The Problem With the “Inside View”

Everybody knows that outcomes are a combination of skill and luck, but we tend not to be very good at understanding the relative contributions of the two.
  • Transcript


Question: Why is the "inside view" not wise when investing? 

Michael Mauboussin: The inside view says that when we face a problem, what all of us typically do is gather lots of information about the problem, combine it with our own inputs and then we forecast into the future. So that could be, you know, "How long is it going to take me to remodel my kitchen? When am I going to finish my term paper? How long will launch a new product for a company?"

The outside view in contrast says I’m going to look at my problem as an instance of a larger reference class. So it allows me to ask a really simple question which is, "When other people have been in this situation, what happens statistically or probabilistically?" There are lots of problems that we face as individuals that are unique to us, but lots of other people have been through before so there’s this thing called the database of humanity waiting for you to tap into it. 

The inside view, by the way, importantly, is the natural way to do things; it’s sort of how your mind’s naturally going to work. So the outside view requires you basically to leave aside this cherished information that you’ve gathered and your own point of view on things in order to forecast. 

So it turns out that we all tend to be too reliant on the inside view which often leads to too optimistic assessments of our futures. So you know the old things, people say "How long’s it going to take you to remodel your kitchen," and people will say to you, "Whatever you think, double the time and double the cost," right? Because you’re relying on the inside, they’re telling you, they’re reminding you to go back to the outside view. But it’s true for the world of investing as well which is we have lots and lots of data about, for example, the growth rates of companies or the return on invested capital patterns or the returns of assets over time and we tend to be caught up very much in the moment with our own points of view, we tend to leave that kind of information, valuable information, to the side. 

So the point is, when appropriate, always seek to look for the outside view and help that inform your inside view and again, more times than not, it tends to damp optimism that’s unduly generated. 

Question: How much should an investor take past performance into account? 

Michael Mauboussin: Everybody knows that the outcomes they observe are a combination of skill and luck, but I think as people, we tend to be not very good at understanding the relative contributions of the two. So one of the ways I like to think about this is there are some activities in life that are pure skill. If you and I have a running race or something like that, you know, the faster person will win almost every single time. And there are some activities that are pure luck. If you go roulette or you roll the dice, it’s going to be randomness and then almost everything else in life is in between. But what’s interesting is they're not right in the middle, they’re almost always leaning toward the skill side or they’re leaning toward the luck side. 

So the question is always when you’re observing a particular activity, where is it on the continuum? One of the things I like to think about are sports, because we have a lot of data on sports and people tend to be very familiar with them. But it probably wouldn’t come as a big surprise that things like baseball or premier league soccer in the UK tend to be a lot more on the randomness/luck side and things like basketball or tennis tend to be much more on the skill side. So it’s an interesting thing that most of us know that skill and luck combine for outcomes, but we don’t have a good sense of where things lie. 

So that has all sorts of important implications. I’ll just mention a couple of them. The first is whenever there’s luck in the system, there’s going to be lots of reversion to the mean. That means if you have an extreme event, either extremely good or extremely bad, you should expect the next thing to be closer to average. So that’s true, for example, for corporate performance. If a company’s been doing really, really well, sales growth has been really rapid, you should expect that the next growth rate should be something closer to average. 

The second thing that’s really interesting is what kind of feedback you give people. Feedback in our society tends to be very outcome-oriented, based on what we have observed, and does it really reckon with this issue of the skill/luck contribution. But if you’re giving feedback properly, it should be focused only on the skill component, only on what people can control and to the best of your ability; you should leave aside sort of the randomness or the luck component.

Recorded on May 14, 2010
Interviewed by Jessica Liebman