Here’s Who Will Buy All That Treasury Supply

By Heisenberg

Now if we only knew the price…

10Y yields are back near their lowest levels since last month’s CPI beat, having given back Tuesday’s Jerome Powell-inspired spike that derailed equities.

They’ll be no shortage of narrative fodder on Thursday with Powell’s second act (this time in testimony to the Senate) and PCE on the docket, but panning out, the question still lingers: how high will yields go? And of course the follow-up that no one can answer: what is the magic number on 10s beyond which equities can no longer pretend not to care?

Here are the monthly yield changes for UST benchmarks from February:

  • 2Y +10.9bp
  • 5Y +12.6bp
  • 10Y +15.6bp
  • 30Y +18.9bp

As a reminder, the two-month rise in real yields (i.e. January plus February) was the largest since the election:


Just to be clear, folks are getting pretty deep into the weeds here when it comes to forecasting yield levels given a set of assumptions. And by “deep in the weeds” I just mean that people are bending over backwards to find a reliable framework for forecasting. It’s not so much that the methodologies being employed are particularly innovative (this isn’t exactly rocket science), it’s just that the amount of time being spent on it is probably some semblance of absurd considering the inherent futility of trying to accurately forecast this. Here’s BofAML’s latest:

An alternative approach to cuffing 10y rates employs a simple averaging of forward rates. This method first breaks the yield curve down into a path of three-month forwards, and then applies a shock to change the slope and endpoint of Fed hikes over the next few years, and then re-computes term rates as an average of the shocked path of forwards. This approach is not concerned with historical relationships between rates and macro variables, but instead gives a simplified mapping of a Fed path onto a 10y rate level. When doing this (Table 2) we first find that the curve is pricing about 3.2 hikes over the next 12 months (using the Treasury curve not the OIS curve), 1.2 hikes in the following 12 months, and 1.1 hikes in the following 12 month period. This is with 10y rates at 2.92% as of midday Tuesday 27 Feb. If we change the path by leaving the 3.2 alone but bumping up the following year to 3 hikes and leaving the next year unchanged at 1.1, which is close to our house baseline forecast, we find a new 10y rate of 3.24%. If the forward curve remains in this configuration over time and we re-compute 10y rates a year from now, we find a value of 3.39%. If instead we shift toward a more hawkish scenario of 4 hikes followed by 4 hikes, and then no further hikes, this approach would imply a spot 10y rate of 3.4% and a year-end 10y rate of 3.6%.


The reason I wanted to highlight that is obvious: Powell’s testimony suggests that an epochal shift may be afoot in terms of the “gradualistic” approach to hikes, although one could could just as easily argue that ol’ Jay just isn’t very good at reading the proverbial crowd yet and that accounts for his errant comment about “his own view” on Tuesday.

But playing out in the background here is a dramatically different supply picture thanks to Trump’s tax cuts and fiscal stimulus push. The supply deluge raises the following obvious question: who’s going to buy it all?

Earlier this week, Morgan Stanley took a stab at answering that. While the bank notes that “will it be absorbed?” is an entirely different question from “at what price?” (the former being a foregone conclusion and the latter not so much), Morgan’s take is (I guess) meant to allay fears about the willingness of price sensitive investors to step up when they’re needed.

While “overseas investors have been adding to Treasury holdings at a much slower pace over the past 5 years than in the 5 years immediately following the crisis, this trend should reverse as Treasury supply picks up and corporate supply wanes,” the bank writes, addressing demand from the rest of the world, before adding that “the US is running a current account deficit with the rest of the world [and] to the extent that foreign central banks have to intervene as dollars procured from trade are converted into local currencies, overseas investment into Treasuries should continue to rise.”


They go on to note that “households own $1.345 trillion of Treasuries, as of last September, more than double their holdings ahead of the financial crisis [and] even if you assume households hold all the Treasuries in mutual funds ($996 billion), the total Treasury holdings of households pales in comparison with their equity holdings which total $17.282 trillion – an all-time high.” From Morgan’s perspective, households “will continue to accumulate Treasuries as the population ages – either directly or through mutual fund shares.”

Speaking of mutual funds, the bank predicts that their Treasury holdings  “should increase dramatically as float-adjusted indexes begin to include both Treasuries no longer purchased by the Fed as well as the deficit-related pick up in Treasury issuance.”


And then here’s what they’ve got to say for pension funds:

Even more so than mutual funds, private pension funds have increased holdings of corporate bonds at the expense of Treasuries and agency MBS since the financial crisis. Many private pension funds, however, operate in a similar fashion to traditional asset managers in that they have a benchmark that they try to outperform or try to match. We expect private pension funds to increase their holdings of Treasuries as Treasuries come to occupy a larger slice of the core indexes. Pension funds will no longer have to compete with the Fed for long duration Treasury supply.

Still – and this is consistent with Morgan’s recent take on the effort to project where 10Y yields will be 10 months from now – the bank notes that simply knowing they’ll be buyers and knowing who those buyers are doesn’t do much to help answer the price question. To wit:

Even if we knew which investors would buy Treasuries this year and which would sell, on a net basis, and we knew the amount each would buy and sell, we would still have to know how all of the other factors would develop to know at which prices those investors would transact. In short, it’s usually a fool’s errand to guess at the demand side of the equation.

Right. So if you’re wondering whether yields are going to end up spiking to whatever your own “magic” number is beyond which equity investors start taking headers off bridges, you’re shit out of luck.

Does that help?

Published by

Gary &