Untitled Document

The Naked P/E

For every company there is a P/E you would expect it to have, given the overall market P/E, the company's size and the sector it's operating in. Then you have the P/E it's actually got, and the two are never the same. The difference? The naked P/E [link to academic paper] is that part of the P/E that is unexplained by the other factors.

With the details I calculated to decompose the P/E, this naked P/E is now exposed for the first time. It turns out that this version of the P/E is excellent for tracking down companies at the extreme end of the value spectrum. It seems that at any one time there are a few companies whose earnings really have been mispriced by the market.

This method only works for small concentrated portfolios of ten shares or less. In fact a six-share portfolio of extreme naked P/E shares would have given average returns of 39% from 1975-2003 (mid-mid, rebalanced annually). Of course, with only six shares the returns are quite variable, but the portfolio lost money in only two years. The portfolio more than doubled in three of the years, and rose by 50% to 100% in seven other years. Many investors might feel that this is the sort of variability that they can live with.

For private investors who would like to try this out themselves, a table of the six most extreme P/E shares can be found here.