Price = Square Root (22.5 x Book Value x EPS)For

**Book Value**, I use Tangible Book Value per share. There are so many odd intangibles (goodwill, customer relations, patents, licenses, contractual rights, etc.) that U.S. companies are now forced to record under purchase accounting that a pure book value could easily skew the calculation. Since Tangible Book Value is usually less than Book Value, the resulting price is generally more conservative. Thus, if a stock is ever selling below its Graham Number, my quantitative analysis will automatically give it a star for being fair valued.

For

**EPS**, I am currently using a trailing 12-months. It has been suggested by some that I should be using an average trailing three years. This would help normalize any unusual swings in earnings, both positive and negative. My inclination would be to use the lower of the two, which would provide a more conservative result.

The

**22.5**is a little more involved. Graham believed that the price-to-earnings ratio (P/E) should be no more than 15 and the price-to-book value ratio (P/B) should be no more than 1.5. He did note that a P/E ratio below 15 could possibly justify a higher P/B ratio. Thus, he used as a general rule of thumb that the product of the two should not be more than 22.5 (15 x 1.5). That is how the 22.5 is derived (15 x 1.5). The 15 P/E is a result of Graham wanting his portfolio to have a yield equal yield to that of a AA bond. Over long periods of time, AA bonds have yielded 7.5%, on average. The inverse of this yield is 1 divided by 7.5%, or 13.3 (rounded up to 15) became the target P/E ratio for his portfolio.

Like all calculations of fair value, the Graham Number should be considered with all other available information and not used in isolation when making a buy or sell decision.

__Related Articles:__