Commentary: Improving portfolio resiliency across risk factors, regimes and time horizons

Diversification is one of the most powerful tools for helping investors navigate the uncertainty of capital markets. It can help limit drawdowns without sacrificing too much return, leading to better portfolio outcomes in the long run. In the context of multiasset-class investing, we believe investors should diversify their portfolios across three key dimensions: risk factors, market regimes and time horizons.

Risk factors are important because investors can expect to earn a risk premium for bearing exposure to systematic risks that are difficult to diversify. By focusing on factors, rather than asset classes, investors can avoid the pitfall of redundant and concentrated risk exposures.

Historically, asset classes have displayed distinct behavior that is highly reliant on the prevailing capital market environment. These market environments, or "regimes," are distinct from one another, shift over time and can be driven by forces such as macroeconomic change, labor market developments and geopolitical events. For example, over the past several decades, we have experienced declining interest rates, low and stable inflation, and a negative correlation between the returns of stocks and bonds, all of which have been supportive of simple, straightforward approaches to diversification, e.g., a balanced portfolio of stocks and bonds. These tailwinds may be abating as we enter an environment characterized by high levels of sovereign debt, peaking globalization, accommodative fiscal and monetary policy, and nominal interest rates near zero. Regime awareness can help protect against an unexpected change.

Time horizon is an important third source of diversification. An effective approach is to combine a long-horizon strategic allocation with a short-horizon active-allocation overlay . This strategy can help protect against the lack of flexibility in a purely strategic approach.

In our view, the four most important systematic risk factors are:

• Growth: Corporate earnings and creditworthiness tend to exhibit positive correlation with broad economic conditions and the growth rate of an economy. Investors willing to bear the risk of a negative growth shock should be compensated for the potential of reduced earnings expectations and increased default risk.
• Inflation: Inflation affects the purchasing power of an asset's future cash flows. Investors willing to bear the risk of losing purchasing power due to rising inflation should expect to be compensated.
• Real rates: Real, or inflation-adjusted, interest rates reflect the effects of inflation and are used by investors to discount future real cash flows. Investors willing to hold long-duration assets expect to be compensated for the risk of higher real rates.
• Liquidity: Liquidity provides investors the option to rebalance or modify their portfolios in response to changes in market prices or prospects. Investors willing to invest in securities that are less liquid and harder to unwind in times of stress should expect to be compensated.

One approach to portfolio construction we have found useful is to create a factor portfolio for each systematic risk factor. To the extent possible, each factor portfolio should isolate exposure to the specific risk factor in question using a broad set of asset classes. An investor can then allocate capital across the four factor portfolios in a risk-balanced manner, as described in the next section.

To guard against being blindsided by an unexpected regime change, we use a machine-learning algorithm to identify distinct regimes and the corresponding return behavior of different asset classes within each regime. We then design a regime-aware covariance matrix (for the four factor portfolios) that seeks to maintain effective diversification irrespective of which regime happens to occur. Allocating capital based on this formulation of "risk" allows us to diversify from typical volatility risk of asset classes as well as the risk of structural behavior changes in asset classes due to underlying regime change.

We believe combining strategies that accrue returns over different time horizons is another effective form of diversification. An approach we have found useful is to complement a regime-aware allocation to the four factor portfolios (described above) with a shorter horizon, actively managed discretionary overlay informed by two key inputs: systematic risk premiums such as carry and momentum, and signals emanating from market sentiment, business cycle and macro outlook. An allocation of about 250 basis points of risk (as measured by volatility) to this overlay can balance risk factor contributions while still benefiting from an active allocation process.

As an illustration of the diversification concepts we have presented, we used them to create a risk-balanced, regime-aware multiasset portfolio. We then conducted a backtest of this hypothetical portfolio from 1998 to the present. We relied on a monthly rebalancing methodology and rescaled the hypothetical portfolio to have 10% ex-ante volatility. In the following table, we summarize the performance of our benchmark illustrative portfolio, absent a discretionary active allocation overlay.

Over this period, our illustrative portfolio had attractive performance characteristics. While there are scenarios, such as a persistent strong bull market in equities or a rapid rise in interest rates where alternative standard approaches might perform better, we believe this approach can deliver attractive downside characteristics and consistent risk factor exposures through time. These features will be critical in achieving resilience to out-of-sample shocks and abrupt changes in market regimes that may well occur in the future.

William Irving is chief investment officer of the global asset allocation division at Fidelity Investments, Avishek Hazrachoudhury is portfolio manager in the division, and Christopher Kelliher is a quantitative analyst in the division, all based in Boston. This content represents the views of the authors. It was submitted and edited under Pensions & Investments guidelines but is not a product of P&I's editorial team.