FolioBeyond Equity Volatility Algorithm
Strategy Summary
The Equity Volatility model is based on the premise that we must model a consistent inefficiency in the market. In other words, we want something that is less likely to be arbitraged away, should it become popular. In general, this means that the derivatives of variables are more likely to produce consistent results rather than the variables themselves, or a change in a variable is more consistent than the variable itself, as a future predictor. Even more consistent is a second derivative, and the most common one is volatility.
Volatility in the equity markets is clumped: periods of high volatility follow periods of high volatility and periods of low volatility follow periods of low volatility. This observation can be used to construct “volatility managed portfolios”. One of the first such attempts was by Tyler Muir and Alan Moreira (https://www.nber.org/papers/w22208)where they showed that “Volatility timing increases Sharpe ratios because changes in factor volatilities are not offset by proportional changes in expected returns. Our strategy is contrary to conventional wisdom because it takes relatively less risk in recessions and crises yet still earns high average returns.”
With this as a baseline test, we constructed a more sophisticated version of a volatility managed portfolio. A range of ETFs spanning the market was chosen. A formula was developed to convert the volatility of each individual ETF into a “weight”. Crucially, the “weight” of the ETF depends only on itself, and NOT (unlike say risk parity) on the performance of any other asset. Each ETF is independently run through the formula and a weight for each ETF is generated. This sum is then subject to whatever constraint is relevant to the portfolio — for instance, if leverage of 10% is allowed, then this sum can range from 0 to 1.1. If no leverage is allowed, then this sum can range from 0 to 1. The sum of the weights is the gross exposure of the portfolio.
Each month, the “weights” for each ETF are calculated via the formula, subjected to any necessary constraints, and then used as the basis for calculating the exposure of each ETF in the portfolio (with each ETF varying independently of all the others). The historical backtested simulation of FolioBeyond’s core equity volatility model consistently outperforms benchmarks - both the S&P 500 and MSCI’s ACWI (“ACWI”). The outperformance versus ACWI is more robust given the dominance of US Equities over the past decade.
Our historical backtested simulations are based on the following assumptions:
Rebalancing is done monthly.
We utilize five major component ETFs spanning different market segments globally.
The allocation model is based on volatility clumping which is not correlated to forward-looking returns.
We utilize a proprietary weighting method based on volatility levels with total exposure ranging between 0% and 110% for moderate risk/intermediate holding periods.
This same technology can be applied to sub-sectors of the equity markets to optimize for higher returns subject to specific constraints. For example, we can run a customized analysis versus MSCI’s EAFE index. Simulations would include the major countries represented in the EAFE Index: U.K, Germany, France, Italy, Japan, Australia, Hong Kong, and Singapore.
