Against duration matching

This is the fourth post in a series on duration matching. The first post covered an introduction to duration, what duration matching is, and why we might want to use it.

The second post looked at why we shouldn’t be afraid of using long-duration bonds, especially if we also have a long-term investment horizon.  

The third post went through how to implement duration matching in practice.

And now we come to looking at a few of the reasons we might not want to adopt duration matching for our portfolio.

But before we start digging in to some of the criticisms of the approach, it’s worth first having a look at the conventional approach to bond allocations. And for most passive investors, this means sticking their bond allocation entirely into an intermediate-term bond fund (typically government bonds) and be done with it.

Why intermediate-term bonds?

Imagine you held a portfolio consisting of 50% stocks and 50% cash. In this scenario, the 50% in cash is acting as a risk reducer, not a diversifier. It’s never going to rise by a huge amount when stocks fall, but it’ll never fall during a crash, either. It’s a risk-reducer.

This is the difference between a risk-reducer and a diversifier.

Risk-reducers merely mute the impact of stocks, diversifiers aim to move in the opposite direction.

All assets typically labelled as ‘diversifiers’ exist on a spectrum. At one end, you have cash. It’s the ultimate risk-reducer because it’ll never move very much. It just reduces the volatility of the portfolio.

Very close to cash on this spectrum are short-dated bonds. They’re slightly riskier than cash, but also have the potential to go up slightly more when stocks fall, as investors pile into safe-haven assets.

Continuing along the spectrum, we come to intermediate-term bonds. Again, riskier, but pack more of a punch when stocks fall.

At the other end of the spectrum from cash sits long-term bonds. They’re much more volatile than cash, but have much better diversification potential.

So why is the typical approach to only use intermediate-term bonds?

Generally, because they’re a happy medium. Compared to cash, they’ll have larger drawdowns, but will also provide better protection when equity markets fall, and higher returns. Compared to long-duration bonds, they won’t provide the same level of oomf in a crash, or have the same level of returns, but they won’t have the drawdowns either.

It’s a classic passive investing approach.

It’s essentially saying, “I don’t know what’s going to happen. And although I know my approach will never be the absolute best decision in hindsight, I can guarantee that it won’t be the worst. And not being the worst is far, far more important than being the best.”

So it has the advantage of being a “good enough” diversifier.

And because it’s the default (and for good reason), this will be our yardstick for measuring the attractiveness of duration matching. Are the net benefits of duration matching enough to warrant adopting it over simply buying and holding intermediate-term bonds?

Intermediate-term bonds vs duration-matching

1. ‘Set and forget’

Using only intermediate-term bonds is a ‘set-and-forget’ strategy.

There’s no need to worry about estimating future spending, working out weighted average durations, or adjusting bond fund weightings to match investment horizon.

I’m a big fan of keeping things simple (hence the name of this blog), and their simplicity is a big advantage for just using intermediate-term bonds over duration-matching in my book.

2. We don’t have a crystal ball

Because we can never perfectly forecast our future spending, we can never be exactly sure of our investment horizon. Life isn’t a spreadsheet.

Given duration matching relies on us being able to gauge our investment horizon so we can match our bond fund’s duration, we’ll never be able to perfectly eliminate interest rate risk. Duration matching can definitely reduce its impact, but because we can’t know our future spending needs with any sort of precision, we can never completely eliminate it as the theory suggests.

3. Long-term bonds are volatile

There’s an added behavioural risk of owning long-term bonds as part of a duration-matching strategy.

Because long-term bonds have a higher duration, they’re more sensitive to interest rates. This results in longer-term bonds having larger drawdowns than short or intermediate-term bond funds. The graph below compares the drawdown profiles for three UK government bond funds of varying durations:

Source: Bloomberg. Long duration bonds are the Bloomberg UK Govt 15+ Years Float adjusted TR Index (the benchmark for Vanguard UK Long Duration Gilt Index Fund). Medium duration bonds are the FTSE Actuaries UK Gilts TR Index (the benchmark for IGLT). Short duration bonds are the FTSE Actuaries UK Gilts up to 5 Years Index (the benchmark for GIL5).

That long-duration fund sure does have some spicy drawdowns.

If we were all investing robots and were able to only focus on the risk/return characteristics of our portfolio as a whole, then we’d all be much better off. But in reality we’re all likely to pay more attention to its constituent parts. Having a line-by-line breakdown of all your positions is going to cause you to focus on whatever has the biggest red minus number next to it, and not what you’re overall portfolio is doing.

So by having an individual position which is liable to suffer from larger drawdowns, it raises the risk of bailing out of the long-term bond fund at the worst possible time – when it’s sitting at a loss.  The intermediate-term bond fund still has much larger drawdowns than the 2-year fund, but with roughly half the maximum drawdown of the longer-duration bond fund it will likely make it a much easier fund to hold.

4. The negative stock/bond correlation isn’t guaranteed

One common argument levelled against duration matching is that the diversification benefit of long-term bonds comes from the negative stock-bond correlation. Long-term bonds are good diversifiers because they’re able to rise more than intermediate-term bonds when stocks fall.

The reasoning goes that although the negative stock/bond correlation has held strong for the last 20 years or so, a negative correlation between stocks and bonds is actually unusual. In fact, for most of history, the stock/bond correlation has been positive. So the recent negative stock/bond correlation isn’t guaranteed.

And if the stock/bond correlation turns positive, and you’re sitting in long-dated bonds, then you could be in for a nasty shock when markets fall and your long-duration bonds fall alongside. They’d fall much further than if you were holding intermediate-term bonds.

I covered the topic of the possibility of a rising stock/bond correlation in this post. The subsequent post, however, ‘Have bonds ever failed?’ pointed out that bonds have done a great job in hedging equity market falls, even during low interest-rate environments and importantly, also when the stock/bond correlation was positive.

We also saw that in the table from the second post on the benefits of long-term bonds, longer-dated bonds have done a great job at mitigating drawdowns going back all the way to 1929. And this 90-year period covers just about as many different correlation environments as you could ask for:

Sources: US stocks: Bloomberg, US bonds: Shiller data, cash: FRED

So I think the first three criticisms of duration matching are valid, but I’m not too worried about the ‘rising correlations’ argument.

Summary

• Intermediate-term bonds are a happy medium between a risk-reducer and a diversifier.

• Compared to cash, they’ll have larger drawdowns, but will also provide better protection when equity markets fall and higher overall returns. Compared to long-duration bonds, they won’t provide the same level of oomf in a crash or the same level of returns, but they won’t have the drawdowns either.

• Compared to duration matching, an intermediate-bond approach has the advantage of being ‘set-and-forget’, doesn’t require any forecasting of future expenses, and might be easier to stick with.

If you’ve followed the series to this point, you’ve seen what duration-matching is, why it’s useful, how to implement it, and now what its drawbacks are.

The final post in the series will (hopefully) tie all these posts together. It’ll answer the question of whether investors should be adopting a duration-matching strategy, and who it’s appropriate for.