Change in historical vs implied volatility?

Discussion in 'Options' started by Hugin, Aug 5, 2008.

  1. Hugin


    The background to the question comes from findings when developing an automated equity trading system. When we analyzed the trades generated by some of the rules we found the signals they generated coincided with breakouts in price and volatility. Now the time has come to take a closer look into it to see if this could be used for trading. Since we do not have information on historical option prices we decided to use the historical 30 days day-to-day volatility as input and also the variable to predict.

    The idea is that an increase in historical volatility would translate into an increase also in implied. Performance is measured as the ability of the system to foresee increases in historical volatility, i.e. if the system signals and the volatility increases during the following 10 day period we consider the signal to be ”good”.

    The initial results are promising but since we’re only in the early stages so before we continue with the effort it would be good to see if anyone out there has some input, especially with respect to the following.

    Is it at all possible to base the creation of a trading system on historical volatility or do you need the implied as input/output?

    To what degree does an increase in historical volatility transfer to a change in implied volatility? Will the market see the increase in historical volatility as an increased risk in the market and increase the implied? Or will the market see it as its previous fears were correct and keep the implied unchanged? Obviously this will vary from case to case so what we’re interested in is to find out whether there is any systematic problem (i.e. the answer we would like is that on average over many trades there will be a corresponding increase in the implied). Has anyone come across any papers on anything like this?


  2. MTE


    There's no direct relationship between historical volatility and implied volatility, as the former is a backward-looking measure and the latter is the forward-looking measure.

    Generally speaking, if the price breakout (increase in historical or realized volatility) was expected then implied volatility gets crushed. On the other hand, if the move is unexpected then the implied volatility jumps.

    A classical example is an earnings release or some similar event. Ahead of earnings implied volatility gets pumped up as the market expects a move, once the earnings are released the implied gets crushed, the realized volatility, though, may increase or not, depending on whether the price jumped or stayed pretty flat.

    You can use historical volatility to develop a trading method/system, but I wouldn't make an assumption that there is a direct relationship between it and implied volatility.
  3. nitro


    I am not going to answer whether it is possible or not. I will tell you that the relations between the two are critical in options trading, especially in equity options.

    I disagree with much that was said in the post immediately above this one.

  4. Hi Nitro

    I got a little idea, but feel free to share with us why ?

    Best regards

  5. nitro


    Hi MasteratWork,

    If you do the work, you will reap the benefits.


  6. MTE


    That's why we have a market, if everyone agreed then noone would trade.
  7. Hugin


    Thank you for your comments. I found some research reports related to the topic on the Internet. It seems like the consensus is that that implied is a better predictor than historical. As nitro indicates, if both are included in the trading system, at least we have the opportunity to not take trades when the relationship between historical and implied is outside our “comfort zone”. As an example we would probably se a higher probability for a successful trade if we take the trade when implied is lower than historical and we have a signal indicating a higher future realized volatility.

    Most researchers seem to be occupied with finding a good forecasting “algorithm” that provides a forecast on the future volatility every day. I found references to models based on GARCH, EGARCH as well as kernel regression and various MA models. None very successful though, which would be expected since they have published their results. So no guidance found yet.

    Next step, if we decide to go forward with this system, will probably be to find information on implied and to make a detailed analysis on the relationship of implied and historical before and after the signal to determine the edge.