To Serve and Collect

My latest paper (co-authored with Michael Makowsky and Thomas Stratmann) is To Serve and Collect: The Fiscal and Racial Determinants of Law Enforcement (forthcoming in the Journal of Legal Studies):

We exploit local deficits and state-level differences in police revenue retention from civil asset forfeitures to estimate how incentives to raise revenue influence policing. In a national sample, we find that local fine and forfeiture revenue increases at a faster rate with drug arrests than arrests for violent crimes. Revenues also increase at a faster rate with black and Hispanic drug arrests than white drug arrests. Concomitant with higher rates of revenue generation, we find that black and Hispanic drug, DUI, and prostitution arrests, and associated property seizures, increase with local deficits when institutions allow officials to more easily retain revenues from forfeited property. White arrests are broadly insensitive to these institutions, save for smaller increases in prostitution arrests and property seizures. Our results show how revenue-driven law enforcement can distort police behavior.

We find that drug arrests, especially of blacks and Hispanics, generate revenues so police have the motive and opportunity to engage in revenue driven policing. What about the means? Arrests for murders or robbery are limited by the number of murders and robberies. Drug arrests, however, are more of a police choice variable, able to be ramped up or down almost at will. Thus, in addition to motive and opportunity, police also have the means for revenue driven law enforcement.

How can we test for this effect? In some states, police get to keep the revenues they collect from forfeitures but these states are not randomly assigned. Thus, we use deficits which are plausibly randomly assigned (relative to our variables of concern) and we identify off of the interaction of the two i.e. the marginal impact of additional budget deficits in states where seizure revenue is retained.

We find that black and Hispanic arrests for drugs, DUI, and prostitution arrests are all increasing with deficits in states where seizure revenues are legally retained while white arrests are broadly insensitive to deficits.

Our identification strategy is somewhat coarse so we are by no means the final word on this issue but surely it is time to forbid police departments from keeping the revenues they collect.

We conclude:

The prospects for justice are dimmed when the probability an individual is arrested varies not only by the character of their transgression but also by the potential windfall they present to the public coffer.

Mike Makowsky has an excellent tweetstorm going into further details.