Quants pitch strategies for when bonds no longer work

Investors are flocking to alternative diversifiers of equity risk

For institutional investors, the language of diversification is changing.

The classic 60/40 rule and voguish ‘all weather’ portfolios of alternative risk premia are giving way to less familiar terms: statistical hedging, outcome-oriented strategies and the defensive frontier.

This is how quants describe strategies designed to replace bonds as a diversifier in portfolios. The common goal is to offset growing exposure to equity risk but without the negative carry of costly derivatives hedges.

Institutional investors are receptive. Some already have put these new-fangled hedges in place, or have invested in specialist tail-risk quant funds. And quants expect more to follow.

Asked what proportion of institutions might ultimately adopt these strategies, Nick Baltas, who heads research and development for quant investing strategies at Goldman Sachs, says: “In one form or another, I expect — every single one of them.”

The truth is, investors have few other options. Bonds cannot be counted on to rally much if markets experience an equity crunch like last year’s. Derivatives hedges are expensive. Quant tail risk strategies, meanwhile, dazzled in the 2020 tumult, in some cases returning more than 60% during the worst of the selloff. Investors are in a bind and quants are one of the few groups offering a way out.

Not everyone is so enamoured. Some investors lost faith in systematic strategies after many of them slumped in line with equities in 2018 and again in 2020. The quant creators of the new brand of hedges acknowledge, also, there must be care in their application.

Nick Baltas

“There’s always the danger a few managers get burnt and everyone says: we told you these strategies don’t work,” says one. “But if we can avoid that, the future is bright for this approach.”

Multiple sources attest to the level of interest and activity from investors so far. A quant at one bank estimates that a “high single-digit” number of pension funds have put statistical hedging programmes in place, and about the same number of asset managers.

The heads of three different bank quantitative investment strategies (QIS) teams say more than half their activity since the Covid market slump has involved setting up tail risk programmes.

In January alone, Investcorp-Tages, an investment firm, fielded almost as many requests for proposals for its tail risk fund as in previous years, says portfolio manager Berouz Fatemi.

Protection and punch

The programmes investors are setting up typically have three components: outright hedges for protection; defensive quant strategies to generate a degree of carry; and more convex strategies to “add punch”, as one quant puts it, during a prolonged downturn.

The toolbox includes long volatility bets where those can be done at a lower cost of carry, option spread trades, and strategies with a history of showing defensive properties, such as intraday and medium-term trend following.

Practitioners describe a spectrum of choice, ranging from costlier hedges that are more certain to work, to more speculative plays that generate positive carry. “There’s a quasi-linear relationship between the consistency of protection versus the cost of carry,” says Chloi Karyda, co-head of global investible index solutions at JP Morgan. 

Contractual hedges, such as put buying programmes, sit at one end of the range. “You have the absolute certainty of a hedge working in a market selloff, but also the certainty that you’re going to lose money to time decay every month,” says Alex Paul, managing director of ArchPoint Investors, an adviser to family offices.

At the other end “you have no guarantee that a strategy will make you money in a crisis, but history suggests it will work and the carry is positive”.

UBS’s quants describe a “defensive frontier”, similar to the efficient frontier in investing that represents the optimal threshold of risk versus return. The defensive frontier describes the point where a client’s chosen level of protection, the desired certainty of a hedge working and the cost of carry meet. UBS has created an algorithm to construct programmes that achieve such a balance from the bank’s menu of strategies.

We are seeing a change in how systematic strategies are used by allocators — towards specific objective-driven investment

Pierre Trecourt, PremiaLab

Both banks and asset managers are busy formulating new offerings. Aberdeen Standard Investments, which launched a tail risk fund in August 2020 has conducted 60-plus meetings in six months with potential investors.

More than two-thirds of strategies launched by banks in 2020 were positively convex, meaning they aim to make big gains during a market correction, according to data from analytics provider PremiaLab. “We are seeing a change in how systematic strategies are used by allocators — towards specific objective-driven investment,” says co-founder Pierre Trecourt.

In a bind

It’s easy to see why investors should want such hedges in place. For many institutions — “you’d be surprised at how many,” says Fatemi — hedging today amounts simply to holding S&P puts. But that’s expensive. Implied volatility for the S&P500 is over 20% compared with around 15% before Covid; rolling a 95% out-of-the-money put option monthly cost -0.8% annualised from 2014 to 2020.

It’s also risky. To minimise costs, investors often try to boost hedges when conditions look set to worsen. The speed of the Covid crash showed how shaky a plan that can be. “No-one is able to predict when these things are going to happen. And often it happens too fast,” says Jaime Martinez Gomez, who runs a tail risk vehicle for BBVA Asset Management.

Investors trying to time a selloff are making “the biggest mistake”, Fatemi says. “By definition they don’t know when selloffs are coming or where. It might be the bond market. It might be mortgages. It could be tech companies. You simply cannot time tail events.”

Markets are fragile, meanwhile. Equity valuations appear frothy. The Shiller price/earnings ratio for the S&P 500 stands at its highest since the 2000 tech boom. “Everyone we’ve traded with has said: listen, I’m just worried about my global equity benchmark,” says a bank QIS head.

Low bond yields have pushed investors to take more risk.

But buy-siders have less stomach than usual for drawdowns. “If you lose 20% of your money today, with zero interest rates or negative bond yields, there’s no way to earn it back. You can’t just stick it back in the bank and wait,” says Andrew Lapthorne, head of quantitative research at Societe Generale.

In the past, bonds provided protection — rallying as a safe haven when stocks sold off. But rate cuts, quantitative easing and the resulting collapse in yields have robbed investors of that source of diversification. When yields are low, they have less room to fall in a crisis.

During the Covid equity selloff, US 10-year Treasuries returned only about 1% compared with 8% when stocks fell more than 10% in the past. “Before, investors bought bonds and they had a hedge and some yield,” Fatemi says. “They don’t have that luxury anymore.”

Risk management has to be based on more than simply a backward measure of volatility

Spyros Mesomeris, UBS

Against such a backdrop, the performance of tail risk hedging products has stood out. The Investcorp-Tages fund was up 21% in mid-March of 2020 and ended the year returning 7%. Intraday trend-following strategies that have since become a popular offering from banks rose more than 20% during the March selloff.

An LGT Capital Partners fund that scales into defensive strategies during market stress beat the S&P from the start of February to March 23 — the index’s low point — by 69%. Aberdeen’s fund, which uses systematic strategies created by BNP Paribas, returned 36% in 2020 despite retaining a negative beta versus the S&P 500 of -0.21. Critically, the cost of carry for such strategies when markets are benign has been either negligible or positive.

Caution

Despite the appeal, investors must tread carefully.

Statistical hedges call for more thorough due diligence than many institutional investors are used to, Karyda says. “It’s paramount that investors use tools like scenario testing and keep in mind the risk of over-parameterisation, and the importance of avoiding simplistic on-off signals. We have had only a few big shock events and they’ve all been different from each other. Expecting the next one to be the same as previous ones is a potential pitfall.”

At Investcorp-Tages, Fatemi warns that hedging algorithms can harbour hidden risks. “We’ve seen algos that don’t have limits in terms of exposure or volatility,” he says. “When you check the rulebook as part of the due diligence exercise, you see that sometimes the underlying positions are huge.”

Hedging strategies can fail for reasons that are hard to account for in statistical analysis. In December 2018, equity markets dropped in the last days of the year but the Vix — the most common S&P hedge — barely moved. Traders were reluctant to buy volatility ahead of the holiday. “You need diversification across regions, asset classes, signals and markets,” Fatemi says.

Quants acknowledge that backtests for tail risk strategies can be even less reliable than usual due to the scarcity of data on periods of market stress.

Then, there are the difficulties of marrying quant tail risk programmes with the objectives of wider portfolios.

“Risk management has to be based on more than simply a backward measure of volatility,” Spyros Mesomeris, UBS’s global head of QIS structuring, says. Many quant portfolios in the past were built to reduce risk when markets become less stable. “For defensive strategies, you don’t want to be reducing their weight when volatility is high. That’s the time when they should work,” Mesomeris says.

Investors also run the risk of inadvertently doubling up unwanted exposures. Trend-following strategies, for example, went long equity markets in 2018 and again in 2020 before markets plunged. “If you had trend following in a portfolio that already was exposed to equities you’d have lost heavily when it mattered and in the part of your investments you were counting on to diversify,” says Sandrine Ungari, head of cross-asset quantitative research at SG.

Sandrine Ungari

Some statistical hedging strategies can start to move in a crisis in tandem with other investments or wider markets, as happened last year. And modelling correlations in multi-asset multi-strategy portfolios is a tricky problem.

Standard quant models assume correlations stay constant whatever the size of a shock, which ignores the financial contagion that often characterises severe tail events. “It’s vital to find a model that considers tail risks, correlations and also tail risks in correlations,” says Julien Turc, who heads research for quant investing strategies at BNP Paribas.

Of course, quants already are wrestling with these problems. In a joint venture with fintech company Raise Partner, SG developed an optimisation solver that “looks through” strategies to their underlying holdings and takes into account any overlapping risks.

And in a recent working paper, quants at BNP Paribas proposed using so-called vine copulas to model correlations in complex portfolios. The technique effectively treats multi-dimensional correlations as a network of individual pairwise links.

“Vine copulas are a technical trick that allows extending pairwise distributions to more than three assets,” Turc says. “The approach first identifies pairs of correlated assets, then moves on to capture dependencies between those pairs, and proceeds to the next stage until the whole structure of dependencies between assets is considered.”

Aberdeen says its ability to actively manage the weightings of strategies allows it to avoid risk overlaps and to dial up or down allocations to contractual hedges to ensure its fund retains a minimum level of protection.

So, can systematic strategies fill the role of bonds in institutional portfolios? Quants will first have to overcome scepticism in some quarters about statistical solutions in general, given the recent poor performance of diversified ‘all weather’ portfolios.

In a 2019 study, Kari Vatanen, chief investment officer of Veritas Pension Insurance Company in Finland, found that many defensive quant strategies were correlated to bonds. “I don’t know if that’s still true in this environment,” he says. “But if it is – and if we don’t believe bonds can offer diversification because rates are low – it’s questionable whether those statistical strategies can offer better diversification. At the least, it is uncertain whether they will work or not.”

Vatanen is broadly skeptical about the idea of using statistical hedges. “I wouldn’t trust their defensive properties in this environment,” he says.

Bank QIS heads concede that inflows into defensive strategies so far have been mirrored by outflows from the alternative risk premia programmes that until recently were their stock-in-trade.

There are also limits on how much these strategies can take in. “The capacity is not there to scale to hundreds of billions in portfolios, or trillions,” Mesomeris says.

And yet quants are bullish about the role of statistical hedges in portfolios going forward.

Investors will have to seek returns in new areas: perhaps international bonds, China bonds, infrastructure and real-estate lending or private equity, Mesomeris thinks. But quant strategies, he says, will be “part of that broader solution”. And most in the industry agree with him.

Correction, February 25, 2021: An earlier version of the article stated that two-thirds of strategies launched by banks in 2020 were negatively convex. In fact, the strategies were positively convex. We have amended the article accordingly.