Well, at least I was able to write a program in Matlab which automatically loads data from yahoo and then calculates month HV using close/close. I also created dummy variables for each month and ran it using a simple ols for XOM. The only month which was signifcant at even the 90% level was June. It was negative. I did the same thing in HAL and the same month was negative and was signficant at the same level. I'll have to do some more research on this and use a heteroskadistic model.