On Yield Curve Rolldown

Imagine a hypothetical scenario where a yield curve is upward sloping, with the 9M tenor yielding 4% and the 1Y tenor yielding 5%. You buy a 1Y bond at 5%, wait three months until the bond becomes a 9M at 4%, and then sell it. If the yield curve stays at the current level (on average), a 100 bp yield rolldown and a Macaulay duration of ~1Y would generate around 1% of extra profit (also on average), in addition to the 5% current yield.

This strategy is called a "yield curve rolldown"

However, how realistic is it that the yield curve will remain unchanged on average, so the odds play in our favor? This post shows that it is harder to execute this strategy systematically than one would naively expect.

Risk-neutral pricing theory dictates that expected future spot rates are equal to today's forward rates. A corollary of this is that an upward-sloping yield curve should experience bear flattening, and an inverted curve should experience bull steepening, where no one on average is expected to make money from the rolldown.

Backtesting historical data indeed shows the tendency of the curve to move in a way that offset gains from the rolldown.

1️⃣ STORY OF 2022

Figure 1 shows an example of an unprofitable rolldown because FED raised rates to fight inflation, and the resulting bear flattener was more dramatic than the curve forecasted by risk-neutral pricing theory.

2️⃣ STORY OF 2025

Figure 2 shows the inverted yield curve. Short-term rates remained high due to FED inaction, while mid-term rates were falling because market anticipated rate cuts. Entering a negative rolldown (aka roll-up) strategy by shorting bonds again turned unprofitable, because the actual curve decline was more dramatic than prescribed by risk-neutral pricing.

3️⃣ THIRD STORY - WHEN STARS ALIGN

Finally, Figure 3 shows a few rare occasions in 2019 when it was profitable to execute a rolldown. In order to do so, at least one of the two conditions needs to be met:

1) A positively sloping curve
2) A curve that does not move upward

Meeting both of these conditions simultaneously is rare-partly because the strategy is self-defeating and partly because of underlying macro factors.

The rolldown strategy works best when nothing is happening and markets don't move. This strategy favors frequent but small gains, compensated by infrequent but larger losses. Its fat-tail profile makes it popular among hedge funds-just like its FX and IR carry siblings.

The rolldown can be seen as a premium that one earns for taking risk. When the market price of risk is high, a hefty rolldown is awarded to speculators per unit of risk. In the absence of MPR, the final yield curve (🔴red line) would, on average, converge into the one rolled under the risk-neutral (Q) measure (🔵blue line).

⚠️DISCLAIMER: Not investment advice or forward-looking trade idea. Opinions are my own and not those of my employer. Read my profile for full disclaimer.

On US repo market

Recently we have seen news in WSJ and FT about the "turbulences" in US repo market after years of relative calm, manifested by widening of the SOFR / EFFR spread (fig. 1) and resulting FED's temporary OMO intervention.

 
I will try to balance the "breaking news" by making this post more educational, explaining why this is happening and why I think the reasons are more prosaic than the fast news suggest.

First, let's define the relevant rates in question:

🔵EFFR (Effective Fed Funds Rate) is the rate at which banks lend money overnight to each other without any collateral.

🔴SOFR (Secured Overnight Financing Rate) is used by many more institutions (not just banks). The loans are made against collateral (typically Treasury bonds). This distinction will become important later.

Because EFFR is unsecured, it is expected to be slightly above SOFR due to a small credit risk premium. This is a textbook knowledge and has indeed been empirically confirmed (2021-2024).

However, in the second half of 2025, SOFR started being jumpy and rose above the EFFR. This stirred the Fed's reaction (fig. 2) and the aforementioned news.

💡Why is SOFR jumpy now?

During quantitative easing, the FED supplied market with plenty of liquidity by buying bonds from the Treasury Department and other participants. Once there is a lender (e.g. the FED), finding borrowers (Treasury and others) is rarely a problem. The result of this dynamics is a smooth, uneventful EFFR-SOFR spread.

However, everyone who has ever borrowed money knows that it is much harder to quickly find another lender to step in, when the original one demands to be paid back. After years of quantitative easing, FED started a period of quantitative tightening. Cash is being drained from the system and flooded instead with treasury bonds.

In other words, there is a scarcity of cash but plenty of collateral. More Treasuries on the market also mean more demand for repo financing (e.g. hedge funds buying bonds and selling bond futures, therefore funding bond purchases by borrowing at SOFR).

💡Why is EFFR not affected?

Remember what we said previously - the FED Funds rate is unsecured and therefore is not affected by the aforementioned repo market dynamics.

💡What is the significance of all of this?

The important thing is the quantitative easing or tightening itself, because it profoundly affects money supply, inflation, employment and long-term economic prospects. The daily noise in the repo market is just a natural second-order effect of reducing FED balance sheet. In grand scale of things, those wobbles don't mean much if managed properly.

The central bank is aiming to smooth this by applying temporary open-market ops, for the first time since mid-2020. This fact alone may indeed sound intriguing. However, in the 2010s, such operations were occuring more frequently (fig. 4).