Alpha vs Beta in fund management

[Directional]

Alpha is a measure of a fund's risk-adjusted return. Alpha can be used to directly measure the value added or subtracted by a fund’s manager. It is calculated by measuring the difference between a fund's actual returns and its expected performance given its level of market risk as measured by beta.

Presumably, the more attractive a fund, the higher the alpha?

How would one accurately calculate the beta of a particular strategy, and from that extrapolate the alpha?

 

[Mathemagician]

Beta is, by definition, relative to a benchmark. I'm not sure that the industry's attempt to apply these concepts to strategies other than long equities or long/short equities within a well-defined universe are entirely useful.

jj

 

[Mathemagician]

FWIW, I believe the calculation is as follows:

Beta = Covariance(Returns, Index) / Variance(Index)

... but I could be wrong.

jj

 

[Jack o'Clubs]

Arb, as Mathemagician suggests its all down to the benchmark you choose to measure performance against. Basically actual return less beta = alpha. In reality there are enough indices to have a stab at most performance, and if you're looking at managed funds, a number of places 'do the math' for you (eg Trustnet) - although whether you trust their calculations or not is another matter. If you go to the Trustnet site and look at index fund performance, it's interesting that the Beta is rarely 1...

 

[Directional]

thanks chaps - most enlightening as always!

 

[MrGecko]

FYI the Alpha of some funds is actually -ve.

after the deduction of fees (usually pretty massive), the return on the funds is less than beta.

on fees....

Fund Manager: "these are the costs associtated with the very best research and getting exceptional talent"

Customer: "F**k that; I could have bought index futures"

there was an article in the economist a few weeks back.

 

[Jack o'Clubs]

FYI the Alpha of some funds is actually -ve.

after the deduction of fees (usually pretty massive), the return on the funds is less than beta.

on fees....

Fund Manager: "these are the costs associtated with the very best research and getting exceptional talent"

Customer: "F**k that; I could have bought index futures"

there was an article in the economist a few weeks back.

That Economist special report on the industry is well worth reading, although whether embittered journos bridling at the unimaginable wealth of the PM's and hedge-fund guys has anything to do with it, I couldn't possibly comment!

 

[TWI]

To be honest this is really just great terminology to use in investor meetings. Whether a fund shows true beta or not is a subject of pure speculation.

 

[Mathemagician]

Oh, I would never argue that the benchmark has to be equity based. Just that it makes more sense to compare a long-only equities manager to the buy and hold performance of his universe rather than a stat arb guy (or any other absolute return strategy) to, um, ... Of course, the fact that it doesn't make a whole lot of sense won't prevent people (even very smart people) from doing it.

jj

 

[grantx]

Arbitrageur,

“given its level of market risk as measured by beta”. Beta doesn’t measure risk as far as I’m aware.

Beta usually expresses the degree of correlation between, eg an index and basket or selection of underlying shares. As stated, Alpha is a measure of over/under-performance by the manager. If an index returns, say, 10% and the fund returns 15%, the manager would seem to have added value. However, if the fund’s gearing is x2, the return represents an under-performance:

15% (fund return) – (10% (index return) x 2 (gearing)) = -5%.

The degree of gearing is the crucial figure here.

Grant.