Security Analysis by Ben Graham and David Dodd is a must read for anyone serious about value investing.
The last several chapters dealt with understanding and correcting past years' financial statements in order to properly reflect a company's true earnings. Now the authors discuss what that past corrected earnings record is worth going forward. This is the most difficult task for the security analyst, because predictions of the future are never reliable. However, armed with an accurate earnings record of the past, the analyst's attempts are not completely futile.
Wall Street is entirely too concerned with current earnings. The authors stress the importance of using average earnings from several years in the past, to avoid misrepresentations due to market cycle fluctuations that play a large role in an individual year's earnings. They also stress the importance of stability, concluding that average earnings with low standard deviations are much more reliable indicators than earnings which jump up and down. Furthermore, projecting trends in earnings should be avoided. Competition and the law of diminishing returns ensure that trends cannot continue indefinitely; here, a qualitative dimension must be employed to help determine the earnings power of the future, which includes the company's industry and its position within it.
When using qualitative judgment, the analyst must be careful not to use intuition, but should instead form his basis in fact. Graham and Dodd take readers through various case analyses, (including one of a depressed stock price of Mack truck) as examples of applying qualitative measures to the quantitative past. Depending on circumstances, qualitative reasons may even be used to disregard past earnings records. Various examples given include the expiring or the signing of contracts, mine depletions, expected changes in the prices of products (e.g. due to oversupply). All of these would render the conservative analyst wary of investment since predictability is too low.