Serving the Quantitative Finance Community

 
User avatar
shunvwu
Topic Author
Posts: 0
Joined: December 12th, 2010, 7:34 am

Delta in FX option

May 4th, 2011, 1:09 pm

Hi,I've some questions about volatility in FX option. When I download the volatility data from Bloomberg, it is represented in terms of delta, with ATM which meaning 0.5delta call. In the textbook the FX option delta=exp(-rf*T)N(d1), thus for ATM it should be 0.5= exp(-rf*T)N(d1). N(d1) has a range of [0,1]. When T is quite long, it is probably that exp(-rf*T)<0.5, then there is no way that exp(-rf*T)N(d1)=0.5. I?m quite confused about that. Could anybody point out the problem included in this reduction? Thanks a lot!
 
User avatar
caperover
Posts: 49
Joined: June 25th, 2005, 4:54 pm

Delta in FX option

May 4th, 2011, 4:44 pm

Shouldn't it be delta=exp(-q*T)N(d1)If exp(-rf*T)<0.5, they won't quote using 50delta. FX option vols are only quoted up to 2Y or 3Y depending on liquidity.
 
User avatar
Leonidas
Posts: 0
Joined: June 6th, 2007, 7:53 am

Delta in FX option

May 4th, 2011, 7:18 pm

Read Reiswich and Wystup. And ignore caperovers remarks
 
User avatar
Venezia
Posts: 0
Joined: April 27th, 2011, 9:58 pm

Delta in FX option

May 5th, 2011, 10:48 am

Market convention is to define the ATM at the currency pair level. For emerging market currencies it usually means ATMF. For G10 pairs it means delta neutral ie the strike at which a straddle has zero delta. Delta is defined also on a currency pair basis. For XXXUSD pairs, it is the amount of XXX to hedge USD value. For all other XXXYYY pairs, it is the amount of XXX to hedge XXX value.
 
User avatar
caperover
Posts: 49
Joined: June 25th, 2005, 4:54 pm

Delta in FX option

May 5th, 2011, 2:04 pm

QuoteOriginally posted by: LeonidasRead Reiswich and Wystup. And ignore caperovers remarksWhy should my comments be ignored?No matter what you read, you will end up with the formula as mine.
 
User avatar
Venezia
Posts: 0
Joined: April 27th, 2011, 9:58 pm

Delta in FX option

May 5th, 2011, 2:06 pm

To clarify further, for longer dates ATM is ATMF for all currency pairs; and you also have the issue of spot vs forward delta. It's fairly convention-heavy.
 
User avatar
Leonidas
Posts: 0
Joined: June 6th, 2007, 7:53 am

Delta in FX option

May 5th, 2011, 9:07 pm

There are three reasons why I said yourremarks should be questioned. You said it shouldbe q instead of rf which shows you are not working in FX. You said FX options are onlyquoted up to two or three years, which showsyou're not working in the market. And the <0.5thing...don't know what you're talking about. There are conventions in the spirit of Venezia'sremarks.
 
User avatar
caperover
Posts: 49
Joined: June 25th, 2005, 4:54 pm

Delta in FX option

May 6th, 2011, 2:45 pm

well, then you tell me what should be the delta for a option on a stock with dividend? Isn't foreign currency interest rate equivalent to dividend rate in BS formula?As mentioned in "Consistent Pricing of FX options" by Castagna and Mercurio, "The FX option market is very liquid, up to relatively long dated expiries (2 years,at least for the EUR/USD exchange rate)." 2Y is already considered to be long-dated. In which market do you see liquid 5Y FX options?If you don't know <0.5 thing, what are you doing here? Someone asked for help on A, you gave him an answer to B? You should really read others' questions and comments before you start.
 
User avatar
Venezia
Posts: 0
Joined: April 27th, 2011, 9:58 pm

Delta in FX option

May 6th, 2011, 2:51 pm

A typical short dated FX Options desk will trade the majors out to 5Y with liquidity dropping off from about 2Y. Beyond that your long dated FX/hybrids desk will take over. They may well trade vanilla options out to the 20Y tenor, thought the market is significantly less liquid out there.
 
User avatar
shunvwu
Topic Author
Posts: 0
Joined: December 12th, 2010, 7:34 am

Delta in FX option

May 18th, 2011, 12:56 pm

It's my fault to consider ATM 0.5 directly without thinking about its meaning. The ATM delta actually means delta neutral straddle, not delta 50, which equals to 0.5*exp(-rf*T).Thanks for all of your reply!