On May 4, the FOMC hiked 50 basis points and restated its dual mandate “to achieve maximum employment and inflation at the rate of 2 percent over the longer run.” During Chairman Powell’s subsequent press conference, he responded to a question that “[a] 75 basis point increase is not something the committee is actively considering.” Stocks and bonds began to quickly rally and closed higher on the day. However, the capital market was hit hard the following day as the market more fully digested the SOMA balance sheet information.

The market may be starting to appreciate the need to be marginal buyers of duration as the Fed will take a much less active role. The hedging needs will be massive.

The Implementation Note issued by the FOMC set reinvestment caps of principal repayments $47.5 billion ($30 billion for treasuries and $17.5 billion for MBS and agency paper) during June through August. This means that the Fed will allow this amount to mature without reinventing the proceeds back into the market to remove duration. The accompanying statement for Reducing SOMA Holdings increases reinvestment caps to $95 billion ($60 billion for treasuries and $35 billion for MBS and agency paper) in September. Those latter levels are “expected to remain in effect until otherwise directed by the FOMC.”

Source: New York Fed
 

Monthly projections of principal payments from the NY Fed,² as well as historical activity of capped purchases during the 2017 to 2019 period, show an interesting perspective. Upcoming maturities are double to six times larger than those from this prior period of capped reinvestment purchases.

During the first year of caps, from October 2017 to October 2018, the 10-year Treasury rate went up almost 100 basis points.

With the Fed being a much less active buyer of duration, the question becomes what will be the profile of the new duration hitting the market.  That will be driven by the Treasury debt issuance and mortgage back security (MBS) production. This depends on the governments financing needs and the mortgage market activity.

Cash has poured into the Treasury coffers, and it must reduce its issuance or risk a more activist role in monetary policy as holding large cash positions technically removes dollars from the financial system. That is the responsibility of the Fed and one the Treasury does not want to affect.

During April, surprise increases in tax revenue receipts, coupled with continued large issuance of coupons and bills, swelled the Treasury General Account (TGA) to nearly $1 trillion. Additionally, the Congressional Budget Office (CBO) Baseline Budget Projections forecast smaller deficits in the future to under 3%, down from nearly 15% in prior years. 

Source: FRED Economic Data
 

With so much cash and smaller deficits, something must go. Historically, bill issuance varied, but today that looks different. Bills as a percentage of overall debt issuance now stand at about 16%, down from 25% two years ago and closer to pre-pandemic levels.

The Treasury seems committed to the market’s need for bills. The announcement of a new 4-month bill benchmark issuance affirms this position. Also, during times of interest rate hikes, the demand for bills increases as investors tend to shun duration. In addition, cash has been hiding out in the Fed’s reverse repo facility to the tune of over $2 trillion dollars as an alternative to bills. It is hard to see the Treasury materially reducing the number of bills at this point.

Source: U.S. Treasury³

The quarterly refunding will reduce its coupon issuance from prior quarters, primarily in the belly of the curve.  The 7-year is especially affected in reduction of issuance of $21 billion dollars. In comparison, the long end – including 10s, 20s, and 30s combined — will see a minor reduction of $12 billion.

Source: U.S. Treasury⁴

Surveying the years from 2002 to 2021, the current level of debt issuance compared to prior years was massive. Fed purchases into SOMA helped to digest that duration. Though reduced, debt issuance will remain at very high levels and concurrently at a time that the Fed is materially capping the purchases it will make from principal repayments.

The key takeaway here is that even though coupon issuance is lower than in recent quarters, it is still very large compared to the last decades, and the Fed is significantly reducing reinvestments in a much larger size.

Duration could increase between $350 to $400 billion in roughly a 5-year duration into the end of the year.

During 2022, the NY Fed will purchase $522.5 billion fewer securities of which $330 billion will be Treasury coupons and $192.5 billion will be MBS. Given timing of cap implementation, lumpiness of maturities, and mortgage production, the timing of the duration increase is dynamic, but much of it will be in the last quarter of this year.  Next year should have even more duration as the caps remain and the paydowns are large.

On current trajectory, the Treasury will issue $23 billion fewer securities per month ($69 billion/3 months) or $184 billion less for the remaining 8 months from May through December. With $330 billion in fewer purchases, this will be approximately $146 billion in net new duration issuance into the end of this year.

The Treasury targets a 5-year duration profile, but it has been issuing securities with much higher duration to actively extend the duration during the recent low-rate environment.  The chart below is from the Treasury presentation to TBAC earlier this year and shows debt weighted average maturity (WAM) of 6 years.⁵ This includes 17% bills, which have a much shorter WAM. It is a reasonable estimate that the go forward WAM of Treasury coupon issuance is 7 or more years.  Indeed, the WAM of the Quarterly Refunding Schedule is slightly over 8 years. A treasury security with a WAM of 7 and 8 years, has a modified duration of 6 to 6.5 years.

Source: U.S. Treasury⁶

Production of MBS is a much harder item to forecast. With rates across the curve rising rapidly, the current To Be Authorized (TBA) market has one of the broadest active coupon markets in recent memory as production stretches from 2.0s to 5.5s. Lower coupons are still moving through the production “pipeline” while newer production in the higher coupons is slower. Nevertheless, the Fed is capping its purchasing up to $17.5 billion of new mortgage production from June to August and then cap its purchases up to $35 billion thereafter. Regardless of where that duration is in new production or as extended MBS in current investor portfolios, it exists and needs to be managed. Market participants have expressed concern that the Fed could outright sell MBS if there is not enough production to hit caps. That concern doesn’t seem to be supported by history, but the risk could be there.

Using a TBAs for July delivery and Bloomberg’s prepay model, 4.0s have a duration⁷ of 6.3. Similarly, 4.5s have a duration a duration of 4.0. Importantly, the negative convexity, a measure of the change in duration as rates move up and down, is -2.7 and -1.7 respectively, which means if rates continue to move higher, these durations will extend as home prepayments will fall. A reasonable simple duration estimates of MBS hitting the market over the coming year is $192.5 billion of a 5-year duration. However, if rates rise by 100 basis points as they did in 2017 when caps were first implemented, the negative convexity will kick in, extending mortgage duration.

Putting together the Treasury duration and mortgage duration we will most likely see $146 billion of 6-year duration from treasuries and $192.5 billion of 5-year duration from MBS.  Grossing up the Treasury notional for 5-year equivalent duration by multiplying it by six fifths, the Treasury duration increases to $175 billion. Together, this is a point estimator of $368 billion of 5-year duration hitting the market over the coming 8 months. With slightly less Treasury issuance and less mortgage production, this could be lower, but with a rise in rates or more issuance, this could increase duration. A reasonable range of duration is $350 billion to $400 billion in a 5-year.

To put this in perspective, this is 3.5 to 4 million more 5-year contracts.  The current 5-Note futures ADV is 1.6 million contracts. This could bring it to nearly 2 million contract or 25% higher in this year alone. Duration will keep hitting the market into the next two years.

Source: U.S. Treasury⁸

Much of this activity will start in June as the market is currently pricing the FOMC to hike another 50 basis points. It could be a very busy summer managing all this duration. Beyond September, as the Fed doubles its caps, things could get even more interesting.

1. This paper used modified duration measured in years a bond’s price sensitivity to 1% change in interest rates. 
2. Lorie K Logan Speech, Federal Reserve Asset Purchases: The Pandemic Response and Considerations Ahead.
3. Treasury Presentation to TBAC Q1 2021.
4. US Department of the Treasury, Quarterly Refunding Announcement, May 4, 2022.
5. Weighted average maturity (WAM) is different than duration as it is the average of all the maturities in the portfolio.  Rough estimations of duration can be made by assessing similar securities durations with similar WAM.
6. Treasury Presentation to TBAC Q1 2021.
7. The Bloomberg mortgage prepay model calculates duration as a 1% parallel shift in par rates.
8. CME Group https://www.cmegroup.com/markets/interest-rates/us-treasury/5-year-us-treasury-note.volume.html.