This is a follow-up on my previous post, where I dealt with the latest Tetlock-Gardner’s book, Superforecasting: The Art and Science of Prediction. As I mentioned there, there are several valuable lessons to learn. Although the best way to hone our forecasting abilities is obviously getting some practice in prediction tournaments, it would be wrong to think that this is the only field where we will be able to apply our new knowledge. As long as we try to answer very specific questions with a reasonable time spam, we keep record of our initial forecasts and we commit to update our beliefs on a regular basis, then this routine qualifies as a useful exercise to improve our forecasting abilities. And it seems that investing in equities meets these requirements.
In particular, I realised on further reflection that there are many similarities between the way superforecasters make their predictions and the way value investors select their investments. So I was wondering whether superforecasters can give us some insights for value investors or not. And indeed they can. Although the comparison may seem a bit awkward at first, we should remember that the purpose of the book is how we can improve the way we make our forecasts, and in this regard value investors have to make forecasts as everyone else. Therefore, although the most succesful value investors already incorporate many of the procedures regularly used by superforecasters, it is worth attempting to make a systematic comparison in order to see what parts of the value investing approach can benefit from superforecasters’ abilities.
Although this is certainly innovative, we should remember that value investors have been increasingly using for the last two decades insights coming from the field of behavioural finance – and, in general, from the field of decision-making. For several years, value investors were cleansed from the economic dogma in academic circles, because according to the efficient markets hypothesis nobody can consistently beat the market. Since the 1990s, however, value investors have found an ally in academic circles: the behavioural guys. Behavioural finance has helped to fill the gap between common observation and economic theory. Therefore, the lessons superforecasters can teach us in decision-making should thus be viewed in this light: complementing and providing additional evidence to what value investors have been doing for several years now.
Superforecasters vs. value investors
In my previous post I highlighted some of the main features displayed by the superforecasters of the Good Judgement Project. Without much effort, we can easily tweak those features for value investors as follows:
- Choose the right questions. This is the superforecaster version of the famous value investing saying “don’t operate beyond your circle of competence”. Try to focus on sectors you can understand, and try to focus on companies whose structure is easy to handle. Superforecasters mostly focus on questions whose time span is less than a year (roughly beyond that, they converge to chimpanzee performance), whereas value investors focus on companies that are expected to have mean-reverting properties. Furthermore, if you are an individual investor with limited time or analytical resources (or limited willingness to analyse big corporations), then focus on companies with simple balance-sheets and with low (or non-existent) levels of leverage; in this way you will get a reasonable “margin of safety” in case things go unexpectedly wrong, and it will leave you a lot of time to understand the underlying business of the company – i.e. not wasting time in debt covenants or in debt repayment schedules.
- Use granular judgements (probabilities), not “black or white”. Many value investors spend a great deal of time trying to figure out what the “normalised” profit of a company should be. Benjamin Graham, for instance, recommended that if, for whatever reason, one does not feel very comfortable at adjusting the current net income figures of a company, then one should take into account the history of the company and to use the average of the net income figures of the last (say) 10 years – approach popularised later by Shiller at the macro level with his CAPE ratio. Because the “normalised” profit is the cornerstone for any serious valuation, it is paramount to understand the underlying business of the company and try to come up with sensible figures, not ball-park (black or white) figures that don’t reflect the true earnings power of the company.
- Keep score. Although in the investment industry now it is very popular (alas, too popular) to provide performance measures for almost every financial indicator on earth, these performance measures obscure the fact that successful stories can also be the product of luck. Learn not only from your mistakes, but also from your successes.
- Fermi-izing. Be honest with yourself and be clear about what you simply don’t know. Fermi-ise the big questions, try to understand the small ones and be clear about where your uncertainty lays. Because you are narrowing the extent of your uncertainty, you will be better prepare to include a suitable margin of safety. For example, playing with a too large margin of safety (because you did not make the effort in thinking where your uncertainty resides) will lead you to miss very good investment opportunities.
- Update continuously. This should be obvious, but there are several ways to improve the investing process. One of them is what the fund manager Bruce Berkowitz called killing a company: “We look at companies, count the cash, and try to kill the company. . . . We spend a lot of time thinking about what could go wrong with a company—whether it’s a recession, stagflation, zooming interest rates or a dirty bomb going off. We try every which way to kill our best ideas”. This is a clever way to update your beliefs.
- Practice and try it. After all, although the value investing literature can actually be quite exciting (well, if you like these things), one must admit that you can get a good sense of what it is about in a relatively short lapse of time. In other words, learning the basic theory is not very time-consuming. As the leading Spanish value investor Francisco García Paramés admits in this video (in Spanish), just do it. But we should add that as in the case of superforecasters, normal practice just does not qualify: the practice that matters is the one carried out with a consistent method and with a strong learning/feedback process.
And my commandments for value investing
Most of the features I listed above could be regarded as “behavioural features”, in the sense that are related with the general philosophy of the approach, but not with the investments tools needed to make an investment decision. It is surprising to check the lack of explicit discussions about the valuation process in value investing. OK, you have Buffett’s very thoughtful digressions on the role of book value and future earnings power in the valuation process, and some similar thoughts by other leading value investors, but beyond that, it seems that the actual analytical process is not spelled out in detail. This may reflect Graham’s conviction that one should be aware about valuation using future cash-flows, because:
The concept of future prospects and particularly of continued growth in the future invites the application of formulas out of higher mathematics to establish the present value of the favored issue. But the combination of precise formulas with highly imprecise assumptions can be used to establish, or rather justify, practically any value one wishes, however high, for a really outstanding issue.
The exception is, of course, Stephen Penman, professor at the Columbia Business School and author of the best analytical textbook on value investing, who has been writing extensively on analytical tools for value investors. In a nutshell, he advocates for accrual-accounting models and in particular for the “book-value-cum-residual-earnings” valuation model. Although I plan to write more about his valuation techniques in the future, for the time being just mention that these models encapsulate many of the superforecasters’ features mentioned above. In other words, these models are a nice way to organise your thinking and to be rigorous.
Adding these “every-valuation-model-ought-to-have” features on top of the behavioural ones, we arrive at a truly list of commandments for everyone who wants to become a value investor. It is worth stressing again that many behavioural features should be implicitly included in the valuation technique, so some overlapping is unavoidable. Furthermore, I have avoided several features that should be obvious enough or even tautological (be contrarian, be patient because prices will eventually return to the fundamentals, etc.) and also other features that have been popularised somewhere else by famous value investors:
- Choose the right questions (a.k.a operate in your circle of competence). Be a business analyst before than a financial analyst: you should focus on the relevant questions of the business.
- Fermi-izing. Separate what you know from speculation and detect where your uncertainty is. In doing so, break down big questions into small ones.
- Uphold a “superforecaster learning process”: keep score, update continuously and try again. Besides, keep looking for vulnerabilities in your own analysis (in Tetlock-Gardner’s parlance, be a “perpetual beta”).
- Anchor the valuation process on what you know or at least on what you think is less uncertain (i.e. book values, normalised earnings, etc.).
- Beware of paying too much for growth (i.e. add a “margin of safety” to your process to account for uncertainty) and beware of paying too much for leverage too (leverage may simply reflect poor business economic performance).
- And when challenging price, don’t use price in your valuation model (e.g. multiples-based valuation).
Granted, I have left out Buffett’s two famous commandments: first, don’t lose money, and second, don’t forget the first commandment. But these are the kind of commandments that only people like Buffett can legitimately say.