On Dec. 31, 2016, implied volatility priced into S&P options, or the VIX, was above 14%. As of Sept. 30, 2017, the VIX has dropped to a near record low of 9.5%.
Active managers may love the new low-risk environment with its low stock-level correlations, which may be conducive to stock picking; however, it also means that to achieve 2-3% active risk in this environment, one needs to take on more aggressive active positioning versus what was necessary to achieve similar levels of active risk at the beginning of the year.
In other words, active managers stick their neck out to outperform the benchmark, but today’s low-volatility environment means active managers must stick out their necks even further to achieve similar levels of outperformance.
The low-risk environment has spilled over into risk modeling forecasts. Risk model forecasting is primarily influenced by the risk environment of the last six to 12 months. Over the long run, risk tends to be easier to forecast than returns due to the persistency in risk trends.
However, risk-model forecasts miss the market during inflection points—they can understate risk during a period of rising risk, and vice versa, until the models have had enough time to recalibrate to the new risk regime.
Figure 6 displays the same U.S.-focused multifactor ETFs profiled in the January ETF.com article but with active risk forecasts as of Sept. 30, 2017.
Figure 6: Active Risk Profiles of U.S.-Focused Multifactor ETFs (as of 9/30/2017)
Now readers will note the magnitude of the drop in active risk forecasts from beginning-of-year forecasts (some as much as half of the beginning of year forecast).
Admittedly, this should be a nervous time for quantitative-based approaches based on these lower risk estimates. We intuitively know that the risk model is likely understating active risk, particularly if we see a pickup in volatility.
Yet if the low risk environment persists (which it has for the better part of the year), then multifactor ETFs should be positioned to deliver two-thirds to one-half the excess return than what was forecasted at the beginning of the year. Presumably, the underlying expense ratios wouldn’t drop to reflect the lower levels of active risk.
The alternative is to crank the portfolio optimizer to ‘11,’ but this would entail taking on greater factor risk to achieve higher forecasted tracking error, which could end up burning investors should the risk regime materially spike to higher levels of volatility. It’s a conundrum, but a more prudent approach would be to just stay the course rather than cranking up factor risk to achieve higher levels of tracking error.