Presentation is loading. Please wait.

Presentation is loading. Please wait.

Understanding the Effects of Anti-Profiling Policies Paul Heaton RAND Criminology and Population Dynamics Workshop June 2007.

Similar presentations


Presentation on theme: "Understanding the Effects of Anti-Profiling Policies Paul Heaton RAND Criminology and Population Dynamics Workshop June 2007."— Presentation transcript:

1 Understanding the Effects of Anti-Profiling Policies Paul Heaton RAND Criminology and Population Dynamics Workshop June 2007

2 Motivation Nationwide movement to enact policies to prevent racial profiling by police Nationwide movement to enact policies to prevent racial profiling by police 26 states with legislation requiring data collection 26 states with legislation requiring data collection Jurisdictions in 21 other states with voluntary data collection Jurisdictions in 21 other states with voluntary data collection Several states, including California and Florida, with mandatory training programs Several states, including California and Florida, with mandatory training programs Use a racial profiling scandal in New Jersey in 1998-1999 to estimate the effect of changes in profiling policy on arrests, offenses, and other behavior Use a racial profiling scandal in New Jersey in 1998-1999 to estimate the effect of changes in profiling policy on arrests, offenses, and other behavior

3 Findings The scandal and policy reforms led to substantial (20-40%) reductions in arrests of Blacks relative to Whites for motor vehicle theft. The scandal and policy reforms led to substantial (20-40%) reductions in arrests of Blacks relative to Whites for motor vehicle theft. Changes in the number and distribution of offenses suggest an increase in vehicle thefts in response to the changing arrest patterns. Changes in the number and distribution of offenses suggest an increase in vehicle thefts in response to the changing arrest patterns. Findings robust to numerous specification checks; similar results are observable in Maryland. Findings robust to numerous specification checks; similar results are observable in Maryland.

4 Why Might Profiling Policy Affect Arrests and Offending? “Troopers are going to be more cautious and are probably much more selective…When it comes time to go further, I'm sure there has to be a conscious decision at some point. Unless it's something very obvious or blatant, [officers] may not go any further.” “Troopers are going to be more cautious and are probably much more selective…When it comes time to go further, I'm sure there has to be a conscious decision at some point. Unless it's something very obvious or blatant, [officers] may not go any further.” Ed Lennon, President of the State Troopers Fraternal Association of New Jersey, in a 1999 interview regarding the profiling controversy Ed Lennon, President of the State Troopers Fraternal Association of New Jersey, in a 1999 interview regarding the profiling controversy

5 Data Sources Agency-level UCR data on arrests and offenses, 1990-2003 Agency-level UCR data on arrests and offenses, 1990-2003 Focus on vehicle theft, closely linked to police stop behavior and well-reported with separate arrest and offense data Focus on vehicle theft, closely linked to police stop behavior and well-reported with separate arrest and offense data Place-level Census data Place-level Census data

6

7

8 Black pre-reform mean arrest rate:.0010 White pre-reform mean arrest rate:.00015 Pre-Reform Difference Black - White 0.0010 - 0.0001 0.00084

9 White post-reform mean arrest rate:.0001 Post-Reform Difference Black - White 0.0005 - 0.0001 0.00038 Black post-reform mean arrest rate:.0005

10 Post-Reform DifferencePre-Reform DifferenceDifference Black - White In 0.0005 - 0.00010.0010 - 0.0001Difference 0.000380.00084-0.00045 ≈ - 45%

11 Estimated Effects of the Scandal on Arrests Estimated % Change in Arrests For: Estimation Approach DDDDDDDDD Motor Vehicle Theft -.371*** -.371***-.195**-.264** (.054)(.093)(.115) No. Obs. 204080 First Difference: Post-PrePost-PrePost-Pre Second Difference: Black-WhiteBlack-WhiteBlack-White Third Difference: Affected Crime- Other Crime Affected Crime- Other Crime Fourth Difference: NJ-US

12 Did the Policy Change Affect Offending? Obstacle: No direct data on the race of offenders independent of arrest data. Obstacle: No direct data on the race of offenders independent of arrest data. Evidence from both the time series and the geographic distribution of offenses suggest increased offending at the time the new policies were enacted Evidence from both the time series and the geographic distribution of offenses suggest increased offending at the time the new policies were enacted

13 Trends in Vehicle Theft, 1994-2003

14

15 Trends in Burglaries and Larcenies, 1994-2003

16

17

18

19

20 Interpretation of Offending Changes Locality at 75 th percentile of Black population distribution would have 17% more vehicle theft than one at 25 th percentile Locality at 75 th percentile of Black population distribution would have 17% more vehicle theft than one at 25 th percentile Policy would result in approximately 3000 additional vehicle thefts annually in NJ Policy would result in approximately 3000 additional vehicle thefts annually in NJ Does not suggest minorities have dramatically different responsiveness to enforcement Does not suggest minorities have dramatically different responsiveness to enforcement

21 Robustness Checks Similar results changing sample, time period, comparison groups Similar results changing sample, time period, comparison groups Placebo tests—policy effects not observed for unaffected crimes, groups, or locations Placebo tests—policy effects not observed for unaffected crimes, groups, or locations Replication using Maryland’s 1995 reforms Replication using Maryland’s 1995 reforms

22 Conclusions Contributions of paper Contributions of paper 1. Adds to profiling literature by estimating policy effects 2. Patterns in arrest/offending data are difficult to reconcile with causal explanations other than changes in profiling behavior by police 3. Provides minority-specific measure of responsiveness to enforcement, an important quantity in evaluating profiling policies


Download ppt "Understanding the Effects of Anti-Profiling Policies Paul Heaton RAND Criminology and Population Dynamics Workshop June 2007."

Similar presentations


Ads by Google