Use murder.csv (or MURDER.dta) to estimate the effect of capital punishment on crime rates. This data set is panel data at the state level, from 1987, 1990 and 1993.mrdrte_s,t is the murder rate in statesin yeart.execs,t is the total number of executions in the past three years, andunems,t is the unemployment rate in statesin yeart.

1. (15pts) Estimate a pooled regression using OLS (i.e. using all three years of data, but don't explicitly account for the panel structure at the state level). Regress the murder rate on the number of executions and controlling for the unemployment rate and year dummies.

Recall that in this pooled specification,s,t contains unobserved time invariant state level characteristicsas and unobserved time-variant characteristics,vs,t (wheres,t=as+vs,t

Interpret the coefficient estimates onexec,unem, and the year dummies.

1. (10pts) Explain how this pooled regression may suffer from omitted variables bias, and posit on the direction of the bias for the coefficient estimates onexec andunem.
2. (15pts) Now account for the panel structure of the data, and estimate the fixed effects regression. That is, you estimate:

and use the fixed effect model to explicitly difference out the state fixed effect,as term (in R, use thefelm command as discussed in class). What types of omitted variables does the fixed effect model account for, and provide an explanation for why the coefficient estimate onexec changed in the way it did.

1. (10pts) Find the state with the largest number of executions in 1993. How much bigger is this than the next highest state?
2. (10pts) Estimate the fixed effect regression dropping Texas from the analysis. What happens to the estimated deterrent effect? What might explain the change in the estimated effect?

here is the link of the data:https://drive.google.com/file/d/1ewxBvByrjjgldARJ4gjgv7vHINW-361l/view

and use R to solve the problem also write down the r code.