A news article in the Edmonton Journal on December 2014 stated that, “The province is looking at cutting 70 of the Alberta Law Enforcement Response Team’s 280 positions to save money.” The cost-cutting proposal is surely related to the Alberta government’s fiscal crunch due to falling oil prices. This suggests that large swings in oil prices might serve as a statistical instrument (or a natural experiment) for expansions or reductions in law enforcement, which could be used to study the effect of policing on crime deterrence.
a. Explain the basic idea behind using a statistical instrument (or “instrumental variable”) for addressing the simultaneity problem, when trying to establish causal effects in economics.
b. Why would swings in oil prices actually NOT be a good instrument to use for determining whether changes in policing have a deterrent effect on criminals? To answer this question, consider the possible sources of simultaneity bias or confounding effects that changes in oil prices may have on the observed relationship between the quantity of policing and the crime rate.