|
Jennifer L. Doleac
Job Market Candidate
Stanford University
Department of Economics
579 Serra Mall
Stanford, CA 94305
[email protected]
|
|
|
Primary Field: Applied Microeconomics (Public Finance and Labor Economics)
Secondary Fields: Law and Economics, Economics of Crime, Experimental Economics
Expected Graduation Date:
June, 2012
|
|
Dissertation Committee:
Caroline Hoxby (Primary)
[email protected]
B. Douglas Bernheim
[email protected]
Ran Abramitzky
[email protected]
|
Research
Working Papers
The Effects of DNA Databases on Crime (Job Market Paper)
Since 1988, every US state has established a database of criminal offenders' DNA profiles. In advocating them, law enforcement officials argued that the databases would (i) allow them to identify suspects and rule out innocent people in cases where DNA was part of the crime scene, and (ii) deter criminals whose DNA profile was collected from reoffending. DNA databases are distinctive among law enforcement tools. First, they feature enormous economies of scale: although setting up a database is expensive, profiling an additional person or conducting an additional search is very cheap. Second, they are hypothesized to work mainly by increasing the probability that a criminal is punished, rather than increasing the severity of the punishment. Unlike enforcement tools such as longer sentences, the effectiveness of databases should not be substantially reduced if offenders have high discount rates. This paper provides the first evidence on the impact of DNA databases on crime, addressing both the partial equilibrium effects of DNA profiling on individuals' subsequent criminal behavior and the general equilibrium impacts of profiling on crime rates and arrest probabilities. To identify partial equilibrium effects, I exploit the fact that if two criminals with virtually the same history are released one day apart, one may be subject to DNA profiling and the other not. These discontinuities in treatment occur every time a state adds a type of offender to its database. The paper shows that profiled violent offenders are more likely to return to prison than similar, unprofiled offenders -- suggesting that the higher probability of getting caught outweighs the deterrent effect of DNA profiling -- but the magnitude of this effect varies with age and criminal history. To identify general equilibrium effects, I exploit variation in the timing of database expansions across states. The paper shows that larger DNA databases reduce crime rates, especially in categories where forensic evidence is likely to be collected at the scene -- for example, murder, rape, assault, and vehicle theft. Back-of-the-envelope estimates of the marginal cost of preventing each crime suggest that DNA databases are much more cost-effective than other common law enforcement tools.
The Visible Hand: Race and Online Market Outcomes (with Luke Stein) - Under Review
Do prospective customers behave differently based on sellers' perceptible race or signals about sellers' socioeconomic class? Does the answer to this question depend on whether a customer lives in an area that is racially segregated or plagued by property crime? We investigate these questions in a year-long experiment in which we sold iPods through local online classified advertisements throughout the United States. Each ad features a photograph of the product being held by a hand that is dark-skinned ("black"), light-skinned ("white"), or light-skinned with a wrist tattoo (traditionally associated with lower social class). We find that black sellers do worse than white sellers on a variety of metrics: they receive 13% fewer responses to their advertisements and receive 17% fewer offers. These effects are similar in magnitude to those associated with the display of a wrist tattoo. Conditional on receiving at least one offer, black sellers receive offers that are lower by 2 to 4%, despite the self-selected--and presumably less biased--pool of buyers. In addition, buyers corresponding with a black seller behave in ways that suggest they trust the seller less: they are 17% less likely to include their name in e-mails, 44% less likely to accept delivery by mail, and 56% more likely to express concern about making a long-distance payment. We find that black sellers suffer particularly poor outcomes in thin markets; it appears that discrimination may not "survive" in the presence of significant competition among buyers. Furthermore, black sellers do worst in markets that are racially segregated and that have high property crime rates. The latter result suggests that at least part of the explanation is statistical discrimination--that is, buyers being concerned about the time and potential danger involved in the transaction, or that the iPod is stolen goods.
Media coverage includes the NY Times Freakonomics blog, AOL News, Stanford Report, Wall Street Journal Real Time Economics blog, USA Today, Financial Times, The Young Turks podcast and a radio interview on Crosscurrents (KALW).
Works in Progress
California's Ten-Year Rule: Spousal Support Laws and Divorce Outcomes
In 1988, the California legislature passed what became known as the Ten-Year Rule, which defines a "long marriage" as one lasting ten years or longer. In a divorce, California courts are instructed to make the higher-earning spouse provide income support to the lower-earning spouse for a "reasonable time" that allows the lower-earning spouse to become financially independent. The default "reasonable time" for a marriage that is not long is one-half the length of the marriage. Courts may deviate from this default in exceptional circumstances, but they rarely do. In contrast, when the marriage has been long, courts have considerable upward discretion in deciding what a reasonable time is, and they may provide lifetime income support to the lower-earning spouse. Media and website advisors on divorce indicate that the Ten-Year Rule is well known among Californians considering divorce. By introducing a discontinuity in the relationship between the duration of a marriage and the expected duration of spousal support, higher-earning spouses are given an incentive to separate before the ten-year mark. Lower-earning spouses are given the opposite incentive. To investigate how spouses responded to these incentives, I gathered and coded a large, random sample of divorce cases filed between 1970 and 2000 in one California county. The preliminary evidence (preliminary since data are still being coded) suggests that couples do not noticeably change the timing of their actual separations, but do alter the reported separation date as part of the property settlement. Lower-earning spouses in ten-year marriages have greater bargaining power than those in shorter marriages, and exchange a backdated separation for a larger share of the property upon divorce.
The case of the Monday jury: A natural experiment in jury pool composition (with Andrew Felton)
Americans generally consider trial by a "jury of one's peers" a Constitutional right, and local jurisdictions go to great lengths to ensure a representative jury pool, but it is unclear how consequential the representativeness of a jury is. Does the racial and socioeconomic composition of a jury pool affect trial outcomes? We address this question using a natural experiment in Washington, DC, in the early 2000s. For several years, the DC Court system allowed residents to defer jury duty to a future date of their choice. Individuals who deferred were more likely to be employed, white, and of a higher socioeconomic status than those who did not, and they disproportionately chose to serve on Mondays (apparently for work-related reasons). The result was that cases that happened to go to jury selection on Mondays drew jurors from a substantially whiter and more affluent jury pool than those whose cases were randomly scheduled for later in the week. This policy was changed in 2006 to ensure a random distribution of deferring jurors. Unlike normal variation in jury composition, which is driven by lawyers' selection of jurors sympathetic to their clients, this "day of the week" variation is not a function of case or defendant characteristics. Using administrative court records, we estimate the effect of a jury pool's racial and socioeconomic composition on case outcomes, by comparing "treated" Monday cases with "control" cases that selected juries (i) on other days of the week, or (ii) on Mondays after 2006. We also test for heterogeneous effects by case characteristics such as the type of alleged criminal offense.
Making the case: Does DNA profiling increase the speed and certainty of convicting someone for each crime?
DNA databases are designed to quickly and accurately identify suspects in criminal cases. The value of this law enforcement tool depends largely on (i) its increasing the speed with which suspects are arrested and (ii) its raising the probability that a suspect, once arrested, is convicted. These effects could occur through multiple channels: quickly identifying the real offender, quickly ruling out innocent suspects, converting a higher percentage of arrests into indictments (which suggesting that the evidence is less ambiguous), indicted individuals being more likely to plea bargain, and individuals who go to trial being more likely to be found guilty. I investigate these channels using a remarkable dataset that includes all arrests in Washington State between 1990 and 2011. The data allow me to follow cases from their origin to their ultimate resolution. My empirical strategy exploits the fact that different types of offenders were added to the database at different times, a process that creates arbitrary differences among crimes in the usefulness of DNA. For instance, an innocent person who is suspected in one crime may quickly be ruled out because the DNA of a DNA-profiled career criminal is found at the scene. An innocent person in a nearly identical case may be arrested because DNA found at the scene does not match anyone in the database--even if the real perpetrator is a career criminal, so long as that career criminal happens not to be have been profiled because of database timing.
|
|