crying baby.jpeg

Poor, poor stock pickers. They say that last year just wasn’t a “stock picker’s environment.” Really?

Take a look at the chart below. These are 2017 returns for S&P500 Index firms, ranked from worst to best for the year.

The green line is the index return. All your poor little stock picker had to do was grab a few stocks from the right side of that line and, BOOM, they beat the market and get famous. I mean, it was only a little worse than a coin flip that one randomly selected stock would have beat the market handily.

How is it that all the geniuses can’t beat the index if they are skilled stock pickers? Could it be a little harder than they admit, to you and themselves?

Screen Shot 2018-06-19 at 10.44.45 AM.png


First, our favorite “Yo’ Mama” joke:

Yo’ Mama is so stupid, she stared at the orange juice carton because it said ‘concentrate.’


But seriously: today we begin an experiment in concentration. Our work – and related work – explains why indexing tends to crush active management. The problem is that active managers have a high probability of picking subsets of the index that miss or underweight the handful of winners that drive index performance. If there is a solution, it is concentration.

I don’t mean concentrating on 25 or 30 stocks. I mean concentrating on 1 or 2 stocks. The very, very, very “best” idea.

So here is our experiment:

(1) Today, Monday, June 18, 2018, we sort the S&P500 firms by last Friday’s (June 15) closing market cap, 1 to 500, from smallest to largest. Note that we will take the S&P500 as constituted at that point in time. No adjustments, for example, for firms that have been acquired if their stock has not yet been taken out of the index or for announced but not yet implemented replacements.

(2) We generate a random number between 1 and 500 using the runif function in R, which generates a random number from a uniform distribution. Today, we got 69.50213. We round that number to 70.

Screen Shot 2018-06-18 at 5.40.26 AM.png

(3) The 70th firm is Under Armour, which has two classes of shares. Wherever we have two classes of shares, we generate another random number between 0 and 2 and pick the alphabetically-lower if below 1 and alphabetically-higher if above 1. Today we got 1.876036, so we pick Under Armour Inc Class C (UA) instead of Under Armour Inc Class A

Screen Shot 2018-06-18 at 6.03.09 AM.png

(4) At approximately 8:45 a.m. central time, we put in an order for as close to $10,000 worth of UA at the market price as we can get. Today that was $21.08 per share for 475 shares (yes, commission is high!):

Screen Shot 2018-06-18 at 8.47.53 AM.png

(5) We then put in an order for SPY – the S&P Index ETF. This is our benchmark portfolio. Today that was 37 shares at $275.425.

Screen Shot 2018-06-18 at 8.46.59 AM

(6) Our Concentrating Chimp follows a simple rule: if his stock has outperformed the SPY over the preceding month (a determination he will make as of market close on Friday, July 13, 2018, then he does nothing and holds the prior pick. If he has underperformed the SPY, he picks a new stock as above and, at approximately 8:45 a.m. the next trading day, sells his previous stock and buys his new stock. And so on.

(7) We’ll post the Concentrating Chimp’s results against his benchmark here every Friday after close. These are real positions in a real, tax-advantaged traditional IRA. We’ll post execution images for each trade as above.

And that is the Concentrating Chimp experiment.

Screen Shot 2018-04-17 at 5.13.56 PM

Given its poor performance relative to passive indexing, it is difficult to explain just how so many active managers who deliver unimpressive performance at high cost relative to other investment alternatives continue to attract and retain investors. In a new paper, Conjecture researchers J.B. Heaton and Ginger Pennington show that potential investors fall prey to the conjunction fallacy of Tversky and Kahneman (1983).  For an interview with J.B. on the paper and its implications, go to the excellent blog, The Evidence-Based Investor.




Excited to be joining Alpha Theory’s Lunch and Learn Webinar focused on our research described below (written while J.B. Heaton was still toiling away in law!).

We develop a simple stock selection model to explain why active equity man-agers tend to underperform a benchmark index. We motivate our model with theempirical observation that the best performing stocks in a broad market indexoften perform much better than the other stocks in the index. Randomly select-ing a subset of securities from the index may dramatically increase the chanceof underperforming the index. The relative likelihood of underperformance byinvestors choosing active management likely is much more important than theloss those same investors take due to the higher fees of active management rel-ative to passive index investing. Thus, active management may be even morechallenging than previously believed, and the stakes for finding the best activemanagers may be larger than previously assumed.