11. You are interested in finding out how getting a college degree (D) affects e
ID: 3320034 • Letter: 1
Question
11. You are interested in finding out how getting a college degree (D) affects earnings (Y) a. Is E(YD = 11-EIYID = 0] a good estimate of the causal effect? Explain. b. Name the three conditions that would be needed for a variable Z to be Briefly explain each in the able to serve as a good instrument for D. context of this question. c. Suppose you found a variable that meets the conditions outlined in part (b). There are two ways to calculate the effect of D of Y using IV. Using the variable D, Y, and Z: i. Write down the reduced from and first stage equations needed to construct the IV estimate -/ ii. Write down the first and second stage equations needed to calculate the IV estimate using 2SLS. 12. Card, Dobkin, and Maestas 2009 paper ("Does Medicare Save Lives?") used a regression discontinuity design (RD) to evaluate whether health insurance (treatment) affects healthcare utilization and overall health. At age 65, all legal residents in the US. are eligible for Medicare, subsidized health insurance for the elderly. This allows for the comparison of health outcomes of people on either side of this cut-off to potentially get a casual effect of health insurance on these outcomes. In this problem we will focus on healthcare utilization as the outcome which is measured by a dummy indicating whether the individual went to the doctor in the past two weeks (md 2wks), age in quarters is the running variable (age_qtr), treatment is a dummy indicating if the individual has health insurance (insured), and the instrument is a dummy indicating if the individual is 65 years (age65). Since the instrument is determined from age in quarters it means that age65 is 1 if age qtr>259, and 0 if age atrs259. All bold words are names of variables. We will be using a fuzzy RD to estimate the causal relationship between health insurance and health utilization. Explain why we are using a fuzzy RD method instead of the sharp? a. Continued on net pageExplanation / Answer
Answer for 12. C
the coeeficient for insured is 0.1016255
ans also the P-value is 0.189(>0.05) and hence it accepts the null hypothesis of "there is no significant relationship with this data set or not identified"
Related Questions
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.