wilmott.com

Hi guys!Could you please provide some insights on what approaches are used in hedging equity portfolio over trading horizon of 5-15 days? It seems simple beta hedging of equity positions with index futures does not work over such a short trading horizon. Are there any other measures of stock-to-market relationship that could be efficiently used in hedging apart from betas?Thanks a lot!Tiger

No matter the time horizon, beta is still, by definition, the ratio of the expected return on the value of a asset/portfolio versus the return of the index, or viewed another way, the statistical "delta" of your asset/portfolio to the factor of index returns.The problem, as you probably know, is that such short time horizons are not long enough to rely on the "average" used in beta which the noise around the beta will wash out, especially if you have a low R^2. If you try to compute rolling beta on such short windows, you will notice it fluctuate as wildly as stochastic vol and other highly exciting charts, which doesn't mean too well for this type of short-term "hedge".Some techniques I have used on similar problems include some representative sampling of a small selection of liquid stocks and indexes to try and get a higher R^2 to the portfolio I am trying to hedge. Even then, I generally have a horizon of 6 months or more. Some other things you might consider, but might still be long stretches, include correlation swaps, or representative sampling of a set of options. Hope this helps...

Some time ago I have used the same techniques as you did to compute my hedge ratio. The bad news (mathematically speaking) really is that over this very short time horizon your signal-to-noise ratio is rather low, however, improves over time. In other words: if you hedge for (e.g.) 3 month, it works rather fine, on shorter time horizons, it doesn't. I do not think that adding more and more futures will really improve your hedge. Of course, you will have in-sample effects, but, unfortunately, the out-of-sample effect is zero. Even using robust estimation of beta does not help much.in short: if you don't find any better product to hedge, you're lost. sorry.

Hi guys, could u explain how to estimate beta (or another statistics) for a short-term index futures cause no enough market price exists for it.

What's been said is accurate. But your problem may be with your equity portfolio.If you have a, say, 5 stock portfolio then your firm specific risk is fairly large relative to your market risk. You hedge the market risk but are left with still volatile returns.If your portfolio is diversified then you hold market risk and very little firm specific risk. Once you hedge this market risk by calculating the portfolio's beta and shorting the market your hedge should work reasonably well (even over 5-15 days). So, to see if this may be the issue, does your portfolio have < or > 20 stocks?