As Prepared for Delivery
Good evening and thank you to the Money Marketeers’ leadership for inviting me to speak with you this evening. I have been to several Money Marketeers events throughout the course of my career and firmly believe in the Money Marketeers’ laudable goal of “establish(ing) greater dialogue between the public and private sector on…substantive issues that move markets.” Consistent with this goal, a broad understanding of the way in which the Treasury Department finances the U.S. government is very much in the best interest of the taxpayer. Tonight, I will seek to demystify that process and explain the core principles that underlie our debt management strategy.
While Congress decides how much money the government spends and how much to raise in taxes, Treasury is responsible for financing the government. Put succinctly, Treasury is the “how,” not the “how much,” when it comes to debt issuance. In this role, our guiding principle, our North Star, is to finance the government at the least cost over time. Treasury, over decades and across Administrations, has sought to achieve this end through three key and interrelated principles: (1) by promoting a broad and diverse investor base to make the Treasury market the deepest and most liquid market in the world, (2) by being “regular and predictable” in our issuance, and (3) by maintaining a robust process to plan for and understand a range of fiscal outcomes.
In my remarks this evening, I will discuss our borrowing objectives, some qualities that make the Treasury market unique, details regarding our approach and process to financing the federal government, and the reasons why we employ such an approach. Hopefully, this will provide some useful insights that will deepen the public’s understanding of Treasury’s debt issuance policy decisions.
Treasury Borrowing Objective
Before I discuss our borrowing objective, it may be instructive to briefly discuss how we determine our borrowing needs.
Borrowing needs are largely driven by three factors: (1) the deficit, (2), the amount of maturing securities that Treasury needs to refinance, and (3) changes in the size of Treasury’s operational cash buffer. Of these three factors, the most challenging to estimate is of course the deficit. Treasury has an office of career staff – the Office of Fiscal Projections (OFP) – which forecasts deficits over both shorter- and longer-term horizons. OFP estimates daily fiscal flows and our cash balance for the next several quarters, and builds these estimates by staying in regular contact with federal agencies to help forecast outlays and with the Office of Tax Policy at Treasury to estimate receipts. Treasury’s Office of Debt Management meets twice a week with OFP to review the latest cash forecast and formulate its issuance decisions.
For deficit forecasts beyond the next several quarters, Treasury looks to both internal information and a range of external forecasts – including those produced by the Congressional Budget Office (CBO), the Office of Management and Budget (OMB), and the primary dealers. By carefully evaluating a range of forecasts and the assumptions used to create them, Treasury can get a better sense of how its forecast is evolving relative to other forecasts. Both the shorter-term and longer-term forecasts figure into our decision making and communications around our quarterly borrowing announcements and our debt management policy statements. It is our aim to develop financing plans that are robust to a range of possible outcomes, particularly over longer time horizons.
As I noted, Treasury’s primary debt management goal – our guiding principle – is to finance the government at the least cost over time. Before describing our strategies, I should clarify what we mean by the phrase “over time.” “Over time” is our recognition that Treasury borrowing is not a one-time event. In thinking about where to issue across the curve, Treasury does not simply find the lowest-yielding instrument at any given point in time and proceed to issue all new securities there.[1] Not only would that be impractical due to the amount of borrowing needed, but also, if it did so, yields would rise in response to the concentration of Treasury issuance that saturates demand for that tenor. Then, when Treasury moved to the next cheapest point, the cycle would repeat. It is worth noting that, among other reasons, Treasury issues a variety of instruments to address its borrowing needs because of a desire to source demand from the broadest possible group of investors. For example, money market funds aren’t interested in owning 10-year notes, just as pension funds hold only a small amount of their fixed-income allocations in short-term instruments like Treasury bills.
Deepest and Most Liquid Market
Turning to debt management strategy, you will often hear Treasury Department officials describe the Treasury market as the deepest and most liquid market in the world. With secondary market trading volume that averages nearly $900 billion per day, investors around the globe use the Treasury market as the safest and most liquid store of value. Among other use cases, Treasury securities are used to express views on the path of interest rates, to provide collateral for loans, to defease liabilities, and to implement monetary policy. Foreign investors play a particularly important role in the Treasury market, holding about 30 percent of outstanding marketable debt.
The depth and liquidity of the Treasury market is not something we take for granted. Treasury works across the official sector through the Inter-Agency Working Group on Treasury Market Surveillance (IAWG) to pursue policies to enhance the resilience of market liquidity, even in times of stress. Most recently, the IAWG – composed of staff from Treasury, the Board of Governors of the Federal Reserve System, the Federal Reserve Bank of New York, the Securities and Exchange Commission, and the Commodity Futures Trading Commission – has made significant progress toward several goals designed to improve the resilience and liquidity of the Treasury market, particularly in times of market stress. For example:
- Treasury launched a buyback program that provides liquidity support for off-the-run securities, led an effort to implement further enhancements to the public release of data on secondary market transactions in on-the-run Treasury securities, and Treasury’s Office of Financial Research finalized a rule to establish ongoing collection of data on transactions in the non-centrally cleared bilateral repo market.
- The SEC adopted new rules requiring central clearing of certain Treasury securities and repo transactions, and amendments requiring certain firms that are significantly involved in market-making of Treasury securities to register as broker-dealers. The SEC also made changes to reporting forms, with Form N-MFP providing more granular information about the activity of money market funds in the Treasury repurchase agreement and Form PF enabling better monitoring of the activity of liquidity funds and drawing clearer distinctions between cash and derivatives activity in the Treasury markets.
This work is purposeful, as we understand that taxpayers benefit from Treasury market liquidity through lower borrowing costs. When the secondary market is resilient, deep, and liquid, investors have a greater willingness to buy Treasury securities, knowing that they can liquidate those securities with minimal transaction costs. This directly translates into higher prices – and therefore lower interest costs – for Treasury when it sells those securities at auction in the primary market.
The History of “Regular and Predictable”
In addition to maintaining a deep and liquid market, Treasury also seeks to follow a regular and predictable issuance framework as a means to finance the government at the least cost over time.
“Regular and predictable” is a phrase that we use often, but rarely clarify. If you will indulge me, I would like to do a bit of a deep dive into the history of this paradigm, what it means, and the benefits of it.
It is probably worth noting that “regular and predictable” issuance, when it comes to bill issuance, is something that Treasury has done since the 1930s, but that has not always been the case for longer-dated instruments. Prior to the 1980s, Treasury issuance of longer-dated securities was largely ad hoc and opportunistic, with debt managers making one-off decisions based on their point-in-time view of market conditions and demand.
For example, in the quarterly refundings of the early 1960s, Treasury offered securities of different tenors with no clear pattern. Looking at 1962, for example, the February refunding included a 4½-year issue, followed by 3¾- and 9½-year issues in May. In August of that year, Treasury offered a 6½-year issue and a 30-year issue, before closing the year in November with 3-year and 9¼-year issues. Following the 30-year bond issue in August 1962, Treasury would not issue such a bond again until July 1964.
In 1972, Treasury started consistently issuing a 2-year note. However, such regularity of issuance did not apply to other coupon tenors, which suggests that Treasury officials at the time wanted to retain optionality around ad hoc issuance for longer-dated tenors. As the federal debt increased through the first half of the 1970s, the feasibility of an ad hoc approach deteriorated and, by 1975, Treasury debt managers first turned to the regular and predictable approach that has been the cornerstone of our strategy for close to 50 years now.[2]
Hallmarks of the “Regular and Predictable” Strategy
But what does it mean for issuance to be regular and predictable? At its most basic level, a regular and predictable issuance paradigm’s main characteristic is that the issuance of securities occurs on a consistent, preset schedule. For example, in the United States, Treasury sells all its benchmark securities at a regular cadence; coupon securities are sold once a month with the 2-, 5-, 7-year tenors being sold near the end of every month and the 3-, 10-, 20-, and 30-year tenors being auctioned in the middle of each month. Shorter-tenor securities (i.e., bills) are generally sold weekly, while Treasury Inflation-Protected Securities (TIPS) and floating-rate notes (FRNs) are sold once a month. Each quarter, Treasury publishes a tentative calendar for the following two quarters with all the tentative auction dates; anticipated coupon auction sizes for the upcoming quarter are published in the quarterly refunding statement.
Similarly, Treasury adjusts issuance sizes and the compositional mix slowly over time. Adjustments are informed by trends in structural demand and Treasury’s assessment of the optimal debt mix to achieve lowest-cost financing over time. One recent example of a shift in structural demand has been the increased demand for bills from money market mutual funds arising from the SEC’s regulatory reforms.
The aim is to minimize borrowing costs over decades, not respond to short-term changes in market pricing or attempt to time the market to capture technical rate dislocations. To provide as much transparency as practicable to market participants regarding Treasury’s issuance intentions, our borrowing plans are communicated publicly via our quarterly refunding announcements. I will get into more detail on this process in a few minutes. This transparency is meant to reduce unnecessary uncertainty that could result in investors demanding an additional risk premium to hold our debt, which would result in higher borrowing costs for taxpayers.
While Treasury makes changes to most issuance only gradually, we frequently adjust bill supply when confronted with rapid, seasonal, or unexpected changes in borrowing needs. As a result, we often refer to bills as an issuance “shock absorber.” Because of their low level of interest rate risk, or duration, changing bill issuance is the least disruptive and most cost-effective way to adjust borrowing quickly or for a short period of time, especially for large changes in borrowing needs.
For example, there is a seasonal increase in bill supply that occurs each February and March during tax refund season, which typically retraces in April as non-withheld and corporate tax receipts rise. Another example is the unexpected change in borrowing needs that Treasury experienced during the COVID pandemic. In 2020, Treasury was able to raise approximately $2.5 trillion in a single quarter by increasing bill financing. This would not have been feasible in longer-term coupon securities.
Treasury also makes significant changes to bill supply during debt limit impasses. Adjustments to Treasury issuance during such periods results both from the constraints on borrowing during the impasse, as well as the subsequent increase in borrowing to rebuild the cash balance once the impasse is resolved. For example, in the four months following the resolution of last year’s debt limit impasse, we increased the cash balance from a low of $23 billion to roughly $700 billion.
Our commitment to regular and predictable practices extends beyond our auctions. In May, we launched a buyback program with the stated intent of following a “regular and predictable” process. At the program launch, we published a quarterly schedule of operations to provide investors with more certainty about the timing of the specific weekly opportunities to sell “off-the-run” Treasuries back to us. We plan to continue to publish such schedules in future quarters.
The regular and predictable paradigm also impacts our thinking about new products. New products are rarely introduced by Treasury and when they are, they are carefully considered. Treasury spends a lot of time thinking about the long-term demand for any new product. We ask ourselves three basic questions: (1) is there likely to be stable demand for such a product over time, (2) would the new product adversely affect demand for existing products, and (3) can the new product be issued in benchmark size?
The key point is that when we introduce a new product, Treasury wants to be sure that it can commit to that product for the long term because we recognize that market participants, particularly intermediaries, will need to commit resources to supporting that product in both the primary and secondary markets. Investors will also want to be assured that a new product will not disappear soon after it is introduced and leave them with a liquidity-impaired instrument. To garner sustained support from investors and intermediaries, we remain regular and predictable.
Introducing a new product, however, is rare. When facing changes in borrowing needs, Treasury tends to first make changes to the auction sizes of existing products. These changes are made in a gradual and transparent manner, and changes to coupon sizes generally are in response to well-formed views on structural changes to demand or to changes in borrowing needs. The increase in auction sizes between August 2023 and April 2024 is a good example of such a shift.
When structural borrowing needs have increased or decreased substantially and changes to auction sizes of existing products have proven insufficient, Treasury has typically met this need by changing the frequency with which it sells existing products. This sort of change is much less frequent and is telegraphed well in advance. An example of this is changes that were made to the TIPS calendar in 2019. A less-frequent option is to add new tenors to existing products, with recent examples including the introductions of 20-year bonds and 6-week, 8-week, and 17-week bills. Least frequent is the introduction of a completely new product type, such as the introduction of TIPS in 1997 or FRNs in 2014.
Benefits of Regular and Predictable
I have spent a lot of time explaining what “regular and predictable” issuance is and how it works. Now, I will focus on the benefits of this paradigm, some form of which is employed by most of the G7. Smaller sovereign issuers typically use a more opportunistic approach to funding, and that approach works for them.
As I noted earlier, the primary reason Treasury uses this paradigm is that it lowers the cost of borrowing over time. A regular and predictable debt issuance program helps to imbue benchmark on-the-run securities with a liquidity premium. Treasury’s regular new issue structure encourages activity in the most recently issued securities, which, because of the extra demand for those securities, lowers Treasury’s borrowing costs. This is evidenced by the yield differentials observed between on- and off-the-run Treasury securities.
A regular and predictable issuance program also facilitates investor planning and encourages broad auction participation because everyone knows when the auctions are and can make investment decisions accordingly. Unlike other approaches, you don’t have to be an “insider” or market expert to invest in Treasury securities. Together with the single-price auction process, where all winning bids are filled at the same clearing level, the regular and predictable schedule is one of the reasons why Treasury auctions get a significant amount of involvement from retail investors. This broad participation benefits taxpayers and investors.
A regular and predictable issuance paradigm also helps to limit the possibility of Treasury supply decisions serving as a source of market uncertainty and risk. If market participants were worried that Treasury might unexpectedly issue additional securities at a given tenor, they would demand a higher yield when buying those securities to account for this risk. In August 2015, Treasury asked the Treasury Borrowing Advisory Committee (TBAC) to look at the costs and benefits of regular and predictable issuance, and their conclusion was that historically it had saved Treasury tens of billions of dollars over the 17-year period that they examined.[3]
Regular and predictable issuance also help us lower rollover risk by spreading out maturities in a consistent pattern. Spreading out maturities of the securities that we issue limits “lumpiness” in the maturity profile that would otherwise occur if Treasury were to concentrate issuance on a smaller number of maturity dates. A lumpier maturity profile would represent significant risks for Treasury, as a great volume of securities would need to be refinanced on a few key dates.
Regular and predictable issuance also helps the Treasury market fulfill its role as the most important global risk-free interest rate. For example, it provides a reliable benchmark for other borrowers (including corporate bonds) and for a host of financial products (including interest rate swaps).
It is also worth noting that government borrowing costs likely would not benefit from Treasury trying to be opportunistic. First, there would be no reason to think Treasury would have a greater ability to predict the direction of interest rates than the market does. Second, any attempt to behave opportunistically and capture perceived “value” opportunities would fail. The market would move well before Treasury could realize the benefits of behaving opportunistically. Treasury announces its activities publicly in advance and relies on market participants to bid competitively in its auctions. Any opportunities that might exist to opportunistically capture value would disappear the moment Treasury announced its intention to capture them. Attempts to move in small increments to capture value would likely have minimal effect given the overall size of the federal debt. Even if there were short-term gains to be had, they would likely be more than offset by greater costs over the long run as market participants priced in an uncertainty premium.
Finally, our regular and predictable approach allows Treasury to gather feedback from a wide variety of market participants. Seeking the broadest possible range of views helps ensure that Treasury makes the best-informed decisions possible.
Debunking Common Misconceptions
I would also like to take a few minutes this evening to address some common misunderstandings about Treasury issuance that we often hear in the marketplace.
One common misperception is that “regular and predictable” means Treasury must never make changes to its issuance plans. This theory rests on the false premise that, once Treasury has a broad issuance plan in motion, it never refines that plan along the way and should ignore market demand signals. History contains many counterexamples; for example, when reducing coupon auction sizes in 2021 and 2022, Treasury reduced 7- and 20-years more, and when increasing auction sizes in 2023 and 2024, Treasury increased those sizes less. Both of these adjustments were made in response to market feedback and our assessment of structural demand. Regular and predictable issuance does not require blind continuation of past practice in the face of new information.
Another example of this sort of misperception occurred in November 2023. At that time, Treasury made an exceptionally modest adjustment to the pace of auction size increases for longer-dated securities. This was driven by Treasury’s views of structural demand and updates to expected borrowing needs, both of which were informed by a rigorous process that I will describe shortly. Total gross monthly coupon issuance was approximately $300 billion per month at the time, so our adjustment – reducing the pace of increase of 10-, 20-, and 30-year securities by $1 billion per month per security – represented a roughly 1 percent change. This kind of modest adjustment is exactly what the regular and predictable framework calls for. This was not outside of any norms – it was consistent with the expectations of many primary dealers and with the recommendations of the TBAC.
Another common misperception that bills as a share of total marketable debt need to remain in a 15–20 percent range. This is not the case. Going back to 1980, bills have been as high as 36 percent of Treasury debt and as low as 10 percent and have been in the 15–20 percent range only about 13 percent of the time during that period.[4]The notion of a 15–20 percent range has its roots in late 2020. In November of that year, the TBAC recommended that Treasury allow the bill share to gradually decline into the 15–20 percent range after it had moved briefly above 25 percent to finance the pandemic response. The TBAC’s recommended range was centered below the long-term average of 22 percent, but above the average during the prior decade. In November 2021, the bill share had dropped to 17 percent, and the TBAC reiterated its recommendation to maintain a 15–20 percent range, noting that cuts in coupon issuance were needed to prevent bill supply from dropping too low, which would present risks to market functioning. However, the TBAC emphasized that “there is flexibility in the TBAC’s recommended range for bills to either fall below 15 percent of outstanding stock (in which case excess cash will likely get absorbed by the RRP facility) or for bills to rise modestly above 20 percent while still maintaining financing flexibility for Treasury.” In August 2023, the TBAC reiterated that it was “comfortable running T-bills in the range of their longer-term historical share of 22.4 percent for some time before returning to the recommended 15–20 percent range, in order to maintain a regular and predictable approach to increasing coupon issuance.”
In short, the TBAC recommendation of a 15–20 percent range is relatively recent, and the TBAC has repeatedly said that Treasury should maintain flexibility on the 15–20 percent range in support of regular and predictable issuance. This is consistent with the “bills as a shock absorber” tenet I described earlier. If the 15–20 percent range were to be considered a hard rule, responding to shocks would require large and rapid changes to coupon auction sizes, contrary to our longstanding approach.
Refunding Process
Finally, I want to focus a bit on Treasury’s financing process and, in particular, our communication with market participants. Not only is this part of Treasury’s adherence to the regular and predictable framework, but it also builds confidence in our decisions via a thorough process with extensive input from professional career staff and market participants.
The quarterly refunding – the term for the quarterly process in which changes in debt management policy are considered, made, and communicated – is the most important way that we interact with the market when communicating policies and other details about debt management. This cycle of activity repeats every three months and culminates in our quarterly refunding statement. But there are many steps that take place ahead of that statement.
Nearly three weeks before Treasury publishes the quarterly refunding statement, career staff at Treasury begin gathering information from market participants that is used as one of several inputs in the development of debt management policy – both for the current quarter and for the longer term. We start by engaging with the primary dealers – institutions that are designated as such by the New York Fed and that are expected to bid their pro rata share in every Treasury auction. We issue a publicly available survey to primary dealers each quarter, the results of which help us to better understand supply and demand dynamics and market expectations for Treasury auctions. In addition to surveying all of the primary dealers each quarter, Treasury meets directly with half of the primary dealers each quarter to gather additional context on their responses.
On the Monday of refunding week, Treasury announces its estimates for how much the government will need to borrow for the next two quarters. As mentioned earlier, Treasury considers a range of estimates of borrowing needs, including those from OMB, CBO, and the primary dealers. These inputs support the development of the point estimates published as part of this release. Such forecasts are of course subject to uncertainty, including from the course of macroeconomy or legislative changes that impact fiscal flows. In designing its borrowing plan, Treasury seeks to preserve flexibility in managing against a range of potential outcomes for realized borrowing needs.
The next day (the day before the quarterly refunding statement release), Treasury meets with the Treasury Borrowing Advisory Committee. The TBAC, a federal advisory committee that has operated in various forms since the 1950s, is charged with providing advice to Treasury on matters related to borrowing and debt management policy. The members of TBAC are senior subject matter experts who represent a broad range of Treasury market investors and intermediaries. The types of firms that are represented on the TBAC include asset managers, primary dealers, money market funds, insurance companies, pension funds, hedge funds, regional banks, and custodial banks. The broad range of firms on the TBAC ensures that Treasury receives input that represents a diverse set of perspectives. Among other things, the TBAC is asked to provide recommendations about Treasury issuance (getting advice about “how” it should borrow) as well as advice on special topics, referred to as “charges.” While Treasury makes the decisions about issuance, inputs from the primary dealers and from the TBAC ensure that Treasury’s decision making takes into consideration the expectations and recommendations of a broad variety of market participants, both from Treasury market investors and intermediaries.
Finally, on the Wednesday of refunding week, Treasury announces its anticipated borrowing plans for the quarter, updates on other relevant topics, and tentative auction and buyback schedules. We release materials from the TBAC meeting and hold a live, on-the-record press conference, both available on Treasury’s website. Consistent with the goals of acting in a regular and predictable fashion, the quarterly publication of these materials provides a great degree of transparency about what Treasury is thinking with respect to policy ideas under consideration at that time.
During each quarter, Treasury also follows a structured process for announcing, auctioning, and settling securities. As coupon auctions are formally announced, the offering sizes are nearly always consistent with the guidance provided at the quarterly refunding.[5] Similarly, Treasury now formally announces buybacks each week, based on the anticipated schedule released at the quarterly refunding.
Bill auctions are typically announced each Tuesday and Thursday. Because bill supply is used as an issuance “shock absorber,” Treasury needs to wait for up-to-date information about fiscal flows before finalizing offering sizes. In all cases, there is a regular process for announcing, auctioning, and settling either the sale of securities or purchases made via buybacks.
Conclusion
I have covered a lot here this evening, and hope that you leave here with a few key takeaways.
Our mission when it comes to debt issuance is to finance the government at the least cost over time. This mission statement has existed for several decades, and we never lose sight of that North Star.
We seek to finance the government at the least cost over time in three key ways—first by promoting market liquidity and resilience, second by issuing securities in a regular predictable manner, and finally by utilizing a clear, regular refunding process.
Market liquidity and resilience benefit taxpayers in the form of lower borrowing costs. Liquidity often comes up in our outreach to investors, and it factors into all of our thinking regarding policies, processes, and regulations related to both the primary and secondary markets. The Treasury market is the deepest and most liquid market in the world, and our goal is to keep it that way.
We issue securities in a regular and predictable fashion as part of our strategy to borrow at the lowest cost over time. This half-century-old issuance paradigm continues to serve us well. As noted earlier, regular and predictable issuance doesn’t mean that Treasury steadfastly “stays the course” even when presented with new information and situations. That would be inefficient because borrowing needs and investor demand change over time. Instead, being regular and predictable provides a framework for making changes in the least disruptive way possible, which minimizes risks to investors and ultimately translates into lower borrowing costs for Treasury.
We implement our strategy with clear, regular processes, after seeking a wide range of inputs to inform our decision-making. Treasury provides transparency about its choices, its reasoning, and, as much as practicable, about its issuance plans while retaining flexibility to respond to the unforeseen events that inevitably arise.
Before I close, I want to recognize the efforts of the many professionals, past and present, who have worked on debt management issues across the Treasury Department over the decades and across Administrations. Most of all, I want to highlight the work of the Office of Debt Management, a team of career staff at Treasury without whom the Treasury market would not be the global standard that it is today.
Many of the concepts we have talked about this evening, many of the processes that we have discussed, were implemented decades ago and have been refined along the way. But one critical idea has not changed: Treasury continues to aim to finance the government at the least cost over time.
Thank you for the opportunity to speak with you today.
###
[1] In addition to the considerations noted in this section, the level of yields at different maturity points is not necessarily the “right” metric. For example, if longer-term yields were at 4.5%, issuing short-term debt today at 5% might have a lower expected cost if the average expected short-term rate over the longer term was 4%.
[2] For more information on the history of regular and predictable issuance, see Garbade, Kenneth, “The Emergence of ‘Regular and Predictable’ as a Treasury Debt Management Strategy,” Economic Policy Review, Volume 13, Number 1, March 2007 (available at https://www.newyorkfed.org/research/epr/07v13n1/garbade/exesum_garb.html).
[3] TBAC Charge on “The Meaning and Implications of “Regular and Predictable” (R&P) as a Tenet of Debt Management” August 2015 (available starting on page 51 at https://home.treasury.gov/system/files/276/August2015TreasuryPresentationToTBAC.pdf).
[4] See “Office of Debt Management: Fiscal Year 2024 Q2 Report,” p. 25, https://home.treasury.gov/system/files/221/TreasuryPresentationToTBACQ22024.pdf.
[5] The only recent deviation occurred in the spring of 2020. Borrowing needs changed dramatically in response to the pandemic, which led Treasury to increase auction sizes ahead of the May refunding.
