Laws of Investing

Jan 20

In both economics and investing, logic is the primary tool. Observation, fact gathering, data analysis, all the empirical techniques so vital for the physical sciences, are strictly secondary in economics and investing and very often misused.

Economists and investors are always touting some measurable relationship from the past, whether the price of two commodities or the spectacular returns earned from some investment formula. Causality is then inferred from correlation (A not only happened with B, but it also was caused by B). Unfortunately, even the correlations, much less the inferences of causation, are necessarily spurious. They contradict the reality of constant change in human choice. People acting, changing their mind, influencing each other cannot be accurately captured on a graph or projected into the future.

As noted in 100 Economic Laws:

“Sometimes the data being studied, graphed, and relied upon are so airy and immaterial that it can hardly even be called data. Federal Reserve Governor Lael Brainard gave a speech in 2016 in which she indicated the Fed was watching changes in a survey of consumer inflation expectations conducted by the University of Michigan. Jim Grant of Grant’s Interest Rate Observer sardonically noted that:
‘Random people, asked to guess where the CPI [consumer price index] might be trending in . . . [that year or five years later] did not [honestly] reply, ‘I don’t even know where it is now. . . Respondents rather vouchsafed an answer . . . [which] was . . . [a quarter of one percent] lower than the guesses previously submitted by previous random people since 2006.’

The “data dependent” Fed supposedly relied at least in part on this “data” in choosing not to increase interest rates.

Investment market analysts and especially sales people commonly both use and abuse data. If an investment firm tells you that it has tested an investment formula against decades of past investment history with spectacular results, what should you make of this? First of all, did the study account for what is called survivorship bias? That is, were all the companies that went bankrupt over the period under review put back into the database? Were any of the stocks too small to be traded? Were costs of trading and managing fully considered? Did the manager resort to any data mining, where every possible approach is tried until one finally succeeds on the back test? Any statement, including even an elaborate study, is illogical if it fails the key logic tests of order, organization, consistency, clarity, and relevance. Many studies fail some or all of these tests, and if the study is intended to persuade you to buy something, caveat emptor (buyer beware).

Even if a study is free of common logical errors, why should we expect the immediate past to tell us anything important about the future? As noted above, markets express human thought and behavior and there is no chance that we will see the same people having the same thoughts or taking the same actions in the future. We may therefore consider it an investment rule/principle/law that past returns cannot guarantee future results, a statement the SEC not coincidentally requires every investment manager to append to a recitation of past results.