Question1)
Suppose you are interested in assessing the value of the statistical life for individuals. You find a dataset on risk and wages. You consider running following OLS regression:
wagesi = || B0 + || B1 f atal_riski + ui .
Fatal_risk ranges from 0 to 100 in this example.
a.) Theoretically, do we expect || B1 to be positive or negative?
b.) Suppose || B1 = 25, 000 (with a standard error of 5,000). What is the 95 % confidence interval for || B1 ?
c.) Compute VSL by using || B1 and an implied compensation associated with the changing fatality risk from 0 to 100.
d.) Recalculate VSL using the ends of confidence interval you calculated in part the b. What is confidence interval for the implied VSL from the above regression model?
e.)What is the potential pitful of using || B1 to calculated the essential compensation to go from a 0 to 100 percent risk of fatality (hint, how much variation in fatality risk are we likely to observe in data)?
question2)
a.) Based on the Ruhm’s paper on recessions and health, what are the age groups where mortality has the strongest percentage response to the increase in an unemployment rate?
b.) What are the Ruhm’s conclusions about potential mechanisms which could explain the negative relationship between aggregate unemployment and mortality? What do you think another mechanism might be and what data would you use test your hypothesis?