Finding genuinely defensive ways to diversify portfolios

Like many things in life, “defensive” assets do not always behave as promised. But these are particularly unusual times. The Covid-19 crisis has triggered the fastest bear market in history, followed by the fastest initial rebound. Traditional defensive assets provided some protection during the March sell-off, but to a lesser degree than in past crashes. This raises the question as to what constitutes a genuinely defensive asset in the era of coronavirus and beyond.

To answer this, we briefly examine asset class performance during the recent market collapse and correlations between security types. We then consider likely market scenarios, what makes an asset truly “defensive” asset, and how investors can diversify portfolios to better withstand future market evolution.

Traditional assets offered protection through Covid-19 crisis, but not enough

Defensive assets should typically have a negative correlation to risk assets. But the success rate for traditional defensive assets during the March sell-off was mixed.

High quality government bonds (e.g. US Treasuries) did correlate negatively to risk during the sell-off. But compared to previous crises, their capacity to protect was somewhat handicapped by the starting level of government bond yields. With yields now at record lows, we would expect the size of any protection from government bonds to disappoint even more should there be another leg down in markets — unless the Federal Reserve decides to introduce negative interest rates, which looks unlikely at present.

Investment grade (IG) bonds, unlike in previous episodes such as the Global Financial Crisis, did not provide a negative correlation to risk. IG bonds have two main drivers of returns: interest rate duration and spread duration. Historically, the interest rate duration impact (i.e. the fall in the risk-free rate) has more than offset the impact from widening spreads, resulting in the asset class posting a positive return in market crises. This time around, however, given the lower starting level of government bond yields, the interest duration impact was insufficient to offset widening spreads.

Currencies generally behaved as expected during the crisis. If we look at the whole of March, the yen had a negative correlation to risk, though intra-month it was volatile until dollar swap lines were put in place. In terms of the size of the move, the yen worked well as it benefited from attractive starting valuations.

Alternatives presented a mixed picture. Real estate, which often takes time to reflect broader economic troubles, was hardest hit by the lockdown as retail tenants struggled to cover rents. Select volatility-driven strategies, however, did provide a negative correlation.

Possible scenarios: Japanification, some growth or stagflation

While traditional defensive assets such as government and investment grade bonds may continue to offer some protection, it is clear that they will not always prevent capital loss. Instead, investors may need to think more laterally about what “defensiveness” means, and consider other ways of introducing defensive exposure.

A helpful way to think about diversification in the current environment is to consider a range of possible scenarios and then combine elements of protection for each. Here are three we believe could unfold over the medium term:

1) Japanification of US and European markets

Going underweight or short Japanese government bonds in the past two decades is sometimes referred to as the “widow-maker trade” as it was the bet that never paid off: yields have not rebounded. If this type of “Japanification” occurs in the US and Europe, then government bonds should play a more defensive role in portfolios, as the risk of persistently rising bond yields would be remote. In this scenario, interest rate volatility should be kept low, and investors can move further along the yield curve and embrace duration.

Yield curves may have flattened already across many major bond markets, but compared to Japan they have further to go. The Japanese government bond (JGB) yield curve is flatter than the US curve (Chart 1), while absolute yield levels are lower in Japan across the curve, including at the very long end (for example 30year Treasuries are at 1.5% compared to 0.6% for JGBs). Finally, any rise in yields should be viewed as a buying opportunity, if one subscribes to the Japanification scenario.

2) Low interest rate volatility, with improving US growth and gradually rising bond yields

If growth prospects improve, the need for defensive assets will diminish. That does not necessarily mean, however, that government bonds would be a negative contributor to investors’ portfolios. If the move higher in yields takes place in a gradual manner, with well-telegraphed guidance from central banks, then returns from government bonds will not necessarily be negative as their coupons will offset some of the negative duration impact.

US Treasuries should fare better in this instance as starting yields are higher in the US than in Europe. The higher government bonds yields move, driven by better growth prospects, the higher their ability to provide better downside protection; as such, their defensive role should be reinforced for every upwards move.

3) Stagflation — rising inflation with no growth

Stagflation (a period of high inflation and stagnant demand) is most likely to be driven by “bad inflation”, via a weakening currency or supply shortages. This is because “good inflation” (i.e. wage growth) usually occurs when the economy is expanding. Stagflation is the worst outcome for government bonds, as correlations between government bonds and equities could rise and diversification benefits would disappear. Central banks would be forced to raise rates as inflation targeting is typically their foremost objective. If they had to lift rates rapidly, government bond returns would be negative.

Central banks may try to avoid raising rates due to fears that higher rates could push corporates and governments into default, given record debt levels. If rates do not rise, there would be a limited negative impact on performance from a duration perspective (as yields will not rise); but investors would still be worse off in real terms as the value of their assets would depreciate every day that inflation ticked higher. While nominal government bonds could struggle in this scenario, inflation-linked bonds or inflation break-evens should offer some protection.

What constitutes a truly defensive asset?

Finding a truly defensive asset that works in any scenario is no easy task. Recent months have demonstrated that, while asset classes exhibit a range of different risk characteristics, it is perhaps insufficient bluntly to label any asset “defensive” without context. Indeed, asset classes may exhibit defensive characteristics at different times, under different conditions and to different degrees. Some offer stronger elements of absolute defensiveness (i.e. the avoidance of capital loss), while others offer relative mitigation (i.e. less pronounced drawdown versus other asset classes).

To build a strategy that can meet objectives over multiple market cycles, we need to look beyond the simplistic “risky/defensive” dichotomy to understand the drivers of individual asset class behaviour, and the effects created by combining them in a portfolio. Combining assets with different triggers for defensiveness, such as moves in rates, liquidity and correlations, can help to increase overall portfolio resilience.

Asymmetry of risk and return is key, so we believe in taking investment positions which exhibit higher upside potential than downside risk. While it is impossible to predict the future, it is possible to hypothesize how the different asset classes might react and interact in the deflation and stagflation scenarios set out above, based on a variety of factors.

Building defensiveness into portfolios with the aim of protecting capital

There are a number of ways to build defensiveness in a portfolio with the aim of capital protection, using asset classes or strategies which all (in their own way) have defensive elements.

Government bonds: Look beyond US Treasuries

For example, if we assume a risk-off

Japan-style scenario, government bonds in developed markets such as the US may continue to behave defensively. And if the Fed changes its mind on negative rates or moves to yield curve control, then the diversification role of Treasuries may become more meaningful than it is now.

But to capture the magnitude of the negative correlation nearer-term, consider other developed markets or Chinese government bonds. Yields for the latter remain attractive compared to US Treasuries and have different drivers.

Leverage: Not just about adding risk

Another way to achieve greater magnitude is to use leverage. Historically, leverage has typically been dialled up to enhance return or yield characteristics, adding degrees of risk to a portfolio overall. But, in future, more investors may consider using leverage to boost the potential defensiveness of traditional asset classes such as Treasuries during an equity and credit sell-off. This type of strategy is not new, though more leverage may be required than before to achieve the desired result, increasing the potential risks if yields spike. Nonetheless, it may allow investors who are comfortable with Treasuries to hold