Benchmark Oddities

Imagine two NBA teams. Team one starts Anthony Davis, Russell Westbrook, Stephen Curry, Kevin Durant, and James Harden. Team two starts Hassan Whiteside, DeMarcus Cousins, Brook Lopez, Marc Gasol, and Nikola Vucevic. Who wins?

The two teams above were determined by a stat called PER (player efficiency rating) with the twist being that while team one is simply the five best players in the league, team two is the five best BIG players (who play the center position). If you think of these as a simple models for building a team, there are two factors: size and player efficiency. In this case, using efficiency alone produces a far better team.

The investing parallel here are style indexes like the Russell 1000 Value or Russell 2000 Growth. Style indexes are very popular, but when you really examine their methodology, they are also rather odd, sort of like team two.

The “strategy” for each style benchmark has two steps. First, determine what stocks are in the index, and second, determine what weight each stock gets. Benchmark constituents (the stocks IN the index) are determined by relative valuation (book-to-price and expected growth), but each stock’s weight is then determined by size (market cap).

The strategy behind these value indexes could be phrased this way: Cheap is good, but big and moderately cheap is better than smaller and really cheap.

For example, Exxon Mobile has a price-to-book of 1.95, but is a huge company so has a weight of 3.3% in the Russell 1000 Value index. Seadrill—a maligned energy stock—has a price to book of 0.28. So it is much “cheaper” than Exxon, but it’s only a $2B company, so its weight in the index is 0.02%, which may as well be zero.

If Exxon went up 40%, it would push the overall index up 1.35% or so. If Seadrill went up 40%, it would push the index up 0.01%. So for the Russell 1000 Value, Exxon is the far more important stock. But if you cared more about value than size, then the weights would be very different. Exxon is the biggest stock, but it’s only in the 50th percentile when sorted by price-to-book instead of by market cap. Seadrill is in the cheapest percentile.

This is odd because the relative merit of one stock over another (and therefore, its weight in your portfolio) is far better determined by its cheapness than by its size. Here is a comparison of the size and value factors in the U.S. Large Cap market, since 1963. These are the historical excess returns for factor deciles, which are rebalanced like the style indexes are, on an annual basis[i].

deciles by cap and value

Relative valuation (measured here by a five factor model using sales, earnings, free cash flow, EBITDA, and total yield) has been far more predictive than relative size, but the index is weighted in the opposite way.

For people running money, building a rules-based index/strategy which weights stocks based on their size is a wonderful business idea, though, because it scales. Bigger names have more liquidity, so you can manage a lot of money in a market-cap weighted strategy. This is especially true in small cap. The Russell 2000 Value ETF (IWN) has nearly $6 billion dollars in it. Even at a reasonable 0.25% expense ratio, that’s still $15,000,000 in revenues for iShares. This level of assets would be totally impossible for most small cap value managers unless their turnover was extremely low (which can be good, but can also be bad because it means that you aren’t rotating into the best new opportunities).

To be fair, value indexes do have advantages: mainly a lower cost structure (fees, trading, taxes), which is huge.  Higher costs are the reason active managers in aggregate do so badly versus indexes.

But still, the strategy itself leaves a lot to be desired. Costs for more “active” strategies will continue to come down, reducing the cost advantage currently enjoyed by indexes.

Benchmarks of the Future

Cap weighting is of course relevant to some extent because Exxon is a far more important stock in general than Seadrill. But cap-weighting also produces weird portfolio outcomes. The smallest 500 stocks in the Russell 3000 Value, for example, have a combined weight of 0.5%. The bottom 1000 stocks have a combined weight of 2.59%. So in the Russell 3000 Value, Exxon (3.1% weight) is more important than the bottom 1000 stocks. That makes most of the stocks near the bottom irrelevant for performance comparisons.

My guess is that with the proliferation of ETFs, we will see more specific “benchmarks” continue to pop up. Deep value managers will have to beat a true deep value benchmark (of say the 100 cheapest stocks in the overall market based on the value metric I used above) rather than the Russell 3000 Value. In these custom benchmarks, the smaller names could play a much larger role. If the future is anything like the past, these more tailored benchmarks will be even harder to beat.


[i] A slight nuance, to avoid seasonality, the stocks in each decile are held for one year like with an annual rebalance, but it’s done on a rolling monthly basis.