Book to price is a bad value factor. It is a decent stock selection factor overall, but relative to the other ways of measuring value (earnings to price, cash flow to price, EBITDA/EV, etc) it is sub par.

I've been interested lately in very concentrated value portfolios, and found it interesting that when running very small portfolios (as little as 1 stock selected per month, with an annual holding period) based on book/price, you wouldn't have done very well (and would have lost money in the one stock versions!)

Here are the absolute and excess returns for concentrated book/price portfolios since 1963. It is not until you get to 25 stocks that the return of the concentrated portfolio edges out the market's average rate of return (11.45% in this period). I also ran concentrated portfolios of the most expensive stocks by book/price and was very interested to find that for the 5 and 10 stock versions, the expensive portfolios outperformed the cheap ones (barely, but still)! Once you get to larger baskets of stocks, then things normalize and cheap beats expensive.

Let's consider the 5-stock version as an example (which underperforms the equal-weighted market by nearly 5% per year). Below are the rolling 3-year excess returns for that portfolio: consistently bad with occasional periods of brilliance.

These stocks tend to be very different from other value stocks. The 5 cheapest stocks at the end of 2014, for example, failed to crack even the cheapest 1/3 of all stocks as measured by a composite value score (e/p, fcf/ev, ebitda/ev, s/p, shareholder yield). Four of the five have negative earnings, so look very expensive by other measures.

Unlike with other value factors, these results suggest that you do NOT want to own a very small basket of stocks that are only cheap based on this mediocre measure of value. Stick to other value factors instead.