Andrew Burkly, Portfolio Strategist at Oppenheimer, discusses the VIX Divergence Indicator and how it can be used as part of an equity risk model.
The VIX Divergence Indicator has had an unusually good run over the last 15 months. As a shorter-term indicator it has given multiple signals over this period and shown results stronger than its long-term average (see chart).
Any indicator which has managed to maintain a bullish signal for a long time has been a good one, given the sustained nature of the rise in stock prices over the last several years. Indeed, our Momentum Indicator has been bullish since December 2011, during which time the S&P 500 has returned 70% cumulatively (19% annualized), but it is a straightforward trend-following indicator which will by nature turn bearish only after stocks have begun to decline.
The VIX Divergence Indicator, though, has been much more active in terms of signal generation, and has performed well over the last 15 months. The indicator is designed to look for short-term discrepancies (divergences) between the level of equity market volatility expected by options traders and the level that would be expected given the market’s recent realized volatility. Thus, when implied volatility (VIX) overshoots realized volatility, it points to excessive optimism or pessimism by options traders and thus gives a contrarian signal: excessive pessimism (implied volatility is “too high”) generates a buy signal for stocks, while excessive optimism (the VIX is “too low”) generates a sell signal. The signals extend for up to three months after an extreme reading unless the opposing signal is given.
Our data shows that in the 310 trading days (15 months) since coming off a neutral signal in October 2013 (through December 31st 2014), over the course of a half dozen signals, the indicator has been bullish (as it is now) for 163 of those dates (53% of the time). Of those days that the indicator was bullish, the S&P 500 rose at an average annualized rate of 43% (0.17% per day). Bearish (or neutral) readings corresponded with modest declines in stock prices on average (-4% annualized). Of course, this recent period has been unusually good. The long-term results are less extreme: since 1990, bullish signals have corresponded to S&P 500 returns of about 23% annualized, while bearish signals have corresponded to stock returns of only about 1% annualized (neutral periods average about 8% annualized).
The VIX Divergence indicator tends to work best in periods of steady to rising market volatility with significant short-term swings, which has been the case more recently than it was for much of 2012 and 2013 when volatility was falling. To the extent that the Fed has ended its QE program and could be raising rates later this year or next year, and the bull market has gone on as long as it has, then more volatility in stock prices and in volatility itself would not be surprising.
We stress, however, that it is very difficult to find any single indicator that works equally well under all circumstances. Hence, a multi-factor approach is crucial. We also tend to avoid the temptation to try to guess which of the 10 indicators we use is likely to do best in the near future, and thus keep all 10 equally weighted in our model (see previous article).