Assessing the Jensen's Alpha significance bw portfolios


I have 10 yr daily data of stock prices for 2 portfolios (all US equities) for the years 2000 - 2010.

I know that the Jensen’s Alpha is a measure of portfolio performance and the portfolio with the greatest Alpha is the better portfolio (in term of return relative to risk).

But is there a need to test for the significance of the difference between the alphas, i.e, to know that the difference is not due to chance, or its enough just to see which Alpha is greater?

Also, I am a bit confused about the risk free rate that need to be used, becasue every year it can change.

Thank you

It depends a bit on the context, but yes, if you are comparing Jensen Alphas between portfolios, you should ask if the difference is statistically significant before concluding that one portfolio manager has been smarter than the other.

In practice, most people don’t do it, in part because the calculation is a bit complex, and probably because people would find that all but extremely large differences are not statistically significant at standard levels, because most track records are likely to be too short.

In fact, people really should be asking if Jensen Alphas (or APT equivalents) are statistically significant in the first place (i.e. different from 0), but I don’t see a lot of that going on, either.