In my previous article on regression to the mean I touched upon the reason for why we should not simply let our intuition guide us in making investment decisions (or any decision, for that matter). Rather, it is best to have some sort of formal decision-making process.
Why is it so much better to use a formula than to simply think holistically? Is not our thinking better? Can we not adapt our thinking to any extenuating circumstances if we think holistically? Consider a stock with a low P/E ratio but a high short ratio. It might be a good idea to avoid the stock, right? Quite simply, no. The P/E ratio is probably ten times more important than the short ratio in determining whether to buy or sell a stock. Sure, the short ratio is correlated with stock market returns, but a lot less so than is the P/E ratio.
When making intuitive judgments it is easy to be led astray by such salient but marginally important information. It is true that our feelings and intuition can be highly useful, such as in our estimation of the quality of a company’s management. However, a formula can incorporate that subjective, holistic information as well. There is no reason why we cannot use intangible information in a formalized investment decision process. We could have a part of the formula where you enter in a ‘1’ if the management is good, ‘0’ if mediocre, or ‘-1’ if bad. In the case of a company with an ongoing proxy fight, we could run the computation twice, with different inputs depending upon who wins. Comparing the two possible outcomes to other investment opportunities would give us a means of deciding whether the possible benefits outweighed the risks. This can prevent emotions and unimportant information from leading us to the wrong decision.
For a further discussion of the problems we humans face in making decisions when faced with complicated and conflicting information, I suggest reading Psychological Study of Human Judgment: Implications for Investment Decision Making by Paul Slovic (1972). (Unfortunately, no free full-text version of this paper is available online.)