Blog

Paratrade Systems Weekly Research News

8 Nov

Using Fear, Adjusted for Volatility, to Improve Timing

We often have a problem comparing one environment with another when it comes to studying fear. Certainly 2008 was an outlier but if we put it aside for a minute we can study the other so-called normal conditions where the market swings between fear and greed. Vix acts as a fair measure of fear. Our problem doesn’t go away however since Vix is bounded at the low end (around 12-13) but easily spikes up to somewhat random levels between 17 and 30.

A Fix

The first thing we’ll do is swap out Vix for a synthetic version which is bounded by zero at the low end and about 10 at the high end (it went higher in 2008). We’ll then deflate it for monthly volatility. This is important since it will allow us to look at fear relative to how much the market is moving every day. The ratio balances the calculation so we can look across time.

blog-fearvol-pic1

In this image, you can see that a few days ago fear (adjusted for volatility) was at a very high level – not that far away from points we saw in February when the market was in great distress.

A Test

Next we need to see how well this indicator works as a predictor of price. If I set a level of .15 as a sort of break point where everything above that level is “high fear” then I can use that as my buy hurdle level.

blog-fearvol-pic2

This chart shows you P&L for one day holding periods if you are long after a reading above .15. The blue line shows you the profit if you buy after the indicator is below .15. There are certainly several sustained rallies where fear was not elevated and the market rallied anyway. Even if we include those moves we can see that the cumulative profit in case 2 was negative 241 S&P 500 points! If we wait for (volatility adjusted) fear to be above .15, we make 1071 points during a period when the market rallied about 800 points.

Avoid High Risk Periods

Let’s say our goal is to be long most if the time but avoid the market when prospective gains are particularly poor. To do this we shall make one change and substitute our definition of volatility from (20- day) Standard Deviation to 20-day Average True Range. You can see the resulting filter here:

blog-fearvol-pic3

I have shown a hurdle level of .1 and you can see there are many days when the indicator is below that level. Let’s say we are allowed to own the market whenever it is above .1. We make slightly more than 1400 S&P points vs a loss of 574 for all the days when it is below that level. That is a huge discrepancy and a massive out-performance. This version may suit us better visually since the degree of opportunity is more properly reflected on the chart which may be preferable to our first completely neutralized series.

Using either filter allows us to identify buying opportunities or at least avoid high risk situations without having to handle the somewhat uncontrollable problems with Vix as a time series.

Add Comment