Monday, 14 November 2011

The VIX - a new research programme

The data bash is finished and the main statistical analysis is now done - all that remains is to write up the results. Then I will be studying portfolio management and position sizing, with the aim of letting leverage rise to perhaps 3 x the current level, with my exposure representing 75% of the days P/L.

In the meantime, I have been thinking about one category of trades that we always seem to come back to - what we refer to as "Taleb trades", trades deliberately set up with the aim of "bleeding" a small amount each day in return for an occasional huge win. This has lead me back to the VIX, the S&P Volatility Index, or the "Fear Index" as it is often referred as in the recent Robert Harris book

When the VIX is examined intra-day, it is a pretty unusual market. A typical current reading is around the 30 level, but a typical day's range can be 6-8 full points, or about 25% of the value of the index. And it has large discontinuities between the close of one day and the open of the next. My first look at trading this was based on applying out intra-day systems to daily VIX data. This was hugely profitable over the period initially tested - Oct 15 to Nov 5 - making around 1500 pts or over 50% unleveraged. However it was also hugely volatile. A whipsaw could cost 400 pts. You could be long at the close and the next day the market could open up 500 or down 500. So the mark to market was very volatile.

So despite a return of 50% in three weeks, the Sharpe ratio is only about 4.0, which is low compared to our current trading. I did try considering just the long trades, but while that did improve the Sharpe ratio a little, it was still not good enough. Maybe I will be led back to intra-day trading?

A typical one week VIX, with a high of 21 and a low of 17, with large discontinuities between close and the next opening.

Also there is a more subtle problem, which is the high correlation between the VIX trading and our trading of US equities - in other words, we already do well on the sorts of days that cause the VIX to rise or fall lots.

Finally, short term trading of the VIX is not really a "Taleb trade" at all. Maybe we should be looking at VIX options, maybe buying 30 calls whenever the VIX falls below 25, or buying the future when it is below 25 and selling it when it goes above 35.

So for the moment, I shall be looking elsewhere for a possible Taleb Trade

No comments: