A defining characteristic of minimum volatility (or min vol) portfolios is that they seek to help smooth out the market’s peaks and valleys by holding low beta stocks — stocks that don’t typically exhibit excessive amounts of volatility compared with the broader market.
After I explain this to clients, I inevitably get asked this question: “How does this low-beta effect behave over time?”
It’s a fair question. Unlike a typical cap-weighted benchmark or other well-known “market effects” such as value and momentum, there are no “live” indices or funds to look at for a quick answer. We can, however, get a flavor for the behavior of the low-beta effect with a simple experiment.
As I’ve explained in previous blog posts, a min vol portfolio tends to overweight low beta securities compared with a typical cap-weighted portfolio. So, let’s look at how low-beta stocks have performed over time.
To conduct this simple experiment, I looked at the stocks in the S&P 500 index and calculated their betas to the market every month since 1959. Why 1959? That is the first full year of the S&P 500 index, making the return of each of these stocks straightforward to collect and easy to compare with the overall return of the market. In addition, I broke out the time periods 1980-2011 and 2001-2011 to get a view of more recent periods in the data.
From here, the experiment is simple: Every month, I took the lower half of the S&P500 stocks by beta and computed their average return. The table below summarizes the results:
*Risk is defined as the standard deviation of excess returns. The data in the table is annualized from monthly returns.
Keep in mind that this experiment did not benefit from the full process of actually constructing a min vol portfolio. But it was able to demonstrate that over the 52-year period, low-beta stocks performed better than the overall market when it came to their level of return per unit of risk. (Remember that this was over a period when the market was up, on average, about 5.2% over the risk-free rate.)
It also helps to demonstrate why investors can view min vol as a core, long-term holding, rather than seeing it solely as a means of “downside protection” when overall markets are performing poorly.
Index returns are for illustrative purposes only. Indexes are unmanaged and one cannot invest directly in an index. Past performance does not guarantee future results.
Is there any research that documents the proportion of minimum variance excess risk-adjusted returns attributable to portfolio-level effects (covariance etc), and the proportion attributable to the low-volatility securitities themselves (absence of benchmark chasing buying pressure erc)?
This question could be reframed as, “Why USMV over SPLV?”
Pingback: Wednesday links: go read a book | Abnormal Returns
great article. what would you consider “low” beta under .7? .5?
Dan Morillo: Depends on the time period, as the dispersion of betas changes over time. Simple scenario analysis suggests, however, that the bottom quarter by beta for US large-cap stocks tends to be in the .7 range.
Daniel Morillo: I am not aware of publicly available work that does exactly this. The vast majority of academic work tends to do simple sorts because this makes reproducibility of results easier and they have fewer moving parts to think about when attempting to do broad empirical work. The disadvantage, of course, is that simple sorts do not capture the nuances of portfolio construction that are present in the vast majority of realistic portfolio implementations, from the exact nature of the risk model used to shorting to cost constraints. My own work and simple empirics on the available data (mostly some of the indices that various providers have been running on these sorts of strategies) suggest that correlations matter — meaning that the effect is generally stronger when minimizing total portfolio risk rather than just individual security risk. I believe this is reasonable given that one of the more solid explanations for the effect may be linked to the delegated asset management effect, which is about beta (where correlations matter), not risk.
Pingback: Hand Me Downs | Prudent Trader