Volatility Arbitrage

Discussion in 'Trading' started by man, Nov 19, 2002.

  1. man


    Little more than a year ago I started some research in a relative volatility play. The idea was to take a sector of rather similar stocks, look at their implied volatilities and trade the top quintile against the lowest quintile by means of strangles. The basket was US banks and all tests indicated good and stable returns, even after assuming reasonable slippage. We used data from ivolatility and started trading. We closed the strategy two months later with a moderate loss, since we had to accept that tradingcosts for rehedging, rolling etc accounted for about 4-5 volapoints. Since our two baskets of stocks differed in their implied volatilities only by about 8 to 10 volapoints, there was no edge.
    I still wonder if our problem was the strategy as such or solveable things like data quality. With options it is always difficult to build strategies, since accurate data is not easy to get. Today we use optionvue, but i still feel that finally we have to save data on his own. If you buy data which averages between expiration dates you are inevitbale biased, believing there is a trade, although the market has already priced it in.

    From my perspective there might be opportunity for that kind of strategy, although I do not anybody really trading such a thing. Anybody experienced with that kind of trading?
  2. It is you data and reliance on an outside model probably ...
  3. Perhaps you could apply a trading model to the underlying that takes advantage of the relative volatility.

    Trend following systems generally work better on markets that move from a low vol state to a high vol state. A few ideas could be along the following:

    1. Apply a trend following system to the banking index, taking signals in the direction of the trend with the bottom quintile and hedging with opposite position in the top quintile.

    2. Rank the shares in the sector by relative performance. For the shares in the lowest quintile (by vol), buy the strongest shares and short the weakest shares. For the shares in the highest quintile (by vol), short the strongest shares and buy the weakest shares.

    The idea of the above is that when the index trends, the low vol group will outperform the high vol group. When it is choppy, profits from fading the trend in the high vol stocks will offset the losses in the low vol stocks.

    I haven't tested the above, but it may be worth experimenting with a little.
  4. nitro



  5. maglia rosa

    maglia rosa Guest

    You were betting that the vol spread was converging to zero (or at least smaller than the existing 8-10 vol spread at the time).
    Maybe the real vol spread between the two baskets is not zero though; if a spread of 8 vol is really fair value for the spread, then you have no edge even before hedging/rolling/execution costs.
  6. man


    maglia rosa,
    agreed that could be and we looked at that, since it does not make sense to trade high vol stocks against low vol stocks, if they never change their position. we looked for each stock not in terms of the basket vol, but in relation to this stock's history. So we shorted the stocks, meaning their options, that were at highs in their volacone and vice versa.
  7. man


    worth looking at it in more detail. thanks for your thoughts.