I remember a discuss with one boss who asked for the beta of an investment in our portfolio. He told me to look it up as if there was a book of betas. I told him we would have calculate it. There was no single number that is used for beta. I still remember the odd look of frustration he gave me; you have got to be kidding me.
The same frustration can be applied with risk factors. There are Fama-French factors that are "standardized", but there are also several proxies and alternative ways of calculating these factors. Some are easier than others, so investors must ask about the specifics for calculating any factor.
Size may be the easiest to measure since it focused on market capitalization. Momentum also is an easy factor to measure with the lookback and delay being the key features. History has proved that the momentum look-back is stable. Volatility is another factor that is intuitively easy to measure and only has a limited number of choice variables. Investors can decide on standard deviation, beta, or risk after accounting for beta. Dividend yield can be as easy as a sort of high dividend; however, there are variations on risk-adjustments.
Complexities increase on when or how rebalancing is conducted for these factors. This structuring can add significant transaction costs.
The more difficult factors to measure are value and quality. While originally these were simple measurements, the level of complexity or sophistication with calculating value and quality through accounting number proxies has increased significantly.
For value, price to book and earnings to price are the standard measures, but there are also value measures with sales, earnings, net profits, dividends, and cash flows. There has been value measurement proliferation especially during these periods of value under-performance.
Quality cannot be measure directly, so there has been a growing number of alternative measures including ROA, ROE, earnings stability, growth, strength of balance sheet, accruals, and cash flows. There is a lot of room for quality interpretations.
So, like the old question of asking about beta, remember that all factor measurements are not the same.
No comments:
Post a Comment