Search
× Search
Tuesday, November 19, 2024

Archived Discussions

Recent member discussions

The Algorithmic Traders' Association prides itself on providing a forum for the publication and dissemination of its members' white papers, research, reflections, works in progress, and other contributions. Please Note that archive searches and some of our members' publications are reserved for members only, so please log in or sign up to gain the most from our members' contributions.

Enhancing volatility predictions using options trading variables

photo

 Daniel Bencík, Energy Trader ve spolecnosti Erste Energy Services

 Wednesday, February 18, 2015

Dear forum, for a research project, I am interested in how one can improve upon a volatility forecast. Usually, volatility modeling/forecasting is being done using lagged volatility, the currently most common formula is that the forecast of volatility depends on previous daily, weekly and monthly volatility (HAR model of Corsi). However, my hypothesis is that this can be improved by some forward looking indicators that I would take from the options market. For example, if the open interest in both puts and calls increases with a negligible change of the underlying close price, then it’s reasonable to expect that the increase in demand is driven by an expectation of larger future volatility. My dataset contains intraday data on the EUR/USD futures over a couple of years. On top of that I have EOD data on options (close, volume, open interest across all strikes and puts/calls) where the underlying is the EUR/USD futures. My question is … what kind of option trading variables would you use if forecasting volatility? In other words, how do volatility traders usually behave when they expect a rise/fall in volatility? Any input is much appreciated, Daniel


Print

4 comments on article "Enhancing volatility predictions using options trading variables"

photo

 Andrey Gorshkov, Algorithmic Trading Analyst, C++ Developer

 Thursday, February 19, 2015



As an offtopic note: looks like in your case a positive mutual feedback possible in the volatility estimation and the open interest as more and more traders will thus buy/sell options according to the estimation growth.


photo

 Nikolay Stoykov, Managing Member at Annapolis Fund

 Friday, February 20, 2015



First of all, the eurusd market tends to be OTC. Not sure what open interest you have in mind - the etf or futures but that data is really not representative of the dynamics of implied vol. So, building any models of partial data is not a good start. Even if you get some data from dealers, again this better be a big dealer, otherwise again only partial.

Second, the idea that open interest is a good input is dubious at best. Yes, you want to see net buying/selling of volatility but open interest can be misleading. Spreads, can increase open interest dramatically while in effect only a small net vega buying.

Finally, you need to incorporate the curve. It is not just buying or selling vol but how far out.The relative cheapness/expensiveness of near vs far options is a very important part of the equation. But I really dont trade currency vol, so cant help you too much more..


photo

 Daniel Benčík, Energy Trader ve společnosti Erste Energy Services

 Friday, February 20, 2015



Thank you both. The open interest I was referring to is related to the option contracts per se, not the underlying.


photo

 Johan K., Senior Investment Manager at Deutsche Bank

 Saturday, February 21, 2015



Let me first say that I have limited experience in FX vol trading. However, as I'm sure you are aware and what Nikolay pointed out above, it is quite difficult to get a grip of the data since the market is OTC.



At desks in large banks it is possible to aggregate a "Gamma Report" of the deal flow that runs through the banks FX options desk. If the bank has a large enough market share of the trading in a specific currency, this can hint of what the aggregated sensitivity is, and give some input to ex-ante direction of the cross and the volatility of the same.



I see that you already use the HAR model ( http://homepage.sns.it/marmi/lezioni/corsi-pisa-2010.pdf ).



In equity it is a bit easier as data are easier to obtain. There I have used relative pricing, like 25 -delta RR as input in forecasting direction and there by volatility as these are highly negatively correlated. These kind of relative option prices show a mean reversion pattern and it is therefore straight forward to find critical values and probabilities for a turn around.

Please login or register to post comments.

TRADING FUTURES AND OPTIONS INVOLVES SUBSTANTIAL RISK OF LOSS AND IS NOT SUITABLE FOR ALL INVESTORS
Terms Of UsePrivacy StatementCopyright 2018 Algorithmic Traders Association