Michael Mauboussin: The Importance Of Utilizing Base Rates

In his recent interview on the Decision Education Podcast, Michael Mauboussin discussed the importance of utilizing base rates in investing. Here’s an excerpt from the interview:

Mauboussin: Yeah, thanks for slowing that down a little bit. So, as I mentioned, rather than thinking about your problem as unique to you, what you want to ask is, what happened when other people were in this situation before? So, you’re gonna now draw from a distribution of experiences versus your own experience. So let me give you an example. I’ll do one from my world that’s actually quite simple to articulate.

Let’s say you’re analyzing a company, and I’ll just make this up, say it has $10 billion in revenues. And you’re going to say, how fast will this company grow its revenues over the next three, five years or something like that? Well, you can either, obviously you sort of analyze the company and think about the industry and its market share and all those kinds of things, and then make a projection.

And alternatively, and I guess what you’d want to also do to come up with a robust forecast, is you might look at the history of all companies of that size. And simply ask what is the distribution of growth rates for companies of that size? And you’ll see, it’s not a perfectly normal distribution, but you’ll see something like a normal distribution. And that will allow you to calibrate where your forecast resides within a historical distribution.

Now, the one thing I should say about base rates is the application varies across various domains. So some base rate distributions are fairly well behaved, things like sales growth rates for companies and the heights of people. There is another part where it’s vastly more complicated where you have these distributions that are much more skewed, much more power law-like, or where there are lots of small observations or very few large observations.

So things that are socially driven, book sales, music, sales, and movies, are much more difficult to do. But in general, this is a wildly underutilized framework. It’s interesting, I guest lectured for a course last night at a college and the students said to me, if you could go back in time and give yourself one framework that would make you more informed going forward, what would it be?

And I said, it would be this idea of base rates. I just think it’s an incredible concept that carries an enormous amount of intellectual freight once you understand it. It’s not natural because you’ve done the work and you have your own experience and hence we think we check the boxes. And second is that base rates are not always at your fingertips.

If I move from here to Philadelphia, I’ve never done that before, so I don’t know what’s going on, but lots of other people have done that before. So there are base rates of results. I just am not aware of what’s going on. So a base rate is basically a distribution of past experiences given the same set of initial circumstances. Just understanding how things have unfolded for other people can be very helpful to understand how things might unfold for you or your object of interest.

You can listen to the entire interview here: