(a) Explain why the linear probability model is inadequate as a specification for limited dependent variable estimation. (b) Assuming that you are regression a dummy variable y (i.e. y only takes to values 1 and 0) on an explanatory variable x. Explain why the regression y = a + bx + u (where u is the error term, and a and b are coefficients to be estimated) is equavalent to regressing prob(y = 1) on x, i.e. prob(y = 1) = a + bx. Notice that since u is the error term, it has zero mean