Not all durations are equal

Modified duration, often simply referred as duration, is the measure of a bond’s price sensitivity to interest rate changes. Duration is expressed in years because for a fixed-rate bond, it is mainly a function of the residual maturity of the bond, which is also expressed in years. There is an inverse relationship between the change in bond prices and the change in bond yields; higher bond yields result in lower bond prices, and the longer the bond duration the more severe the bond price decline in response to a 1% increase in bond yields.

Not necessarily the case. For example, short-term interest rates tend to rise faster than longer-term rates, and they tend to exhibit higher volatility during a monetary tightening cycle than in stable market conditions. As such, a certain amount of duration from a bond maturing in two years versus the same amount of duration from another bond maturing in ten years can exhibit different volatilities. Fig 1. The relative volatility of 2-year bond yields versus 10-year bond yields is typically a function of the stage of the monetary cycle.

Fig 1: Yield volatility of US 2-year government bonds vs US 10-year government bonds

Source: Bloomberg and Eastspring Investments as of 20 Sep 2024

It certainly does. In Fig 2, we can see that the volatility of Singapore 10-year government bond yields was on par with the volatility of US 10-year government bond yields during the 18 months following the start of the Covid-19 pandemic. This was when global central banks including the US Federal Reserve and the Monetary Authority of Singapore provided synchronised monetary stimulus that kept interest rates low. However, in the pre and post Covid-19 period, US bond yields were consistently more volatile than Singapore bond yields amid less synchronized global monetary policies.

Fig 2: Yield volatility for US 10-year bonds vs Singapore 10-year bonds

Source: Bloomberg and Eastspring Investments as of 20 Sep 2024

The utility of duration is well captured in our Asian bond strategies which hold both US dollar and Asian local currency bonds. Typically, when we construct a bond portfolio, we match the risk characteristics of the benchmark index, plus/minus active positions that we think will help a strategy outperform the index. Usually, the strategy tends to have a duration profile that is comparable to the benchmark.

In Fig 3, you can see that one particular strategy had an overall duration of 4.8 years with 3.35 years coming from US dollar bonds and the rest from Asia local currency bonds. The benchmark index, which is 100% in US dollar bonds, had a duration of 4.45 years. The active duration, the difference between the portfolio duration and the benchmark duration, is 0.35 years with a duration overweight in Asia local currencies, especially in Singapore dollar, and a duration underweight in US dollar.

Fig 3: Duration breakdown by currency

Source: Eastspring Investments as at end Dec 2023

In Fig 4, we decompose the portfolio, benchmark and active durations by residual maturity; the portfolio has an overweight to intermediate and long maturities and an underweight to short-dated bonds. The total active duration is the same as in Fig 3 (0.35 years), but we use a different dimension to decompose the number i.e. residual maturity as opposed to currency.

Fig 4: Duration breakdown by residual maturity

Source: Eastspring Investments as at end Dec 2023

As per the above duration breakdown examples, looking only at the headline active duration number can lead one to conclude that the strategy is unequivocally bullish due to the duration overweight. Such a conclusion would be incorrect, and a more granular approach is required.

As we saw from historical volatility charts, one unit of duration risk from a single bond may result in more or less volatility risk depending on the time to maturity¹, the stage of the monetary cycle and the bond’s currency.

By adjusting for volatilities and correlations between tenors and currencies, one can better assess the risk of the portfolio, particularly its risk relative to the benchmark index. Ignoring those relationships may lead you to ill-informed investment decisions such as having excessive interest rate sensitivity or lacking that interest rate sensitivity.

In Fig 5, we show how the active duration can change from a positive number (+0.35) when using the conventional duration measure, which we discussed previously, to a negative number (-0.33) after adjusting for long-term volatilities and correlations between tenors and countries.

The negative adjusted duration is mainly driven by the strategy’s allocation to Singapore dollar-denominated bonds, at the expense of US dollar-denominated bonds, given the relatively lower volatility of Singapore government bond yields vs the US government bond yields as we saw earlier.

Fig. 5 : Conventional duration versus beta-adjusted duration

Source: Eastspring Investments as at end Dec 2023

Our proprietary adjusted duration measures are typically calibrated using long-term covariances to better capture expected relationships between tenors and currencies over a full economic cycle. This means that assuming those long-term relationships hold, the above-mentioned strategy is actually more defensive (-0.33) than what the conventional active duration measure (+0.35) suggests.

Over the past year, USD rate volatility has increased significantly. Market participants initially anticipated policy easing from the Fed, only to recalibrate their expectations in response to prevailing US economic conditions, which influenced the Fed’s policy trajectory. The resilience of the US labour market and persistent inflation in the services sector have particularly complicated the long-duration trade—traditionally seen as a strategy to capitalise on rate rallies when the Fed begins easing.

Duration in other currency markets is expected to perform well alongside USD duration as the Fed cuts rates or signals an upcoming cut. Investing in the duration of these markets offers the added benefit of reducing portfolio volatility when expressing a policy easing view, which can be advantageous given that the rate trajectory may not follow a linear path.

As we incorporate non-USD duration into our portfolios to reflect our constructive duration outlook, a beta-adjusted duration analysis provides a more precise way to assess volatility contributions. This approach will support active rebalancing as market conditions evolve.