Find the data for one of the AER papers that Alwyn Young analyses in his paper that has an instrumental variables regression without clustering, with a single instrument and a single endogenous regressor. Submit the sample used for your regression as part of the problem set, and report which paper you selected.
Replicate one of the instrumental variables regressions in the paper.
Construct a 95% confidence interval based on the regular bootstrap and the pivotal bootstrap based on the t-statistic. Compare the p-values for all three standard errors.
Do the same for the two reduced form regressions.
For the two reduced form regressions calculate the leverage for each observation.
Do a small simulation study to assess the credibility of the results reported in the paper. Describe the choices you made in the simulations (the distributions of the variables chosen, the parameter values, the sample sizes, etc). Discuss the results from the simulations. Keep the simulation study small.
Discuss whether you find the results as reported in the paper for this regression credible,
AER paper + Young Paper
and what results you recommend should be reported instead of, or in addition to, the results currently reported.
