Wednesday, August 15, 2012

Traders Edge 8/15/12- the VIX is a warning gauge, unless it isn't

You have undoubtedly seen commentary discussing the low and declining level in the VIX index or the S&P volatility index, and the insinuation that this is a precursor for a fall in the stock market. Now these forecasters may be using more sophisticated models than I am using, but the historical evidence does not necessarily bear this out. In fact, a low level on the VIX, in a historical context, actually suggests improved future performance relative to a buy-and-hold scenario.

The price chart of the S&P 500 and the VIX since 1990 are shown below. One item that stands out to me is that in the early 1990's and mid 2000's, a low VIX level was subsequently followed much higher stock prices.


Now you know your eyes can be deceiving and maybe you think that the data would say something different. Well, you would be wrong. I created a simple trading strategy showing the performance of the S&P 500 following any day since the beginning of 1990 when the VIX was at or below 15 over various time periods. The results are presented below.

The Buy-and-Hold Base Case



1w 1m 3m 6m 1y 2y
Avg -0.1% 0.7% 1.9% 3.9% 8.1% 17.2%
Median -0.3% 1.0% 2.5% 4.6% 9.5% 18.6%
High 22.5% 22.4% 38.8% 50.2% 69.3% 93.7%
Low -16.0% -29.8% -41.8% -46.5% -48.8% -51.5%
# Up 2509 3475 3709 3903 4211 3899
% Up 44.0% 61.2% 65.8% 70.0% 77.3% 75.0%
Stdev 2.5% 4.7% 7.9% 11.7% 17.5% 29.1%
Sharpe -0.04 0.14 0.24 0.33 0.46 0.59  

The Performance of the S&P 500 following a VIX at or below 15


1w 1m 3m 6m 1y 2y
Avg -0.1% 0.8% 2.5% 5.2% 10.7% 21.7%
Median -0.2% 0.9% 2.5% 4.5% 9.1% 20.6%
High 5.5% 9.3% 11.6% 22.1% 41.1% 71.8%
Low -3.4% -6.5% -13.6% -9.1% -16.8% -49.3%
# Up 637 990 1107 1273 1319 1288
% Up 42.2% 65.5% 73.3% 84.4% 87.5% 85.6%
Stdev 1.3% 2.3% 4.0% 5.7% 10.3% 29.3%
Sharpe -0.07 0.35 0.63 0.92 1.04 0.74

Now if this simple model was an effective warning gauge, you should expect the forward price performance on the simple trading strategy to underperform the buy-and-hold case. It is apparent that this is far from the truth. In all periods except for the very short-term (1 week), the market's outperforms following a low VIX across all time frames.

So does this mean we ignore the VIX? Probably not. Maybe a low VIX in conjunction with some other (for instance the VIX in conjunction with bond yields, profit expectations, etc.) measure will indicate a market top or bottom. Maybe a more sophisticated model may provide more informational content. This analysis does not change my view that the market is in a short-term topping process, but the VIX is not one the reasons I will readily cite in my thesis.

No comments:

Post a Comment