recently i've been looking into volatility arbitrage in the index options markets (Nikkei 225) as part of a vol forecasting (MSc) project that i'm just starting. based on conversations, emails, a quick search of the forum, etc. vol arb seems to be of significant interest in those markets. that's why my initial focus is on index options.was wondering, though - how about in the fx options market (talking G7)? is vol arb significant there as well? why or why not?feel free to PM me, if you'd rather.

Remember that the fx options market is huge and very liquid - daily notionals are around 100 billion USD/day, which is roughly the same as the daily notional of equities traded globally (not equity options - equities).A lot of market inefficiencies have been ironed out because of that.On the other hand, bid/offer spreads are super-tight compared to equity options markets (~0.2 vols vs ~3 vols for S&P500 index options), so if you have a view on relative value it doesn't cost you much in transaction costs to put on a position.

In addition to mghiggins excellent post on the subject, the FX options markets also have many forms of liquid exotic options, which might offer additional opportunities for vol arbitrage to the astute trader. It can be done through a combination of simple product with complex options or complex with complex options. But the good part is that most options are liquid and the spread is usually tight (depend on the product of course).

kapital asks "why?" One reason for the liquidity in FX options is that the FX market is nearer to Black-Scholes "ideal" market. It trades nearly 24 hours, bid/ask spreads are tiny, liquidity is very deep, and the stochastic process is nearly a brownian motion (ie there are few jumps). IMHO none of these factors are true of equity markets!

Indeed the wholesale FX option market is so close to the B-S ideal that it quotes it's levels in vol rather than premium. Traders then convert these vol levels to premium when a trade is closed. This is only possible with broad agreement over models.

One reason for the liquidity in FX options is that the FX market is nearer to Black-Scholes "ideal" market. It trades nearly 24 hours, bid/ask spreads are tiny, liquidity is very deep, and the stochastic process is nearly a brownian motion (ie there are few jumps). IMHO none of these factors are true of equity markets!q1: when you say that it's close to the black scholes "ideal" market, do you mean that vol is pretty much constant over time? it doesn't tend to vary much with the level of the currency or "cluster", shifting from low to high (except in a tight range around one number)? you don't have the vol skew?...quotes it's levels in vol rather than premium. Traders then convert these vol levels to premium when a trade is closed. This is only possible with broad agreement over models.q2: so what are the standard models in FX?q3: i've read some (fairly recent) papers by Andersen et al. which demonstrate successful vol forecasting over short time periods in the FX markets. just makes me wonder, why aren't there traders out there picking up on the idea and trading it until the market is efficient again? i when i asked my original question, i sort of expected to hear "oh, yeah, that old trick. well, we (or someone we know) traded the hell out of that n-years ago, but it's all out of steam now."

The standard pricing model that I mentioned is an off the shelf system called FENICS , the company is now owned by an interdealer broker GFI. I believe that the precious metals options market also uses the same system as its standard too. Obviously traders will have their own in-house models running too but the FENICS model is close enough for most purposes.

To detour quickly,Would fx Options be a good place to move ones career ? is there interesting work ? Back to the main topic,Do you have any links to those papers Kapital, i would be interested in reading them.

FX is cool from a quant perspective because it's such a mature market: it's one of the few places where you can work with things like stochastic vol models and traders actually care (because the bid/offer spreads are narrow and OTM options trade liquidly - as do simple exotic barrier options like knockouts and one touch binaries).OTOH, less mature markets are also fun because there are no standards: you make it up as you go along.

I answered my own question I think, here is a link to some of Andersen's papers AndersenThanks mghiggins. I may consider staying in FX a while longer. I think as i learn more and get to apply it in a market that more closely approximates the assumptions, i will be able to move into less developed markets.

Originally posted by: kapitalq1: when you say that it's close to the black scholes "ideal" market, do you mean that vol is pretty much constant over time? it doesn't tend to vary much with the level of the currency or "cluster", shifting from low to high (except in a tight range around one number)? you don't have the vol skew?It is true that in FX, vol is not very constant over time. However * the absolute range of implied vol levels is quite small compared to other asset classes* the skew is not very pronounced and does not have a consistent directional bias * vol doesn't tend to vary with the level of the FX rate* vol doesn't tend to cluster

Has anyone come across creating synthetic fx options from buying and selling vol on an interlisted name? How did you find was the best way to measure and manage the correlation risk?For example buying vol in market A in USD and selling the ADR in Market in Yen for example..both options are identical with maturity and strike, however, the trader is left with an fx position that behaves like an fx option.Any thougts? (most traders I believe just assume zero correlation?)Currently, I run a correlation between the USD asset price and the USD/Yen...and apply that to the current implied fx vol from a similar fx option contract to help adjust for vols differences between to the options...Any feedback would be niceSam