It’s official. Or as official as it can be considering we’re talking about something that’s impossible to measure with anything like precision. The term premium is negative. Again.
I won’t pretend anyone cares. The term premium, like r-star, isn’t even exciting when it’s undergoing a sea change. And anyone you might impress by talking at length about it isn’t someone you’d want to marry, with apologies to all the single economists out there searching for a soul mate.
And yet, I’d be remiss not to mention that the ACM model is now back in negative territory.
Recall that the term premium’s rapid ascent out of negative territory (beginning in late July) was responsible for the back-up in long-end US yields which, in turn, tightened financial conditions and prompted the Fed to consider forgoing the final rate hike tipped by the September dot plot.
In theory, the term premium shouldn’t be negative (although nothing says it must be positive). You generally want to be compensated for the risk associated with locking up your money over long periods (versus simply rolling shorter-dated debt). But it was negative, and habitually so for quite a while. There were a number of explanations for that phenomenon. I won’t trouble readers with the notoriously mind-numbing discussion, but suffice to say that beginning in early August, the combination of supply-demand concerns (i.e., increased coupon supply against a shifting buyer base for Treasurys in favor of price-sensitive investors) and questions around the read-through of endemic political dysfunction for America’s fiscal outlook, drove the term premium rapidly higher.
That repricing was the talk of the proverbial town. When Treasury’s refunding announcement tipped smaller-than-expected coupon increases this month, it was widely viewed as a polite nod to the term premium. That nod, in conjunction with a run of soft macro data and assumptions about the end of Fed hikes, drove US yields sharply lower this month, erasing declines which threatened to make 2023 the third straight year of losses for US debt, an unprecedented event.
From the October highs to the November 17 low on the ACM model, the 10-year term premium was down 50bps. 10-year yields are around 60bps lower over the same period.
Why does this matter? Two reasons, at least. First, it negates a meaningful portion of the tightening impulse cited by the Fed for holding off on another rate hike. Whether that puts one final increase back on the table will largely be determined by the next round of top-tier US macro data due during the first full week of December. Second, it suggests, to some anyway, that recession risk has increased meaningfully.
For BMO’s Ian Lyngen and Ben Jeffery, the receding term premium “add[s] to the case for deemphasizing the auction size increases as the driver of outright Treasury yields.” “Positive term premium had a brief and eventful stay to be sure, and we’re reminded that it will soon return in a more sustainable way once the Fed begins cutting rates and the yield curve commences its traditional bull steepening,” they wrote.
One thing the decline back into negative territory certainly doesn’t suggest is that the political dysfunction which contributed to the Fitch downgrade and the Moody’s warning has in any way abated. Things are every bit as fraught inside the Beltway today as they were four months ago, I can assure you of that.




Admittedly, I’m way out of my depth when it comes to term premiums and r-star, but couldn’t this be construed as evidence that rates in developed economies are still in a long-term downtrend that ends in either zero or near zero interest rates except when we get inflationary shocks like the pandemic? With demographics flattening out and the Fed put, I don’t see why rates or r-star or risk premiums wouldn’t trend down over time and end near zero with some brief exceptions.
Add in what I think will be a significant deflationary impulse from AI and it won’t surprise me in the slightest if we are soon back in the familiar world of the 2010s.
Good comment
Judging by various term premium estimate series on FRED, term premium and yield trended down from 1990 until recently, very roughly but not always moving together. Periods of negative term premium seems associated with periods of ultra-low yields (<1% for five year, <2% for ten year). If yields are going to be stable-ish from here, or at least not headed to ultra-low levels, then naive chart staring suggests term premium should tend to be positive. Low positive, but positive. That said, the charts also show these relationships do get stretched for longer than wrong-footed traders can remain solvent.
See charts
Five year
https://fred.stlouisfed.org/graph/?g=1bGNW
Ten year
https://fred.stlouisfed.org/graph/?g=1bGNZ
Thinking about the reasons for a term premium, the list of possibilities includes uncertainty (of inflation, rates, economic); demand for bonds from foreign and domestic buyers; bond supply – and I guess we need to include borrower creditworthiness. I don’t know how those sum up.
Example list
https://www.federalreserve.gov/pubs/feds/2005/200533/200533pap.pdf