The paradox of fixed-income diversification

In the AFR this weekend I argue that one of the most popular yet flawed homilies in fixed-income investing is that diversification is a universally good thing. Whereas a diversified global equities fund run by Magellan might include 25 to 30 global stocks, active bond (or credit) portfolios typically comprise hundreds, if not thousands, of individual securities. Click on that link to read the full article or AFR subs can click here. Excerpt enclosed:

And yet the apparent safety of having risk spread across so many names failed investors in the external shock wrought by the COVID-19 crisis. Notionally well-diversified credit portfolios suffered from extreme illiquidity and months of inferior returns that resulted in an enormous increase in exit costs (if any such liquidity was actually available).

This also contributed to havoc in the passive, exchange-traded fund (ETF) world where index-tracking products that were meant to be diversified across hundreds of Aussie bonds suddenly endured extreme illiquidity and price declines that were much larger than the losses recorded by the benchmark the ETF was supposed to be hugging.

This begs the question as to why there is such a striking difference in the best practice diversification approaches in active equities relative to non-government bonds.

One obvious response is that the tolerance for volatility and losses in equities is much greater than what fixed-income investors are prepared to accept given the high return targets required from assets sitting at the bottom of the corporate capital stack.

A rudimentary corollary is therefore that one needs a far larger number of assets in a credit portfolio than you would ordinarily find in active equities to reduce investors’ total portfolio risk to more acceptable levels.

There are, however, severally flawed assumptions that underpin this logic. The most important is the presumption that this diversification actually works.

In order for diversification to ameliorate the probability of loss, the underlying assets have to move in uncorrelated ways. And yet when an economy is subject to a large internal or external shock, such as the crisis in the US in 2008 or COVID-19 in 2020, which then precipitates a deep recession, that event, by definition, imposes adversity on a large number of businesses at the same time.

And so we tend to observe correlations converge to one during these crises as investors rationally price in systematic (or non-diversifiable) downside risks that will afflict many companies concurrently.

This is the little-appreciated paradox of fixed-income diversification whereby it becomes counter-intuitively risk increasing—rather than risk reducing—in times of extreme duress.

If retail investors were to open-up a well-diversified credit portfolio and examine the holdings line-by-line, some might discover that anywhere from one-quarter to one-half of all the assets have been issued by medium-sized businesses they’ve never heard of.

These are precisely the sorts of companies that struggle during recessions. In the current episode, we have seen larger household names fail around the world, including Virgin Australia, Neiman Marcus, JC Penney, J Crew, Hertz and, in Europe, Wirecard.

There has likewise been a surge in defaults on racier “high yield” (or sub-investment-grade) corporate bonds to the highest levels since the global financial crisis, which is projected to continue for some time.

This is one reason why so many “daily liquidity” credit portfolios have suffered from unexpected illiquidity, which has made it very costly for investors to recover their capital. The market knows that many corporate bonds have experienced a massive increase in their risk of default and has not, as a consequence, been willing to trade them in large (or any) volume in the absence of central bank support.

To take one case-study, the AusBond Floating-Rate Note Index carries an ultra-strong AA- average credit rating and is assumed by many to represent a high-grade portfolio of extremely liquid Australian bonds. And yet the ETF that tracks it suddenly suffered much larger intra-month losses in March than the index itself theoretically reported.

That might have something to do with the fact that the real-world investors buying and selling the ETF knew that only 68 per cent of the index includes bonds from relatively safe Australian businesses despite the fact that it is the premier benchmark for floating-rate Aussie credit.

Almost 20 per cent of the index is made-up of lesser known companies based in China, South Korea, Singapore, Japan and the Middle East. The bonds issued by these entities were especially illiquid in March and much more difficult to price. While the index did not reflect this (since valuations froze), the price action displayed by the ETF might have.

The illiquidity and risk induced by excessive diversification in fixed-income can be further exacerbated by the popular fad of “bar-bell” investing. This involves running a notionally investment-grade credit portfolio with a solid average credit rating in say the BBB or single A band that conceals two very different underlying exposures.

Half the portfolio might comprise high-grade and liquid AA and AAA rated assets while the other half encompasses riskier low BBB and sub-investment-grade BB, B and CCC rated securities (or indeed loans with no rating at all).

The problem is that in any recession the higher-yield assets can freeze and de facto undermine the portfolio-wide liquidity. When faced with redemptions, the investor is often forced to sell the highest-quality assets, leaving the portfolio with greater exposures to the remaining “illiquids”.

These complications can be amplified by the fact that it is very difficult for even institutional credit investors to truly understand the risks inherent in their individual holdings when they are overseeing hundreds if not thousands of securities with idiosyncratic default hazards at both the issuer and security level.

Consider the time and effort that Magellan’s massive analyst and portfolio manager team commits to understanding the risks in their comparatively pithy 25 to 30 stock global equities fund. Imagine if you asked them to try and run the same strategy with the same level of diligence and care covering 500 securities. The irony is that the fixed-income teams doing precisely this tend to be much smaller than their equity counterparts because of their relatively skinny fee budgets.

To be clear, there are professional investors out there who are happy to take the illiquidity and downside risks outlined above across their active credit exposures. They don’t need or expect daily liquidity and are very comfortable with the idea that a diversified portfolio of high-yield bonds are inevitably going to experience downgrades and defaults, which is the inevitable trade-off they accept for the loftier returns these strategies can generate during the good times.

But there may also be other investors who assumed that diversifying across hundreds of bonds was a source of safety and security, which enhanced rather than detracted from both their daily liquidity and probabilities of loss.

The obvious question that flows from this analysis is what is the alternative? If daily liquidity is the requirement, my personal preference is to focus on constructing portfolios of assets issued by unquestionably strong businesses that have negligible-to-no risk of default in any serious recession. These are often entities that benefit from implicit or explicit government guarantees and which have unassailable monopoly or oligopoly rights over their markets.

As one illustration, I would prefer to buy a sub-investment grade (ie, high yield) hybrid issued by Macquarie Bank than a senior unsecured bond issued by Virgin Airlines with a superficially similar BB band credit rating.

In March, the ASX hybrid market experienced a huge increase in liquidity, with north of $120 million trading on individual days while many investment-grade corporate bonds with ostensibly superior credit ratings were completely frozen. And investors trying to short Virgin’s bonds early in the month could not find a single bid.

There is a valuable role to be played by all the assets discussed above in retail and institutional portfolios. But to help navigate the tricks and traps, one should consider obtaining high quality financial advice.