Serving the Quantitative Finance Community

 
User avatar
woodsdevil
Topic Author
Posts: 10
Joined: March 29th, 2004, 2:12 pm

Using FX heuristics in risk management ?

August 2nd, 2012, 6:54 am

As you know it's quite common for FX exotic desks to develop their own "heuristics" (aka "pricing rules") to price first generation barriers. These heuristics are some sort of ad-hoc approximation to the value of the trade based on some sort of intuitive approach, i.e. these are not the result of self-consistent term structure models such as stochastic volatility models. These are used only for pricing, and the usual statement is not to use them for risk-management. I have some idea why, but would you be able to enumerate them please ? What's so bad to use them in computing delta,vega, etc... and hedge w.r.t. those risks ? It clearly depends a lot on the heuristics, but the general idea is all I'm after (for now).Thanks.
 
User avatar
farmer
Posts: 61
Joined: December 16th, 2002, 7:09 am

Using FX heuristics in risk management ?

August 24th, 2012, 10:47 am

These models only produce good prices when the operator has an incentive to produce good prices. If he has an incentive to create the impression of low risk, he can do that also.
 
User avatar
spv205
Posts: 1
Joined: July 14th, 2002, 3:00 am

Using FX heuristics in risk management ?

August 24th, 2012, 1:00 pm

what about sticky strike / delta in equities --- i would say there you have risk management heuristics ( eg you price everything in local vol and then say outside model that smile changes with delta...)
 
User avatar
MCarreira
Posts: 64
Joined: January 1st, 1970, 12:00 am

Using FX heuristics in risk management ?

August 24th, 2012, 3:07 pm

Demand plays a large role, e.g. customers usually buying BRL Call USD Puts RKOs when somewhat close to a level the govt seems to be defending.What I remember is that a particular pricing software provider quite popular in FX used vanna-volga and some heuristics to arrive at the bid-offer (only 25d RR and Fly, no 10d), and traders used to check it when pricing exotics, which had a kind of self-reinforcing dynamic.As vanna-volga was not used in interpolation of vanillas, and therefore was not used in risk management, then you'd just shift the barrier and store 1st day P&L for the inevitable time decay (unless the option was KO).spv205 is correct, we had RM numbers with a sticky delta calculation and we'd calculate another delta that was calculated with a co-movement of spot and smile.