Two Decades of Factor Returns

The Dow Jones U.S. Thematic Market Neutral Indices are, according to the methodology page for them, “a family of weighted return strategy indices that each reflect the performance of offsetting long/short allocations in two component indices.”  They were launched in August 2011 (the vertical line in the charts), but backtested history starts at the end of 2001.

Each index shown below is a long/short blending of two components, rebalanced quarterly to equally weight two hundred stocks long and two hundred short.  Sector neutrality is maintained between them.

This chart shows the two factors that have received the most attention over time.  The Value Index uses three valuation measures to form the long (cheapest) and short (more expensive) portfolios.  The Size Index is long the smallest stocks and short the largest.

The factors each outperformed over the first dozen years of the chart, but then underperformed significantly until bouncing in mid-2020.  Broadly speaking, their trajectories reflect the rise in popularity of (and asset flow into) “smart beta” strategies — and the subsequent disappointment as performance deteriorated.

The other three factors in the group are shown here.  The scale of the chart is the same as on the first one, so you can see that these were generally less volatile than value and size.  They didn’t have a similar migration up and down over the two decades, with the most noticeable moves occurring during periods of market trauma in 2002, 2008, and 2020.

The Quality Index (a factor for which the definitions tend to vary quite a bit across the industry) is sorted by return on equity and debt-to-equity ratios.  The Momentum Index is long high momentum and short low momentum, while the Beta Index is long low beta and short high beta.

It is worth noting that there are different approaches to displaying factor returns, so other calculations will vary, but the profiles are similar.

Two decades is not that long in market time, even though lots of products are marketed with much shorter track records and investor memories fade more quickly than they should.  Using this time period leaves out some important earlier patterns.

Regarding this particular family of indices, Cliff Asness wrote that their starting point “means they miss what was a very strong period for factors in the mid-to-late 1990s, a very bad period for the value factor and for most reasonable combinations of quant factors during the tech bubble, and a big part of the recovery from the tech bubble craziness from 4/2000–12/2001.”

The factor story of twenty years ago — and the one of ten years ago — differed from today’s.  As with everything, the evidence changes over time.