Cashflow-matching credit – room for improvement?

The idea behind cashflow matching in defined benefit (DB) pension schemes is straightforward: line up asset cashflows from gilts and corporate bonds* with liability cashflows (and use liability-driven investing, LDI, to take care of gaps and any inflation sensitivity).

But can it be improved? In this blog, I outline the case that short, or at least shorter-dated, credit could act as a powerful diversifier to cashflow-matching credit, helping DB schemes to meet pensions more efficiently.

Matching gilts

The great feature of cashflow matching with gilts is that once the match is made, the path of yields in the future ‘doesn’t matter’ – the destination remains the same. The picture below illustrates this for a 10-year zero-coupon gilt. Regardless of what happens to interest rates between now and maturity, you can rely on receiving the same end cashflow. Align it with a liability cashflow also due in 10 years and – congratulations – you’re matched!**

Matching credit

Not all schemes can afford to invest purely in gilts (and swaps), so they may also look to corporate bonds. To a decent extent, cashflow-matching credit investors can also ‘look through’ moves in credit spreads, like the gilt investor who can look past moves in gilt yields. This powerful property forms the basis of cashflow-driven investment (CDI) strategies. Unlike for gilt yields, however, the path of credit spreads does matter for two reasons:

• Spread widenings tend to coincide with downgrades. Investors are unlikely to tolerate a sustained worsening of the quality of their portfolio, so downgrades may result in rebalancing trades into higher-rated and lower-yielding bonds. Such rebalances reduce the size of the contractual cashflows ultimately payable. The longer until the cashflow is payable, the larger the reduction will be (so this is important for long-dated bonds but less so for short-dated ones).
• Spread widenings also tend to coincide with higher default rates, even if the average credit rating is preserved.

There is a negative relationship between the path of spreads and outcomes: if spreads are elevated over the horizon, you should anticipate worse outcomes.

Rolling credit

There’s another possible strategy for meeting a liability cashflow in the distant future: roll short (or at least shorter)-dated credit. This needs to be combined with some LDI (interest-rate swaps) to hedge interest rate risk. Assuming that’s taken care of, what do the risks look like compared with cashflow matching? Answer: a different kettle of fish …

• Although default risk on shorter-dated bonds is the same (for the same rating), and the chance of a downgrade is also the same, downgrades don’t matter nearly as much should they happen. If a bond is downgraded the investor can just wait a relatively short period until the bond has matured, for example. Even if they rebalanced into higher-quality bonds, the loss would be small.
• Unlike cashflow matching, there is reinvestment risk. The proceeds of the investments need to be reinvested multiple times before they can finally be used to pay pensions. The concern is that if spreads are low over the investment horizon, the ultimate return of the strategy will be poorer than expected.

You’ll note that the relationship here is positive: if spreads are elevated over the horizon, you should expect better outcomes.

A powerful diversifier?

You can see where this is going. The long-term returns from cashflow matching and rolling credit both depend on the path of spreads – but in opposite directions, potentially leading to a negative correlation on long-term returns. Low or negative correlations can be extremely potent. The upshot is that rolling short-dated credit could complement cashflow matching, helping schemes to meet cashflows more efficiently.

How much to diversify into shorter-dated credit requires consideration of other factors, such as:

• Expected returns, which I discussed in an earlier blog
• The fact that spreads tend to be range bound, which limits reinvestment risk (with multiple reinvestment times/points, you’d have to be unlucky to buy only at low spreads!)
• A high but imperfect correlation between moves in short and long-dated spreads

Putting all this together in a model, our analysis suggests that investing as much as a third of the credit strategy in short-dated instruments could be optimal. All investment models need to be taken with a generous pinch of salt, of course. But it is interesting that you can get such high allocations under plausible assumptions.

Other diversifiers?

The golden source of the strong diversification between long-run returns on rolling short-dated credit and cashflow matching is that a ‘high-spread path’ over the investment horizon is generally bad for the latter but good for the former. For simplicity I focused on investment-grade UK credit in my model, but the logic also applies to international credit including (hard currency) emerging-market debt. It can also work, albeit to a lesser extent, with:

• High-yield bonds for the rolling strategy. For low-duration instruments, the impact of downgrades (with an associated increase in spreads) is smaller, so credit quality may be less of a concern. We know of some schemes balancing high-rated long-dated credit with lower-rated shorter-dated credit.
• Medium-dated credit. So long as the duration is shorter than for CDI there will still be a diversification benefit, although it will be weaker.

I hope you found this interesting! If you’d like to find out more, please read our in-depth paper here or get in touch.

* Equity dividends can be useful too. However, there are drawbacks. For example, eventually the equity must be sold, leaving huge price risk at the point of sale.

** A similar principle exists for inflation-linked gilts and real cashflows.

Assumptions, opinions and estimates are provided for illustrative purposes only.