Like other big structured products issuers, UBS had concerns about the state of the equity markets coming into 2018. An 18-month bull run in stocks had given rise to concerns over toppy valuations, and a house view that investors were being lulled by abnormally low volatility. Unlike other houses, UBS also had a clear view of where its exposures lay should the rally hit the skids – and a plan to respond when it did.
The bank didn’t have to wait long. On February 2, 2018, global stock markets suffered a sudden lurch, with the S&P 500 falling 2.2%. The following Monday morning, February 5, that became a full-blown selloff. Having opened at 2,762, the S&P 500 suffered intraday falls of 1.7% by early afternoon accompanied by a huge spike in the Vix, which tracks implied volatility on the US equity benchmark.
So far, so expected, but what happened later that day was anything but. After staging a modest recovery rally, just after 3pm, US equities suddenly suddenly nosedived 2% in the space of seven minutes before rebounding almost immediately to close at 2,648. The Vix did the reverse: opening at 17.31, spiking to 35.73, before gapping to 30.71 in the space of a few minutes in late afternoon trading, and closing the day at 32.12.
Even equity market veterans call it one of the strangest trading days of their careers. “The correction that happened that Monday, I’ll never forget it – not least because I was in a meeting with a regulator at the time. The moves were much greater and faster than everyone expected: you saw clients buying the dips, markets stabilising and then dropping in the last 40 minutes of the day,” says Dan Persson, global head of treasury and market risk at UBS.
Extreme end-of-day moves were attributed to dealers hedging short exposure to gamma – the rate of change in an option’s delta as the underlying market moves. Banks with short gamma exposure are forced to continually sell into a falling market, potentially exacerbating those moves.
Some dealers freely admit they’re still trying to understand what drove the market moves that day. Persson, a veteran of technically amplified market corrections past, having spent several years in Asia-Pacific as the bank’s head of equities and emerging markets risk, had formed a pretty shrewd idea by early afternoon trading.
“We weren’t certain whether the move we were seeing was technically driven, or whether it was fundamentally driven. I think it materialised quite quickly that it was technically driven,” he says.
Like most structured products issuers, a huge chunk of the bank’s volatility exposure comes from its autocallables book – instruments that automatically knock out when an index rises to a pre-set level, or knock in when a downside barrier is breached. The instruments leave the investor inherently short volatility and the bank long. The bank typically hedges this exposure with vanilla options on the underlying index, says Persson, tailored to the expected tenor of its autocallables and estimating when these are expected to knock out.
When spot markets fall, the bank becomes significantly longer vega, and its hedges move out of the money. As a rough rule of thumb, says Persson, every 10% market move downward increases the bank’s vega exposure on the relevant index by around 50%.
With the S&P heading for its largest one day fall in almost seven years on February 5, the bank had a tightrope to walk; it could simply sell off its long vega position and realise a gain from the rise in volatility. But that would also mean shedding the protection afforded by the exposure should markets fall further and threaten the downside barriers on its autocallable books.
Losses on structured products books often occur when banks sell off their vega exposure, only to be forced to buy it back when markets rally.
Unlike some on the Street, UBS maintained that vega exposure and was significantly long gamma coming into February. That provided a cushion, but as markets fell on the morning of February 5, around half of the long gamma exposure was erased, says Persson.
“It was at that point, towards the end of the day, that we decided to get longer gamma, to protect ourselves, and to protect our books,” he says. “We replenished about a quarter, because we weren’t particularly certain how the markets were going to move over the following days.”
That meant buying more vanilla S&P 500 put options for additional protection. But managing options hedges in a low volatility environment is particularly difficult, he adds: “Getting long gamma, particularly very localised gamma, which dissipates with the markets going down, is a very tricky hedge.”
This refers to gamma’s sensitivity to market fluctuations within a narrow range. Once markets move outside of that range, gamma exposure deteriorates dramatically.
“We were well protected over the period because we’d taken a forward-looking view; we’d identified where we had tail risks, and where we were particularly vulnerable. We were able to look at the day’s events against a topographical map of our risks, and quickly find the hotspots,” says Persson.
Once the bank had worked out where its pain points were and how it wanted to respond, however, it still faced a basic problem. Since scaling back the size of its investment bank in the wake of the crisis, the value-at-risk limit it runs on its equities business has been relatively low – offering little headroom when markets turn suddenly and its exotic exposures balloon.
“We took the decision to increase the VAR limit on the derivatives book – which sounds a bit counterintuitive when you have a volatile backdrop and markets selling off. But at the same time we were getting longer vega, due to the exposures on our autocallable books. If you reduce your vega exposure, then it applies further pressure on your VAR and stress limits,” says Persson.
As an accelerant to the day’s fire sales, inverse-leveraged Vix products such as Credit Suisse’s XIV and ProShares’ SVXY – estimated to have some $4 billion of largely retail investor cash in them at the time – were rapidly plummeting in value as volatility exploded.
The collapse of both products sparked a rush for the exits, with dealers falling over themselves to replenish their gamma by buying Vix futures. A record 282,000 contracts changed hands that day, amplifying the moves on volatility.
Unlike its Swiss neighbour and other large issuers, UBS does not offer inverse leveraged exchanged-traded products tracking the Vix, wary of the exposures the strategy places on the issuer to hedge itself should volatility suddenly turn.
“Being short vol in a low-vol environment was not a space we would want to be in, particularly as we run long vega exposure on our structured product books,” says Christian Bluhm, the bank’s chief risk officer.
Based on his own experiences of the 2013 Uridashi blow-up in Japan, and the 2015 HSCEI-linked structured products rout, which followed the huge selloff in Chinese equities in August 2015, Persson has no doubt other dealers felt significant pain on the day.
“A lot of people had been buying the dips. Markets then stabilised and actually rose over a 15-minute period – and then you had that drop in the last 40 minutes of the day. That really made it difficult for a lot of investors’ assets to recover. Many got caught short there, that’s for sure.”
Getting long gamma, particularly very localised gamma, which dissipates with the markets going down, is a very tricky hedgeDan Persson, UBS
Keeping a cool head on a day while markets were tumbling was only half the story, however; the bank was better prepared than most thanks to policies instituted several years beforehand. Since 2015, UBS has moved to significantly to dilute the risk inventory that comes with managing a big structured products business.
While a build-up of big, chunky positions that can be difficult to manage comes with the territory, the bank has boxed clever to avoid the kind of situation it faced on its Japanese structured products business in 2012, applying concentration limits to areas of the books where it sees exposures building up, and looking to diversify the index pairings on popular retail products such as ‘worst-ofs’, which play gains on one index off against another.
In its first-quarter disclosures, the bank reported a significant jump in capital required to support its market risk, to Sfr22.4 billion ($22.6 billion) – virtually double the previous quarter’s minimum requirement. Unlike many of its peers, the bank suffered no VAR backtesting exceptions in the first quarter – meaning realised P&L was in line with modelled estimates, despite the quarter’s wild swings.
UBS has good reason to be leery of back-testing exceptions: under Switzerland’s strict market risk framework, if it suffers more than four within 250 trading days, it gets slapped with a multiplier, which ratchets up the amount of capital it must hold, to account for the perceived risk that its models are not accurately capturing the risks the bank faces.
Aside from being a bellwether of sound risk management, a lower multiplier translates into very real capital savings for the bank and its shareholders. Bluhm is proud of UBS’s record in this department, and says it is no mean feat now that it operates a smaller, nimbler investment bank.
“Since we scaled down the investment bank significantly, we have not had a very high absolute VAR number across the bank. If you run a smaller VAR, you can get hit by larger P&L swings and suffer backtesting exceptions more easily as a result. That’s the other side of the equation. But I was very happy we managed to actually keep backtesting exceptions under control during the quarter, despite the fact you had big swings in P&L and a much higher VAR,” says Bluhm.
Another tool brought across from Asia – the central risk book – has also helped bolster the bank’s equity business this year. An idea that started life on the bank’s equity desks in Hong Kong, it encourages the systematic netting off of exposures internally; typically, recalls Persson, when the derivatives desk was looking to hedge out its delta, it would go out to the Street and end up crossing the bid/offer spread – despite its cash desk being a natural holder of the requisite exposures. A review of its inventories was enough to tell the bank it could do more in the way of recycling risk internally, says Persson.
Porting the idea to London and applying the same logic to second-order derivatives exposures has helped bolster the equity division’s bottom line, says Persson.
“It’s been particularly useful for the corporate derivatives books, where we’re running a few collar financing positions in size on certain names. It allowed us to hedge out a lot of the vega and skew exposure on those positions that we otherwise would have had to find some syndicate partners to take.”