Equity derivatives house of the year: Bank of America

Risk Awards 2020: Focus on risk spurred invention and made space for $2 billion Natixis deal

Hichem Souli

Bouts of intense volatility in 2018 left dealers and investors facing two big, hairy problems in 2019.

First, scorched by the year-end stock market sell-off, the supply of cliquet options that powers the $70 billion fixed index annuity market – popular with US retirees – started to dry up. Second, with yields elsewhere stubbornly low, investors continued looking for ways to bet against volatility, while also asking for protection against a swift reversal.

Bank of America found answers to both. It was able to continue writing cliquets by rewiring the way the underlying indexes target fixed levels of volatility – a “fantastic solution”, according to one client. A short-vol strategy that decouples from the market at times of stress, meanwhile, raised more than $1 billion in just 10 months, with one investor describing it as “a game-changer”.

Innovations like these have been powering the growth of BofA’s equity derivatives business. Its book added $385 billion in notional during the 12 months to mid-2019, according to US regulatory disclosures – growth of 26%, which led all other big US dealers. At $1.85 trillion notional, the portfolio has more than doubled in five years.

But this isn’t the wild, bells-and-whistles invention of the industry’s past, insists Hichem Souli, global head of equity derivatives structuring at BofA.

“Ultimately, wherever we are innovating, what we care about is that we can hedge all of these innovations. We’re not creating innovation for the sake of it,” he says.

Keeping risk on a tight leash can also create opportunities to grow, as the bank proved in helping clean up one of the year’s trouble spots.

In January 2019, BofA completed a transfer of autocallable risk from Natixis. Market sources say the transfer totalled $2 billion notional, though BofA declined to comment on size. A month earlier, the French dealer had revealed a €260 million loss on Korean equity-linked notes and was seeking to cut its exposure. BofA found itself with spare capacity, having dialled down its own share of the $58 billion Korean structured product market from 6% at the start of 2018 to just 2% by year-end.

“We had shrunk exposure a lot at the end of 2018,” says Souli. “Internally we decided those products weren’t correctly priced when embedding all of the risk. So, when we saw a competitor coming out of that business, we were ready to benefit from pricing taking into account the risk embedded, so we positioned in the right manner and refilled our book.”

The trade paid off, with the majority of the portfolio being called by late 2019 as markets rebounded, locking in gains.

Vol target index

The US annuities work is worth a closer look. Structurally, the business is built on volatility controlled indexes, on which dealers sell huge amounts of option strips, known as cliquets – typically forward-starting one-month calls.

Because the volatility of the underlying is, theoretically, known in advance, the cliquets generate relatively thin premiums, but they still have the potential to trip banks up – in times of stress, realised volatility can blow through the target level. Last year’s big moves in US stocks brought these risks to the fore, prompting some dealers to hike premiums on the instruments or to shrink cliquet capacity altogether.

When one US carrier came to BofA, worried the FIA market could be strangled, Souli and his team went to the heart of problem – the underlying vol target index itself.

“The vol target itself is a simple and powerful mechanism, but there are still some risks in it and those have been neglected,” he says.

One of those hidden risks stems from the daily rebalancing of the products, which takes place at each market close. Over a year, volatility in the products is likely to realise with little error, because there are 252 closing prices on which to draw. But over a month, say – just 20 trading days – the same may not be true. Put another way, the vol-target mechanism stabilises long-term volatility, but fails to account for short-term moves.

Ultimately, wherever we are innovating, what we care about is that we can hedge all of these innovations

Hichem Souli, BofA

BofA’s fix is a high-frequency rebalancing technique – so-called ‘fast convergence’ technology – that swells the number of observations and bolsters the rebalancing opportunities from just 20 to hundreds, or even thousands depending on the preferred fixing period. This allows the index to quickly lever up or down in response to intraday shifts, stabilising short- as well as long-term volatility.

There’s a balance to be struck between the number of observations and trading costs – the first live iteration uses multiple intraday fixes, while a big data algorithm aims to maximise execution quality. Back-tests show the approach can slash short-term vol-of-vol three-fold, while gap risk can be reduced by as much as 70% depending on the speed of the algorithm, Souli says. The new approach also delivered 1.7 percentage points of improvement when compared to a traditional vol target index – again, in back-tests.

The client was pleased with the result. “Their engineers put together a fantastic solution,” says the firm’s product head, noting a triple-whammy of benefits. In addition to addressing the capacity issue, the innovative way of managing volatility translates into cheaper options for the client. An unexpected, though welcome benefit of the vol control was an improved the Sharpe ratio of the underlying.

Souli says the approach could be applied widely across derivatives markets: “Anything that embeds a vol target and anything that embeds a gap risk related to close-to-close rebalancing can be enhanced through fast convergence technology. It opens up a wide scope of usage.”

Volatility conundrum

Since the Vix-plosion of February 2018, when the US volatility benchmark made its highest one-day gain in history, there has been no shortage of attempts to reimagine the short volatility carry trade. Here, BofA was again able to show off its structuring smarts, as the dealer created one of the fastest-growing strategies of 2019.

A widely popular strategy since global rates tanked, systematic harvesting of the volatility risk premia has often been likened to picking up pennies in front of a steamroller – investors build up small incremental gains through the good times, only to have years of profits wiped out by a single shock.

Defences intended to guard against these losses have been varied. Recent innovations include the use of signals to leverage short volatility up or down, intraday delta hedging to reduce risk exposure near to the close, buying back out-of-the-money put options to cap losses, or diversifying the trade across asset classes.

Most still struggle to cope with the most extreme moves. Short volatility has an implicit positive correlation to equity market beta – in essence, if you’re short big moves, then you’re long small ones – and attempts to hedge can be difficult, given the tendency for assets to move in lockstep during periods of stress.

“Whenever you do short vol, you’re long delta so you tend to inherit a tail risk in a severe market selloff. It’s a problem we’ve seen at different points in history,” says Souli. “We saw a huge unwind in vol carry strategies in February 2018, but people want to return and create some alpha. In a situation where yields keep shrinking, it was clear to us we needed to find a product to settle that issue.”

That product had to combine short volatility with negative equity correlation – an apparent paradox that the team resolved by looking under the bonnet of alternative parts of the trade. Eventually, they struck gold with an innovative, yet surprisingly simple, delta-hedging structure.

Whenever you do short vol, you’re long delta so you tend to inherit a tail risk in a severe market selloff. It’s a problem we’ve seen at different points in history

Hichem Souli, BofA

The result is Synthetic Uncorrelated Volatility (SUV). Packaged in an index, the strategy sells strips of vanilla options on a daily basis, harvesting the implied-versus-realised risk premium through benign times without incurring losses in a sell-off.

Souli is cautious when describing the mechanics, but was willing to shed more light on it for Risk.net when speaking off-the-record.

“We managed to create this uncorrelated aspect of the strategy by using an innovative delta-hedging mechanism to minimise the gamma losses from buying high and selling low. By using this mechanism we managed to make sure if markets go down, the strategy is completely uncorrelated. As long as the market is range-y, you keep harvesting the risk premium,” says Souli.

In essence, the hedging ratio increases when spot moves closer to the strike, and falls when further away. The structure allows the maximum juice to be squeezed out of the trade in benign times, while ensuring decorrelation from equity markets in times of stress.

Clients say it’s an effective twist – and surprisingly elementary.

“What surprised me is that once you understand the idea, it’s one of those cases where you’re like ‘Oh yes, of course!’ but nobody else has come up with it. It’s an example of something that could be a game-changer,” says one pension client.

There’s no such thing as a free lunch, of course, and under some conditions the strategy could underperform – the structure effectively transfers risk from the left-tail of the return distribution (risk-off events) to the right tail (aggressive market rallies). Investors may be swapping gap-down risk for gap-up risk, but according to one satisfied customer, “it’s a very good late-cycle trade”.

Like fast convergence, application of the SUV magic dust could be far-reaching. The delta-hedging mechanism can be tailored to improve performance for asset owners’ bespoke call-selling activities. It’s a level of inventiveness rivals haven’t been able to match, according to another pension client. “They came up with something that is very, very clever and others haven’t. This gave us better performance while still maintaining the spirit of what we are trying to do, so it’s still very much a hedge.”

Momentum robots

With an $11 billion annual IT budget, it’s little surprise BofA is also giving machine learning technologies a spin under the watchful eye of the digital innovation group. The first crop of products to market include the Dynamically Diversified Momentum (DDM) index. Live since mid-2018, the strategy aims to address another conundrum – the so-called ‘Curse of Markowitz’. Nobel prizewinning economist Harry Markowitz quantified the benefits of diversification in his pioneering work in the 1950s, which remains the basis of modern portfolio theory.

As asset pools get larger, however, this diversification benefit can erode as short-term momentum signals create noise that can work against the strategy. DDM addresses this using clustering techniques to divide the investment universe into buckets of similar risks, loosening dependency on correlation estimates and uncovering hidden alpha.

With back-tests showing 4.6% annualised returns over 10 years, the bank is pitching the strategy to insurance clients. One who had seen the product described it as ‘extraordinary’.

It is this kind of innovative thinking that helped BofA buck a losing trend in bank quantitative investment strategies. While other dealers have seen outflows from their QIS business following a period of poor performance for alternative risk premia, BofA grew assets under management in its equity derivatives investable index business.