My First Principles
The purpose of an investment is to reap profitable returns. A good investment is one that results in healthy returns and, by logical extension, a good investor is one who can choose these good investments with some degree of certainty and regularity beyond the purely random.
But how do we know whether an individual investment is good or bad? And how can we identify an effective process for choosing investments?
The first question is far more difficult than the second. The vast majority of good investors could no more guess the direction of a particular stock in a particular year than a dart throwing chimpanzee (if you are unconvinced, I suggest you examine the results of the recommendations from last year’s Sohn Conference).
But a good investor need not be able to discern whether any individual investment is good or bad to be a good investor. In this, investing is like science. The philosopher Friedrich Hayek argued that science is not a method for ascertaining particular facts in particular instances. “In the case of very complex phenomena the powers of science are limited by the practical impossibility of ascertaining all the particular facts which we would have to know if its theories were to give us power of predicting specific events,” he wrote. “All scientific knowledge is knowledge not of particular facts but of hypotheses which have so far withstood systematic attempts at refuting them.” [1]
Fewer systems are more complex and dynamic than modern financial markets, and obtaining all of the particular facts of a particular investment are complicated because of the system itself and the dependence on unknown facts—unknown, because they will be the product of unpredictable future events.
When evaluating the skill of the prophets of profit, it is far more valuable to look at the hypotheses that guide their decision-making than particular investment picks. The only way to know whether an individual investor is good or not is to attempt to evaluate whether the hypotheses that underpin his investment philosophy are good. The best way to do this is by using long-term historical market data for as many different markets as possible, rigorously testing and attempting to falsify each hypothesis.
This is why academic finance is useful, because what academics do is take a hypothesis (e.g., buying companies with low EV/EBITDA multiples will generate returns in excess of the broader market) and test that hypothesis statistically in the complete data set of historical stock returns. Those hypotheses that prove statistically significant could possibly be true, while those that failed in their predictive power we can dismiss as falsified.
For the past eight years, I have been systematically evaluating different investment rationales. In addition to what I learned at Bridgewater and Bain Capital, where I used to work, I have interviewed dozens of managers, read deeply the academic research on investing, and run several large scale quantitative studies with my research partner, Brian Chingono.
My goal has been to determine which rationales function effectively as universal principles. For example, if an investor says that they bought Company A’s shares because Company A’s historical revenue growth had been excellent, we could universalize that rationale into a principle: buy all companies with strong historical revenue growth. And we could test that principle using quantitative methods to discover that it was a bad one. In contrast, if an investor said they bought Company B’s shares because Company B traded at a very inexpensive enterprise value to EBITDA multiple, we could universalize that rationale to the principle “buy all companies with inexpensive enterprise value to EBITDA multiples” and conclude that this was a good decision.
Before investing in Company A for Reason A, we should evaluate whether Reason A would be a good universal law. We can use quantitative studies to determine whether individual principles work as universal laws, and, through this process, come to separate the bad rationales (e.g., don’t buy declining businesses, invest only in companies with wide moats and strong competitive position, rely on 5-year financial models to determine net present value, etc.) from good ones (e.g., buy cheap, avoid companies with increasing debt and declining asset turnover, buy small, use leverage).
This process has led me to an approach that I believe represents the best absolute return strategy in equities: investing in leveraged small value equities that have a high probability of deleveraging. Brian and I published our research to seek refutation of our ideas or improvements on our methods, and we have been flattered to see our ideas replicated in academia.
I do not understand why investment managers would prefer case-by-case analysis to principles-based investing, nor do I understand why business schools (alone among the academic disciplines) prefer case studies as the prime method of instruction. As Immanuel Kant wrote, “Nor could anything be more fatal to morality than that we should wish to derive it from examples. For every example of it that is set before me must be first itself tested by principles of morality, whether it is worthy to serve as an original example, i.e., as a pattern; but by no means can it authoritatively furnish the conception of morality."[2]
Anecdote, case study, experience—these are unstable stepping-stones to truth. Rather, we must use rigorous logic, paired with testing (as universal as possible), to deduce the principles that govern the way markets work. And then we should trust the managers whose principles are supported by evidence, rather than those who either don’t invest based on principle or invest based on principles that are unproven (or proven to be fallacious, like growth investing).
This form of investing is preferable not only on methodological grounds, but also on account of fundamental flaws in human psychology. It is a recurring paradox of human psychology that even in the face of contrary evidence we tend to believe that our immediate, deliberative capability in particular circumstances is superior to adherence to considered rules deliberated in advance of a particular scenario. As the famous legal philosopher Joseph Raz observed in his lesser known work on the subject of rules, “The advantage of normally proceeding through the mediation of rules is enormous. It enables a person to consider and form an opinion on the general aspects of recurrent situations in advance of their occurrence. It enables a person to achieve results which can be attained only through an advance commitment to a whole series of actions, rather than by case-by-case examination.” [3]
We are in search of principles, not case studies, universal knowledge, not anecdotes, and we are far more likely to find the simple true laws in the largest possible samples than in small and non-representative samples—where the powers of randomness might yield sophisms, spurious relationships, and false truths. We can only determine whether an investment manager is good or bad based on their principles, not based on their outcomes. And the best way to evaluate principles is by universalization and then rigorous testing of those hypothetically transcendent rules. Only longevity of persistence in the face of intense attempts at refutation will suffice. As Francis Bacon wrote, “demonstration longe optima est experientia.”
Acknowledgement:
I would like to thank Nick Schmitz for his help in ensuring the correctness of my philosophical references and the integrity of my interpretations of these philosophers. His advice and editing were essential to this research.
[1] F.A. Hayek “Law, Legislation and Liberty” 1982. Pg. 16
[2] Immanuel Kant, "Groundwork for the Metaphysics of Morals," 1785. Accessed here.
[3] Joseph Raz, “Authority and Justification”, Philosophy and Public Affairs, vol. 14, no. 1, Winter, 1985. p. 23. See also, F.A. Hayek, The Constitution of Liberty, Routledge, London, 1960. p. 59: “We all know that, in the pursuit of our individual aims, we are not likely to be successful unless we lay down for ourselves some general rules to which we will adhere without re-examining their justification in every instance. In ordering our day, in doing disagreeable but necessary tasks at once, in refraining from certain stimulants, or in suppressing certain impulses, we frequently find it necessary to make such practices an unconscious habit, because we know that without this the rational grounds which make such behaviour desirable would not be sufficiently effective to balance temporary desires and to make us do what we should wish to do from a long-term point of view. Though it sounds paradoxical to say that in order to make ourselves act rationally we often find it necessary to be guided by habit rather than reflection, or to say that to prevent ourselves from making the wrong decision we must deliberately reduce the range of choices before us, we all know that this is often necessary in practice if we are to achieve our long range aims.” See also generally, John Rawls, “Two Concepts of Rules”, The Philosophical Review, vol. 64, no. 1., January 1955.