A Tidal Wave — Beginners Guide to FX Volatility Cycles I

We, as people, take some aspects of life for granted. The fact that there are exactly 24 hours every day, or that the sun will rise and set tomorrow. Even though we can’t be 100% sure it will happen, we believe it is an indisputable truth. Another aspect of life we take for granted is the idea that we can exchange currencies for goods and services. The idea of exchanging tokens (coins, paper notes, bank transfers) for physical goods and services has always fascinated me. As a child, I was puzzled by the idea that people use pieces of paper with some numbers written on them in exchange for a toy or a video game, as the paper note looked pretty worthless to me because it’s just a piece of paper. Later, when I grew up, my grandmother taught me the monetary system and even went as far as teaching me about the Foreign Exchange mechanism. That experience has drawn me into foreign exchange (FX), as I found this notion to be more than the sum of its parts.

Unlike in most markets, where participants trade yielding assets (dividend flows in equities, physical usage in commodities, and stream of coupons in Fixed Income), Currencies are essentially worthless unless exchanged against something. A $1,000,000 in our bank account (or in our safebox) is worthless unless we want to put it to use. In FX, 1 EUR has no value unless exchanged against another currency. In essence, currencies are a transmission mechanism (or the plumbing of the market) when we need to move from one segment of the market to another one. A US investor looking to buy a European corporate bond needs to exchange USD into EUR to pay for the bond (and exchange the coupon paid in EUR back to USD), or a Chinese refinery needs to exchange CNH for USD to buy barrels of oil.

Clearly, FX is a derivative of other markets, which makes the FX derivatives market a second-order derivative of the global capital market. So, trading FX volatility (and higher dimensions like skew/convexity) becomes almost impossible to understand unless we familiarise ourselves with the market microstructure and the dynamic that affects the volatility cycle in FX. In this two-part write-up, we will dip our toes into the strange world of the FX volatility market and how the dynamic in that market creates a volatility cycle.

Meeting The Players

One of my previous write-ups elaborated lengthyly on the different players in the FX derivatives market, so if you are interested, feel free to check this introduction.

In short, there are mainly three types of participants in the market:

1. Interbank Dealers/Market-Makers — These are the market liquidity providers that warehouse risk and provide quotes on a non-volunteering basis. They aim to facilitate the buy-side users' needs to hedge their exposures or gain exposure/risk. The market-maker aims to keep its risk (options books’ Greeks) under check and enjoy the bid-ask spread (sell high, buy low).
2. Speculators (CTA, Macro, Vol RV Managers) — A diverse community of players aiming to take currency market exposure as part of their strategies. They are an opportunistic group that would trade only when opportunities present themselves so that they will provide the market with liquidity on a volunteering basis. Within the group, we can find more directional players (CTA/Trend-Following/Macro), as well as Relative-Value (RV) vol players, that aim to take advantage of mispricing of volatility (whether that’s volatility surfaces or volatility risk-premium).
3. Hedgers (Real-Money/Corporates) — Corporate/Real-Money participants use the FX derivatives market to hedge cashflows and cross-market funding needs. Given that the purpose of their trades is to hedge their exposures, they are less sensitive to the price of volatility, and when they need to hedge, they do not make their decision based on that. Due to the varying nature of the hedgers’ cashflow timing/duration, their flow covers the entire term structure (expires) from 1 day up to 20 years (and even longer sometimes).

The interaction between the different market participants, with their different utility functions, is the underlying driver of the FX volatility dynamic. Hedgers’ need to hedge their exposure creates a supply/demand need in options space that dealers can’t always warehouse, and thus, speculators act as secondary underwriters (mostly Vol RV managers). A macro fund manager looking to express her direction view on a policy divergence between two economies will trade options (or structures) to express the view and create demand for volatility (or higher order aspects like skew and convexity), which will translate into the way implied volatility is repriced (given the supply/demand imbalance).

Options Flow, Dealers Positioning, and Anything in Between

To understand the basics of how implied volatility surface evolves, we need to understand the underlying forces that drive that dynamic.

Generally speaking, implied volatility represents the premium the end user is willing to pay/receive to own/sell a risk. Like with the insurance business, (implied) risk/probability of occurrence is mainly priced at a premium to the actual/physical probability. As noted before, the fact that some participants in the market don’t look at the risk from an implied/actual volatility perspective (those could be hedgers that need to hedge or directional players looking to gain optionality) implied volatility can vary significantly from the actual/realized volatility.

Our base assumption, when analyzing the FX volatility dynamic, is that, for the most part, dealers actively hedge their option-related Greeks (sensitivities), trying to keep them under their risk limits (usually set by the dealers’ risk departments). Among the Greeks that dealers mostly manage, we can find four very important ones:

1. The Delta (Δ) — the option value sensitivity w.r.t the underlying spot change
2. The Vega (ν) — the option value sensitivity w.r.t the underlying implied volatility change
3. The Gamma (Γ)— the option delta sensitivity w.r.t the underlying spot change
4. The Theta (Θ)— the option value sensitivity w.r.t to the time passage

There are other Greeks that dealers monitor, but for now, we will focus on them before we go deeper into the belly of the beast…

As we noted before, the buy-side landscape is a well-diversified group of end users, which makes it harder for us to analyze. Therefore, we assume the end-users don’t actively manage their option-related Greeks exposure. That simplifies the analysis significantly as it lets us analyze the dealers' side of the market, which we can do by utilizing risk matrixes like shocks/slides (spot ladder) and spot/vol matrix.

Given that dealers trade between themselves via interdealer brokerage desks and most end users trade in the FX OTC market via FX prime brokers, we can assume that, on average, all dealers tend to have the same exposure profile. Some might be more exposed than others in some currency pairs and some less, but, on average, we can treat dealers’ exposure profile as “the streets’ exposure.”

The biggest obstacle we face when trying to understand how dealers are positioned is the OTC nature of the FX, where about 75% of the transactions/flow is done “Over The Counter” (directly with the banks), as opposed to other markets, where most flow is carried on exchanges. This means that we have low visibility of the flow and trades done in the market. The US regulators made trade reporting mandatory for US-based entities and US banks so we can observe the option trades done. Still, the scope of reporting is limited (for example, it doesn’t disclose whether an option was bought or sold to the client, which is crucial in any analysis).

Assuming we can distinguish between trades bought by clients (sold by banks) and sold by clients (bought by bans), we could build a “Greek Profile” that will help us understand how dealers, as a collective, are likely to hedge their exposure to various types of moves in the underlying market.

The immediate Greek that affects the volatility dynamic is the “Gamma”, and the “Gamma Profile” generally determines the ranges where dealers are “net long gamma” and “net short gamma”.

When dealers are “net long”, they get a shorter delta as the underlying spot moves down and a longer delta when the underlying spot moves up. As dealers aim to be “delta neutral,” or, in other words, indifferent to the underlying spot direction, they will sell the underlying when the spot moves up and buy the underlying spot when the underlying spot moves down. This dynamic hedging by dealers creates “vol compression” as the hedging activity in the underlying asset keeps the underlying spot dynamic in a range (and exhibits mean-reversion tendency). On the other hand, when dealers are “net short gamma”, they will exhibit the exact opposite behavior and essentially “chase the market” on moves up and down of the underlying spot. This dynamic will create a “vol expansion” period with high realized (and implied) volatility. A cross-sectional analysis of dealers' gamma exposure and the change in realized volatility reveals a negative correlation between dealers’ net gamma exposure and realized volatility.

As we can see, dealers’ positioning has, to some degree, predictive power of the future realized volatility, which, as a result, affects the way implied volatility is priced.

The spectrum of end users in the FX derivatives market spans from hedgers to speculators with different utility functions and trade horizons, and different forces govern the shape of the implied volatility surface. In the absence of external factors (strong macro narrative/view, spillover from cross-market hedging), the short-dated implied volatility will be controlled mostly by systematic/RV players who “supply” dealers with gamma as they try to “collect” theta. On the other hand, the back end of the implied volatility curve is naturally governed by supply/demand forces, resulting in the hedging needs of corporate users and long-term investors like pension funds. In between the front end and the back end of the implied volatility curve, the nature of the flow driving the middle of the volatility curve varies from systematic/RV managers who take a view on the shape of the volatility surface to macro investors trying to express their view on the currency pair direction (trajectory).

Adding Exotic Flavour

So far, we’ve explored the basic dynamic in the derivatives market, where supply/demand by the different market participants affects dealers’ exposure profile and, as a result, determines the shape of the volatility curve. The additional layer of complexity we need to acknowledge when analyzing the volatility dynamic in FX is the significant portion of exotic derivatives flow in the market.

Unlike other exchange-traded markets, where most flow that goes through the market is in “vanilla space”, i.e., vanilla call/put options, the OTC nature of the FX derivatives allows participants to use exotic features (knock-outs/knock-in barriers, conditionality, etc…) in their products to cheapens the cost of the structure (and create more leverage). That additional complexity introduces dealers with optionality/exposure in higher dimensions of the volatility surface (skew, convexity). It allows them to recycle risks they incur from their “vanilla book” using exotic options.

Without expanding too much on the FX exotic market (feel free to go through my previous write-up on the subject), we can distinguish between three types of exotic derivatives:

1. Barrier Options — Vanilla options with some contingency (will either cease to exist or come to life once a specific level is touched/breached). The knock-Out (KO) option will act like plain vanilla as long as the KO level is not touched, while a Knock-In (KI) option will ONLY become a vanilla option if the KI level is touched.
2. Payout Option (Cash-or-Nothing) — Options with predefined/fixed payout contingent on a condition (trigger) being touched/breached. The common types of Payout Options are One-Touch(OT)/No-Touch(NT) and digital options (at-expiry barriers, where the barrier is observed ONLY at expiry).
3. Strikeless contracts — contracts that allow investors to gain a purer exposure to higher dimensions like realized/implied volatility, volatility curve shape, correlation, skew, and convexity using products that are not bounded by actual strikes of the underlying asset.

The existence of exotic derivatives serves both buy-side participants and dealers. From the buy-side perspective, they can use exotic options to express their view in a more efficient and leveraged way (using path dependency and conditionality), while from the dealers' perspective, it allows them to recycle existing exposure and gain exposure to higher dimensions risks that otherwise would have been more costly.

As the portion of exotic derivatives is significant in the context of the entire flow traded in the FX market, their presence makes the analysis of volatility surface dynamic more complicated (yet far more interesting).

If we explore the most common type of exotic option traded by end-users in the FX market, the Reverse Knock-Out (RKO), we can see it has some interesting features. In short, an RKO option is an option that its chance of being triggered increases as its intrinsic value grows. In other words, the more the option becomes in-the-money (ITM), the closer it gets to the trigger level.

To understand it better, let’s analyze a typical RKO option: 3-month USD/CNH RKO call with a strike of 7.25 and Reverse Knock-Out trigger at 7.45.

As we can see, the RKO has a significant discount to an equivalent vanilla struck at 7.25 (RKO costs about 0.24% while the equivalent vanilla cost ~0.84%), with initial Greeks that are pretty neutral (small long vega and gamma exposure). If we look at the Gamma/Vega profile w.r.t. spot move, we can see that from the dealers’ perspective, the option gains both gamma and vega as the underlying spot moves towards the barrier.

In short, RKO-type options essentially make dealers LONG optionality out-of-the-money (usually on the “expensive” side of the implied distribution), so an accumulation of barriers should act as a compressor of higher dimensions of the volatility surface like skew and convexity, as long as the underlying spot price doesn’t move rapidly towards the barrier level.

On the other hand, once the RKO barrier is triggered, dealers need to unwind their hedges (Delta, Gamma, and Vega hedges they held against the original option) as the underlying exposure ceases to exist. This discontinuity of the risk around the barrier level creates an interesting dynamic, as dealers need to reassess their exposure profile and restructure their option books, which can change the volatility surface dynamic.

In short, supply/demand by market participants in the “vanilla space” primarily affects the shape (and dynamic) of the ATM volatility curve, while the exotic derivatives’ exposure primarily affects the shape of the “volatility smile”, or, in other words, the premium OTM options over ATM volatility.

The next chapter will discuss the anatomy of a volatility cycle, its expansion and compression, and the way the market shifts between the different regimes.

Stay Tuned….

Harel.