The Few Rule the Many — Power Laws in Market Returns (2015 Case Study)

As index investing has grown in popularity, investors focus more and more on the market’s overall return and less on the return of its component parts (individual stocks). But underneath the hood of each market index we find many inequalities. The top 20% of stocks represent 85% of the overall market’s size in 2015. Similarly, the top 20% of the stocks account for 85% of the market’s total sales. These ratios have been constant through time.

pareto

A similar pattern emerges in the distribution of market returns. Again, we find that the top 20% of stocks represent, on average, 80% of the market’s total return in any given year.

2015 Case Study

The easiest market story to tell in 2015 was the dominance of F.A.N.G. and 5-10 more dominant stocks. Without them, the market would have been negative. It is because of these stocks that so many have described the market as “narrow.” As we shall see, what was unique about 2015 was not so much the contribution of the top 10 stocks, but rather the fact that the top ten contributed more than the bottom 490 stocks (which, for 2015, were down in aggregate).

There are several different ways we can put the performance of the top 10 stocks from 2015 in historical context. First is to take the total contribution to return[i] for the top 10 stocks every year and compare that to the contribution from the bottom 490 stocks in the S&P 500[ii]. You can see in the chart below the historical results in blue. Any time the blue line is above zero, it means the top 10 stocks out-contributed the bottom 490 stocks, which has been rare (in calendar years, anyways). NOTE: I’ve only included years in which the market’s total return was positive here[iii].

In this same chart, I’ve also included, in orange, the ratio of the top ten contributors to the total market’s return (which contrasts with a simple spread up top in blue). If the ratio is greater than 100%, which it was in 2015, it means that the rest of the market was net negative. The return spread (blue) was very wide in 1970, 2011, and 2015, but a big reason was because the rest of the market was negative. 1999 stands out (as it always does) as a bizarre outlier. The spread between top 10 and the rest was the widest ever, but those top 10 (including Cisco, Microsoft, Oracle, Sun Microsystems, Intel, and EMC) still represented less than 100% of the total market return—crazy times.

all years

The opposite end of the spectrum (wide performance breadth) would be a year like 2013, where the bottom 490 (which added 26.2%) contributed +20.5% MORE than the top 10 (which contributed 5.7%). To many, this might seem obvious: of course 490 stocks added more than 10 stocks! But in a world governed by power laws, the top 10 stocks have usually accounted for between 40-60% of the market’s total return. More often than not, the few rule the many.

Best vs. Worst

The best stocks are one thing, but the worst stocks play a huge role as well. There is a built in skewness to market returns: stocks can only go down -100%, but can go up any amount (although, as Jim Chanos says, there are a lot more stocks that go to zero than infinity). On average, the top 10 contributors overpower the bottom 10 detractors. When you add up the total contribution from the best and worst 10 stocks, you get the blue line in the following chart. Usually, the top stocks pull the market up MORE than the bottom stocks drag it down. The notable exceptions are during the major bear markets.
top and bottom 10

What’s interesting in this cycle is that while everyone is talking about how much the top ten helped stocks, what they haven’t mentioned is how much the bottom ten have hurt them. Exxon, Chevron, Conoco, Kinder Morgan, Berkshire Hathaway and others negated a sizeable chunk of F.A.N.G.’s meteoric rise.

When we account for the bottom 10 alongside the top 10 stocks, 2015 doesn’t look peculiar at all: just a little below the long-term average.

Power laws, again

In 2015, all the talk was about the top 10 stocks, but let’s step back to our 80/20 paradigm again.

If you isolate all 12-month periods historically (1964 on) where 1) the overall market is up and 2) both the top 20% and the bottom 80% of stocks in the S&P 500 are up (that is, both groups have a positive total contribution), then you find another remarkable tendency. In these periods, the top 20% of stocks in the S&P 500 account for 79.5% of the markets total return—almost exactly matching the 80/20 rule.

If you isolate all 12-month periods where the market is down, the worst 20% of stocks in the S&P 500 account for 81.2% of the markets total loss, closely hugging the 80/20 rule yet again—you can’t make this up.

Active Management and Power Laws

Given that stocks at the tail ends of the distribution explain an outsize chunk of the overall market’s performance, active managers can try to harness the power-law tendencies of the market in a few different ways. Many managers emerged as stars because of the tech stocks that they didn’t own between 2000-2002. Howard Mark’s often points out that one great way to succeed in investing is to view it as a “negative art.” By simply not owning (or being significantly underweight) the stocks on the left (bad) side of the return distribution, you can do very well.

Conversely, if you overweight stocks in the right (good) side of the return distribution, you can do extremely well. This is, of course, much easier said than done. The popular “active share” metric, which measures what percent of any portfolio is unique or distinct from its benchmark, is a great way of seeing how much of a potential advantage a given manager has. The more different a portfolio from the overall market, the wider the potential range of excess returns, both good and bad.

My next post will investigate whether or not these stocks at the tails of the return distribution share any common attributes which may be useful for active managers.

[i] A quick word on the simple math for return contribution here. On every starting date, I calculate each stocks weight in the overall market (market cap (float) for stock x, divided by the sum of all market caps). I then fetch the return over the next year (or, if a stock is acquired or goes to zero, the return over the stub-period less than 1-year). I then multiple the starting weight to the forward 1-year return to get each stock’s contribution to total return. So, for example, Microsoft started 2015 as 2.2% of the market’s overall weight, it went up 22% or so, leading to a total contribution of +0.5% (2.2%*22%=~0.5% contribution). Get a contribution for every stock, add them all up, and you have the market’s return.

[ii] A quite caveat/disclaimer: none of these numbers are “official,” they won’t exactly match the results for the S&P 500, but they will be darn close. When I don’t have actual SP 500 constituents, I use the top 500 stocks by market cap.

[iii]  (the top ten show huge excess return when the market is down which muddles the message)