Q1

Suppose a micro-finance institution wants to evaluate the impact of taking a loan (D, ) on an mdivldual' s income one year later (Yil), uith a sample of 200. Individuals are randomly assigned to a treatment group treatment group (Z = 1) is offered a loan, but not evelyone who is offered a loan pgtY411y_ es it. Offered Loan Not Offered Loan Took Loan 80 10 Did Not Take Loan (D = 0) 20 90 Average income pre-experiment (EIY,nl) S975 Sl,025 p-value of difference 0.25 Average income_ one year later S3,ooo Sl,ooo 0.01 In an earlier problem set, we considered the regressions Di = ao + alZi + i and Yil = YIZi + z.'i (where i and are the error terms) Now consider the regression Yit = + 1POSTr + 2Zi + 3POSTr Zi + Ei_ Yo + (a) Can we test the hypothesis that 132 = O? If so, what is the p-value? Why might we be Interested in testing this hypothesis? (b) Compute 133 _ How does it compare to your answers to pans (b) and (c) from the earlier lem set?