Weird moves in markets related to popular structured products suggest many dealers struggled to cope with a brutal fourth quarter, when stock indexes tanked around the world, and dynamics more often seen in Asian equity markets leapt to Europe.

The Eurostoxx 50, which underlies an estimated $120 billion equivalent of autocallable bonds, shed more than 11% in the last three months of the year – “a disaster” for structured products issuers, according to one strategist, because their sensitivity to volatility, dividends, correlation and convexity rapidly changes as spot drops.

At least one dealer admits it now has lessons to learn: “It was very tough in Q4. We’re not sitting on our hands, we’re analysing the situation, looking at what we could have done better and what we did well. We’ll draw our conclusions after that,” says the bank’s equity derivatives head.

The signs were there. In October, implied dividends on the Eurostoxx 50 plummeted, far outpacing equity index losses – an early sign of stress. Two months later, normal hedging patterns in index options markets – in which banks sell volatility as markets fall towards crucial options barriers – began to break down, a phenomenon some participants attribute to fears of a scary, shadowy inflection point, known as ‘peak vega’.

When a sliding Hang Seng China Enterprises Index (HSCEI) hit this inflection point in the third quarter of 2015, dealers lost an estimated $300 million.

“Some people in the market were not clear whether peak vega is real or just a myth, but we have been testing it,” says Vincent Cassot, equity derivatives strategist at Societe Generale.

Instead of selling volatility in tandem with the falling Eurostoxx, some dealers began buying in December, he claims – as they would if their vega profile had flipped.

How well dealers managed the episode should become clearer in the coming weeks, as large issuers queue up to deliver fourth-quarter results. Early indications suggest they struggled.

Some people in the market were not clear whether peak vega is real or just a myth, but we have been testing it

Vincent Cassot, Societe Generale

On January 22, UBS announced a 23% collapse in equity derivatives revenues for the fourth quarter – it is the only big dealer to break derivatives out from the wider equities business – and a $47 million dollar quarterly loss at its investment bank. Five days earlier, Societe Generale, warned revenue for its global markets business would be down 20% year-on-year due to a “challenging environment”. It is due to announce earnings on February 7, with Natixis following on February 12 – the latter has already revealed a €260 million ($296 million) hit linked to the sale of autocallables in Korea.

BNP Paribas will deliver its results on February 6 and Credit Suisse on February 14.

US banks reported fourth-quarter earnings in January without any obvious blowups, although structured products typically make up a smaller share of their business. Year-on-year equity revenue ranged from flat at Morgan Stanley to a 17% surge at Goldman Sachs. Many of the US houses, including Bank of America Merrill Lynch (BAML), Goldman and JP Morgan were comparing last year’s revenues against a rough fourth quarter in 2017 when they were hit with a combined $1 billion in losses on margin loans to South African retail giant, Steinhoff.

Scramble for vega

Structured as zero-coupon bonds with upside and downside barriers, autocallable bonds see investors receive an above-market coupon if spot remains within those bounds. The structure knocks out when spot hits the upside barrier, returning principal and an improved coupon to investors. The downside barrier, typically set 40–50% below the initial strike, triggers a knock-in put option and eats into investor principal if breached.

The structure leaves issuers long volatility. As spot heads south, increasing the probability of downside barriers being breached, dealers’ sensitivity to volatility, or vega, increases. To balance books, traders are forced to sell puts to flatten their risk.

This causes spot and vol to head down in tandem – inverting their usual relationship. Analysis from BAML shows this phenomenon in full swing during October, when correlation between two-year out-of-the-money put strikes and the Eurostoxx hit a 10-year high of 40%. Call options maintained their inverse correlation with spot – a sign the dislocation was driven by autocall hedging.

“Long-dated put vol [on Eurostoxx] actually fell, but that wasn’t the case for long-dated calls. Call vols behaved normally. That isolates the impact of dealer hedging of autocallables as their exposure is more prominent on the put side,” says Abhinandan Deb, equity-linked analyst at BAML.

As spot moves closer to downside barriers, the peak vega inflection point is reached. Beyond this point, vega exposures collapse and dealers collectively switch to buying back volatility. The scramble for vega causes implied volatility to jump, ultimately locking in losses for dealers who are forced to buy back hedges at a higher level.

Industry estimates of peak vega on the Eurostoxx 50 range from 2,500 to 2,900 – around 10% above downside barriers. The index began October at 3,414, and hit its year-low of 2,937 on December 27.  

Although the index never quite hit the top end of those estimates, some dealers started behaving as they would if their exposure had flipped, says Societe Generale’s Cassot – possibly in anticipation of spiralling costs.

“When Eurostoxx went down sharply in December, people started to buy volatility. It happened earlier than some were expecting but some players may be hedging earlier, perhaps because this dynamic is better known and they don’t want to be squeezed. To avoid what happened on HSCEI in 2015 no one wants to be the last one to buy,” he says.

Long dividend exposure

Other European index parameters suffered extreme dislocations through the fourth quarter of 2018. In October, Eurex-listed Eurostoxx 50 dividend futures fell at a faster rate than the underlying equity benchmark. Dividends typically move in sympathy with the underlying equity market but with lower volatility. This is because corporate payouts are more stable than equity prices, determined by corporate policy rather than the whims of investors. The closer to the dividend payment date, the greater the certainty of payouts and the more stable the contract.  

“We saw outsized moves in Eurostoxx 50 dividends relative to the market, the likes of which we haven’t seen since 2011 – and to some extent 2008. We saw a lot of selling pressure that was not in keeping with the fundamentals of the market. It created some very interesting opportunities for clients looking to buy,” says BAML’s Deb.

On October 24, front-year dividend futures expiring in December 2019 fell to 116.5, down from 122.6 at the end of September. Five-year contracts expiring in December 2023, which align more closely with European structured products hedging needs, fell to 96.6, down from 112.9 at the end of September.

This saw one-year contracts fall at exactly the same rate as the wider market, or with a market beta of 1. The contracts have a historic average beta of around 0.25, meaning they usually drop just a quarter of a point for every one-point drop in the wider equity market. Five-year contracts dropped faster than the equity market with a beta of 1.5.

BAML analysis shows those levels were extreme, even compared to previous crash scenarios. In February 2018, five-year dividend beta remained below 1, despite a one-month Eurostoxx 50 drawdown of 9.4% – similar to index losses sustained in October. During the global financial crisis in 2008, five-year dividends traded with a beta of 1.5 alongside a 27% index drawdown.

We saw outsized moves in Eurostoxx 50 dividends relative to the market, the likes of which we haven’t seen since 2011 – and to some extent 2008

Abhinandan Deb, Bank of America Merrill Lynch

Again, market participants believe the phenomenon is indicative of extreme selling pressure from autocallable issuers, which see their long dividend exposure increase rapidly as spot falls. Autocall dealers are implicitly long dividends on products which offer investors forward exposure to a price return index. As spot falls, the chance of the product being autocalled decreases and the expected maturity of the trade extends. Dealers become longer dividends as they hedge their forward exposure, which they flatten out by selling dividend futures, typically out to five years to align with European structures, though selling pressure was seen across the curve.

“On the way down, dealers’ long dividend exposures have increased significantly. You don’t have the liquidity in the longer term contracts so many desks have hedged that with the shortest, most liquid maturities, causing beta to explode on the way down. There’s no fundamental reason why the short-term dividend future should be destroyed like that as the amount of uncertainty on a one-year dividend contract is low,” says Antoine Porcheret, equity derivatives strategist at Citigroup.

Dealer exposures to Eurostoxx 50 parameters have been growing since 2015 as lacklustre index performance prevented autocallable bonds from being redeemed via their upside triggers. Market estimates put the outstanding stock of Eurostoxx 50 autocalls at $120 billion.

“The existing stock of products has been extremely high as exposures have been building while the market didn’t move much. Products that were struck in 2015, 2017 and 2018 haven’t redeemed as the market didn’t rally,” says Porcheret.

Cutting exposures

Despite heavy selling pressure, some analysts believe the latest turbulence should have been manageable for most structured products desks, particularly those that handled similar bouts of dividend volatility through the October 2014 flash crash and August 2015 equity selloff.

“There was a bit of an overreaction in dividends in late October but they quickly bounced back along with the broader market. We’ve seen this time and again when markets take big tumbles. You see an outsized beta from dividends followed by an outsized recovery. Issuers have been here before, and they generally expect these kinds of moves during spot weakness,” says Pete Clarke, head of equity derivatives strategy at UBS. 

With hedging costs threatening to eat into revenues, dealers have stepped up efforts to lighten burgeoning structured exposures by packaging exotic risks into attractive investment opportunities that can be recycled to hedge funds and other sophisticated investors.

Product innovations may also help mitigate rapidly changing sensitivities in the structured products book. For example, fixed dividend versions of Eurostoxx 50 and other structured products underlyings have been developed by banks including Societe Generale and Natixis. The alternative indexes save dealers from having to dynamically hedge dividend exposures, though some critics believe the indexes can detach from their benchmarks. Some banks, including BNP Paribas, sold autocalls referencing the dividends rather than the index – ultimately creating a dividend exposure that perfectly matches the book. Different iterations of the product included instruments on the equity index whose coupon and returns were based on dividends, to contracts with dividends as the direct underlier.

“It’s not a widely traded product but it’s an interesting way of doing it. Dividend vol was very low for a while so this wouldn’t have been attractive a year ago, but dividend vol has popped somewhat now, so it starts to look a bit more interesting,” says UBS’s Clarke.