Autocallable issuance upsets Euro Stoxx volatility market

The ‘Asianification’ of the Euro Stoxx 50 volatility curve is bringing a similar dynamic to the index as that witnessed on the Nikkei, HSCEI and Kospi, presenting dealers with new risks and opportunities – not least in the form of corridor variance swaps

Dislocations in Euro Stoxx implied volatility

At the start of the eurozone's growth wobble last year, when the Euro Stoxx 50 index fell by almost 2% on October 7, equity derivatives traders noticed something odd. Three-year fixed-strike implied volatility, instead of rising as spot plummeted, actually fell to a touch under 20% - highly unusual for a daily equity decline of that magnitude.

A week later, exotics desks found popular relative value trades, such as a long Euro Stoxx/short S&P 500 volatility pairing, were transacting at all-time lows, as implied volatility on the former index continued to move in tandem with spot – defying the logic of traditional market dynamics.

It was not a one-off event. On July 6 and 7 this year, the Euro Stoxx 50 sold off more than 2% on successive days. Yet Delphine Leblond-Limpalaër, director equity derivatives sales at Societe Generale, says calls to exotic desks to sell one- to three-year volatility increased, again against received wisdom. Two-year volatility on a Euro Stoxx 50 option struck at 3,300 on July 1, when the Euro Stoxx was at 3,496, stood at 20.9%. On July 7, when the index closed at 3,294, the implied volatility on the same strike was 60 basis points lower.

The trigger for this dislocation is investor demand for autocallable structured products tied to the European blue-chip benchmark. The nature of these products means dealers stack up weighty volatility positions on key indexes in times of market stress that they all need to hedge at the same time, and then unwind at the same time when the market rebounds.

See-sawing indexes have resulted in expensive hedging losses for dealers on these products in years past – up to $500 million in 2012 alone – and continue to distort the term structure of volatility on the Japanese Nikkei 400 and South Korean Kospi 200 indexes, as well as Hong Kong's HSCEI (see box: The HSCEI rollercoaster).

The pace of issuance out of Asia tied to the Euro Stoxx has been strong in recent years, and is having an important impact on how long-dated volatility, forwards, and dividends are behaving

Today, a surge in structured product issuance – much of it to Asian investors fed up of low returns on products tied to local bourses – referencing the Euro Stoxx 50 is bringing the same dynamic to European shores. What is happening to volatility on European stocks is a consequence of what is being dubbed the 'Asianification' of the Euro Stoxx.

Abhinandan Deb, head of European equity derivatives research at Bank of America Merrill Lynch (BAML), says: "Structured products have always been big in Europe, but what's interesting is the pace of issuance out of Asia tied to the Euro Stoxx has been strong in recent years, and is having an important impact on how long-dated volatility, forwards, and dividends are behaving."

Issuance of autocallables referencing the Euro Stoxx 50 out of South Korea were a modest €1.1 billion ($1.19 billion) in 2013, but rocketed to €9.5 billion in 2014 and are on course to top €20 billion this year, outstripping estimated issuance in Europe itself – which will be €13 billion, according to research by BAML.

And it is not just single-index products making waves. Correlation autocallables linked to the worst-performing of a basket of indexes – such as Euro Stoxx 50 and HSCEI, or S&P 500 and Kospi – are also popular. Dealers involved in the sale of these products are typically international investment banks, with Credit Suisse, Morgan Stanley, Societe Generale and UBS among the biggest players.

The current verve for Euro Stoxx underlyings is a by-product of the insatiable appetite for high-yielding instruments among Asia's sprawling retail investor base.

Robert Pitcher, head of exotics trading at Citi, says: "Europe has been a no-go for many investors outside the eurozone for the last few years. In Asia you also don't have people who know the European stock market. However, the autocallable product is very popular over there, and now with quantitative easing in the eurozone there has been a pickup in interest from Asian clients and even US clients as well, so the retail investors have started to dip their toes in."

From the retail investor's perspective, these autocallables are bets on an upward movement of the stock markets they reference. Investors can reap a coupon of typically 5–10% or more from the product knocking out above an upside barrier, often set at 100–110% of the spot level of the index. This is paid for by selling the dealer a downside knock-in put option struck at a lower barrier, typically struck at 45–60% of spot. The dealer's sensitivity to volatility between these two barriers is highly dynamic and impossible to hedge perfectly (see box: Autocallable anarchy).

With the frenzied issuance of recent years, dealers have built up enormous pools of vega which must be hedged. In October 2014, Credit Suisse estimated that in aggregate, the retail trade forced banks to sell $6 million in vega for every 1% downward move in the Euro Stoxx 50, meaning up to $66 million could have been sold during that month's sell-off, when the index plunged 12%.

"The fourth quarter was tough for many banks with big autocallable books," says Alain Dublin, head of equity and commodity exotic trading and index flow trading for Europe at BNP Paribas. "It was tough because of the violent trend that moved one way down, then up [so] big deformations had to be cut at some point, both ways. We are not among the top players on those products in terms of market share due to our conservative pricing models, so we managed to get through without too much damage."

Structural problems

There is a structural problem underlying dealers' hedging woes – namely that natural institutional buyers of downside puts, such as insurance companies, have shrunk their equity allocations and, in turn, their appetite for downside puts to protect these investments, reducing demand for long-dated volatility positions on the Euro Stoxx. The result is a classic supply and demand imbalance, leading to a flip in the shape of the implied volatility curve.

This effect became particularly pronounced in early July when concerns as to the status of Greece in the Eurozone reached their peak. Speaking to Risk.net on July 7, Pete Clarke, global head of equity derivatives strategy at UBS, said: "Euro Stoxx 50 term structure is currently inverted at the front end, troughs out at a maturity of about three years, and then starts to pick up again as you go further out. This dip in the term structure reflects vega supply from primary issuance of yield-based structured products, and also more importantly the vanna impact of existing inventory. In a balance-sheet-constrained world, we can expect this negative vanna to continue to pressure longer-dated implieds on any meaningful spot dips, especially after periods of concentrated inventory build-up."

He added that although concerns and large daily spot moves relating to Greece and China had given investors good reason to bid shorter-dated implied volatility higher, there had still been a net supply of volatility at longer-dated maturities throughout the decline. The situation was constantly evolving as the Euro Stoxx yo-yoed amid the unfolding Greece crisis.

Clarke said: "The rally since January 2015 has led to Euro Stoxx product autocalling throughout the first half of this year, which led to a bid for longer-dated vol and encouraged investors to reload. The sell-off since April has then taken markets down by 10% or so and this has definitely been enough of a move to see vega building up across existing inventory."

Then again, between July 7 and July 13, the index rallied 8%, depleting these vega inventories. Where the market will move from here is unknown, demonstrating just how dynamic dealers' exposure can be.

Stanislas Bourgois, head of equity derivatives strategy at Credit Suisse, says: "Since the beginning of the year there have been two different impacts offsetting each other. The first is a big rebound in the Euro Stoxx 50. Even over the last few weeks at the height of the Greece crisis we still were more than 10% higher than where we were in October. This means that the bulk of the autocalls inventories, which still consists of structures issued pre-2015 with knock-outs typically around 3,000/3,200, is deep in-the-money for the investor and their vega effects aren't felt as much."

He estimates that banks now have to shift $4-5 million in vega for every 1% drop in the SX5E, suggesting vanna has been cut from the fourth quarter of 2014.

"The second impact is that there seems to be less Euro Stoxx volatility buying interest against that. Now in October to December last year we did see buyers of put options. What seems to have changed is even since Greece started making headlines is we didn't see as many panic buyers of put options. Additionally we didn't see vol-arbitrage buyers of SX5E/S&P variance spreads until very late as the spread had reset to a different regime," he adds.

However, the dislocation does offer opportunities to those canny players familiar with the experience of the Nikkei, Kospi, and HSCEI.

With implied volatility artificially depressed at the long end, there is an arbitrage opportunity for those willing to bet that realised volatility will outperform in the long term. One way for dealers to offer this exposure is through the issue of variance swaps and spreads, a trade that has kicked off in Asia over the course of the last year or so, with Credit Suisse leading the way. The time is right for similar deals to be struck on Euro Stoxx volatility, according to dealers.

"This is a regular topic of discussion. The one question I'm asked more than any other is what is the impact of structured products on the long-dated part of the Euro Stoxx volatility curve and how it evolves with the recent market moves we have seen in Europe and Asia," says Leblond-Limpalaer at Societe Generale.

"Institutions that are long put on the Euro Stoxx are finding that when markets go down, while they would make money on delta, they wouldn't profit from volatility because exotic desks are selling long-dated volatility and therefore skew underperforms," she adds.

Corridor structures

The corridor variance swap structure is one Asian innovation tipped to make a splash on the Euro Stoxx: "I've no doubt it will happen here because for dealers it is a neat way of passing this risk off to other people. When they do these kind of trades, the need for them to dynamically hedge their positon starts to be reduced because they are literally flipping the positon they are getting from these autocallable notes and giving that position directly to the client," says John Moffat, senior portfolio manager at Capstone Investment Advisors, a multi-asset volatility manager.

Yet despite Credit Suisse's much trumpeted success with these instruments in Asia, the market is split on whether they are truly the panacea they are made out to be. The mechanics of the product mean that the swap buyer receives payments based on realised variance only while spot moves within the boundaries of the corridor, which is typically set in reference to the barriers of the autocallable product - between 60% and 110% of strike, for instance. Therefore, if the referenced index either vaults or crashes the buyer loses his exposure to variance.

Moffat explains: "[With a corridor swap] you don't get the tail [exposure] and typically what you want to own in Asia is the wings because that's where crazy things start to happen - you get these outsize moves and these wings come into play. Equally, you're not selling [the wings] on the leg you're selling, so you could argue that it's swings and roundabouts."

Mark Mehtonen, portfolio manager at Ilmarinen, a Finnish pension insurance company, says: "Almost all of the investment banks operating in Asia came to us with corridor structures, but we don't do them because for me it doesn't make sense as you're taking a few bets that the reference index will stay within the corridor. What we've been looking at instead is doing outright fixed strike or variance, spread against another index like the S&P 500 or FTSE 100. You don't have to do corridor trades to access the dynamics of this".

On the other side, the dealer selling the swap could suffer if the spot market trends below the lower barrier, but then rallies back into the corridor. In this situation, the dealer's put option embedded in the autocallable would be triggered, and their volatility exposure wiped out. However, they would still be required to pay fixed payments to the buyer in event of a market climb - essentially paying for volatility protection they no longer require.

One source at an international bank says this dynamic means the corridor structure favours range-bound indexes and underperforms where spot moves erratically. The Kospi 200, for instance, has bounced along between 240 and 270 for the past three years, whereas the Euro Stoxx has surged from 2,200 to 3,800 over the same period.

"In an environment where spot is rangy, the vega profile remains more stable but means convexity positions are more stable as well. On Euro Stoxx, the index has been rallying aggressively and where spot moves a lot, your position is unstable and you're carrying a short convexity position, which is more risky," says the source.

On the dealer side, the experience in Asia favoured some firms over others. While first movers did well, laggards found the going tougher.

One former trader who worked in the region says: "Typically the people who end up buying these corridor variance swaps are hedge funds. They are very few, and very sophisticated, and understand exactly the situation with respect to implied volatility versus realised. What we experienced was that it became a very competitive market, because on the one hand there were a lot of participants issuing autocallable notes, and on the other hand very few investors."

Societe Generale was one bank that found the going easier. Timothee Bousser, head of equities and derivatives for Asia-Pacific at the bank, says the risk of this product class is its dependence on strong liquidity. "A lot of firms have very short memories. When there is liquidity people pile in massively, but it does not happen all the time. There were large transactions that were pushed by less-prepared banks on the Nikkei, which led to massive losses. They had to cut these off, which kept them away from the liquidity move in the first quarter of 2013," he explains.

It is a market for cool heads and fast operators. While some dealers of Euro Stoxx variance will keep in mind the lessons of Asia, others may well fall into the same traps as those laid low by the market turmoil of 2012.

 

Autocallable anarchy

Autocallables are structured investments linked to the performance of a single stock or equity index, typically between two and five years' duration. The retail investor is entitled to an above-market coupon if the underlying exceeds a predetermined upside barrier, typically set at 100-110% of strike, on fixed autocallable dates. If this barrier is breached, the product automatically knocks out, returning principal and coupon. On the flipside, if the value of the underlying drops below a downside barrier (set at 40–60% of strike), the investor's capital is at risk.

Essentially, retail investors buy from dealers a strip of digital options with maturities set at each autocallable date, funded by selling the dealer a deep out-of-the-money knock-in put.

Dealers are therefore axed long volatility, because when the underlying is close to the upper barrier they would prefer high volatility to shift spot downwards. Similarly, when the underlying is close to the lower barrier, high volatility would increase the chance of the downside put knocking in. In addition, they are long skew - since as the probability of autocall increases, the demand from issuers for upside calls to hedge the investor coupon grows, while on the other side, as the knock-in grows more likely, the issuer would sell downside puts in advance to hedge.

To hedge this position, dealers need to sell volatility on the open market, but the uncertain maturity of the autocallable makes this hedge highly dynamic. As spot falls the product is less likely to knock out at the next autocallable date and dealers' sensitivity to volatility – known as vega – extends in duration, while their exposure to vanna – the change of vega in relation to a change in spot – increases, requiring them to sell long-dated volatility to stay within risk limits.

Conversely, when spot rises they need to buy back these long-dated vega hedges as the expected maturity of the product shortens. Dealers also have to reel in their longer-dated vega hedges if spot falls towards the downside barrier and it becomes increasingly likely that their knock-in put will be activated. Furthermore, if an American downside barrier is used and knocks-in, this long vega exposure is wiped out as the dealer reaps the value of the in-the-money put.

This dynamic between barrier strikes means dealers are particularly sensitive to the point equidistant between the two, where their knock-in put exhibits maximum reactiveness to volatility. Hence this is the strike that typically attracts the greatest amount of vega hedging.

This dynamic has been affecting the Euro Stoxx term structure of volatility over the course of the past 18 months or so, says John Moffat at Capstone. "Dealers are long the two-year vega bucket, and because there is less liquidity out there, they have to move forward to hedge at the 18-month or one-year point, so this depressive effect feeds through to the front of the curve till it gets to a point where the options are sufficiently short-dated that the effect of the two-to-three-year distortion is limited."

 

The HSCEI roller coaster

One index particularly susceptible to the distorting impact of autocallable issuance at the moment is the Hang Seng China Enterprises Index (HSCEI). Morgan Stanley estimates $11 million of issuance on the index in June alone, with total outstanding vega on the index at $95 million.

This pressure on HSCEI volatility meant that although in early July the index crashed up to 25% from its May 26 high, astonishingly one-year volatility fell. This goes against all market logic, as a rout of this magnitude would be expected to cause implied volatility to spike massively. Morgan Stanley attributes the positive correlation to dealers' need to offload their long vega positions, which grew by $50 million in a month.

Despite the crash, however, the bulk of outstanding issuance is yet to knock in on the downside, meaning dealers are still exposed to vega swings. The highest HSCEI downside barriers are struck at 8,000–8,500, which is 28–33% lower than the index closed on July 7. But the decline did make it less likely that certain product generations would knock out on the upside, meaning issuance in the coming quarter may be down on recent trends.

Frank Drouet, head of global markets for Asia-Pacific at Societe Generale, says: "This correction on the HSCEI has partially shifted the vega of autocallable products referencing baskets of indexes onto the HSCEI, with, the latter becoming the worst-of in some cases. This vega shift could have pressurised the implied volatility of the HSCEI index in other market conditions, but the sharp correction in China coupled with the Greek situation has created a risk-off sentiment and a volatile environment. That is particularly true for this index, so different types of client came to buy vega-plus products linked to the HSCEI, absorbing the increase of vega generated by the autocallable products."