•  
• Tweet
•  
•        Print
•  
•         Track Keywords
•  
• 

Bryan Goh (Bryan.Goh@DrKW), who recently joined Dresdner Kleinwort Wasserstein in London from Singapore based Oaks family office, wrote to Opalesque because he wanted to "raise a couple of points" on "this thing about alpha and beta", which is "all very muddled". Bryan referred to last week's article in the NY Times "Hey, You Have a Problem Paying Alpha Fees and Getting Beta Returns? (see Opalesque archive)"

"First of all some definitions. To define alpha (a) and beta (b), you need a reference framework. This is entirely cast in the linear regression. If you do not have a linear relationship, you do not have well defined betas or an alpha.

This is the simple regression model where y are the returns of the fund, x the returns of the market (haven't said which market) and e is an error term which on average is zero.

y = a + b.x +e

The garden variety fund presentation will have you believe that if we call x the returns of the S&P500 or the MSCI World Equity Index, then you have a good specification and the alpha and beta are consistent. You can do this even if the fund invests in credit or in fixed income, in multiple spins of the roulette wheel or in a lottery ticket pool. Because of misspecification, the beta will be low. The worse the misspecification the lower the beta, and the more significant the alpha.

What about this specification?

y = a + b1.x1 + b2.x2 + b3.x3 + b4.x4 + e

Here could be the proper s......................

To view our full article Click here