(b) The authors concluded that “ on the basis of the statistical data (...) when the nonsmokers and the total high smoking classes of patients with lung cancer are compared with patients who have other diseases, we can reject the null hypothesis that smoking has no effect on the induction of cancer of the lungs ” (p. 333).

Do you think that the paper provides a definite proof that smoking causes lung cancer?

You can discuss any aspect of the study, especially the empirical strategy. Your answer

should relate (and be guided by) the potential outcome framework. (15 marks)

(c)  If you were a policymaker, and taking into consideration your knowledge of applied microeconomics, would you take action against smoking based on the evidence      presented in the paper? Moreover, would you be more interested in the average

treatment effect on the treated (ATT) or on the untreated (ATU) and why? (5 marks)

(d) Which modifications of the research design would you suggest to the authors of the     paper to enhance the chances that the estimated parameters reflect a causal estimate?

(10 marks)

(e)  Discuss what would be your ideal study design that is feasible in the context of studying the connection between smoking and lung cancer. (10 marks)

2. A recent paper by Goldin, Lurie and McCubbin, “Health Insurance and Mortality: Experimental Evidence from Taxpayer Outreach”  published in December 2019, looks at the effect of providing health insurance – one of the most important questions in Health           Economics. You do NOT need to read the paper to answer this question.

The authors sent informational letters to 3.9 million households randomly selected among 4.5

million who had failed to obtain mandatory health insurance under the direction of the

Affordable Care Act.

This is also known as an “encouragement design” where some randomly-selected individuals

are incentivised to take part on some programme.  Note that this is different than

randomising the programme itself. More specifically, the health outcome (yi ) is explained by  the endogenous variable (the take-up of insurance zi , where zi  = l is i has insurance, and     zero otherwise), and instrumented by the random provision of the informational letter (:i  = l if i got the letter, and zero otherwise).

Write the reduced-form regression, the endogenous and the instrumental variable

specifications, and briefly explain what you can learn from each.  Define the Local

Average Treatment effect (LATE) and provide an interpretation for that estimate. How would the interpretation change if an email had been sent, rather than a letter? (7 marks)

Figures 1 and 2 below show the effect on mortality over time and by age bracket.

Describe the effects of the policy, and interpret them in light of your answer to item (a).

Compare to the results you know from what you learned from Oregon and the RAND

Health insurance experiments.  In particular, describe why conclusions might be different

across studies.(18 marks)

Figure 1: Mortality Effect Over Time