can be explained by negative spillovers to those who remain that mask gains to those who travel. Open enrollment apparently benefits those students who take advantage of having access to vocational programs without harming those who do not.

Abstract:
In previous research, a substantial gap in test scores between White and Black students persists, even after controlling for a wide range of observable characteristics. Using a newly available data set (Early Childhood Longitudinal Study), we demonstrate that in stark contrast to earlier studies, the Black-White test score gap among incoming kindergartners disappears when we control for a small number of covariates. Over the first two years of school, however, Blacks lose substantial ground relative to other races. There is suggestive evidence that differences in school quality may be an important part of the explanation. None of the other hypotheses we test to explain why Blacks are losing ground receive any empirical backing. The difference between our findings and previous research is consistent with real gains made by recent cohorts of Blacks, although other explanations are also possible.

Abstract:
A wide range of social indicators turned sharply negative for Blacks in the late 1980s and began to rebound roughly a decade later. We explore whether the rise and fall of crack cocaine can explain these patterns. Absent a direct measure of crack cocaine's prevalence, we construct an index based on a range of indirect proxies (cocaine arrests, cocaine-related emergency room visits, cocaine-induced drug deaths, crack mentions in newspapers, and DEA drug busts). The crack index we construct reproduces many of the spatial and temporal patterns described in ethnographic and popular accounts of the crack epidemic. We find that our measure of crack can explain much of the rise in Black youth homicides, as well as more moderate increases in a wide range of adverse birth outcomes for Blacks in the 1980s. Although our crack index remains high through the 1990s, the deleterious social impact of crack fades. One interpretation of this result is that changes over time in behavior, crack markets, and the crack using population mitigated the damaging impacts of crack. Our analysis suggests that the greatest social costs of crack have been associated with the prohibition-related violence, rather than drug use per se.

Abstract:
Previous empirical studies have typically uncovered little evidence that police reduce crime. One problem with those studies is a failure to adequately deal with the simultaneity between police and crime: while police may or may not reduce crime, there is little doubt that expenditures on police forces are an increasing function of the crime rate. In this study, the timing of mayoral and gubernatorial elections is used to identify the effect of police on crime. This paper first demonstrates that increases in the size of police forces disproportionately occur in mayoral and gubernatorial election years, a relationship that had previously gone undocumented. After controlling for changes in government spending on other social programs, there is little reason to think that elections will be otherwise correlated with crime, making elections ideal instruments. Using a panel of large U.S. cities from 1970-1992, police are shown to reduce crime for six of the seven crime categories examined. Each additional police officer is estimated to eliminate eight to ten serious crimes. Existing estimates of the costs of crime suggest that the social benefit of reduced crime is approximately $100,000 per officer per year, implying that the current number of police is below the optimal level.

Abstract:
Previous empirical studies have typically uncovered little evidence that police reduce crime. One problem with those studies is a failure to adequately deal with the simultaneity between police and crime: while police may or may not reduce crime, there is little doubt that expenditures on police forces are an increasing function of the crime rate. This paper attempts to use the timing of mayoral and gubernatorial elections to identify the effect of police on crime. It is first demonstrated that increases in the size of police forces are disproportionately concentrated in mayoral and gubernatorial election years. Having controlled for the state of the economy and other types of social spending, however, there is little reason to think that elections will otherwise be correlated with crime, making elections plausible instruments. Using a panel of large U.S. cities from 1970-1992, it appears that additional police reduce crime: the elasticity of violent crime with respect to sworn officers is found to be approximately -1.0; for property crimes the elasticity is -0.2. Because of the weakness of the instruments, however, the impact of police is imprecisely estimated. Consequently, the null hypothesis that the marginal social benefit of reduced crime equals the costs of hiring additional police cannot be rejected.

Abstract:
Over the last two decades the punitiveness of the juvenile justice system has declined" substantially relative to the adult courts. During that same time period juvenile violent crime" rates have grown almost twice as quickly as adult crime rates. This paper examines the degree to" which those two empirical observations are related, finding that changes in relative punishments" can account for 60 percent of the differential growth rates in juvenile and adult violent crime" between 1978 and 1993. Juvenile offenders appear to be at least as responsive to criminal" sanctions as adults. Moreover, sharp changes in criminal involvement with the transition from" the juvenile to the adult court suggest that deterrence, rather than simply incapacitation important role. There does not, however, appear to be a strong relationship between the" punitiveness of the juvenile justice system that a cohort faces and the extent of criminal" involvement for that cohort later in life.

Abstract:
Over the last two decades juvenile violent crime has grown almost twice as quickly as that of adults. This paper finds that changes in relative punishments can account for 60 percent of that differential. Juvenile offenders are at least as responsive to criminal sanctions as adults. Sharp drops in crime at the age of majority suggest that deterrence (and not merely incapacitation) plays an important role. There does not, however, appear to be a strong relationship between the punitiveness of the juvenile justice system that a cohort faces and the extent of criminal involvement for that cohort later in life.

Abstract:
A strong, negative empirical correlation exists between arrest rates and reported crime rates. While this relationship has often been interpreted as support for the deterrence hypothesis, it is equally consistent with incapacitation effects, and/or a spurious correlation that would be induced by measurement error in reported crime rates. This paper attempts to discriminate between deterrence, incapacitation, and measurement error as explanations for the empirical relationship between arrest rates and crime. Using a modified version of the techniques of Griliches and Hausman (1986) for dealing with measurement error in panel data, this paper first demonstrates that the presence of measurement error does not appear to explain the observed relationship between arrest rates and crime rates. To differentiate between deterrence and incapacitation, the impact of changes in the arrest rate for one crime on the rate of other crimes is examined. In contrast to the effect of increased arrests for one crime on the commission of that crime, where deterrence and incapacitation are indistinguishable, it is demonstrated that these two forces act in opposite directions when looking across crimes. Incapacitation suggests that an increase in the arrest rate for one crime will reduce all crime rates; deterrence predicts that an increase in the arrest rate for one crime will lead to a rise in other crimes as criminals substitute away from the first crime. Empirically, deterrence appears to be the more important factor, particularly for property crimes.

Abstract:
This paper attempts to discriminate between deterrence, incapacitation, and measurement error as explanations for the negative relationship between arrest rates and crime empirically. Measurement error cannot explain the observed patterns in the data. To differentiate between deterrence and incapacitation, the impact of changes in the arrest rate for one crime on the rate of other crimes is examined. Incapacitation suggests that an increase in the arrest rate for one crime will reduce all crime rates. Deterrence predicts that an increase in the arrest rate for one crime will lead to a rise in other crimes as criminals substitute away from the first crime. Empirically, deterrence appears to be the more important factor, particularly for property crimes.

Abstract:
This paper attempts to discriminate between deterrence, incapacitation, and measurement error as explanations for the negative relationship between arrest rates and crime empirically. Measurement error cannot explain the observed patterns in the data. To differentiate between deterrence and incapacitation, the impact of changes in the arrest rate for one crime on the rate of other crimes is examined. Incapacitation suggests that an increase in the arrest rate for one crime will reduce all crime rates. Deterrence predicts that an increase in the arrest rate for one crime will lead to a rise in other crimes as criminals substitute away from the first crime. Empirically, deterrence appears to be the more important factor, particularly for property crimes.

Abstract:
The number of prisoners incarcerated on drug-related offenses rose fifteen-fold between 1980 and 2000. This paper provides the first systematic empirical analysis of the implications of that dramatic shift in public policy. We show that the increase in drug prisoners led to reductions in expected time served for other crimes, especially for less serious offenses. Reductions in time served, however, increased other crimes by no more than a few percent. Moreover, incarcerating drug offenders is found to be almost as effective in reducing violent and property crime as locking up other types of offenders. We estimate that cocaine prices are 10-15 percent higher today as a consequence of increases in drug punishment since 1985. Based on previous estimates of the price elasticity of demand for cocaine, this implies a reduction in cocaine consumed of as much as 20 percent.

Abstract:
The Hawthorne effect, a concept familiar to all students of social science, has had a profound influence both on the direction and design of research over the past 75 years. The Hawthorne effect is named after a landmark set of studies conducted at the Hawthorne plant in the 1920s. The first and most influential of these studies is known as the Illumination Experiment. Both academics and popular writers commonly summarize the results as showing that every change in light, even those that made the room dimmer, had the effect of increasing productivity. The data from the illumination experiments, however, were never formally analyzed and were thought to have been destroyed. Our research has uncovered these data. We find that existing descriptions of supposedly remarkable data patterns prove to be entirely fictional. There are, however, hints of more subtle manifestations of a Hawthorne effect in the original data.

Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.

Abstract:
Profit maximizing behavior on the part of firms is a fundamental, but rarely tested, assumption of economics. In this paper, I analyze the decisions made by an MIT trained economist running a company that delivers bagels and donuts. The simplicity and transparency of the business (e.g. marginal cost is easily observed) allow for direct tests of profit maximization in the quantities delivered each day and the prices that are charged. Using thirteen years of data representing more than 80,000 deliveries, I find that the company is extremely adept at determining how many bagels and donuts to deliver to a particular customer on a given day. In stark contrast, the company appears to price on the inelastic portion of the demand curve for the entire period, thereby foregoing a substantial share of available profits. I argue that these results generalize well beyond this particular case study: firms are likely to be close to the efficient frontier on dimensions for which there is frequent and informative feedback regarding profits, but absent that feedback, systematic deviations from profit maximization are more likely.

Abstract:
Private expenditures on crime reduction have potentially important externalities. Observable measures such as barbed-wire fences and deadbolt locks may shift crime to those who are unprotected, imposing a negative externality. Unobservable precautions, on the other hand, may provide positive externalities since criminals cannot determine a priori who is protected. Focusing on one specific form of victim precaution, Lojack, we provide the first thorough empirical analysis of the magnitude of such externalities. Because installing Lojack does not reduce the likelihood that an individual car will be stolen, any decrease in the aggregate crime rates due to Lojack is an externality from the perspective of the individual Lojack purchaser. We find that the presence of Lojack is associated with a sharp fall in auto theft in central cities and a more modest decline in the remainder of the state. Rates of other crimes do not change appreciably. Our estimates suggest that, at least historically, the marginal social benefit of an additional unit of Lojack has been as much as 15 times greater than the marginal social cost in high crime areas. Those who install Lojack in their cars, however, obtain less than ten percent of the total social benefits of Lojack, causing Lojack to be undersupplied by the free market. Current insurance subsidies for the installation of Lojack appear to be well below the socially optimal level.

Abstract:
Private expenditures on crime reduction have potentially important externalities. Observable measures such as barbed-wire fences and deadbolt locks may shift crime to those who are unprotected, imposing a negative externality. Unobservable actions, on the other hand, provide positive externalities since criminals cannot determine a priori who is protected. Focusing on one specific form of victim precaution, Lojack, we provide the first thorough empirical analysis of the magnitude of such externalities. Because vehicles equipped with Lojack are not identifiable to criminals, any decrease in crime rates associated with Lojack is an externality from the perspective of the Lojack purchaser. We find that Lojack has large crime-reducing effects. Each one-percentage point increase in the Lojack installation rate in a market is associated with a 20 percent decline in auto theft rates in large cities and a 5 percent decline in the rest of the state. Rates of other crimes do not change appreciably. Our estimates suggest that the marginal social benefit of an additional unit of Lojack is 15 times greater than the marginal social cost. Those who install Lojack in their cars, however, obtain less than ten percent of the total social benefits of Lojack, causing Lojack to be dramatically undersupplied by the free market. Current insurance subsidies for the installation of Lojack appear to be well below the socially optimal level.

Abstract:
We develop an algorithm for detecting teacher cheating that combines information on unexpected test score fluctuations and suspicious patterns of answers for students in a classroom. Using data from the Chicago Public Schools, we estimate that serious cases of teacher or administrator cheating on standardized tests occur in a minimum of 4-5 percent of elementary school classrooms annually. Moreover, the observed frequency of cheating appears to respond strongly to relatively minor changes in incentives. Our results highlight the fact that incentive systems, especially those with bright line rules, often induce behavioral distortions such as cheating. Statistical analysis, however, may provide a means of detecting illicit acts, despite the best attempts of perpetrators to keep them clandestine.

Abstract:
Game theory makes strong predictions about how individuals should behave in two player, zero sum games. When players follow a mixed strategy, equilibrium payoffs should be equalized across actions, and choices should be serially uncorrelated. Laboratory experiments have generated large and systematic deviations from the minimax predictions. Data gleaned from real-world settings have been more consistent with minimax, but these latter studies have often been based on small samples with low power to reject. In this paper, we explore minimax play in two high stakes, real world settings that are data rich: choice of pitch type in Major League Baseball and whether to run or pass in the National Football League. We observe more than three million pitches in baseball and 125,000 play choices for football. We find systematic deviations from minimax play in both data sets. Pitchers appear to throw too many fastballs; football teams pass less than they should. In both sports, there is negative serial correlation in play calling. Back of the envelope calculations suggest that correcting these decision making errors could be worth as many as two additional victories a year to a Major League Baseball franchise, and more than a half win per season for a professional football team.

Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.

Abstract:
This paper demonstrates that rising crime rates in cities are correlated with city depopulation. Instrumental variables estimates, using measures of the certainty and severity of a state?s criminal justice system as instruments for city crime rates, imply that the direction of causality runs from crime to urban flight. Using annual city-level panel data, our estimates suggest that each additional reported crime is associated with a one person decline in city residents. There is some evidence that increases in suburban crime tend to keep people in cities, although the magnitude of this effect is small. Analysis of individual-level data from the 1980 census confirms the city-level results and demonstrates that almost all of the crime-related population decline is attributable to increased outmigration rather than a decrease in new arrivals to a city. Those households that leave the city because of crime are much more likely to remain within the SMSA than those leaving the city for other reasons. The migration decisions of high-income households and those with children are much more responsive to changes in crime than other households. Crime-related mobility imposes costs on those who choose to remain in the city through declining property values and a shrinking tax base.

Abstract:
This paper demonstrates that rising crime rates in cities are correlated with city de-population. Instrumental variables estimates, using measures of the certainty and severity of a state's criminal justice system as instruments for city crime rates, imply that the direction of causality runs from crime to urban flight. Using annual, city-level panel data, our estimates suggest that each additional reported crime is associated with a one person decline in city residents. There is some evidence that increases in suburban crime tend to keep people in cities, although the magnitude of this effect is small. Analysis of individual-level data from the 1980 census confirms the city-level results and demonstrates that almost all of the crime-related population decline is attributable to increased out-migration rather than a decrease in new arrivals to a city. Those households that leave the city because of crime are much more likely to remain within the SMSA than those leaving the city for other reasons. The migration decisions of high-income households and those with children are much more responsive to changes in crime than other households. Crime-related mobility imposes costs on those who choose to remain in the city through declining property values and a shrinking tax base.

Abstract:
Measurement of seat belt and air bag effectiveness is complicated by the fact that systematic data are collected only for crashes in which a fatality occurs. These data suffer from sample selection since seat belt and air bag usage influences survival rates which in turn determine whether a crash is included in the sample. Past researchers either ignored sample selection or adopted indirect estimation methods subject to other important biases. We propose a simple, but novel, solution to the selection problem: limiting the sample to crashes in which someone in a different vehicle dies. Under relatively weak conditions, consistent estimates can be obtained from this restricted sample. Empirically, we find seat belts to be more effective in saving lives than most previous estimates. Air bags, however, appear to be less effective than generally thought. If our coefficients can be generalized to all crashes, the cost per life saved with seat belts is approximately $30,000, compared to $1.6 million for air bags.

Abstract:
The market for sports gambling is structured very differently from the typical financial market. In sports betting, bookmakers announce a price, after which adjustments are small and infrequent. Bookmakers do not play the traditional role of market makers matching buyers and sellers but, rather, take large positions with respect to the outcome of game. Using a unique data set, I demonstrate that this peculiar price-setting mechanism allows bookmakers to achieve substantially higher profits. Bookmakers are more skilled at predicting the outcomes of games than bettors and systematically exploit bettor biases by choosing prices that deviate from the market clearing price.

Abstract:
On tests of intelligence, Blacks systematically score worse than Whites, whereas Asians frequently outperform Whites. Some have argued that genetic differences across races account for the gap. Using a newly available nationally representative data set that includes a test of mental function for children aged eight to twelve months, we find only minor racial differences in test outcomes (0.06 standard deviation units in the raw data) between Blacks and Whites that disappear with the inclusion of a limited set of controls. The only statistically significant racial difference is that Asian children score slightly worse than those of other races. To the extent that there are any genetically-driven racial differences in intelligence, these gaps must either emerge after the age of one, or operate along dimensions not captured by this early test of mental cognition.

Abstract:
In most settings, it is difficult to measure discrimination, and even more challenging to distinguish between competing theories of discrimination (taste-based versus information-based). Using contestant voting behavior on the television game show Weakest Link, one can in principle empirically address both of these questions. On the show, contestants answer questions and vote off other players, competing for a winner-take-all prize. In early rounds, strategic incentives encourage voting for the weakest competitors. In later rounds, the incentives reverse, and the strongest competitors become the logical target. Controlling for other observable characteristics including the number of correct answers thus far, both theories of discrimination predict that in early rounds, excess votes will be made against groups targeted for discrimination. In later rounds, however, taste-based models predict continued excess votes, whereas statistical discrimination predicts fewer votes against the target group. Empirically, I find some evidence of information-based discrimination towards Hispanics (i.e., other players perceive them as having low ability) and taste-based discrimination against older players (i.e., other players treat them with animus). There is little in the data to suggest discrimination against women and Blacks.

Abstract:
The Ku Klux Klan reached its heyday in the mid-1920s, claiming millions of members. In this paper, we analyze the 1920s Klan, those who joined it, and the social and political impact that it had. We utilize a wide range of newly discovered data sources including information from Klan membership roles, applications, robe-order forms, an internal audit of the Klan by Ernst and Ernst, and a census that the Klan conducted after an internal scandal. Combining these sources with data from the 1920 and 1930 U.S. Censuses, we find that individuals who joined the Klan were better educated and more likely to hold professional jobs than the typical American. Surprisingly, we find few tangible social or political impacts of the Klan. There is little evidence that the Klan had an effect on black or foreign born residential mobility, or on lynching patterns. Historians have argued that the Klan was successful in getting candidates they favored elected. Statistical analysis, however, suggests that any direct impact of the Klan was likely to be small. Furthermore, those who were elected had little discernible effect on legislation passed. Rather than a terrorist organization, the 1920s Klan is best described as a social organization built through a wildly successful pyramid scheme fueled by an army of highly-incentivized sales agents selling hatred, religious intolerance, and fraternity in a time and place where there was tremendous demand.

Abstract:
Over the last thirty years, the use of child safety seats in motor vehicles has increased dramatically, fueled by well publicized information campaigns and legal mandates. In spite of this movement, there is relatively little empirical evidence regarding the efficacy of child safety seats relative to the much cheaper alternative of traditional seat belts. Using data from the Fatality Analysis Reporting System (FARS) on all fatal crashes in the United States from 1975-2003, I find that child safety seats, in actual practice, are no better than seat belts at reducing fatalities among children aged 2-6. This result is robust to a wide range of sensitivity analyses, including controlling for sample selection that arises because the FARS data set includes only crashes in which at least one fatality occurs.

Abstract:
Donohue and Levitt (2001) argue that the legalization of abortion in the United States in the 1970s played an important role in explaining the observed decline in crime approximately two decades later. Foote and Goetz (2005) challenge the results presented in one of the tables in that original paper. In this reply, we regretfully acknowledge the omission of state-year interactions in the published version of that table, but show that their inclusion does not alter the qualitative results (or their statistical significance), although it does reduce the magnitude of the estimates. When one uses a more carefully constructed measure of abortion (e.g. one that takes into account cross-state mobility, or doing a better job of matching dates of birth to abortion exposure), however, the evidence in support of the abortion-crime hypothesis is as strong or stronger than suggested in our original work.

Abstract:
Young children are required to use child safety seats, and the age threshold at which children can legally graduate to seat belts has steadily increased. This paper tests the relative effectiveness of child safety seats, lap-and-shoulder seat belts, and lap belts in preventing injuries among motor vehicle passengers aged 2-6. We analyze three large, representative samples of crashes reported to police, as well as linked hospital data. We find no apparent difference in the two most serious injury categories for children in child safety seats versus lap-and-shoulder belts. Child safety seats provide a statistically significant 25% reduction in the least serious injury category. Lap belts are somewhat less effective than the two other types of restraints, but far superior to riding unrestrained.

Abstract:
While it is widely believed by academics, politicians, and the popular press that incumbent congressmen are rewarded by the electorate for bringing federal dollars to their district, the empirical evidence supporting that claim is extremely weak. One explanation for the failure to uncover the expected relationship between federal spending and election outcomes is that incumbents who expect to have difficulty being reelected are likely to exert greater effort in obtaining federal outlays. Since it is generally impossible to adequately measure this effort, the estimated impact of spending is biased downward due to an omitted variable bias. We address this estimation problem using instrumental variables. For each House district, we use spending outside the district but inside the state containing the district, as an instrument for spending in the district. Federal spending is affected by a large number of actors (e.g. governors, senators, mayors, and other House members in the state delegation), leading to positive correlations in federal spending across the House districts within states. However, federal spending outside of a district is unlikely to be strongly correlated with the strength of that district's electoral challenge. Thus, spending in other districts is a plausible instrument. In contrast to previous studies, we find strong evidence that non-transfer federal spending benefits congressional incumbents: an additional $100 per capita in such spending is worth as much as two percent of the popular vote. Additional transfer spending, on the other hand, does not appear to have any electoral effects.

Abstract:
While it is widely believed by academics, politicians and the popular press that incumbent members of Congress are rewarded by the electorate for bringing federal dollars to their district, the empirical evidence supporting that claim is extremely weak. One explanation for the failure to uncover the expected relationship between federal spending and election outcomes is that incumbents who expect to have difficulty being re-elected are likely to exert greater effort in obtaining federal outlays. Since it is generally impossible to adequately measure this effort, the estimated impact of spending is biased downward because of an omitted variable bias. We address this estimation problem using instrumental variables. For each House district, we use spending outside the district but inside the state containing the district as an instrument for spending in the district. Federal spending is affected by a large number of actors (e.g., governors, senators, mayors, and other House members in the state delegation), leading to positive correlations in federal spending across the House districts within states. However, federal spending outside of a district is unlikely to be strongly correlated with the strength of that district's electoral challenge. In contrast to previous studies, we find strong evidence that federal spending benefits congressional incumbents: an additional $100 per capita in spending is worth as much as 2 percent of the popular vote. The only category of federal spending that does not appear to yield electoral rewards is direct transfers to individuals.

Abstract:
This paper tests several theories of the effects of congressional representation on state economic growth. States that were represented by very senior Democratic congressmen grew more quickly during the 1953-1990 period than states that were represented by more junior congressional delegations. We find some, but weaker, evidence that states with a high fraction of their delegation on particularly influential committees also exhibit above-average growth. We also test partisan models of distributive politics by studying the relationship between a state's degree of political competition and its growth rate. Our findings support both nonpartisan and partisan models of congressional distributive politics. In spite of our findings with respect to economic growth, we can not detect any substantively important association between congressional delegation seniority, the degree of state political competition, and the geographic distribution of federal funds. The source of the growth relationships we identify therefore remains an open question.

Abstract:
We document and analyze the emergence of a substantial gender gap in mathematics in the early years of schooling using a large, recent, and nationally representative panel of children in the United States. There are no mean differences between boys and girls upon entry to school, but girls lose more than two-tenths of a standard deviation relative to boys over the first six years of school. The ground lost by girls relative to boys is roughly half as large as the black-white test score gap that appears over these same ages. We document the presence of this gender math gap across every strata of society. We explore a wide range of possible explanations in the U.S. data, including less investment by girls in math, low parental expectations, and biased tests, but find little support for any of these theories. Moving to cross-country comparisons, we find that earlier results linking the gender gap in math to measures of gender equality are sensitive to the inclusion of Muslim countries, where in spite of women's low status, there is little or no gender gap in math.

Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.

Abstract:
Over the past 40 years the fraction of mixed race black-white births has increased nearly nine-fold. There is little empirical evidence on how these children fare relative to their single-race counterparts. This paper describes basic facts about the plight of mixed race individuals during their adolescence and early adulthood. As one might expect, on a host of background and achievement characteristics, mixed race adolescents fall in between whites and blacks. When it comes to engaging in risky/anti-social adolescent behavior, however, mixed race adolescents are stark outliers compared to both blacks and whites. We argue that these behavioral patterns are most consistent with the "marginal man" hypothesis, which we formalize as a two-sector Roy model. Mixed race adolescents -- not having a natural peer group -- need to engage in more risky behaviors to be accepted. All other models we considered can explain neither why mixed race adolescents are outliers on risky behaviors nor why these behaviors are not strongly influenced by the racial composition at their school.

Abstract:
We present a methodology for measuring the risks posed by drinking drivers that relies solely on readily available data on fatal crashes. The key to our identification strategy is a hidden richness inherent in two-car crashes. Drivers with alcohol in their blood are seven times more likely to cause a fatal crash; legally drunk drivers pose a risk 13 times greater than sober drivers. The externality per mile driven by a drunk driver is at least 30 cents. At current enforcement rates the punishment per arrest for drunk driving that internalizes this externality would be equivalent to a fine of $8,000.

Abstract:
This paper provides a theoretical analysis of violence as a mechanism for allocating scarce resources in a nonmarket setting. We demonstrate that the efficiency with which resources are allocated in that context are strongly positively related to the predictability of fight outcomes. The lethality of the weapons used, in contrast, has an indeterminate impact on the costs of violence, except at very high or very low levels of lethality. Our results suggest that the observed link between guns and homicide rates may not be primarily attributable to the lethality of guns, but rather to the lack of ex ante predictability of the winner when guns are involved in a fight.

Abstract:
In standard models of optimal deterrence, which assume perfect information on the part of the social planner, fines dominate jail sentences as an instrument for punishing crime. In the real world, however, punishment using prisons is quite common. This paper provides an explanation for that empirical observation. If criminals either have private information about their own wealth, or have a substantial portion of their wealth in the form of human capital, the social planner cannot simply impose a fine on a criminal, who can always claim to have insufficient wealth to pay the fine. Rather, because of the need for incentive compatibility, it is as if the social planner offers the criminal a choice: either pay the fine or go to jail. That additional constraint on the social planner dramatically reduces the effectiveness of fines vis-a-vis a perfect information world. Mandatory sentencing guidelines exacerbate the problems associated with incentive compatibility by restricting a judge's ability to tailor penalties to individual criminals.