Question 1: Is the Pareto the only distribution that is likely to be
asked in calculating the variance of the parameters for the method of
moments?
Question 2:
If Pareto
lamda = 500,000
alpha unknown
I believe you need to know that Var[g(y)] = (dg/dy)^2Var(y)
How do you get from that formula to
Var (estimated alpha) =
{(-1/500,000)*(estimated alpha -1)^2}^2Var(observed mean)