Investors spend a lot of time trying to determine a company's future earnings. For stable companies in mature industries, the best clue to future earnings is often the earnings record of the past. But past earnings records are full of line items that obscure a company's true earnings power. As such, investors are often left wondering what components of past earnings are sustainable, and which elements are non-recurring or "one-time" in nature. Penman and Zhang sought a systematic way to identify sustainable earnings, and the results suggest they have succeeded.
In the paper they published subsequent to their findings, Penman and Zhang describe the process for determining sustainable earnings. Using only financial statement data, the process eliminates the components of earnings that are suspect (e.g. through reductions in deferred revenue or allowance for bad debt accounts).
Companies are then given an S-score, with high S-scores suggesting strong earnings sustainability. The authors then track the portfolio of high and low S-score companies to see how the earnings look several years out. High S-score companies generated higher returns on assets and strongly outperformed low S-score companies in portfolio returns as well.
Further information on the methodology for identifying sustainable earnings is detailed in the paper, which is available as a free download here.