1. Suppose you are interested in whether the death penalty serves as a deterrent to murder. You have

data for each of the 50 states and DC for 1990 and 1993 and have the following population model in

mind: mm'rteﬁ = ﬁg + ﬁgexecﬁ + ﬁgunemﬁ + 60D93+ a,- + 31,-; where mrdrre is murders per 100,000 population, exec is total executions over the past 3 years, unem is

the annual unemployment rate and D93 is equal to 1 for 1993. a) [5 points] Suppose you have run an OLS regression to estimate it. Are the OLS standard errors correct? What other concerns, if any, might you have about these OLS estimates? Explain your

answers. b) [5 points] You decide to estimate the model using first differences ((2 implies change), and obtain the

following: Source I 55 df MS Number of obs = 51 + F( 2, 48) = 2.96 Model | 6.88790282 2 3.44395141 Prob > F = 0.0614 Residual | 55.8?24805 48 1.16401001 R—squared = 0.1097 + Adj R—squared = 0.072? Total | 62.?603834 50 1.25520T6T Root MSE = 1.0T89 cmrdrte I Coef. Std. Err. t P>ltl [95% Conf. Interval]

+ cexec | —.1038396 .0434139 —2.39 0.021 —.1911292 —.01655 cunem | —.0665914 .1586859 —0.42 0.677 —.3856509 .252468 _con5 I .4132665 .2093848 1.9? 0.054 —.00?7299 .8342628 Interpret the coefficient on cares. Based on this model, does the death penalty seem to serve as a

deterrent to murder? Are there any problems with this interpretation? Explain. c) [5 points] Suppose you estimated a ﬁxed effect model. How would the estimated coefficient on exec compare to that on cexec above? Can you predict what the coefﬁcient on D93 would be in this new

regression? If so, what is your prediction? If not, why not? Explain.