Alt risk premia chasing 'tail beta' – again

Quant strategies that failed in the coronavirus crash face a reckoning

When markets fell into a pandemic-provoked tailspin in March, investors in alternative risk premia strategies got an unwelcome reminder that ARP won’t always help in a crisis. These alternative strategies – intended to move independently of wider markets – fell as if they were tethered to plunging stocks. Some portfolios are down by over 15% year to date.

The crash exposed technical flaws in the way some strategies are constructed. In many, it has shown up so-called tail beta – vulnerability to moves in stocks or bonds during a period of stress – even if their beta to those markets as an average over the long-term remains low.

“Clients are asking how alternative these risk premia are,” says Spyros Mesomeris, who heads a unit constructing quant strategies at UBS.

For investors, the 2018 episode of ARP tail beta is still a vivid memory. AQR's Cliff Asness evocatively dubbed it the sector’s ‘Ragnarok’, a reference to the cataclysmic battle of the gods in Norse mythology – a world-ending event.

Recent performance has been even worse.

A Societe Generale index of ARP funds, for example, crashed nearly 12% between late February and late March this year. In 2018, the funds fell less than half that amount.

Strategies that investors see as defensive – and would expect to perform better in a downturn – simply haven’t.

The broad collection of quant hedge fund strategies used in ARP, which are usually bundled together with the hope they diversify one another, range from trading long/short equity factors such as value or momentum, through cross-asset carry or trend-following to things like merger arbitrage.

“This is not what you want from a long/short fund,” a pension fund portfolio manager says. “Even if it only tracks the market or even underperforms, you’d expect it to have less risk.”

Banks and quant funds contend alternative risk premia was never meant as a crisis hedge and that performance is expected to suffer in tough times.

But many asset owners see things differently. A poll by consultancy bfinance found a third of investors that used alternative risk premia were “not satisfied” with Q1 performance.

Clients are asking how alternative these risk premia are

Spyros Mesomeris, UBS

Antti Suhonen, an adviser at investment consultant MJ Hudson Allenbridge, talks of a “repeat re-evaluation” of the sector by asset owners. A fund manager says he expects redemptions to start in the coming months: “There have been too many years in a row of disappointing results.”

The correlation of long/short equity low-volatility strategies to wider stocks reached around 80%, according to Unigestion Asset Management’s Joan Lee, who runs the firm’s ARP fund.

Equity value, which investors might hope would at least hold up when frothy stocks tumble, recorded eight daily drops of more than two standard deviations in March, according to UBS data. Such down-days should occur about once a month.

“People thought that aggregation of these risk premia would offer some kind of protection from market risk, and that has not really functioned,” says Jean-Philippe Bouchaud, chairman of Capital Fund Management.

The existence of tail beta was not unknown, practitioners say, but perhaps was taken too lightly.

“Some trades have tail beta that you don’t see in the history that much,” says Mesomeris at UBS. But for many of these strategies, to perform when markets are going up, “there’s always some residual beta hiding somewhere”, he says.

“They may be market neutral over the length of a cycle, but at any particular point in time these strategies can actually have some directional bias.”

Already, strategy creators are changing their models to avoid similar slip-ups in future – adding more strategies seen as defensive to the mix – or plugging tail risk products that could go alongside alternative risk premia investments.

All kinds of pain

Covid-driven de-risking was the common cause of strategies stalling, but individual trades have broken down in different ways.

In some cases, technical hitches were at fault. Indiscriminate selling of stocks wrecked the beta-adjustment of equity low volatility strategies, which go long low-volatility and short high-volatility stocks.

The strategy works – normally – because long-only asset managers favour riskier stocks that allow them to reach their portfolio’s target volatility without leverage. Banks and quant funds lever up the low-vol leg to make the trade market-neutral. But in March, the beta of both legs converged to one. As the exposure of the long-beta leg exceeded the short, the strategy fell with the market.

The effect of this hedging miscue can be seen in how low-vol as a factor behaved in its vanilla state, says Societe Generale head of quantitative research, Andrew Lapthorne. In un-adjusted form, the strategy beat the market. But beta-hedged versions did the opposite (see figure 1).

 

In equity value, strategies that typically neutralise exposures at sector level turned out to have unwanted tilts at a deeper level. Practitioners found themselves neutral industrials but long transportation, or neutral IT, but long software services – industries that were negatively affected by coronavirus.

Elsewhere, vulnerability to thinning liquidity was spotlighted.

Previously successful strategies that trade implied-index volatility against volatility on single stocks suddenly lost money when investors rushed to hedge against falling markets, prompting a spike in implied volatility on the S&P. Trades that trade volatility across indexes suffered also.

These strategies arbitrage price dislocations that usually make S&P implied volatility relatively expensive and arise from variable annuities hedging and the sale of retail structured products by banks. But liquidity can drain quickly in instruments where it was already thinner.

“If you’re trading some sort of liquidity mismatch and a liquidity shock hits, it passes through different parts of the market at different speeds,” says Adam Singleton, head of investment solutions at Man FRM. “We saw that in 2008 and we definitely saw that last month.”

Even in strategies like merger arbitrage, where firms are trading in two similarly liquid instruments, the trades are often crowded because there is one obvious way to be positioned. When arbitragers unwind those bets, spreads widen rapidly, he says.

In other instances, the crisis proved the safest-looking strategies can go wrong when market moves are extreme.

Dividend premia strategies crashed when companies U-turned on planned payouts, causing a 55% drop in the December 2020 Eurex-listed dividend contract during the month. “That alone could have blown you out of the water,” says one practitioner.

The violence of the moves was incomparable with anything in back-tests. It prompted Unigestion to liquidate a dividend premia strategy mid-month. Henrik Nordestgaard, head of portfolio construction at Danish pension fund PFA, thinks banks selling out of the trade may have exaggerated the slide.

Slumping returns on high-yield CDS in the US left a popular trade – going long high-yield, short investment grade CDS – with 5-6% losses, even if credit beta had been neutralised.

And a flattening investment grade credit curve upended strategies that go short 10-year CDS and long five-year. The CDX five-year point widened 40 basis points but the 10-year point widened only 15. Ordinarily, the strategy generates a premium because the roll-down is steeper at the short end of the curve compared to the long-end.

These trades are considered super-safe because they take positions along the same curve and could be levered as much as 10-times, practitioners say. Strategies that sought to neutralise sensitivity to underlying credit spreads by going long more of the five-year than short the 10-year point suffered bigger losses.

It was the worst of times 

Of course, strategy creators argue that the times have been exceptional.

Anthony Lawler who heads GAM Systematic says members of the team who traded these strategies through the financial crisis felt as though history was repeating itself. Returns at Partners Group, where GAM’s Lars Jaeger ran an alternative risk premia programme in 2008, fell in a similar pattern at that time (see figure 2) before recovering steadily in the following months.

 

“Part of the risk you take in any strategy is that if the system has a massive moment of panic or crisis you are unlikely to make money at that turning point. The only way to make money at that point is to be long volatility. And most investment strategies are not long volatility,” Lawler says.

Practitioners will say the misery now is the reason strategies earn a premium when conditions are normal. “In alternative risk premia you get compensated for taking risks that most investors don’t want in their portfolio,” says Cyril Lureau, chief investment officer at ERAAM, a boutique quant manager in Paris.

The strategies often go long assets that are unloved or less liquid – like cheap stocks, or less-traded options. But unfavoured assets sell off hardest in a crisis. And liquidity often dries up in the first phase of a panic as investors retrench.

As CFM’s Bouchaud puts it, risk premia have to express their risk at some point. “This is that point.” At the same time, the strategies that hurt most when they go wrong often have higher long-term returns, he says. “That is the trade-off.” 

Stretch the argument to its logical conclusion and recent events could be portrayed as a real-world stress test, from which alternative risk premia programmes emerge bloodied but viable.

“In the past four years, we’ve seen everything: liquidity events, trend reversals, the natural gas shock in the autumn of 2018, value strategies suffering. We’ve seen every type of shock you can witness in the portfolio,” says Lureau.

Practitioners appear sincere in saying this – but clients will be tough to persuade. That’s the problem, concedes one manager: “To convince investors that now is the time to get in when it’s so cheap – it’s impossible.”

Turning tail

Already, then, firms are moving to reduce tail risk in their programmes.

NN Investment Partners hurried through a change to its momentum signal in March, adding a shorter-horizon signal. The update had been planned for some time, but was yet to be implemented. The change will make the fund’s trend-following strategies more reactive to market reversals.

GAM Systematic rebalanced its alternative risk premia portfolio mid-month rather than wait until month-end as it usually would. Its models cut risk by 10-15%. The firm is now looking at writing that capability into code so the same would happen automatically in similar episodes in future.

Part of the risk you take in any strategy is that if the system has a massive moment of panic or crisis you are unlikely to make money at that turning point

Anthony Lawler, GAM Systematic

At Unigestion, Lee says the firm is speeding up plans to add additional defensive risk premia to its mix, starting with fixed income value. And Bouchaud plans to allocate within CFM’s risk premia fund to ISTEC, the firm’s beta-protected trend-following programme, which trades short-term trend with long-volatility hedges.

Looking ahead, some commentators reckon firms could make a virtue of their readiness to overrule models in a crisis or add discretion to the investment process.

“We’ve already seen a move towards combining systematic macro with discretionary elements,” Suhonen says. “Possibly we will see more discretionary management being applied to systematic strategies.”

Milind Sharma, a quant fund manager and chief executive at QMIT, a provider of quant research and signals, advocates making discretionary allocations to groups of strategies, one that does well in risk-on environments, a second that does better in bear markets.

To the end of March, risk-off factors defined by QMIT delivered 17.5%. The risk-on group lost 5.8%.

Unbundling

One possibility is that clients will demand a different strategy mix.

“It’s the end of the fallacy that you can put together any portfolio of risk premium strategies and be market neutral. When strategies with tail-beta exposure are present, the concept of market neutrality works a lot better in calm markets than in volatile ones,” Mesomeris says.

At ERAAM, Lureau thinks asset owners will want to invest exclusively in those strategies “they believe in” and make sure they fit with wider portfolios. ERAAM is considering launching a fund combining strategies based on fundamental signals alone – certain equity long/short factors, credit strategies and merger arbitrage. Already ERAAM offers a separate fund of equity strategies like this.

Investors may also allocate to defensive portfolios and alternative risk premia in combination.

Quants at Deutsche Bank say firms will have to revise how they assess the defensiveness of strategies, no longer relying on volatility measured over long periods as a proxy for risk.

“Volatility is not the right estimate for tail conditions,” says Caio Natividade, head of quantitative investment solutions research. Instead, the bank advocates testing the improvement in expected shortfall – the “convexity” – when a defensive overlay is added to a portfolio.

It’s a massive tail risk you are selling and very few, if any, things pay off in that scenario to offset that risk, causing the large losses of such strategies

Stan Verhoeven, NN Investment Partners

LGT Capital Partners is offering a fund that scales into safe-haven assets and long-volatility strategies when signals indicate market stress. “Protection” can be thought of as a style premium much the same as value or carry, Pascal Spielmann, head of the investment team, says. Except in this case it pays off when trouble strikes. Spielmann calls it a “pressure valve” for portfolios.

A few strategies will doubtless fall out of fashion. Newer funds and the next generation of bank offerings will have to back-test using a history in which low volatility struggled in a crisis. Value and short-volatility – another big loser in March, but not a surprising one – will likely play a smaller role in portfolios.

What now?

Ironically, for investors the opportunity now in alternative risk premia may be the best seen in a while.

Systematic strategies have done badly in the early stages of past crises when markets experienced a liquidity shock. Later, when a real-economy downturn set in, the strategies recovered.

In the liquidity-driven quant “quake” in 2007, statistical arbitrage funds – which traded some of the strategies the alternative risk premia sector trades now – saw big falls. Those same quant managers went on to beat the market in 2008 and 2009. Could the same happen again?

Some of the trades that veered lower in March have stabilised. SG’s merger arbitrage premia index was down 4.6% year to date at the time of writing, having been down more than 17% in March. SG’s hedged equity low-volatility index has recovered to be down about 5% year to date at the time of writing, from a low of -13%.

As one portfolio manager says: “The outlook for the sector doesn’t look good. I wish it looked better. But then again the crisis has maybe just begun.”

What ARP funds don’t do

“Some alternative risk premia firms have done well considering the environment,” says Antti Suhonen, an adviser with investment consultancy MJ Hudson Allenbridge. “Most have lost money. But some really got the wrong end of the trade.”

Why such big differences? One factor is what firms choose not to do.

“Clearly there are different strategy choices,” Suhonen says. How strategies are implemented has played a role: “Some of it could be due to active risk management.”

Dr Markus Ebner heads multi-asset investing at Quoniam Asset Management in Frankfurt. “Our strategy is strictly restricted to those premia that are not linked to equity or fixed-income market performance. And we build the premia in a way that ‘hidden’ beta exposure is avoided,” Ebner says.

Quoniam shuns trend-following because the strategy picks up varying exposure to stocks and bonds. The firm separates developed- and emerging-market currencies when implementing forex-based premia to avoid over-exposure to emerging-market currencies, Ebner says.

Elsewere, LGT Capital Partners is not investing in equity low-volatility or similar low beta strategies. “Low-vol is short-vol,” says Pascal Spielmann, head of the investment team.

The strategy is a bet on investors’ leverage aversion, which financial theory says is a premium that can be captured by building a portfolio that buys less-risky stocks and sells the riskiest ones. To keep the portfolio beta-neutral, the long leg must be levered in nominal terms, making investors vulnerable to bigger losses in a crisis, when individual stock correlations spike and historical betas decompress due to indiscriminate selling by investors. It is a “falsely defensive” approach, says Spielmann.

NN Investment Partners trades equities in index form only. Investing in single names doubles up on risk that clients run in their wider portfolios.

Meanwhile, so-called structural premia strategies have an equal share of fans and critics. These strategies such as dispersion trades capitalise on liquidity imbalances in derivatives markets, often created by banks selling structured products. They were among the worst hit in March.

“It’s a massive tail risk you are selling and very few, if any, things pay off in that scenario to offset that risk, causing the large losses of such strategies,” says Stan Verhoeven, lead portfolio manager for factor investing at NN IP. Quoniam, likewise, chooses not to trade option-based premia on equities and bonds, Ebner says.

Another factor is how far firms are willing to override systematic strategies in a stress. Unigestion Asset Management manually raised risk scores in its models in response to short-selling bans and strained liquidity in March, effectively halving its allocation in equity factors. But many managers distrust human judgment in a market crisis and are reluctant to make such adjustments even when things look bleak.

Different funds are targeting different levels of volatility, so some dispersion in performance is natural. But how much of their respective fortunes could firms have foreseen? “Luck was a huge factor,” says one manager at a firm that has fared relatively well. Luck aside – this latest episode underlines that in alternative risk premia, no two funds are the same.

Correction, May 5, 2020: An earlier version of this article stated incorrectly that GAM was already coding a permanent change to its ARP model. In fact, GAM is considering the change. 

 

Editing by Louise Marshall