Stress testing a new normal for investors

Consultants and pension fund executives across the globe are putting more emphasis and focus on stress and scenario testing of portfolios following a year of geopolitical surprises and all-time highs in markets.

While stress and scenario testing has been present since the fallout of the financial crisis, sources say shocks in 2016 — such as the U.K.'s vote to leave the European Union and the election of Donald Trump as U.S. president — have been a catalyst to put increased weight on the outcomes of these tests.

It is not so much an increase in the frequency of stress testing, “but putting more focus on the actual outcomes or projected outcomes,” said Matt Maleri, partner, asset allocation at Rocaton Investment Advisors LLC, Norwalk, Conn.

While the worst outcomes were easier to dismiss in the past, Mr. Maleri said they now receive attention. Outcomes often are “getting elevated to committee level, board level discussions, whereas years past it was perhaps just a discussion among staff at Rocaton. We have learned that the so-called one-in-100-year event seems to happen more frequently — we don't find that surprising at all,” he said.

All-time highs in markets also help to heighten awareness of what can go wrong, he added.

Some countries' regulators encourage periodic stress and scenario testing. The U.K. Pensions Regulator's DB code of practice encourages trustees to use a variety of qualitative and quantitative approaches to risk management, such as stress and scenario testing. In the U.S., the Office of Financial Research said in relation to risks in non-financial corporate credit that broader stress testing could reduce financial stability risks. “At a minimum, regulators could require stress tests of the largest exposed groups of firms,” including pension funds, said its 2016 Financial Stability Report. The European Insurance and Occupational Pensions Authority also conducted a Europe-wide test in 2015.

“We would have said it is best practice anyway,” said Alasdair Macdonald, U.K. head of advisory portfolio management at Willis Towers Watson in London. “It is not always the case that regulators are pushing a best practice open door, but this is a case where the regulatory imperative aligns with best practice and with where we are in markets: the world is changing. To the extent that past performance is not a great guide to the future, and models derived from past performance, we do have to rely more on stress testing or scenario modeling.”

While stress and scenario testing is not necessarily new to consultants and pension fund executives, the developments of 2016 have shifted the goal posts in terms of how they are conducted.

In the past, these tests have been driven by a specific event such as the technology boom or the Asian financial crisis. “It tended to be one thing at a time, (and clients) wanted to see stress against this crisis or that crisis,” said Edmund Walsh, director at Meketa Investment Group in London. “But as we are coming into this "new normal,' it is going to be particularly important trying to think about how to handle those different types and variety of risks. There is an increase in the sheer number of (risks), but also the interconnectedness of those — how they play off against one another, and how (to) deal with multiple crises.”

Meketa has been investing in technology and big data, incorporating machine learning into its work to understand the interconnections of risks. “The only way to get these complex, multifaceted moving parts was to really scale up the type of analysis we were doing,” he said.

Aon Hewitt's John Belgrove, senior partner in London, said he is a “massive fan” of stress and scenario testing. “What's interesting is the kind of economic scenarios you might imagine ahead of you keep changing. At any one time in the cycle there are always genuine scenarios that you can see that will have a positive or negative effect in different ways, and also scenarios that you don't know about that can come along and bite you,” he said. Scenario and stress testing is also useful to check biases in a portfolio.

WTW's Mr. Macdonald said while Brexit and the U.S. election were surprises, they were at least defined for specific stress testing. “It is very rare that you get these clear and present dangers, so they have been thought about as scenarios. But a eurozone breakup or a crisis in China, it is not quite clear when they will happen. The problem with scenarios is you are limited by your imagination.”

Brexit was a good example for the use of scenario testing by APG Asset Management, said Willem Huson, senior financial risk manager at the e443 billion ($473.9 billion) firm in Amsterdam. “We have always engaged in stress testing and scenario analysis. It is important to us to apply also assumption-free methods, and that is why it is an attractive way of looking at our portfolios.”

APG executives run regular stress tests based on historical scenarios and generic stresses, as well as regulatory stress tests. “For particular events we try to design custom stress tests to look at the potential impacts.” Regarding Brexit, executives evaluated the outcome of a severe and moderate scenario, forecasting how the vote would affect investments. “In this particular case, because Brexit was eventually decided, we were also able to backtest our forecast scenarios,” said Mr. Huson.

APG analyzes the results of its stress tests, and results may be elevated to the risk committee and communicated to relevant departments or portfolio managers.

Understanding the potential impact of different economic and capital market scenarios is a regular part of the risk management framework at the $306.6 billion California Public Employees' Retirement System, Sacramento, said Wylie Tollette, chief operating investment officer at the fund.

Executives have refined the approach to looking at the impact of stress events over longer time periods. “For example, we used to use 10-day value-at-risk and conditional VAR statistics. We moved that to a 30-day last year and are now more focused on one-year VAR and CVAR statistics, as they reflect the risks that investors with a longer-term time horizon should be more focused on.”

He added stress testing allows decision-makers to anticipate pain and loss, helping them to “determine if the actual risk exposure is aligned with the actual risk appetite.”

Making the experience of a stress or scenario test as real as possible is also something adopted by Rocaton. “We try to put things in dollars, put real money on paper,” said Mr. Maleri. It is one thing to tell clients they may make a 25% loss on a portfolio in a certain event, but more tangible to say they may lose $100 million. “Putting it in real dollars is more impactful, especially with pension clients paying out benefits, or endowment clients meeting spending needs. It is easy to talk in (percentages) but that doesn't translate into how they are using the money.” Clients can then question whether they can live with that outcome, and if the answer is no, rethink asset allocation, he said.

However, use of stress and scenario testing does come with caution, in that “all investors need to carefully examine what decisions or changes to their investment strategy they expect to derive from their stress testing results,” said CalPERS' Mr. Tollette. “Unless it has real relevance to the investment decision-making process it is simply interesting — and perhaps expensive — information.” He also warned that scenarios rarely — if ever — happen the same way twice.

Raj Mody, global head of pensions at PricewaterhouseCoopers in London, agreed stress and scenario testing is a good idea, so long as executives are clear on their reasons.

“A pension fund is an interesting beast, with a long-term horizon, and assets that don't match liabilities, so it is right in principle to understand your vulnerabilities to stress or volatility. But I would caution against doing that in such a micro, granular way that you don't see the wood for the trees,” he said. Rather, he sees the role of stress testing as “a periodic check-in with whether contingency plans (designed to deal with certain stress events) are fit for purpose.”