This bipolar period has seen both extremely elevated and very depressed phases for volatility markets in equities, fixed income and currencies. Notwithstanding the extraordinary market conditions, the distinctive GCAV strategy profited in 2019, 2000, 2021 and in 2022 at the trading level; in 2022 long exposure to forward rates volatility was particularly helpful. This has been achieved while keeping volatility, drawdowns and equity market correlation inside targets and limits. GCAV is distinguished by several features including its focus on several asset classes, a blend of systematic and discretionary strategies, optimization based on equity market tail risk, and by using both exchanged traded and over the counter instruments for implementation. 

Cross asset 

“GCAV sits inside Quant & Overlay, which covers DBAM’s multi-asset liquid alternatives as well as quantitative equity strategies. These strategies play an in important role in the broader multi-asset solutions, which is a strategic focus area for DBAM. In particular GCAV has strong diversifying properties for traditional equity/bond portfolios, which makes it a valuable addition to any multi-asset portfolio” says Thomas Gade, Head of Quant & Overlay.

GCAV currently trades volatility in equity, short term interest rate, long term interest rate and currency markets, mainly in the US and Europe. Though “cross asset” strategies sometimes express relative value views on relationships between asset classes, this would be a fairly unusual trade. “We do cross asset trades from time to time, but it is not a cornerstone to our strategy,” says Jensen. GCAV is mainly expressing views on implied and realized volatility within asset classes. Individual strategies are usually independent, and trades are booked separately (even on rare occasions when there can be opposing and offsetting positions).

Beyond academic risk premia

The volatility space has evolved so fast that their current suite of strategies is very different to their prior experience in pension funds. One enduring quality however is that some of the strategies are partly inspired by academic research into areas such as volatility risk premia and volatility clustering; “ATP has a very strong academic backbone,” says Jensen. Though some of the GCAV strategies have familiar names, they are more sophisticated than the simplest volatility risk premia approach that underly some products such as certain ETFs and some alternative risk premia strategies. The devil is in the detail of how the strategies are implemented and traded, and the managers have made their own judgments when designing strategies, for instance to calibrate trading, hedging, rolling and rebalancing frequencies to the behaviour of the different markets. “The actual implementation and trading are always optimized to the actual market dynamics and microstructure,” says Jensen. For instance, strategies consider the dynamics of underlying markets, such as momentum and mean reversion, and how this can interact with delta hedging. 

A variety of exchange traded and OTC instruments are used to optimize liquidity: equity is exchange traded but bonds and currencies are OTC, bilaterally traded and cleared. Options are plain vanilla rather than exotic. The manager is conscious of keeping the portfolio highly liquid and thus mainly trades options that are actively trading in the markets, e.g. shorter dated equity and FX options and interest rate options across the expiry spectrum. “Volatility trading is a niche area, so knowledge of the specifics around the microstructure is paramount in order to achieve optimal implementation of the strategies,” stresses Jensen. 

This has proved its worth: the strategies have considerably outperformed generic volatility carry approaches while maintaining low equity beta. 

Systematic strategies complement tactical trades

The systematic strategies are generally taking a multi-year view on risk premia and other inefficiencies, whereas the tactical (i.e. discretionary) strategies are more opportunistic in seeking mis-pricings over a multi-month horizon.

The typical split is 70% systematic and 30% tactical, and the manager could only envisage tactical becoming (significantly) larger if there was a radical change in market environments, though given the surprises seen since 2018, it is hard to imagine what sort of change would in fact prompt a reduction of the systematic book. 

Given the manager’s confidence in the persistency of alpha generation from the systematic strategies, and its internal diversification including a mix of long and short volatility and carry strategies across the asset classes, investors might well ask why the fund is not fully systematic. “The answer is that the systematic strategies would then need to be more dynamic in nature, but we do not want to take a view on timing them,” explains Jensen. 

Views on timing are instead mainly expressed in the tactical book, though the manager can very occasionally also use discretion to modify the systematic strategies. One example was that some FX strategies were adjusted during Brexit negotiations. 

Tactical strategies have more freedom than systematic ones: they can be long, short or relative value, and may or may not delta hedge. Taken together, the systematic book also does all these things, but the parameters for each systematic strategy are more narrowly defined.

There have been some years when both systematic and tactical books profited, and others when they moved in opposite directions. In 2022, tactical trades profited while the systematic book lost money, whereas in the first five months of 2023 the systematic trades have been the main driver of returns, in line with the fund’s history. In 2021, systematic strategies also did better, profiting in every quarter against tactical losses in every quarter. “It is in more volatile markets when the two can become more negatively correlated,” says Jensen.

Strategy weights inverse to equity tail risk

The risk-adjusted return target – a Sharpe ratio of 0.5 – is relatively modest and conservative for this sort of strategy but given the minimal equity market correlation it can still be a powerful diversifier for portfolios. More importantly the strategy is not optimized to maximize the Sharpe or minimize standard deviation, though there are caps on it. The Value at Risk (VaR) limit (3.5% weekly for 95% confidence) needs to be published. Unlike some risk models, this risk model is based on historical realizations of all risk factors instead of assuming a specific joint distribution of risk factors, e.g. a normal distribution. “As the risk model is based on actual real-world events, it is a strong tool to gauge tail risks,” says Jensen. This mindset feeds into the strategy allocations. The systematic strategy weightings are substantially fixed, based on their returns relative to equity market drawdowns (of over 5%), which could include both “corrections” and more serious crises such as fully fledged bear markets. Returns are viewed relative to one unit of tail risk, rather than their absolute volatility. (Strategies with a slightly negative expected return could even be considered if they profit under tail risk scenarios.) Jensen believes that, “This is a more robust method of optimization than more commonly seen mean/variance optimization based on correlations and covariances”. 

The worst peak to trough drawdown using daily data was about 10% within March 2020 (the strategy ended that calendar month down 1.7%) but what differentiates GCAV from some short biased volatility strategies is that the drawdown peaked after the Fed’s “bazooka” response, and not when the VIX peaked. This is perhaps not surprising given that GCAV has a mix of long and short volatility strategies, and when its equity market correlation has been both positive and negative at different times. “GCAV does not target a constantly zero or negative equity market correlation, but rather aims to avoid downside equity market correlation,” says Jensen. 

Incidentally, in light of 2022, the manager might be open minded about broadening the tail risk definition to include a blend of equities and bonds, such as a 60/40 portfolio, though for the time being tail beta is defined in terms of equity risk.

Strategy inspiration and development 

Strategies can take months or years of development and testing before being rolled out live. They start with a firm economic hypothesis around premia for convexity or volatility and then design a trading strategy around it.

At the highest level, market habitat segmentation theory provides one explanation for some of the inefficiencies. “We run strategies that try to monetize different risk preferences across different investors,” says Jensen. Other strategies may be based on a structural oversupply or undersupply of volatility in particular markets or geographies, which may in turn be based on structured products selling options or hedgers buying them. 

Curve roll down strategies could pick up a form of “carry” from implied volatility curves. Term structures could also be used to predict volatility clustering and spikes. They might of course throw up some false alarms, but a systematic strategy is looking for a high hit ratio over the cycle. Some strategies are always short or long of volatility while others switch directionality and can also be flat at times, though they may have a short or long bias over time.

FX strategies do not include some Nordic currencies for any sort of “home bias” reason. The analysis that feeds into the basket is based on covariances being more persistent than volatility or returns. 

Some strategies have a “risk on” bias but others have a more “risk off” bent. The fund is mainly active in developed markets across Europe and in the US. The only emerging market exposure is a bearish biased emerging market FX options volatility strategy that could profit from a crisis situation. The emerging market FX strategy has detracted from returns during the risk on climate of early 2023 but is well positioned to profit from an adverse change in market sentiment.

Strategy diversification

Since 2018, the portfolio has been steadily ramped up to eight systematic and three tactical strategies and is probably close to a full complement. Jensen expects that, “Around ten strategies is optimal and there are diminishing improvements from each further addition at this stage”. That said, he does expect to add one or two new strategies over the next year or two and has not deleted any (though some were naturally downsized as the number grew).

A strategy could be deleted if its economic thesis ceases to be valid. “If the economic rationale for a strategy disappears, e.g. because of a structural change in supply/demand then we will most likely discontinue said strategy,” says Jensen.

Corporate credit could be added as a logical extension of fixed income and volatility, but GCAV does not plan to trade commodity volatility.

Outlook

Returns in 2022 demonstrated the diversifying properties: GCAV was flat against a 60/40 portfolio that had its worst returns in decades. This is helping to spur strong interest in the strategy. 

Returns in 2022 were somewhat handicapped by a 1.9% mark to market loss on fixed interest bond collateral, which should be recovered upon the bonds’ maturity. Given management fees below 1%, interest income is now a meaningful source of returns. Year to date returns to May 2023 are up 5%, with contributions from both interest income and trading profits.

The manager is prepared for both continuity and change in financial markets: structural and behavioural anomalies documented over decades provide the foundation of the return stream while tactical trades can sometimes be nimble enough to take advantage of new regimes and new paradigms in some areas of the volatility markets.