Power Ranking Investing

By Steve Tepper, CFP®, MBA


Past performance is not a predictor of future returns.

Ever heard that one before? We all have. But most investors invest like they haven’t. Or perhaps more accurately, most investors invest like they don’t believe it. As I’ve said before, and will probably say again, every investor who picks a fund because of its Morningstar “5-Star” rating is picking what they think will be a winner based on the fund’s past performance.

This is one of the most curious phenomena of my industry. Why? Why do smart people and experienced investors delude themselves by denying the first disclaimer of every financial industry advertisement? Why is the majority of the investment community aligned with the belief that past performance somehow holds the key to picking winners?

An answer came to me this afternoon as I was completely not goofing off but doing important job-related research. I was perusing the upcoming season’s NFL power rankings. This list was slightly different from other 1-to-32 team rankings I’ve seen. The list was divided into quartiles—a fancy way of saying four tiers—each with eight teams. The top quartile was populated with the analysts’ predictions for the eight best teams in the league this coming season.

Something about that word, quartile, sounded a bell in my brain. Then again, that bell might have been the ice cream truck. It comes by the office about that time of day. But let’s pretend it was a cogent thought. I remembered a recent article I wrote about mutual fund performance, specifically a study of top-quartile performers that looked to see if they stayed at the top in the next period. To no surprise to me, anyone in the office, or any (I hope) of our clients, they didn’t. Performance was completely random, with only about a quarter of the top performers remaining there in the next period.

So now I’m staring at a list of NFL teams, and there are some familiar names: Rams, Saints, Chiefs, and of course, that team from Boston. And that’s when the light bulb went off. So now I’ve got bells and light bulbs. It’s like Harrah’s Casino in my head.

Somebody thinks past performance is an indicator of future performance! These were the top teams last season, and now all the experts think they will be again. I did a quick search for the year-end 2018 power rankings, and there it was: The top six teams from last year were all in the 2019 upper quartile.

Of course, this is summer and not a single down of football has been played. By the end of the season, that list will look very different when, you know, the Cleveland Browns (ranked No. 32 in 2017) go 14-2 and win the Super Bowl by 63 points. Or perhaps I’m putting a little too much faith in Baker Mayfield.

But the bells and light bulbs sent me on a little research expedition: to look at the last eight 1-year rolling periods and see how many of the top eight NFL teams repeated in the top eight the next year. Here’s what I found:


If results were random, then we’d expect that, on average, two teams would repeat each year (and the six others would be evenly divided among the other three quartiles). It might not be exactly two each year. Maybe one one year, three the next, even four or zero once in a while—but the average would be right around two.

The results in the table are clearly not random. Most years, half or more than half of the teams repeat at the top. And it’s not just year to year. Over the course of the entire eight years, each team should appear on the list about twice. But a handful of teams are tops way more than that. The Seahawks and Steelers appear five times. The Saints, Packers, and Broncos make the cut four times. And yes, that team of deflated balls and Spygate made the list every single year.

Perhaps the disclaimer should be “Past performance is not a predictor of the future unless Tom Brady is on your team.”

The Takeaway

Does this mean we’ve got it wrong? Does past performance mean anything? Well, that’s two questions, and the answer to the second question is “Definitely!” Yeah, I said it. Past performance means a lot! But also, to the first question, “No!” We don’t have it wrong.

The takeaway is this: It’s easy to understand why investors so often ignore the basic truth of investing and look to “power rankings” to pick winners going forward. We are conditioned into it by a lifetime of such lists, lists that make complete sense and actually mean something in so many other contexts.

Of course past performance of a sports team can be used to predict their future fortunes. If they had a bunch of great players and great coaches last year and did very well, and if they still have those players and coaches this year, why wouldn’t they continue to do well? It’s not a 100% rule (the Jacksonville Jaguars, top-quartile performers in 2017, fell to the bottom group in 2018), but even a 50%–60% rule is significant.

We fully and correctly assume past performance is a valid predictor of future performance not just in athletes but in actors, musicians, employees, cars, appliances, neighborhood crime rates, and on and on and on. We use rankings and reviews every day, to pick a movie, to buy a bottle of wine, to hire an employee, to hire a contractor, to make a major purchase.

We don’t always pick right, but who among us would prefer to leave every choice to pure random chance? No way. We are going to do our research to maximize the probability of making a good decision (or at least scan the Yelp reviews).

So, too, we expect great companies that have done well in the past to continue to do well in the future. Do they still have great products and distribution, strong financials and management? If those things hold true, their good fortunes are far more likely to continue than a company with none of those things. So why does investing (and only investing) have that ugly caveat, that nothing we know about a company, good or bad, can help us pick a winner?

The answer is simple: We aren’t trying to pick winning companies. We are looking for outperforming stocks, and there’s a world of difference in that distinction. If your decision is to buy the company’s products, or hire a key employee away, or enter into competition, then heck yeah, find out everything you can on past performance. But when buying stock, expectations of future performance are not meaningful, as the price of the stock already reflects those expectations.

A great company’s stock is usually very expensive relative to less successful companies, and that high price lowers the probability that the stock will “beat the market” going forward. That is a pretty special and unique feature of capital investing.


We spend our whole lives looking at past performance as the best predictor of future performance in myriad circumstances. But a great investing experience depends on unlearning that behavior. So here is one more reminder: Don’t ignore the disclaimer.