Please answer part d e f only

2. This question is about the paper "Labor Market Returns to Two- and Four-Year Colleges", NBER Working Paper 4268 by Kane and Rouse (1993), who investigate the difference in wages between individuals who went to a two-year college versus those who went to a four-year college. The relationship of interest is: Yi = Bo + educiB1 + raceiB2 + experiB3 + partiB4 + Wi, where Yi are log hourly wages for individual i, educi is a binary variable that equals 1 if individual i went to a four-year college, and equals 0 if they went to a two year college (everybody in the sample went to one of the two), race; is individual i's race, exper; is the individual's number of years with work experience, and parti is an indicator variable that is 1 when individual i's job is part-time. a) Why do we care about the causal effect of educi on Yi, as opposed to the correlation between educ; and Yi? b) What is in u;? Give examples specific to the context of this question. c) Explain why E(uileduci) # 0. The authors use an instrumental variable approach. As an instrument for educi, they use distance, the "distance to the nearest four-year college minus distance to the nearest two-year college for individual i when individual i was school-age" d) Why is this instrument relevant? e) Defend the validity of this instrument, i.e. convince me that E(uildistance;) = 0. f) Attack the validity of this instrument! Explain why E(uildistancei) # 0