The following topics cover common situations that exotic FX derivative traders come across. Issues around risk management and pricing are investigated, along with how exotic derivatives are used within FX hedging and investment strategies.
Exotic FX derivative contracts are primarily risk managed using the same Greek exposures (delta, gamma, vega, etc.) as vanilla FX derivatives contracts. Therefore, exotic and vanilla FX derivatives in the same currency pair are often risk managed within the same trading position.
When vanilla and exotic contracts are risk managed together it is important that their valuation and Greeks are aligned as closely as possible. Where possible, exactly the same volatility surface (including, e.g., ATM event weights) should be used for all options. If it isn't, risk management becomes more challenging, particularly when exotic risk is hedged with vanilla options close to maturity.
The main additional complication when risk managing exotic FX derivatives comes from barriers (both European and American) close to maturity. For exotic FX derivatives traders the key decisions come around how to hedge barrier risk. There are essentially two possible approaches:
In practice, traders use both approaches. Prior to expiry, when the risk on the exotic is mainly vega-based and Greeks evolve fairly smoothly, vanilla hedges can work well. However, coming up to expiry, if the barrier risk is large and spot is fairly close to the barrier level it may be preferable to try and hedge the barrier risk itself rather than transacting increasingly large vanilla hedges.
The bid–offer spread on exotic FX derivative contracts comes from two primary factors: volatility surface exposures and barriers. The most important volatility surface exposure is often vega.
The vega exposure on exotic options can be multiplied by the ATM volatility spread to the deal expiry to obtain the exotic bid–offer spread from vega. For example, if a 1yr knock-out barrier option has 0.20 CCY1% vega and the 1yr ATM volatility bid–offer spread is 0.4%, the bid–offer spread from vega on the exotic would be 0.08 CCY1%.
This approach keeps an important link between exotic and vanilla bid–offer spreads, but note that exotic contracts may have zero vega at current spot but significant vega exposures in the wings (for example, a European digital option with digital level at the forward). Bid-offer spread from exposures to the volatility smile could either be quantified using rega and sega exposures or volga and vanna exposures could be used via a VVV-esque methodology (see Chapters 12 and 18).
American and European barrier risk within exotic options must be individually considered in terms of the barrier size (using intrinsic value), the equivalent one-touch bid–offer spread (for American barriers) or European digital bid–offer spread (for European barriers), and the probability of the barrier needing to be risk managed (i.e., less bid-offer spread should be taken on barriers very far from spot).
If the exotic option contains a delta gap at an American barrier, that should also be included in the bid–offer spread. As discussed in Chapter 23, barrier delta gaps can lead to slippage, which increases P&L volatility. Additional spread must therefore be charged to cover this risk, perhaps using expected slippage multiplied by the barrier delta gap size.
Having said all this, traders quickly learn that summing the bid–offer spread from all these different elements of an exotic trade usually results in a bid–offer spread that is wider than can be shown in practice. For this reason, certain elements are often omitted or a scaling factor is applied at the end of the calculation.
For a trader it is most important to understand the exotic bid–offer spreading methodology used by the desk pricing tool, keep any input parameters (e.g., one-touch or European digital spread grids) updated, and work with quants to ensure the spreading methodology remains appropriate as market conditions change.
The interbank broker market is the primary source of liquidity in exotic FX derivatives because there is no direct market (i.e., traders do not directly call each other for prices on exotic contracts). In the interbank broker market, it is possible to get prices on almost any exotic contract but the most commonly traded contracts are knock-outs, reverse knock-outs, one-touches, and double-no-touches. Volatility swaps, variance swaps, and correlation swaps are also quoted fairly frequently. The majority of trading in the exotic interbank broker market takes place over chat systems and by voice.
Exotic contracts in the interbank broker market are generally traded vega hedged. Within the AUD/USD broker details in Chapter 18, the letters “vh” indicate that the contract is traded with a vega hedge; that is, when trading the exotic option, an appropriate notional of ATM to the option expiry is also transacted (priced at the implied volatility level used to generate the exotic contract TV) such that the initial vega on the transaction is zero. This means that, for small changes in the ATM implied volatility, no significant P&L will be generated in the trading position. Hedging the vega in this way helps isolate the exotic (i.e., barrier) risk within the contract.
Some exotic contracts are priced and traded with a rho hedge, denoted “rh.” This involves trading different notionals of spot and forward contract to delta hedge the trade (rather than either a spot hedge or forward hedge) such that the initial rho exposure on the transaction is zero. Again, this is done to isolate the exotic risk on the contract. Rho hedges are particularly relevant on long-dated exotic options.
Some exotic contracts are priced and traded with delta exchange, denoted “dx.” This involves the two counterparties within an American barrier transaction agreeing that when the barrier level hits they will exchange an agreed notional of spot at the barrier level. This agreement reduces the amount of spot both parties need to transact in the market and therefore reduces slippage and P&L volatility. Delta exchange is particularly relevant on exotic options with significant barrier delta gaps.
FX derivatives contracts are often combined into strategies that clients use to hedge their FX exposures.
Corporate clients could simply use forward contracts to hedge their future FX flows. However, by using options, particularly exotic FX options, hedging strategies can be customized to closely match the client requirements. In general, clients either want to beat the forward rate, in which case they must accept some additional risk, or are willing to accept a worse rate than the forward in return for some additional potential upside.
There are numerous FX hedging strategies available, many of which are covered in Uwe Wystrup's book: FX Options and Structured Products (John Wiley & Sons, 2006). One of the simplest strategies is a forward extra. Within a forward extra contract the client has the right, but not the obligation, to buy a currency in the future (at maturity) at the pre-agreed strike providing spot hasn't touched a barrier level. If the barrier level trades at any time, the right becomes an obligation.
A forward extra to buy EUR and sell GBP is constructed using two separate derivatives contracts with equal notionals:
If spot at maturity is above 0.8000, the client will use their EUR call option to buy EUR/GBP at 0.8000. If spot at maturity is below 0.8000 (and spot has never traded below 0.7500), the client can buy spot in the market at the prevailing rate. If the 0.7500 barrier has knocked, the client will, again, buy spot at 0.8000.
Clients can choose to enter this forward extra strategy rather than entering into a EUR/GBP forward at 0.7900. In this example, the client would be accepting a worst-case hedge rate worse than the forward (0.8000 versus 0.7900) in order to potentially buy the currency at market rate between 0.7500 and 0.8000 if 0.7500 never trades.
The client's short FX exposure is shown in Exhibit 25.1. The payoff of the forward extra strategy is shown in Exhibit 25.2 and the client's FX exposure plus the forward extra strategy are combined in Exhibit 25.3.
The strike and barrier levels within the strategy are often set such that the package is transacted for zero premium. Within hedging strategies there is invariably a balance to be struck. In this case the worse (higher) the guaranteed hedge rate, the more the client can benefit from favorable spot movement down to a better (lower) barrier.
Structured forwards often have one bought leg and one sold leg, so the vega risk is partially offset. This EUR/GBP forward extra is net short vega for the client and hence long vega for the trading desk. This is common; clients often net sell vega within hedging strategies. It is easier for many clients to believe that the current market situation will continue, rather than any particular scenario occuring. Intuitively, the client wants a less volatile spot because this makes it less likely that the right to buy at 0.8000 becomes an obligation (i.e., it is less likely that spot hits the barrier).
FX derivatives contracts are also often combined into investment products. The simplest investment instrument is a structured deposit, in which the client deposits money and rather than receiving a guaranteed coupon based on the money market interest rate, the coupon is instead generated by an embedded FX derivative contract.
When interest rates are high in the deposit currency, it is common for the deposit to be principal protected. This involves taking the interest earned on deposit up front and using it to purchase options that may pay out more at maturity. The client cannot lose their investment, although they have forgone the interest they would otherwise have earned. As an increasing amount of principal is put at risk, increasingly large option notionals can be purchased.
For example, a client deposits AUD1m for one year within a structured FX deposit. Within the deposit, the 5% AUD interest rate should earn approximately AUD50k at maturity but this amount is present valued and used to purchase a 1yr European digital option that pays out AUD100k (i.e., approximately 10% coupon on the deposit) if spot at maturity is higher than the initial spot rate.
Shadowing barriers is a risk management technique that is conceptually similar to writing-off vanilla risk. It involves moving barrier levels within the trading position in a way that costs a small amount of money but creates large potential windfalls. This technique is also known as smoothing or bending barriers.
Example: A trading book is long a downside USD/CAD one-touch option in USD1m with a 1.0500 barrier. This barrier can be shadowed (i.e., moved) to 1.0450, which will cause a loss in the trading book as the value of the option has reduced. However, if the one-touch has a low value initially, it may not cost much to move the barrier a long way.
To be clear, the barrier on the client transaction is not changed, but for risk management purposes the barrier level has been changed. If this functionality is not available within the risk management system, the same effect can be achieved manually by booking a barrier spread between the main trading book and a shadow barrier trading book. Within the USD/CAD example, the main trading book would sell the 1.0500 barrier and buy the 1.0450 barrier at zero cost. The risk within the shadow barrier trading book should be left untouched until either maturity or if the shadow is unwound.
The 1.0450 one-touch barrier can now be risk managed as usual, but a windfall equal to the barrier notional is generated in the shadow barrier trading book if spot touches the real long 1.0500 barrier but does not touch the 1.0450 shadow barrier. Traders sometimes use part of the initial bid–offer spread cross on exotic transactions to establish shadow barriers at the deal inception, and/or they may shadow barriers further away gradually over time.
Like writing-off vanilla risk, shadow barriers work best when traders are engaged with their position:
A similar barrier shifting approach can also be used to generate bid–offer spreads for barrier products by looking at the value change produced by flexing barriers up and down a certain amount. This approach works particularly well in structures containing many different barriers.
Traders on bank trading desks have the ability to access the bank's client base in order to complete transactions at better levels than would be possible within the interbank broker market. For example, a trader may want to sell USD/MXN USD100m 6mth ATM. The current midmarket level is 10.2%, the “support” (i.e., the generic starting rate) two-way volatility in the interbank broker market is 10.0/10.4%, and if the trader sells within the broker market, the expected transaction level is 10.1%. However, a 10.2% offer for USD100m 6mth ATM could first be shown to clients (usually via the sales desks) who may be interested in purchasing the ATM contract at a midmarket level.
The 8.2% offer is called an axe: a special, better-than-usual bid or offer that the trader is showing because they are especially interested in completing the transaction in a specific direction.
The FX derivatives market can often get “one way” (i.e., all market participants simultaneously want to buy or sell the same types of contract). Therefore, showing axes to clients is not always applicable, but when it works it creates a win-win-win situation. The trader transacts at better levels than would probably be possible in the interbank broker market, plus they have not revealed anything about their positioning to the market, the client has transacted the deal at midmarket, and sales have executed a deal, improved their relationship with the client, and potentially earned commission, too.
For vanilla options, the process of axing out prices to offset the client flows is sometimes called recycling risk. This must be done with care. For example, generic contracts must often be used rather than the exact strike and maturity originally traded by clients.
Within exotic options, risk can be recycled in more complex ways. For example, if a reverse knock-out option has been bought by a client, the trader is left with long one-touch risk at the barrier level as described in Chapter 24. This one-touch risk can then be offered out to clients at an improved rate. Alternatively, if the exotic position contains short one-touch options, they could be hedged by selling reverse knock-out options to clients. In the same way, the digital risk on European barrier options can be recycled using European digital options.
Buying or long the American barrier means that on the barrier deals, positive P&L change results from spot trading through the barrier level. Long barrier positions can be obtained in a number of different ways, e.g., buying a one-touch, selling a double-no-touch, selling a knock-out, selling a reverse knock-out.
Selling or short the American barrier means that on the barrier deals, negative P&L change results from spot trading through the barrier level. Likewise, short barrier positions can be obtained by, e.g., selling a one-touch, buying a double-no-touch, buying a knock-out, buying a reverse knock-out.
Importantly this is not the same as the whole trading position making or losing money due to the hedges which are in place. Market participants generally prefer to sell American barriers for a number of reasons:
In some currency pairs, spot often follows a dynamic whereby it usually moves with low volatility but it jumps as economic data is released or other external events occur. This creates a fat-tailed distribution that is reflected within the volatility smile, but importantly it is specifically the spot jump dynamic that leads to a preference to sell American barriers. Selling high TV (i.e., expensive/close) one-touch barriers is an effective strategy when spot follows a jump dynamic. If realized spot volatility is low and the barrier does not trigger, one-touch options can lose value rapidly over time.
This effect is particularly important in pegged or managed currency pairs, where there is often an extreme jump dynamic and traders must be very careful pricing exotic contracts with American barriers. Consider a USD/HKD 1yr 7.7490 one-touch option. USD/HKD spot is currently kept within a 7.7500/7.8500 range by the HKMA (Hong Kong Monetary Authority) so the down barrier is positioned slightly below the lower band. The TV of this one-touch is 92%, but the market price will be far lower.
The same preference is observed in the vanilla market. In liquid currency pairs there is common market preference to sell short-dated wings (i.e., short-dated vanilla options away from current spot).
As described in Chapter 23, within delta hedged trading portfolios, long American barriers produce a stop-loss spot order and short American barriers produce a take-profit spot order. There is usually a larger chance of losing money from stop-loss orders than take-profit orders. With a stop-loss order, the risk is that spot trades through the barrier and continues before the stop-loss order can be fully executed. The difference between the order level and the execution level is slippage and it can cause large negative P&L. Therefore, short American barriers are more attractive than long American barriers.
Example: Long an AUD/USD 1.1000 topside one-touch generates a stop-loss order to buy AUD150m versus USD. Spot trades through the barrier and then even higher before all the spot can be bought in the market. In the end, the order is transacted at an average of 1.1012—a painful loss of USD180k.
Consider what happens in emerging market currency pairs with restricted spot market open hours. When the spot market is closed, the NDF market is still active and it is possible to accurately imply a spot rate from the level of the NDF. The NDF market may imply a spot rate at which the American barrier would trigger but the barrier can only officially be knocked once the spot market reopens.
If the NDF implies a spot rate through a topside barrier level but the stop-loss order is not executed, if spot continues higher, the stop-loss spot order might have to be executed far through the barrier level and hence cause a P&L loss when the spot market reopens. If the NDF implies a spot level through the barrier and the stop-loss order is executed by trading an NDF, but the market then retraces below the barrier, the barrier will not have triggered and unwinding the NDF trade will cause a P&L loss.
With stop-loss orders in currency pairs which are not constantly tradable it is possible for traders to be damned if they transact the hedge and damned if they don't—a difficult situation to risk manage. Therefore, in currency pairs with restricted spot market open hours there is a particularly strong market preference to sell American barriers.
The ATM curve represents the volatility of the forward (with a changing maturity), but American barriers knock out on spot. Therefore, if the forward is significantly more or less volatile than spot, any pricing model that does not have a stochastic interest rate component will not correctly capture the barrier knock probability and therefore structures with American barriers, particularly at longer tenors, will be mispriced.
Often the ATM curve rises at longer tenors but that does not imply that the market expects spot to get more volatile over time; rather it means that there is a significant interest rate component that must be taken into account when trading long-dated American barrier options. An example long-dated USD/JPY ATM curve is shown in Exhibit 25.4.
Generally, forward volatility is higher than spot volatility and therefore the market price for one-touch options will be lower than prices generated using a smile pricing model without stochastic interest rates.
3.138.124.135