A future of rising uncertainty demands investors fundamentally re-think the way they assess risk when building resilient portfolios, argued a panel of experts from MSCI and Singapore sovereign wealth fund GIC.
A joint research paper focussed on building balanced portfolios in a dramatically changed market environment. Its authors outlined five forward-looking scenarios future markets could face, and argued investors should assess the resilience of their portfolios against these scenarios, rather than relying on backward-looking assessments of risk such as mean-variance optimisation.
The paper, Building balanced portfolios for the long run: A new framework for incorporating macro resilience into asset allocation, was co-authored by Grace Qiu and Ding Li, both senior vice presidents, total portfolio policy & allocation, economics and investment strategy, at GIC, and Peter Shepard, managing director and head of analytics research and product development at MSCI.
All three authors spoke with Amanda White, director of international at Conexus Financial, in a panel discussion at Conexus Financial’s Fiduciary Investors Symposium held in Singapore.
The paper argues investors face two major shifts in the investment environment. The first is that private assets have gradually moved from peripheral alternatives to lie at the core of many asset allocations.
The second, more sudden development is the period of heightened macro uncertainty markets are now facing, with higher inflation and interest rates adding to the challenge of secular forces such as climate change, geopolitical tensions and de-globalisation.
Private assets: higher returns, lower risk
Grace Qiu said private assets promise higher returns and lower risk, but they can be hard to evaluate and incorporate into the asset allocation process alongside public markets. Investors need a measure that can be consistently applied across public and private markets.
“We all know that simple statistical-based volatility or drawdown measures will potentially understate private market risk due to accounting smoothing or lagged valuation phenomena,” Qiu said. “Using a public market proxy can be a dirty and quick fix, but it lacks fundamental linkage to private markets, and therefore can be quite difficult to communicate to your private market colleagues and receive buy-in.”
Macro-economic risks that could dominate the coming decades include supply-driven inflation, a less-credible central bank, rising real rates and slowing productivity growth, the paper argues. It uses five potential macro scenarios to test a framework for allocations that is more resilient against long-term risks, at the cost of shorter-term volatility.
Peter Shepard said investors have largely been limited by “two ways of approaching risk,” one being very statistical and short-horizon, and the other a purely qualitative approach to the long-term.
“That’s really unsatisfying,” Shepard said, noting investors either operate “like a casino with known rules, or you give up on a more data-driven process altogether.”
The paper presents a middle ground that models the long horizon and then allows investors to apply judgement to the more uncertain components, he said.
A lot can be predicated about the long term
Investors have long assumed that it is difficult to forecast for the short term, and even harder to forecast for the long term, Shepard said. But the team realised there is actually a lot that can be predicted about the long term.
“It’s hard to say how much it will rain next week, it’s harder to say how much it will rain a year from next week, but it’s actually easiest to say how much it will rain next year,” Shepard said.
Cash flow shocks and discount rate shocks are painful on a short term horizon because prices drop, but “discount rate shocks are less and less painful” on a long horizon because expected returns go up, he said.
Conversely, trend growth shocks to the trajectory of markets are among the biggest risks facing long term investors, Shepard said, unlike macro volatility which is less of a systemic risk and more of a “bump in the road.” These risks cannot be managed by looking in the rear view mirror, he said.
Ding Li said mean-variance optimisation has been key in past decades for portfolio construction because it provides a useful perspective, but it is not the optimal choice for long term investors.
The study shows, for example, that with a ten year horizon, bonds are resilient to only one scenario–a demand shock–but not for other scenarios. Infrastructure, on the other hand, shows balanced resilience across different scenarios. But mean-variance approach will tell investors that bonds are a very good all-round diversifier based on the past 20 years of negative correlation between equities and bonds.
“Our portfolio does show relatively better resilience across different scenarios, but the cost is, if you’re focussed on a short-term volatility risk profile, our portfolio has higher volatility than a traditional mean-variance portfolio,” Li said.