Can you profit from relative valuation?

(MoneyWatch) Sector or asset class rotation is a strategy many active managers employ in an effort to enhance performance. As a strategy, it certainly has appeal -- assuming you can identify the best-performing sector/asset classes in advance. It would seem that a logical way to do that would be to rely on relative valuations -- when a sector/asset class looks cheap relative to others, rotate into it.

Martin Fridson and John D. Finnerty's paper, "Is There Value in Valuation," studied whether a valuation-based rotation strategy worked for high-yield bonds. They first showed that a rotation strategy seemed to have great appeal. For example, for the period 2007-2011, while the mean returns for BB-, B-, and CCC-rated bonds were 9.9 percent, 8.3 percent and 15.2 percent, respectively, choosing the best-performing sector would have produced a return of 21.0 percent. Clearly, there's a large opportunity to add value.

• How chasing yield affects returns
• Know the data before buying high-yield bonds
• Does a high-dividend strategy help or hurt returns?

Before getting into their findings, it's important to understand that you should generally expect that the lowest-rated bonds (CCC) will outperform when the risk premium (the incremental yield over Treasuries) for high-yield bonds decreases and will underperform when it increases. This should certainly be true if the changes in risk premiums are large. (If the risk premium widens by a small amount, it's possible that the higher yield of the lower-rated bond would provide enough of a cushion for it to still outperform.) This is all intuitive. If a risk premium falls, the riskier asset class should outperform, and vice versa. If this is true, then we don't need information about relative valuations to inform us about which sector/asset class will outperform. We need only to forecast whether the premium will widen or not. (The evidence on active management demonstrates that managers don't have much success at that endeavor.) Studying the quarterly results for the period 1997-2011, this is exactly what Fridson and Finnerty found.

Using what is called the option-adjusted spread (OAS) -- to take into account the call risk inherent in corporate bonds -- they found that this relationship held true 76 percent of the time for B and BB bonds, 83 percent of the time for B and CCC bonds and 86 percent of the time for BB and CCC bonds. That finding calls into question a valuation-based trading strategy -- to know which sector to go to, you need only accurately predict whether the premium for the entire asset class will rise or fall. If it's going to narrow, own the lowest-rated bonds. If it's going to widen, own the highest-rated bonds.

To test whether a valuation-based approach was likely to add value, they built a five-factor (credit availability, capacity utilization, industrial production, speculative grade default rate and yield to maturity on the five-year Treasury) valuation model. If the OAS spread was greater than the model predicted, the asset was cheap (you should rotate into it). The asset would be expensive (you should rotate out of it) if the spread was less than fair value.

The model explained a high percentage, between 74 and 80 percent, of the variance in performance. Unfortunately, using various tests, the authors found no statistically significant benefit from a sector rotation strategy based on the valuation model. It's also important to note that high-yield bonds are an expensive asset class to trade. Thus, even if there appeared to be some theoretical advantage to a rotation strategy, that advantage might be ephemeral, with trading costs creating a large hurdle to overcome.

The bottom line is that while valuation-based strategies have intuitive appeal, the evidence suggests that they don't truly provide opportunities to generate excess returns.