4 . Hill estimates that fracking increases the prevalence of low birthweight babies by about 25%. Assuming we believe these estimates, that low birthweight babies typically incur costs of $500,000, and that fracking generates $5 million in GDP, the government should do nothing to regulate fracking and reduce this negative externality. . Suppose that you can prove that the cost of all angioplasty performed in the United States last year totaled $200 billion but that the benets of this same treatment (in terms of longer lives) totaled $150 billion .If we wish to maximize net benets ,this benet -cost ratio suggests that we should stop performing angioplasty procedures . Suppose that it costs a rm $5,000 per employee to provide health insurance (and suppose that the implicit tax subsidy for benefits is repealed ). It must be preferable that the rm provide health insurance rather than paying employees an additional $5,000 in cash because the firm can purchase health care for a lower cost per employee than they would be able to purchase on their own. . Abouk 85 Adams (2013 ) look at how accident rates change when a state bans texting while driving . Their estimation strategy can be found in the paper , or in Lecture 13 slides on Bcourses . The results can be found in Column 2 of Table 3 in the paper (also in the slides ). Here you'll use the dataset used in that paper to replicate their results . (a) As a rst step, run a xed effects model with no additional controls . As we'll discuss in section ,xed effects are another way to think about difference in difference regressions To generate one dummy variable for each state and for each time period, use the following commands in Stata : e tabulate t, gen (t_) o tabulate st, gen (st _) To include all of these dummies in your regression, you can simply include 15\" and st} ll . ll tune as variables. Think of these as our typical and "location variables in a diffs - indiffs regression, then add in treatment variables for both strong and weak bans to estimate the difference in difference regression . Include [aweight=lpop] at the end of your regression to weight states by population. Report and interpret your coefficients on the two treatment variables. (b) Now run the same regression , but include all of the controls that Abouk and Adams used : the percentage of the population that's male , log of population , unemployment , and gas tax. Report and interpret your coefficients on the two treatment variables (Note, these should match Column (2) of Table 3 in the Abouk and Adams paperl). Have they changed much from your first regression