Question

3. Does putting criminals in jail reduce crime? Suppose you regressed crime rates (crimes per100,000 people) on incarceration rates (prisoners per 100,000 people) and some control variables for economic conditions, demographics, etc. a) Briefly explain why there may be simultaneous causality bias. b) Levitt (1996) suggests that lawsuits aimed at reducing prison overcrowding is a valid IV. Briefly explain why the rank/relevance and order/exogeneity condition should hold for this variable. c) Suppose instead he used number of lawyers per capita as an IV. Briefly explain why the rank/relevance and order/exogeneity condition should hold for this variable.

Fig: 1

Fig: 2

Fig: 3

Fig: 4