Time for a new toolkit to de-risk portfolios

Cross-asset investors are questioning the efficacy of traditional approaches, having witnessed the failure of risk models over the 2008 crisis.

Asset class diversification, the basic tenet of cross-asset investing, did not work because the risk factors embedded in traditional asset classes turned out to be highly correlated, and provided no shelter in turbulent markets. Investors are now realising that in order to deliver a truly diversified, risk-controlled portfolio, it is necessary to diversify the risk factors themselves – and this means looking outside the traditional asset class toolkit.

Diversification is an often-used, but much-misunderstood term. Examining the risk of a typical balanced portfolio, one can see that contribution to total portfolio risk from bonds is minimal and that it is equity variance that dominates portfolio risk.

For the average balanced portfolio, overall returns will still depend on the vagaries of the equity market. More sophisticated investors have intuitively countered this problem by diversifying the typical multi-asset portfolio into corporate and emerging market bonds as well as further-flung equity markets. Yet still, at its core, the portfolio has exposure to only three broad risk factors, namely equity, interest rates and credit quality. With negative and rising real yields on government bonds currently pushing investors to further overweight equity and credit – two highly correlated risk factors – the risk-mitigating qualities of such a portfolio is in question as the bull market matures.

Some investors have taken a different approach, moving away from traditional assets into ‘alternatives’. However, the term is nondescript with the alternatives spectrum encompassing everything from property and infrastructure funds at one extreme to opaque hedge funds at the other. This amalgamation of alternative assets into one ‘asset class’ is unhelpful in the quest for risk factor diversification, especially considering the terrible, and correlated, returns from property and some hedge funds over the crisis.

However, a closer look at the alternatives space shows that the ‘asset class’ can be decomposed into two broad camps – those asset classes that are ‘illiquid’, with investments that are typically associated directly with the physical world – property, infrastructure etc – contrasted with those alternative asset classes that are ‘systematic’, with investments linked to liquid trading strategies in financial instruments.

Access to alternatives of both types has been a problem in the past, with liquidity, fund structure and cost of the available vehicles all proving prohibitive. Yet it need not be: risk factor-based investing has shone a light on structures such as hedge funds, to expose not only how returns are generated, but also that elements of these returns are systematic, and therefore replicable in low-cost, liquid format.

Just as long-only equity fund returns are decomposed into style risk exposures as well as broad market (beta) and stock selection skill (alpha), so too have hedge fund returns been dissected. While many hedge funds give access to the returns of truly skilful managers, much of the industry’s returns can be explained in ‘systematic’ terminology.

These systematic risk factors, collectively named ‘risk premia’, represent an alternative source of return, distinct, liquid and most importantly uncorrelated with traditional risk factors. The ‘style’ risk premia, for instance, while traditionally associated with the equity asset class, are also found across others. They represent returns that accrue to investors that systematically exploit market behavioural effects such as valuation biases (value and low volatility), herding tendencies (momentum), or survivorship bias (quality). Long-only equity funds have long since tilted portfolios towards these factors, but many hedge funds – such as ‘equity quant’, global tactical asset allocation (GTAA) and commodities trading (CTA) funds – also isolate and exploit the same factors but in market neutral format.

Likewise, with the advent of liquid derivative markets came ‘structural’ risk premia. Where a liquid options market exists, the volatility risk factor has been observed, with investors effectively being paid an excessive premium for insurance against sudden market moves. Likewise, asset classes that exhibit term structures have seen strategies develop that systematically exploit the shape of their curves and, similarly, where there is a yield differential, there is a carry trade to be made.

All of these risk factors represent a widely expanded toolkit for the cross-asset investor. While exposure to risk factors individually may deliver good Sharpe ratios as stand-alone investments, the true power of risk factor investing comes at the portfolio level, where low correlation between alternative risk factors can significantly reduce portfolio volatility and catastrophic downside risk from rare events (tail risk). When compared to traditional risky assets, correlations between alternative risk factors have remained low and stable, especially over the 2008 crisis.

In a world where volatility targeting is now ‘de rigueur’, the addition of alternative risk factors to a traditional portfolio brings more stability to covariance estimates and therefore represents the simplest and most reliable methodology to forecast and control volatility.

Risk factor investing is not without its pitfalls. The model of strategic asset allocation with tactical overlays is settled as the standard framework for traditional multi-asset portfolios. Yet this approach struggles to cope with the vastly expanded opportunity set of alternative risk factors. Also, risk factors are expected to generate a positive premium and therefore must have a sound economic rationale for their existence. As exposure to many risk factors is gained by ‘design’ of systematic trading rules, the very existence of the risk factor can be questioned when back testing and data-mining are the only proffered evidence. Similarly, model risk aside, risk factors can also be cyclical and dependent on market regimes of volatility growth and inflation as well as also being capacity-constrained. All of these issues make design, selection and forecasting a non-trivial issue when including systematic factors in the portfolio, so significant research and resources is still required when allocating to these factors. In this brave new world, this at least, is one constant and similarity with more traditional asset allocation that has not been washed away.

Risk factor investing is no panacea for cross-asset investors, but it does represent a seismic shift in portfolio design and philosophy. The decomposition of portfolio risks on a factor basis rather than on an asset class basis will often require an about-turn of the mindset of the cross asset investor, but when achieved, will allow for more targeted portfolio objectives, realigning expected risks and rewards across the portfolio and finally giving true meaning to the word ‘diversified’.

Toby Hayes is a portfolio manager at Franklin Diversified Income Fund

Key features

Cross-asset investors are questioning the efficacy of traditional approaches, having witnessed the failure of risk models over the 2008 crisis.

Some investors have taken a different approach, moving away from traditional assets into ‘alternatives’.

Risk factors are expected to generate a positive premium and therefore must have a sound economic rationale for their existence.