What is under the hood of US 10-year bond yields?

Bond yields contain a wealth of forward-looking assessments around the real economy. Very often, it is only when disaggregating overall yields into sub-component parts that one is able to gain “line of sight” with regards to these embedded assessments of economic prospects.

At the core of theoretical descriptions of the term structure of interest rates is the Expectations Hypothesis. According to this hypothesis, the expected return from holding a long-dated bond of a given maturity should be the same as the expected return from rolling over a series of shorter-dated bonds for the same total maturity, or, more concretely, there should be no additional expected compensation (for example) priced in for the investor in a 10-year bond, than for an investor who embarks on a strategy to buy a series of 1-year bonds, rolling them over ten times consecutively.

Whilst the Expectations Hypothesis provides an intuitively appealing interpretation of the term structure of interest rates, it ignores duration risk. Very simply, duration is a measure of price-yield sensitivity, and longer-dated bonds have greater duration (i.e. they have prices which are more sensitive to changes in prevailing yields) than shorter-dated bonds. Traditionally viewed within the framework of the expectations hypothesis, the term premium encapsulates deviations from the hypothesis as it accounts for the uncertainty surrounding future short-term interest rates and economic factors that might impact long-term bond returns. Therefore, term premia are pivotal for central banks and financial institutions, as they provide insights into market expectations and the risk landscape within long-dated securities.

The modern term structure literature provides econometrically robust ways of decomposing bond yields into and expectations component and a term premium. Very often, the starting point is the assumption that bonds are priced in a way that precludes arbitrage opportunities across all maturities. In other words, the pricing is assumed to make it impossible to form a portfolio consisting of bonds with different maturities that generates a riskless profit.

Reza Ismail, Head of Bonds at Prescient Investment Management

Before we delve into the changes in the underlying components of US bond yields, recall for the most part of last year, there was a widely held view among market participants that US real growth would enter recessionary territory in 2024; a product of the lagged effect of previous tightening monetary conditions in an environment of broader global tightening. The key assumption embedded in this view is that although financial conditions adjust immediately to reflect expected and actual changes in monetary policy, the full adjustment in output, employment, and inflation occurs with a considerable lag.

In contradistinction to this, there was a second camp (ourselves at Prescient included) who believed instead that recent empirical experience as well as econometric modelling reveals that the peak drag on economic growth from a tightening in financial conditions occurs without considerable lags - after a mere two or three quarters, on average. Given that the most significant tightening in financial conditions occurred in the middle of 2022 - when the Federal Reserve pivoted sharply toward more aggressive rate hikes – the second camp contended that the maximum drag on growth had already occurred quite early on (perhaps in early 2023 at the latest), meaning simplistically that major headwinds were very likely to have already been “in the rear-view” mirror at that point.

What has occurred factually over the course of the year 2024 had in fact been a situation very close to the latter of the camps described above - i.e., that US aggregate demand and consumption has in fact remained robust and resilient, and if anything continues to surprise to the upside with regards to its contributions to the real economy.

Turning now to a decomposition of the US 10-year bond yield, from mid-September 2024 to late January 2025, we can make the following assertions:

Sources: Prescient Investment Management, Bloomberg, January 2025

• Overall, the 10-year US bond yield is 96 basis points (nearly 1.0%) weaker since September 2024 (navy-blue bar shown above is the change in the US 10-year bond yield, not the level of the yield itself)


• This increase is chiefly driven by a 0.73% or 73 basis point (bps) increase in the nominal term premium (darker green bar) – this means that the market has re-assessed the compensation it requires for taking duration risk.


• Decomposing the nominal term premium increase of 73 bps, we find that this in turn is driven largely by a 67 bps increase in the real term premium (purple bar), or the term premium embedded in the inflation-linked (real yield) bond market.


• Proximate reasons for this material increase in real term premia are likely due to the US labour market proving much more resilient than initially estimated. Fiscal transfers during COVID bolstered household incomes in the US and created savings stockpiles which were able to guide and buffer consumers through what would otherwise have been a very brutal and punitive hiking cycle.


• Prospects for economic growth have therefore been revised to the upside to such an extent that real yields have to adjust upwards to reflect an increase opportunity cost for locking capital away instead of employing it in the real economy.


• What is also clear is that there has been an upward revision of 22 bps in average expected short-rates (lighter green bar). This revision has occurred because market participants have priced in expectations for average policy rates (in this case the Federal Funds rate) which espouses a shallower cutting cycle than what was the case in second and third quarters of 2024.


• Since the increase in real yields in the US have largely repriced due to a revision in growth prospects and resilient consumption, this would ordinarily be met by US and foreign capital “seeking out” risk-assets in pro-cyclical destinations like emerging markets.


• However, since the third quarter of 2024 this has not happened, because the stronger US labour market, together with progress with disinflationary traction has in some sense “cross-subsidised” the FOMC’s latitude to keep policy rates relatively elevated, and in so doing, led to yet another upward re-pricing of the front-end of the US term structure – making it a very attractive investment destination.


• Therefore, the stronger the US economy grinds on with continued disinflationary progress, the longer the risk-free front-end of the US yield curve can afford to remain elevated, which in turn creates a high hurdle to scale before global capital will look for yield elsewhere. (The opportunity cost of moving away from the US has again proved to be a steep hurdle to scale.)


• The inflation risk premium (light purple bar) has increased by 6 bps, because now, with a very likely disinflationary trend gaining traction, the risk of inflation surprises (as assessed by the bond market) has reduced, with investors accepting lower compensation for bearing this risk.

Given the decomposition above, it is clear then US bond yields have become weaker largely on account of re-pricing of real yields, which, in turn, occurs when the market requires additional compensation for parting with capital and delaying consumption in an environment where the real economy is able to offer viable alternatives.

Importantly, the observed weakness in US bond yields has not been driven by a material re-pricing of breakeven inflation. This means that the traded market’s assessment of future average inflation has not been at the heart of the recent sell-off at all.  This should be a positive signal for the Federal Reserve, who have previously expressed concerns around the resilience of the labour market possibly leading to an upward revision of consumption-driven inflation-expectations once more. DM