P/B is predictive in the aggregate, but possesses qualities that are unattractive to me, for example, preferring highly leveraged balance sheets. LSV (among others) examined P/B in theirContrarian Investment, Extrapolation, and Risk paper, and found that it was predictive (Of course this finding needs to be tempered by O’Shaugnessy’s observation that the smaller deciles are uninvestable because they are too small). For me, it’s a matter of finding a metric that avoids P/B’s unattractive qualities. I have some other research on another metric that produces returns in the same magnitude as P/B, but avoids the highly leveraged balance sheets (it actually prefers a unleveraged, cash-rich balance sheet). I would still not, however, use it in isolation. LSV note in the same paper discussed above that using another metric (cash flow-to-price (C/P), earnings-to-price (E/P), and 5-year average growth rate of sales (NYSE:GS) alongside P/B improves its returns. I think that’s the obvious solution to avoid the temptation to swing from assets to earnings etc because you’d be using it all the time.
And below is my reply.
Yes, Indeed.Of course a better way is to know where we are and use the “right” formula accordingly. But I guess few have this capability.Then a couple of ideas popped up while I was typing.
- Statistically bear market is 1/3 in length of bull market. So maybe we use both, but assign a higher weight to income and a lower weight to asset.
- Focus on income, and use asset as a cut off screener. This will offer some downside protection during bear market. Actually I have a SeekingAlpha article about this “Cloud Computing: Design a Portfolio for the Best, Normal and Worst”.
- Hedge. A simple one is to hedge with a short position of SPY when SPY is below 200 MA. Or rather stay out of market when SPY is below 200 MA.The list can be expanded further. I think the key is that we don’t want a formula to handle all problems for us.