Don't Forget About Base Rates

"Steve is very shy and withdrawn, invariably helpful but with very little interest in people or in the world of reality. A meek and tidy soul, he has a need for order and structure, and a passion for detail. Is Steve more likely to be a librarian or a farmer?"

Helpful... meek... tidy... little interest in people... I have to go with librarian.

This quote from Kahneman's Thinking, Fast and Slow is meant to illustrate the importance of base rates. After reading the description and making an intuitive judgment based on quick mental associations, most assume that Steve is a librarian. This is, of course, without considering that male farmers outnumber male librarians more than 20 to 1. No matter what situation you are presented with, remembering to consider base rates, or the past statistical precedent for a given outcome, is likely to improve your forecasting ability.

As an investor, you are faced with constant judgment. What does this news story mean for the stock? What happens if the Fed raises rates? What's priced in? What could be underappreciated? How much will this regulation impact sales? Base rates offer the historical answer to the question you're facing. As I'll go over, there are situations where base rates are more applicable than others, but in all cases, they serve to reframe the question from what do you think will happen? to how likely is this to diverge from historical results?

As Michael Mauboussin constantly reminds us, the investor's objective is to find the gap between the financial performance implied by an asset's price and the results that will play out in reality. Because that price reflects expectations about future performance, making an investment requires an understanding of what is "priced in" and a view on the company's future performance.

One way to derive that forecast is to gather all the information you can, feed it into your internal pattern matching system, and extrapolate, making adjustments based on judgments you may or may not be conscious of. Kahneman calls this the inside view.

As you can imagine, the inside view is susceptible to all the biases and distortions that are the subject of his book. You might artificially inflate sales expectations because you like the founder. You might exaggerate the time it will take to launch a new product because you find it complex. In general, the tendency is to overestimate the uniqueness of a given forecast and base our judgment too heavily on our individual experience.

The outside view considers a forecast in the context of a larger "reference class." Instead of asking what do you think will happen? the outside view asks what has happened historically when others were in this situation?

When analyzing a $2B market cap company that sells enterprise software and estimating future operating margins, you might look at companies of a similar size in the same industry and find the base rate of operating margins for that reference class. This reframes the question and forces you to think about why and to what degree this company will be different than its historical peers. The outside view shouldn't simply displace the inside view - they should both be considered and weighted differently depending on the nature of the environment.

Of course, the practicality of base rates varies depending on what you are trying to forecast. In environments governed by extreme randomness, the base rate tells you little of future expectations and can usually be ignored.

Mauboussin has an enormous body of work concerning the effects of skill and luck on outcomes in domains from sports to music to investing. He says that we can come up with relative weights for the inside and outside views based on where a domain falls on the luck-skill continuum.

The rule of thumb is: for activities where skill dominates, overweight the inside view; for activities where luck dominates, overweight the outside view.

He classifies NBA basketball as the sport that is the closest to pure skill on the continuum. If a team is having a great season with a winning record, the odds of them winning the next game should be higher - the base rate of wins for all teams historically has less bearing on the result. By contrast, investing is heavily influenced by luck, especially in the short run. If a manager has a massive year, it's reasonable to assume the next year will revert closer to the mean.

The base rate is another word for the mean. Reversion to the mean is more powerful in domains dominated by luck than those dominated by skill. When there is strong reversion to the mean, it makes sense to more heavily weigh the outside view, as future results are more likely to mirror historical averages.

Regardless of where your decision falls on the luck-skill continuum, I think it's useful to consider base rates. It seems so obvious that it hardly bears mentioning, but I find that I constantly forget the concept and often overweight my inside view. Especially as an investor faced with constant forecasting and group decisions, base rates are a simple and powerful tool to round out your decision-making.