The duration deliberation - to extend or not to extend?

In normal market conditions, fixed-income investors might choose to increase the duration of their portfolios when they expect an upcoming economic contraction. Despite the persistent buoyancy of the US economy, some cracks are now beginning to show; far fewer rate cuts are now expected for 2024, unemployment has started to climb, and credit card and mortgage delinquencies have increased. Investors seeing these signals might feel inclined to pile in to the long end of the yield curve, anticipating that rates will be cut due to the effects of previous hiking cycles feeding through to economies. However, with yield curves still inverted, in our view, the normal playbook of extending duration at this stage in the cycle is different this time around. Yields on shorter dated bonds are still currently higher than those with longer maturities, meaning there is often no additional yield on offer for taking additional capital risk in much longer dated credit.

At the end of 2023, the US Treasury, Gilt and Bund yield curves were all steeply inverted, with over 150bps separating the front of the Treasury curve and the five-year point. This meant investors were giving up significant yield opportunities by buying longer dated bonds As a result, our portfolios were firmly focused on the short end of the yield curve and were managed with a notably shorter duration when compared to our longer term stance. 

Fast forward to today, yield curves have seen significant steepening and whilst the short end has remained anchored, there has been notable yield rises amongst longer dated bonds. The steepening this year has brought devastating losses for sovereign bonds, with major capital erosion in the first few months of 2024. By the end of April 2024, the FT Gilt Index had fallen by 4.59%; the FT >15-year Gilt Index was down by 9.34%; and the >20-year Treasuries by 10.71%.

Though the curves for Treasuries, Gilts and Bunds are still currently inverted, the gap in yields between shorter and longer-dated bonds is now far less pronounced. This, combined with the yield rises seen, has tempted us to start modestly increasing duration in our short-dated strategies for the first time in more than two years, moving back towards our longer-term average duration stance. However, there are many things to consider when it comes to expressing a duration view, including pinpointing where on the yield curve is best to take duration, as well as sector and geographical considerations.

To help shape our views, we predict that the Fed will cut rates once per quarter, starting at the end of the third quarter of 2024. Our current expectation is that they will deliver six 25bps cuts in total between now and the end of 2025, which would bring the effective federal funds rate down from 5.5% currently to 4%. If these cuts occur, a more normalised, upward-sloping yield curve for Treasuries looks likely. We believe we’ll go back to more normalised yield curve shape because quantitative easing (QE) is unlikely to return, and the increased supply of Treasuries will keep pressure on yields and the steepness of the yield curve.

We may see something like 100bps of steepness between the very front end and the 10-year point, which is roughly the average level of steepness over the last five decades. Such a scenario for Treasuries would result in a pivot at the five-year point of the curve. Up to that mark, investors might expect capital gains on top of attractive yields, giving attractive total returns. Once you go beyond the five years of maturity, investors might expect a modest rise in yields along with some small capital losses, but still, an overall positive return. As a result, shorter dated bonds up to five-years maturity will provide the best overall return opportunities for investors, which is why we now think we’re at a much more attractive entry point to add three, four and five year bonds to our portfolios.

To date, we have primarily increased our overall portfolio duration through extending our government bond exposure, albeit expressing this view on different curves depending on the underlying macro data. For example, earlier in 2024, we switched our entire Gilt position into 5-year Bunds, given German economic weakness and the expectation that the ECB would be the first to cut rates. However, there is a limit to how much we could increase the overall duration of the portfolio from rates alone.

“We recognise that curves have moved a long way and we have therefore become more tolerant of interest rate duration, but in our view  it’s not the time to extend credit spread duration given how tight spreads are.”

When it comes to extending credit spread duration, we are proceeding with caution. Credit spreads in certain sectors, such as non-financials, look tight. Whilst we have done some name specific extension trades within credit, mainly financials, the overall credit spread duration of our portfolios has remained consistent and still focused on shorter dated bonds. To give us more confidence to add duration via credit, we would need to see credit spreads widen sufficiently from here. Should economic weakness intensify later this year, leading to spread widening, it could present an ideal opportunity to add spread duration to the portfolios. However, for now, given current spread levels, we think it’s an expensive trade.

The inversion of yield curves complicate the conventional approach of increasing portfolio duration in anticipation of an economic slowdown. The average level of yield in fixed income is very attractive, and rate cuts are coming, albeit there are still some question marks around the ultimate timing. As a result, we retain our yield curve steepening bias as we believe shorter dated bonds still offer the best opportunities – they have yield plus capital gain potential. Longer dated bonds still have attractive yields, but could experience yield minus small capital losses as curves normalise. We are concerned that spreads are looking tight in some credit sectors, therefore we are less bullish on credit spread duration but have been adding duration selectively in the rates space.