Asset managers and owners are deploying novel bank-quality modeling and alternative datasets as they take aim at risk-informed portfolio management.
The global COVID-19 pandemic is changing the risk management vocabulary and toolkit for investors of all stripes. With the outbreak’s early months rocking markets to their core, the world’s largest asset managers and institutional investors saw their investment process and risk frameworks tested in ways even the 2008 credit crisis did not fathom.
The pandemic did not happen in a vacuum, though. In fact, many risk teams view the events of 2020 as an accelerated and perhaps more intense version of movements already afoot on the buy-side—towards broader quantification of exogenous, extreme events, and more sophisticated risk assessments being plugged into the portfolio management process earlier, and in a more automated fashion. A fusion of risk and investment.
Likewise, remnants of the 2008 crisis continue to cast long shadows over today’s events, whether in global prudential responses, investment styles, linkages between different risks now under scrutiny, or indeed new regulations that continue to be foisted upon investment managers that govern how, and by how much, they can be exposed to certain trading activities or markets. After braving a year of cacophony and chaos, firms are using the lessons learned to project a path forward into 2021—be it on their data spend and technology or buy-in strategies and quality controls.
As part of November’s Risk Live 2020 event, Risk.net convened a panel discussion on this topic with three buy-side experts taking stock of the pandemic’s consequences as well as looking ahead.
The early pandemic found financial markets moving “totally irrationally” as one speaker put it, but developments in the years after the credit crisis had set the scene for a particular kind of response; investors moving into larger risk premia and focusing on smaller short-lived crises like the 2011 Ukraine conflict that had strong downward movements, but were followed by stronger recoveries.
Pascal Traccucci, who heads up risk management at La Française Asset Management, says that his firm adjusted some of its risk models based on a similar environment to 2008. The difference here was not only the drop in the equity markets, but also what happened in the volatility asset class. “That trade had become quite popular, where many in our industry decided to go short to get the volatility premium,” he says, “and during the early pandemic that was clearly a big challenge for many of the players.”
One of the key lessons lies in using risk metrics to better manage these shock “structural breaks,” panel members said. The events unfolding in late February and early March moved incredibly quickly and with greater volatility than in 2008. “It was something like the fastest ever correction in the S&P 500 from a new high into bear market territory,” explains Christopher Reeve, risk director at systematic investment manager Aspect Capital. “Things can happen quickly and, along with all the volatility and increased correlations in the markets, we were challenged by new regulations. So you need to know how to respond in advance; the time to be figuring out how to do that, and who should be doing that, is not right in the teeth of the crisis, when you’re in a hurry.”
The pandemic changed the perception of what's possible and revealed new risk management requirements.
Risk managers are building out greater stress testing capabilities as a result. However, as Boryana Racheva-Iotova at FactSet puts it, they also recognize that the pandemic changed the perception of what’s possible and revealed new risk management requirements. “They acknowledge that simply modeling the fat tails, i.e., predictable extremes, is no longer sufficient,” she says.
“On the asset owner side, we now see them preparing for possible asset allocation disruptions,” she explains. “We also see a strong need for tactical nimbleness on the portfolio management side, driving managers to find ways for much more granular dissection of company attributes; how to measure supply chain risks, and finding ways for cheaper execution of tactical reallocations.”
The nature of buy-side stress-testing programs is evolving right along with the universe of risks and opportunities created by volatility. But how do you build effectively with these shifting priorities?
One way of doing this, says Reeve, is for risk teams to generate different perspectives on how a scenario will play out. “Risk is not just about the impact of a stressor, but also the probability of that event happening,” he says. “Today that is a large part of my role: making sure our risk strategies are built into the pre-trade part of the strategies, rather than being reactive afterwards. We developed an integrated approach, what we call our ‘extended’ or reverse stress test, which is defined as a triptych: start with scenario, loss, plausibility, or start with plausibility.”
This effort, which borrows tenets from beefier investment banking programs, is also more relatable to both regulators and portfolio managers, Reeve adds. Traccucci’s team worked with La Française’s hedge fund unit on a similar program, engaging with managers to synthesize and unpack the kinds of scenarios they see as plausible, and then building them into a testing framework that incorporates expected outcomes on both the downside as well as the upside. “We developed an integrated approach, what we call our ‘extended’ or reverse stress test, which is defined as a triptych: start with scenario, loss, plausibility, or start with plausibility,” he explains.
But developing this institutional knowledge remains a challenge. Racheva-Iotova at FactSet says achieving this level of sophistication means tests need to incorporate more complex concepts such as causality, on top of the kinds of exogenous factors that can make for a once-in-a-century event. “Clients are looking for multi-step stress tests and ways to make them actionable, translating findings into the investment decision-making process itself,” she says. “This makes the entire process of stress-testing construction more challenging, not only because of the need to identify and quantify those endogenous factors, but also how to embed them through causality, not simple correlation, and how to assess the impact of their transmission into the financial markets.”
That transmission is also filtered by human behaviors and elements that create linkages between what has been traditionally viewed as independent risk flavors, creating yet another front for analysis pre-trade or allocation.
For instance, suddenly dimming performance scares investors, who may begin redeeming their shares. “For an asset manager, that’s when the situation gets worse,” Traccucci explains. “On one side, your assets are no longer as liquid but at the same time, the client wants to get out, and how you work your way through that can really affect your reputation in the marketplace. One of the new issues, then, is how this market risk and liquidity risk interconnect, which could lead to some higher correlations than the ones you have observed in the past.”
The same was proven true of liquidity and regulatory risk, as some markets that are traditionally highly liquid—like S&P futures—showed stresses during the early pandemic period. Panel-members said that lower volume at the top of the order books caused a withdrawal and slightly wider bid-ask spreads as well, which can have an impact depending on both the strategy and the rules in play.
The complexity, says Reeve, becomes a technology issue too, as systematic managers like Aspect now run much more sophisticated strategies, such as relative value plays, driven largely by machine learning. “During 2020 we were affected by the short-selling bans that were introduced in a hurry by a lot of the European regulators. So, while on the face of it those markets remained liquid, we were severely restricted in what we could do in our trading. If you’re constrained from taking one of the positions that the model wants to take, for liquidity or regulatory reasons, you’re left with an unbalanced portfolio. So you have to retrain your model in a hurry in order to react, and to get back to a more sensible position,” he says.
Finally, participants pointed out that new or alternative datasets feeding into stress tests will form a key pillar of these programs going forward—whether social, climate, or political—and ranging from totally unstructured to newly-emerging benchmarks in themes like ESG, which will affect strategic asset allocation and asset-liability management.
Once ‘alternative’ datasets are already mainstream.
Here the pandemic era may not have created new pressures, so much as proving that affinity and governance for these types of datasets is already a necessity. “We are improving our capabilities in long-horizon modeling, things like social impact and climate risk,” says Racheva-Iotova. “Incorporation of these kinds of factors into the risk and portfolio management is critical, and if this year proved anything, it’s that these once ‘alternative’ datasets are already mainstream.”
Reeve, whose operational risk committee at Aspect is now focused on non-financial risk “war games,” says these new considerations can cut several directions, but particularly to the other side of the trade. “We’ve thought a lot about counterparty failures [and] the operational impacts—be it a trading counterparty or a service provider—making sure that we have backups in place.”
Indeed, the possibility of a pandemic was a global blind-spot. For the buy-side, which has spent much of the last decade chasing efficiency, it was a reminder: derive value from risk and don’t get caught on autopilot.
“I think stress tests need to be tailored to the portfolio you’re running; how frequently you run your stress tests depends on how quickly the portfolio changes,” Reeve adds. “What we like to do is to automate these stresses once we’ve built them. But that’s actually the really important part [in this new environment]. It’s easy to automate a risk check and then forget about it. You’ve got to also have someone looking at it, and saying, “Hang on, this looks interesting; this looks concerning.”
And increasingly, while portfolio managers anticipate the next investment decision, that is the question now being asked.