I am having trouble understanding the rationale of the following related to the MGF:
Function Mx(t)=E[exp(tX)], the expectation exists for all t in a neighbourhood of zero, and X has mean mx, show that (1)logMx(t)/t>=mx for t>0 (2) the above inequality becomes equality as t tends to 0.
Here is what I have so far: For (1), I applied Jensen's inequality, so that LogE[(exp(tX)] >= E[log(exp(tX)) = E[tX] = tE[X] = tmx. So the first problem was solved.
I think I stuck here, so I have two questions: (1) was I on the right path to solve the first problem? (2) how should I proceed to solve the second problem?
Thank you very much.