Friday, April 30, 2010

The problems with beta

In a recent note to clients, BofA analyst Savita Subrmanian made a case for a shift in stock market leadership from high-beta to low-beta stocks, based on extensive research into prior "beta rallies."

From CNBC:
"Over the last 25 years, there have been three other beta rallies like what we have seen recently," wrote Savita Subramanian, a Bank of America Merrill Lynch quantitative strategist, in a note to clients. "After each of these periods, value strategies consistently took the lead in the subsequent leg of market performance."
"The most pronounced tilt of quant funds was toward statistically inexpensive companies," wrote Subramanian, who analyzed a sample of U.S. Large Cap Quantitative Equity Funds for their holdings. "A tilt toward valuation and away from beta is currently appropriate, in our view."
I don't disagree with her analysis. I wouldn't be surprised to see cheaper, steadier, large cap value stocks outperform in the months ahead. The problem though, is what these large cap quantitative funds will own from shifting from "high" to "low" beta stocks, because beta changes. (A stock's beta is basically how much it moves in relation to the market. A stock with a beta of 1.2 will go up 1.2% for each 1% move up in the S&P 500. A stock with a beta of .8 will go up .8% on a 1% jump in the S&P500, and it works the same on the downside.) Beta changes because all it is is a backward looking measure of how a stock moves relative to the market.

I looked at two examples, Microsoft and Priceline. Back in March07, Microsoft had a beta of .99 and Priceline's was 3.15. Had you been a quantitative hedge fund and loaded up on high-beta stocks, you'd pick PCLN. Fast forward to today, and BofA's Subramanian suggest rotating out of high beta stocks and into low beta stocks. The only problem of course, is that PCLN now has a beta of .89, even lower than stodgy Microsoft! That's simply a function of its recent rise tracking the market closer than the rise in the year prior to March07:

A much larger problem with beta's is their impact on the equity cost of capital, in relation to Discounted Cash Flow Analysis. The Equity Cost of Capital is calculated as

(Beta X Equity Risk Premium)+(Risk-Free 10yr Treasury) = (.045 X 3.15)+.045 = 18.7%

So. The enormous changes in PCLN's beta cause enormous changes in the perceived value of 2011 Free Cash Flow. You did all this work to estimate that FCF, and it's now wrecked b/c of stock price movements. Even stupider: the less debt a company has, the more its stock price action affects the present value of future cash flows. It's even worse when discounting cash flows that are further away:

The Capital Asset Pricing Model is fatally flawed by its dependence on beta, and the "equity-risk" premium that boosts cost of capital by 4.5% for no other reason than it's not debt. You can spend all day telling Microsoft that their cost of capital is 8%, but in the end they'll laugh at you and go borrow in the debt markets for ten years at 4.2%...

Copyright 2010 TheAlphaNinja

1 comment: