2. Suppose you are interested in the effect of police on crime. You download dat

Question

2. Suppose you are interested in the effect of police on crime. You download data for an 1.1.d. sample of n cities, {Pi, ci), where ci s the crime rate in city i (total number of crimes per 1000 people) and p is the total number of police in city i. Now, suppose we run the regression: (a) Suppose we wish to adopt the causal interpretation of A. (For instance, suppose 01. What does-.01 mean if we you run OLS on your sample and find - adopt the causal interpretation.) (b) Do you think EUPI = 0 here?

Explanation / Answer

a) if we find beta_hat - -0.01. It can be interpreted as the Increase in number of police by 1%(Since change in log is interpreted as percent change) will cause drop in crime rate by 0.01.

b) E[U|P] is not zero since there are many unobservable characteristics of the cities in the U which effects crime rate and are correlated with number of police. Eg, Low level of average education in the city will lead to more crimes and will also effect how many people enter the police force.

Related Questions

Navigate

• Browse (All)
• Browse 2
• Subjects

---

---

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.