Externalities and Crime

Transcription

1 Externalities and Crime Marco Gonzalez-Navarro University of Toronto Workshop Internacional de Avaliação de Impacto de Políticas Públicas March 20, 2013

2 What is property crime? Person A takes person B s property without B s consent (home burglary, auto theft, etc.) This is not only a simple transfer of goods. Often theft involves violence: Very high human life cost Psychological damage

3 What is property crime? The high costs of crime lead to victim precautions affecting: How cities develop (i.e. gated communities, highrises) How homes are built (walls and fences around every home, electrification) Where people spend their time (shopping malls vs public spaces)

4 What are externalities? In a mutually beneficial transaction, person A pays B for a good. Both parties are happy to trade. But sometimes a third person is affected. For example: A factory sells cloth to people. Business is great. However, while producing, the factory dumps pollution into the river. The fishermen that use the river are affected by the transaction, because their fish are dead! This pollution is an externality. An externality is a harm imposed on a third party when A and B transact

5 Externalities in Crime Crime victim precautions can impose lots of externalities on others! Example: Steering wheel locks used to prevent auto theft (The Club) Imagine 2 cars, one has a Club, the other does not. When one person purchases the Club, this increases the theft risk of the other vehicle. This is an externality!

6 So what? The social value of spending to divert crime is low! We care about reducing aggregate crime, not displacing it.

7 So what? When crime is displaced, this affects estimates of program impact Example: Lets say we study the effect of UPPs on crime. A naïve design compares neighborhoods with UPP to others without. But if opening an UPP in the treatment neighborhood leads to criminals engaging in crime in non-upp neighborhoods, the estimate will be too large! Once you treat the whole city, the effects may be much smaller than they are initially

8 Is this really important? Gonzalez-Navarro (2013, AEJ: Applied) documents substantial geographical externalities in auto theft Lojack (stolen vehicle recovery service) In Mexico was sold in certain models of all new Ford vehicles in some states, but not others Thieves knew which cars had Lojack and which did not

9 Geographical externalities So they stopped stealing Lojack equipped vehicles (50% reduction) But surrounding places, where Lojack was not available, experienced large increases in theft risk In the case of Mexico, many cars are stolen for their parts, if it became difficult to steal cars in one place, they simply stole them in another Auto theft is easily displaced!

10

11

12 What could have been done? The problem was that thieves could tell which cars were protected and which were not If no decals were allowed, and sold to all cars, then Lojacks would deter theft of all cars, not just those protected

13 What could have been done? This is exactly how Lojack was sold in the US and Ayres and Levitt (1998, QJE) find large reductions in auto theft, both for protected and unprotected vehicles! The same device created a positive externality! In the US the problem is too few Lojacks were installed

14 In general what can be done in the presence of externalities? Treat the whole unit (so that it covers potential displacement areas) Example: In Germany, a large percentage of motorcycle theft was for joyriding (just for fun, by amateurs, motorcycle abandoned after a few hours) When obligatory helmet laws were introduced, motorcycle theft fell substantially: You must be using a helmet not to be stopped by police and joyrider thieves were not walking around with helmets

15 Conclusion Consider potential externalities when doing Impact Evaluation studies (positive and negative) Thinking of these is challenging, but can lead to better public policy design, and better impact evaluation studies