6) Consider the general regression model y - B, + Pix, + B2x2 + B3x3 + u. a) (4 points) How you would test the null hypothesis that B, - B2 - 283 using a t-test? In your answer, be clear about the t-statistic you will use, how you would get the standard error, and how you would decide whether to accept or reject the null hypothesis. b) (4 points) Assume the errors are homoskedastic, and you are interested in testing the null hypothesis that B3 0. Evaluate the following statement "the logic of the F-test is that if you impose this restriction on your model and the R2 does not change much, then the null hypothesis is probably false and you should reject it." c) (4 points) For the OLS estimate of B, to be unbiased, you can make one of two assumptions: 1) E [u(x1, X2, X3] = 0 (mean independence of the error), or 2) E[ulx1, X2, X3] = E[u|x2, X3] (conditional mean independence of the error). Explain the key differences between these two assumptions