Tuesday, August 9, 2016

What if Factors Rarely Matter?

Back in December I wrote that It's Generally Smart to Avoid Credit Risk outlining that more than 100% of credit's excess performance over time has come when the level of credit spread was extreme.

What if the same were true for well known investment factors?

Taking a Look at the Small Cap Premium

The chart below takes the average market cap of the 30% largest companies within Fama French data and divides it by the average market cap of the 30% smallest to form a ratio.

The next chart takes the above ratio and ranks them by percentile. For example, the 50th percentile is a 172x ratio, which is slightly above the current 148x ratio.

The Market Cap Ratio as a Predictor of Small Cap Relative Performance

Now is where things start to get interesting... taking the ranks in the above chart and plotting the average excess small cap forward performance by starting decile bucket (i.e. all ranks 0% to 9.99% are within bucket 0... 10% to 19.99% are within bucket 10....etc...), we can see a definitive pattern. When the ratio is low (i.e. there is not a large gap in the market cap between small and large firms), there is perhaps only a minimal opportunity for small companies to catch up to large caps leading to small cap underperformance given their higher risk; when the gap is high, there is perhaps a lot of room to catch up - leading to small cap outperformance.

The next chart shows the stand-alone forward performance of small and large stocks within the same buckets as the chart above, rather than the variance between the two. Note the relative performance by market cap ratio has been driven mostly by the performance (bad and good) of small caps when the ratio is low or high, rather than the performance of large cap stocks.

Potential Implication: What if Extreme Gaps Drive the Whole Premium?

Now is where I believe things get very interesting. An investor making an allocation to large stocks in 97% of all instances (i.e. making an allocation only when the market cap ratio was north of 425x which marks the highest 3% of all instances and only includes a portion of the Great Depression and the Internet Bubble), would have received 75% of the small cap premium since 1926 with significantly less volatility. An investor making an allocation to small caps only when the month-end ratio was north of 175x (the median) would have done even better than small cap stocks alone, returning 13.1% vs. the 11.8% of small caps since 1926.

Additional analysis is needed before any claims can be made, but if investment factors generally lead to consistent excess performance only at extreme levels, it questions the benefit of having a strategic allocation to these factors irrespective of the market environment.