Portfolio Construction: Preparing for all probabilities

Related images

  • Portfolio Construction: Preparing for all probabilities
  • Portfolio Construction: Preparing for all probabilities

Related Categories

Emma Cusworth discusses strategic asset allocation weights adjusted for the economic cycle

Faced with an uncertain global economic outlook, right now a key question for those responsible for investing Europe's €5.7trn of institutional assets is how to position portfolios if growth disappears for a long time and inflation continues rising - the dreaded ‘stagflation'.

A growing body of research analyses how returns, correlations and volatility change for different asset classes during each stage of the economic cycle and one result of this is that rebalancing portfolios around a quadrant matrix of fundamental economic factors, most commonly growth and inflation, is gaining popularity.

The greater the equity exposure a pension scheme has, the more its fortunes are hostage to medium-term economic cycles. Since the economy jumped from expansion to recession in 2007, effectively missing the slowdown phase, the realisation has been growing that the simple buy-and-hold philosophy is no longer adequate.

Analysis from Morgan Stanley of annualised returns since January 2001 shows the extent to which returns for a portfolio invested 65% equity, 29% fixed income and 6% cash changed as the economy moved between four different quadrants of the economic cycle characterised as high growth/high inflation (16.5%), low growth/high inflation (-9.4%), low growth/low inflation (-2.2%) and high growth/low inflation (12.7%%). As figure 1 demonstrates, asset-class performance differs greatly over that cycle.

Ed Peters, co-director of global macro at First Quadrant, says: "The dominant approach of taking a long-term average and applying it in a buy-and-hold strategy is simplistic. One of the key things about the economic quadrant approach is that, not only are returns different across the quadrants, but so are correlation and volatility - the actual dynamics of the market change. Taking a long-term average washes that out."

Few expected the extent to which those market dynamics would shift in 2007-08, resulting in a greater focus on tactical overlays designed to rebalance as the economic cycle shifts.

As Andrew Soper, head of multi asset class solutions for State Street Global Advisors in the UK, explains: "If dividing up the quadrants is a good way of finding areas where returns differ, then, by identifying where you are in that economic cycle would be a good way to allocate."

Providers apply their individual definitions of the stages, the factors that should be used to determine them and the inflection points between them, but most agree on a two-dimensional matrix consisting of GDP growth and inflation. However, not only do markets move ahead of the general economy, but GDP and inflation data are also backward-looking, making it difficult to determine a turning point until several months, or even a year after the events.

"Many factors tend to be lagging indicators, whereas most market returns are leading indicators, so that's not going to provide brilliant signals for tactical asset allocation in terms of precise timing," as Soper puts it. "[But] if cycle quadrants run for two years, rebalancing doesn't have to be perfectly timed to achieve more than if you'd stuck rigidly to your strategic asset allocation."

While historical data plays an essential role in identifying commonalities in how assets perform during the different quadrants and the indicators of an impending shift, many models rely on leading market indicators to predict where inflation and growth are likely to be over the following six months.

The extent to which this predictive element ensures investors move early enough to protect their portfolios is demonstrated in figure 2. Research Affiliates' business cycle forecast model (blue line) indicated a roughly 50% chance of the US entering recession from the second half of 2006. "Accordingly, our model cut the allocation to equity-like asset classes, such as stocks or high yield bonds," explains director of research Feifei Li. "While our forecast is done in real time, the National Bureau of Economic Research (NBER) did not announce December 2007 as the start of the recession until a year later, in December 2008." During that one-year delay, the S&P500 fell 41%.

Uncertainty is one indicator of where the underlying economy is going. Periods of increased volatility, as measured by the options-based VIX index, correspond to the economic cycle. Valuations can also be predictive of change. In 2007, credit markets were pricing-in a forthcoming recession, while equity markets were still priced for growth. "When different markets are pricing in radically different scenarios, there is an inflection coming," says Keith Guthrie, UK head of investment management at Cardano.

However, given that inflation is difficult to measure and opinions on the most appropriate method for doing so vary, determining which quadrant the economy is currently in is hard enough. Predicting where it might go next is even more complex. Each model uses a different and highly resource-intensive process taking account of a large number of indicators, including unemployment, corporate health, debt levels, fiscal policy, and more qualitative factors such as confidence levels. The result is a diverse set of views and resulting strategies.

"We can look at history to see whether there are relationships between the market and macro-economic variables, but the answer is highly complex and there is a lot of divergence of opinion," says Lisa Goldberg, executive director for MSCI's Analytics and Risk Research business. "When inflation is rising, inflation-linked bonds naturally do better than nominals, but the relationship is not repeated so well for other phenomena. It depends what is driving that inflation. As soon as you step away from a simple model, it becomes truly granular and more complicated."

Others question whether economic factors are a good driver of asset allocation decisions, particularly in continental Europe where institutions are traditionally more conservative, with considerably less equity exposure than their UK or US counterparts. Peter Bänziger, CIO at Swisscanto is not convinced: "GDP growth and implied inflation do not tell us much that would influence portfolio decisions. I don't know of any Swiss funds that are applying this approach."

Whatever the factors and indicators used, even when these models correctly point to a change in the economic environment, they cannot predict how profound that change will be and therefore the extent of rebalancing required. "It is impossible to predict the degree of change," Peters concedes. "All you can do is prepare yourself."

Still, in a time of considerable macroeconomic uncertainty, being prepared for any eventuality has clear advantages. A global survey of 800 institutional investors and corporate executives by the Economist Intelligence Unit (EIU), sponsored by BNY Mellon, showed that 47% believe euro-zone sovereign debt default is likely, 85% expect further political turmoil in the Middle East and 53% expect the oil price to hit $150/bbl. However, only 25% thought that a double-dip recession in the global economy was likely (46% unlikely, 29% neutral).

"The range of potential outcomes in terms of growth and inflation are wider than ever before," says Mark Taborsky, head of asset allocation product development at PIMCO. "Long-term investors need to have a view on growth as a driver of returns and inflation in order to maintain their purchasing power. The risk premium is the only way to offset liabilities so a systemic shock impacts the ability to cover liabilities. Taking a tactical approach is the only way of getting returns without taking on leverage or liquidity risk. In a low-return environment, being tactical matters even more."

Using the economic quadrant matrix to analyse how portfolios would perform in each of the wide range of possible economic outcomes is an increasingly valuable risk-management tool. Portfolios can then be rebalanced based on the probability of each different scenario taking place.

"Different markets price future economic scenarios with different probabilities," says Guthrie. "By considering how it would fare in each scenario and the probability of that scenario developing, it is possible to come up with a robust portfolio."

While this probability-weighted approach would naturally underperform a strategy that perfectly predicted future growth and inflation, it does mean investors can be agnostic regarding which scenario develops.

"Clients are using the economic quadrant matrix more as a way of understanding potential scenarios than as a forecasting model," Goldberg says. "They are then stress-testing their portfolios for those scenarios, which doesn't require them to know where the economy is going, but rather what their concerns are and how they translate into the portfolio. Playing ‘what if' is a greater part of risk management."

By continually and incrementally rebalancing as the probabilities of different scenarios gradually shift, investors would also benefit from a smoother, less painful process. The majority of institutions tend not to rebalance until the pain is extreme. Markets can absorb a steady shift by big investors, but significant moves at inopportune moments exacerbate that pain.

Furthermore, the current quadrant is widely expected to last longer than previous comparable periods as concerns remain about US debt and spending, euro-zone sovereign defaults, the ‘Arab spring', a further refinancing wall in 2014 and little end in sight for the energy crisis.

As Joe McDonnell, head of portfolio solutions for Morgan Stanley Investment Management, says: "The question is what happens when growth disappears for a long time. The next expansionary cycle will take longer than normal to reach and be subject to bouts of doubts reflected in market volatility. The behaviour of the markets is somewhat of an issue. With less consistency of growth globally and interest rates also moving at different times and speeds, correlations of risky assets are gradually breaking down."

Even if these economic quadrant models don't precisely predict the turning point, faced with continued low growth and rising inflation, the cost of not rebalancing to account for the fundamental economic drivers of risk could be far greater than being exposed to an extended period of poor or volatile returns, particularly so for schemes with high equity allocations. They present a valuable opportunity to stress test portfolios and prepare for any economic outcome.

Have your say

You must sign in to make a comment


Your first step in manager selection...

IPE Quest is a manager search facility that connects institutional investors and asset managers.

  • IN-2454

    Closing date: 2018-08-01.

Begin Your Search Here