ZIRP, And The Factors That Launched 1,000 ETFs

The rise of smart beta–or more broadly, factor investing–has coincided with a 6 year period of zero interest rates. During this period, factors have been particularly ineffective relative to longer term results. Using publicly-available data (Ken French) we can explore the recent results for the most popular stock selection factors and compare them to longer-term periods of both rising and falling rates.

While Ken French publishes a wealth of useful data, the most important are data on “core” stock selection factors: book/price (value), size, operating profitability + investment (quality), and momentum (price trend). These are the factors that launched a thousand (ETF) ships.

Before we get to the results, it is important to note how difficult it is to pin down isolated, causal relationships between macro variables (like interest rates) and stock selection factors (like value). There are many things going on that affect the value factor, including relative popularity (value vs. growth fund flows), valuation spreads, sector specific trends (think energy), adaptive markets (where past alphas morph into future betas), and so on. With that in mind, remember that the below is meant to be educational, not prescriptive.

Here are the three regimes across which we can measure factors:

three interest rate regimes

There are many ways to evaluate the effectiveness of a stock selection factor. To keep it simple, I’ll use the annualized long/short return spread (best-worst decile) for each factor. This is by no means comprehensive, but its a good start. Stocks in the best and worst deciles are equal-weighted (French also provides cap-weighted versions).

Across the five factors, the average long/short spread was 10.3%/year for the entire period (1964-2015), 10.6% for the period of rising rates (1964-1981), 11.8% for the period of falling rates (1981-2008) and 2.3% for the ZIRP period. Clearly, things have tailed off.

factor spread averages

Here is the factor by factor look. Some immediate caveats: momentum looks so poor only because of 2009, which I’ve explored before. If you remove 2009, the spread jumps to +17.7%! So absent 2009, pure momentum has worked well. Even still, the average spread has come down considerably during the ZIRP period. Also, the spreads for operating profitability are not that impressive, but I include it here because its part of the Fama/French five factor model. The average spread across the five since 2010 (so removing 2009 from the calculation) is 4.9%, still a big drop.

factors by period

The big market story this year has been growth over value. Across the five factors, value looks particularly bad in the last several years. But, during this period of zero interest rates, every factor has had a spread equal to, or below, the factor spread across the entire period.

One final view helps contextualize the performance of these factors. Below is the rolling 10-year long/short spread for each factor, show across the three regimes.

rolling 10 year long short spread

This tells a similar story: these particular factors have worked far less well in the recent past.

Is this drop a result of interest rates, or the result of some other cocktail of market influences/dynamics? It is very hard to say. I should also note that while Ken French’s data is the “gold standard,” I believe these factors can be improved upon (for example, book/price is among the worst performing value factors). These are also all individual factor results, meaning stocks in each best/worst decile are being measured by just one factor. I believe better use of factors is selecting stocks which look attractive across a broad spectrum of factors (quality, value, momentum, shareholder yield all at once).

Most people interested in this topic already know: factors are not a panacea. They come into and out of favor. Assuming the Fed raises more in the future, we will get a better sense of how factors perform during the first rising rate environment in years.

More to come, including more granular looks at interest rates cycles, and including different factors. I hope this becomes an open inquiry and topic of interest for all of you. Please email me or comment below to get involved.