When the day's action is over and VIX futures have settled I will look at various VIX futures metrics (found on the VIX Futures Data page) and make a determination if I should buy, sell, or hold various VIX ETPs. To better communicate my outlook on these products I've rolled up most of the metrics I use in making decisions into Daily Forecast gauges.
I first introduced these gauges last month. Today I've expanded the daily forecast to include ZIV and I've added some charts to track the daily forecasts against the actual performance of the securities.
For the VXX forecast there are two components to the VXX forecast - the Bias and the Spike Risk.
This gauge measures the bias of VXX and is largely determined by the roll yield (more on the roll yield in this post), which can be conceptualized as a headwind or tailwind. The gauge values are on a scale from -10 to +10, with a more negative reading indicating a stronger negative bias (headwind), and a more positive reading indicating a stronger positive bias (tailwind).
May 3, 2013 Update: Note that the gauge's range was expanded yesterday to cover +4 and -4 standard deviations from the past 9 years of data for a total of 8σ (previously was 6σ). This tweak to account for a wider range of values has the effect of making the numbers on the scale smaller than they were previously but everything is still on a scale of -10 to +10. The neutral zone is now at +/-1.
March 2, 2014 Update: The Bias indicators now include inputs for the direction of short-term price movements. See this post for details.
VXX Spike Risk
The second forecast gauge reflects the probability of a sharp spike in VXX in the very near term (1-2 days). As inputs, it takes data from the term structure of the front two months of VIX futures, the premium of front month futures over (or under) spot VIX, historical volatility, and recent moves in spot VIX to identify various risk conditions. All risk conditions are then weighed together to provide a single output to represent the total risk. The higher the reading, the greater the risk of a large magnitude spike.
The gauge's output is a scale of 0 to 100%. Values below 25% indicate a low risk of a spike in VXX. Values between 25% and 60% are moderate risk. Values above 60% are high risk. Typically any reading above 50% is cause for some concern.
Using the Bias and Spike Risk Gauges Together
The Bias gauges provides a view of the direction of VXX and ZIV over a medium timeframe. There may be occasional moves in the opposite direction, but over the course of several weeks, the price of VXX will typically move according to the Bias of the gauge. Red means it's going down, green means it's going up, and yellow means sideways and generally choppy conditions.
The Spike Risk Gauge gives information about the probability of moderate or large spikes on a more short timeframe (1-3 days out). When long XIV (or short VXX/UVXY), this can serve as a signal to make sure you are adequately prepared for such a move just in case, i.e. reduce positions, exit positions, buy protective VXX calls, set stops, or maybe just get mentally prepared to see the value of your holding go down.
Obviously there are market/world events which will cause VIX and VXX to spike dramatically without warning. While this model cannot predict the unexpected, it does provide the ability to detect shifts in investor sentiment of known macro risks.
ZIV Bias Forecast
The ZIV forecast only has a bias gauge because the VXX spike gauge can be applied to ZIV. The differences being that ZIV will move in the opposite direction and generally at a smaller magnitude. Similar to the VXX bias gauge, the ZIV bias gauge estimates the direction and strength of ZIV based on the shape of the term structure, but also includes an adjustment based on the shape of the front end of the curve under certain conditions. The scale is also from -10 to +10 with the neutral zone at +/-0.5. As you can see below, the reading based on Thursday's forecast of 2.7
Tracking Forecast Results
To give this some context I've included a six month view of data comparing the values of the forecast to the price of ZIV at the bottom of the Daily Forecast page. For an extended look at the ZIV forecast, below are the results from April 2011 to present (forecasts on the left axis; ZIV price on the right axis (log scale)). Generally long ZIV is a good play when forecast values are in the green zone on the bias gauge (above 1).
Similarly for VXX, the 2 year chart shows that being short VXX when the bias is below -1.0 (red zone on the gauge) and long VXX when the forecast is above 1.0 (green zone on the gauge) are very good trades. Note the choppiness in the price of VXX when the forecasted bias in in the neutral zone (forecasts on the left axis; VXX price on the right axis (log scale))
The 2-year chart of Spike Risk vs VXX % Change is a bit harder to read on this time frame but I've included it below anyways. A well-formed model will have a reading of 50% or higher (left axis) prior to any significant daily spikes in VXX (right axis). You can alternatively view the 6 month chart on the Daily Forecast page for a cleaner view on a shorter time frame.