Description
DEPARTMENT OF THE TREASURY
“CONSIDERATION OF A PROPOSED
TREASURY SECURITIES LENDING FACILITY”
. The Department of the Treasury (“Treasury”) is considering whether it should make available an
additional, temporary supply of Treasury securities on rare occasions when market shortages
threaten to impair the functioning of the market for Treasury securities and broader financial
markets, and, if so, how Treasury should accomplish this. This document is intended as a
vehicle to facilitate public discussion. Treasury has not taken any position on the basic question
of whether it should establish a securities lender of last resort facility (SLLR), or, if it does so,
how Treasury should implement such a facility.
Comments must be in writing and received by August 11, 2006.
Please submit comments to Treasury’s Office of Debt Management, Attention: Jeff Huther,
Director, Office of Debt Management, Room 2412, Department of the Treasury, 1500
Pennsylvania Avenue, NW, Washington, DC 20220. Because postal mail may be subject to
processing delay, we recommend that comments be submitted by electronic mail to:
debt.management@do.treas.gov.
All comments should be captioned with “Comments on Securities Lending Facility.” Please include your name, affiliation, address, e-mail address and telephone number(s) in your comment. All comments received will be available for public inspection by appointment only at the Reading Room of the Treasury Library. To make appointments, please call the number below. FOR FURTHER INFORMATION CONTACT: Jeff Huther, Director, Office of Debt Management, 202-622-2630 (not a toll-free number). 2 .
1. Introduction1 A safe, liquid and highly efficient U.S. Treasury securities market is an invaluable national asset. Treasury securities play a key role in financial markets as risk-free assets, and the extraordinary liquidity in the Treasury market has also led to a role for Treasury securities as pricing benchmarks for a broad array of private financial assets.
Moreover, market participants can execute and manage large positions in the Treasury market with relatively low costs, making Treasury securities the instruments of choice for many in managing interest rate risk. Market participants are willing to pay a premium price for these special attributes of Treasury securities, which in turn allows the U.S. government to borrow at the lowest possible cost over time. Confidence in the safety and liquidity of the Treasury market is supported by the efficient settlement and clearing of Treasury transactions.
This underscores the importance of safeguarding, and enhancing where possible, a well-functioning Treasury market. The Treasury market generally operates very well—but there have been a few instances in which market functioning has been impaired by forces such as attempted market manipulation, catastrophic operational disruptions, and complications associated with historically low short-term interest rates. Some of these episodes have been associated with elevated levels of settlement fails as outsized demands for particular Treasury securities have outstripped the available supply.2,3 Adverse market outcomes in these cases have included one or more of the following—distorted prices in the Treasury cash, derivative and collateral markets, and deterioration in dealers’ market-making activities.
Left unaddressed, such developments could pose risks to efficient Treasury market functioning and result in higher borrowing costs for the U.S. Treasury. In August of 2005, Treasury announced at its Quarterly Refunding that it had concluded that the concept of a backstop securities facility warranted further consideration, and indicated that further advice from market participants would be sought on this idea. At subsequent Quarterly Refundings, Treasury indicated that it was continuing to study the desirability of a 1 This notice was prepared by the staff of the Office of Debt Management, U.S.
Department of the Treasury, in consultation with the staff of the Markets Group, Federal Reserve Bank of New York. It has benefited greatly from comments provided by colleagues in the Division of Monetary Affairs at the Board of Governors of the Federal Reserve System. 2 Settlement failures occur when the party selling a security fails to deliver the security on the agreed upon settlement date. Settlement failures occur for a variety of reasons including errors and miscommunications.
These failures, often called frictional failures, are small and are generally resolved quickly. Larger, more chronic fails can occur due to wide-scale operational disruptions. In addition, under current market conventions, the costs incurred by market participants in failing to deliver securities fall with the level of the market repo rate.
The potential for chronic fails episodes thus increases in a very low interest rate environment such as that prevailing during the summer of 2003. 3 In the collateral market, market participants borrow securities by lending funds against Treasury collateral, typically through the use of repurchase agreements. At the inception of the transaction, the dealer “borrows” the security and lends funds at the repo rate. When the transaction matures, the security is returned and the loan is repaid with interest.
Although sometimes described in economic terms as a collateralized loan, a repurchase agreement consists of a purchase of securities, followed by a sale at the unwind of the transaction. 3 . standing, nondiscretionary securities lending facility. This concept was also discussed at meetings of the Treasury Borrowing Advisory Committee in August and November, 2005. To assist in further consideration of this issue, Treasury is publishing this notice to seek comment on the question of whether establishment by Treasury of an SLLR would be an appropriate response to the potential threat of financial market duress described above. Assuming that an SLLR was seen as an appropriate response, we further seek comment on how we could accomplish this goal. Treasury has not taken any position on whether a SLLR would be beneficial, or, if so, the way in which an SLLR should be structured or implemented. In order to focus the discussion, however, and to solicit meaningful reaction and comments, this notice also outlines one potential structure for an SLLR. To foster discussion and feedback on these basic questions, the notice identifies some important policy considerations underlying these questions.
Section 2 begins by discussing basic issues associated with chronic settlement fails and also notes some of the related history of past proposals to establish a securities lending facility. Section 3 contains some basic lender of last resort principles that might apply to the design of a securities lending facility. Section 4 summarizes some of the potential benefits and costs of a SLLR.
Section 5 then outlines one of several possible structures for a prototype SLLR and evaluates many critical design features, and section 6 addresses legislative changes that may be needed and other implementation issues. Section 7 concludes with a summary of critical questions for public comment. We invite comment on any and all aspects of this notice. 2. Chronic Settlement Fails When settlement fails become acute and protracted, the smooth functioning of the Treasury market is undermined. Such episodes can lead to increased and unintended credit exposures, and can also hamper efforts by investors to liquidate positions.
In these circumstances, resources are diverted from productive activities to the monitoring, controlling and clearing of unsettled trades.4 Protracted acute fails may also shake investors’ confidence in the safety and liquidity of U.S. Treasury securities at precisely those moments when bolstering public confidence is most needed. In such situations, the reliability and effectiveness of Treasuries in their benchmark and risk management roles could be compromised, with potential adverse spillover effects on the functioning of broader capital markets.
This was precisely the situation encountered in the second half of 2003, when persistent and chronic settlement fails plagued the May 2013 ten-year note.5 The risk of acute and protracted settlement fails could potentially be alleviated by a temporary increase in the supply of Treasury securities. While market participants may be able 4 Garbade, Kenneth, D. and John B Kambhu, “Why is the U.S.
Treasury Contemplating Becoming a Lender of Last Resort for Treasury Securities?,” Federal Reserve Bank of New York, Staff Reports, No. 223, October 2005, revised April 2006. 5 For a detailed discussion of this episode, see Fleming, Michael J. and Kenneth D.
Garbade, “Repurchase Agreements with Negative Interest Rates,” Federal Reserve Bank of New York, Current Issues in Economic and Finance, Volume 10, Number 5, April 2004. 4 . to implement changes in market conventions that improve the availability of securities in high demand, only the U.S. Treasury can increase the aggregate supply of securities.6 There are other options available to Treasury to address impaired Treasury market liquidity, including permanently increasing supply through a reopening or, in some cases, the issuance of a Large Position Report.7 However, these options may be limited in their effectiveness, disruptive to Treasury’s “regular and predictable” issuance patterns and costly to Treasury’s commitment to stability of supply.8 The 1992 Joint Report on the Government Securities Market identified a SLLR as a preferred option to reopenings in addressing acute supply shortages. The report states that “the securities lending approach has some significant advantages over auctions and taps. It would be a temporary measure to deal with a temporary market problem.
It provides for a better possibility for the Treasury to capture some of the pricing anomaly and thus in effect make money for the taxpayer. Finally, like a tap, it is a more flexible approach than auctions to ending a squeeze.”9 3. Objectives and Principles We anticipate that the structure and operation of a securities lender of last resort would embody many of the basic objectives and principles that underlie traditional lender of last resort facilities. Fundamentally, a well-designed SLLR would act as a form of “catastrophe” insurance in the Treasury market—in normal circumstances, its impact would be minimal, but it would play an important role in mitigating the impact of very rare but potentially very costly events that weaken investor confidence and threaten the overall functioning of the Treasury and broader financial markets.
Consistent with this broad objective, we anticipate that the design of a prototype SLLR could incorporate a few key principles listed below. 6 Garbade and Kambhu (2005/2006) posit that “forestalling chronic settlement fails by introducing a lender of last resort of Treasury securities is conceptually similar to forestalling systemic bank suspensions by introducing a lender of last resort of money.” Pg.2. 7 Under 15 U.S.C. § 78o-5(f), Treasury may require persons holding or controlling large positions in certain Treasury securities to report their positions for the purpose of monitoring the impact in the Treasury securities market of concentrations of positions. 8 Following the post 9/11/2001 reopening of the August 2011 ten-year note in October 2001, then-Treasury Under Secretary Peter Fisher made it clear that reopening securities on an ad hoc basis to address shortages was not something that would be utilized frequently to address shortages because of the impact on borrowing costs. In remarks to the Futures Industry Association on March 14th 2002, US Fisher stated “.
. . the unscheduled reopening of the 10-year note last October was undertaken because of concerns about the long-term consequences of systemic failure in our credit markets - even though the uncertainty it engendered may have added to our borrowing costs in the short run.
For that reason, unscheduled reopenings will remain the exception - the exceedingly rare exception.” 9 See Department of the Treasury, Securities Exchange Commission and Board of Governors of the Federal Reserve System, The Joint Report on the Government Securities Market (January 1992). The report also identified other options and stated that there were advantages and disadvantages of each option. 5 . • The SLLR would provide additional, temporary supply on rare occasions when market shortages threaten to impair the functioning of the Treasury and broader financial markets. The SLLR would be intended to act only as a backup source for Treasury securities during the rare episodes in which Treasury market liquidity and functioning has become impaired. The terms and conditions should ensure that program usage is confined only to those instances in which markets are not operating normally. • Usage of the SLLR would be determined by market forces rather than Treasury discretion. Crisis events can occur with little or no warning, and administrative discretion in determining whether the SLLR should be available could result in delayed access and in speculative uncertainty about its availability. The pricing of Treasury securities would be less certain in this environment and policymakers could be perceived as acting in an arbitrary or capricious manner or engaging in favoritism. (We note that such concerns led the U.K. Debt Management Office to establish a non-discretionary securities lending facility.10 ) In addition, a transparent program that is driven objectively by market forces would be in keeping with the Treasury’s commitment to a “regular and predictable” debt management policy. • The availability of the SLLR should strengthen investor confidence in the continued safety, liquidity and efficiency of Treasury markets. In many cases, the potential for a substantial decline in market liquidity can be selffulfilling—market participants fearing a deterioration in market conditions may pull back from market activities such as securities lending, thereby exacerbating the situation.
An effective SLLR should work to prevent this by bolstering confidence among market participants that an additional, transparent supply of highly sought after securities would be available. 4. Potential Benefits and Costs of Proposed SLLR Market analysts have observed a number of possible benefits and costs that might be associated with an SLLR. Among the potential benefits, an effective SLLR might bolster overall Treasury market liquidity, even in normal circumstances, by insuring against extreme shortages 10 The U.K. Debt Management Office obtained the authority to lend securities in the late 1990s.
However, market participants were unsure about the criteria that would inform the DMO’s decisions to influence the supply of securities in this way, and this uncertainty was a source of concern. To address such concerns, the DMO proposed a non-discretionary facility in 1999 that was implemented in June of the following year. Under the terms of the facility, eligible institutions could borrow securities at any time.
However, the securities were made available at a penalty rate that effectively discouraged borrowing except in those cases when market conditions were extremely tight or disrupted. Since its inception, the non-discretionary facility has been utilized quite infrequently and reportedly has had little, if any, adverse impact on the normal operations in the gilt cash, repo, and futures markets. 6 . of particular securities. Moreover, an SLLR could contribute to greater confidence during a financial crisis by assuring investors that additional supply of scarce Treasury securities will be available in periods of extreme market disruption. If this effect were significant, the SLLR could be an effective crisis management tool.11 Finally, by guarding against widespread settlement fails, a SLLR could substantially reduce expected operational and regulatory costs associated with settlement of Treasury transactions. Weighing against these possible benefits, some observers have pointed to the potential for significant adverse market effects. In particular, some have argued that a SLLR could contribute to moral hazard by effectively “bailing out” investors with short positions.
The increase in moral hazard, in turn, might contribute to excessive risk-taking in markets. In addition, some have pointed to the potential for a SLLR to be “gamed” by market participants in a way that would be detrimental to investor confidence and that could impair the overall functioning of the Treasury cash and repo markets. Such an outcome would ultimately feed back in higher borrowing costs for the U.S.
Treasury. An even broader conceptual question is whether there is a clearly-defined weakness in the market structure sufficient to warrant the involvement and intervention of the federal government in the market through a SLLR, and whether such an intervention would undermine or reduce private sector incentives to better (and perhaps more efficiently) resolve the issues that the SLLR is intended to address. Quantifying the potential benefits and costs associated with a SLLR is inherently difficult. Other countries have implemented securities lending facilities, apparently without significant adverse effects.
On the other hand, the level of activity in the U.S. Treasury market dwarfs that in other sovereign debt markets, so drawing inferences from the experience of other countries on this point may not be appropriate. 5. One Possible Structure--Terms, Conditions and Other Operational Details The critical design features for the SLLR are the basic distribution mechanism and the various terms and conditions of securities loans, including rate, maturity, and delivery conventions.
A number of other parameters, such as eligible borrowers, available securities, borrowing mechanics and transparency, collateral valuation, margins and rights of substitution, borrowing limits, and reporting and administrative criteria are also important. Each of these design features is discussed in greater detail below. The terms and conditions that are presented below are not being recommended by Treasury and are being provided solely as a vehicle for more focused comment and discussion. Treasury has found in conversations with market participants and the public that a “straw man” model is extremely useful in eliciting views that are ultimately applicable to any of the many possible models on which an SLLR could be structured. The substantial detail presented in this 11 When faced with unprecedented levels of settlement fails that persisted for weeks after the 9/11 terrorist attacks, the Treasury Borrowing Advisory Committee “overwhelmingly felt that Treasury should expand their ability to enhance liquidity in the Treasury market.
To accomplish this, they could set up a repo facility to help alleviate protracted shortages, in particular, large and persistent fails when for some reason emergency reopenings, large position reporting, and debt buybacks do not work.” Report of the Treasury Borrowing Advisory Committee (October 30, 2001). 7 . particular SLLR model should not be construed as an endorsement by Treasury of either the general concept of implementing an SLLR, or, of this model. • Distribution Mechanism: Auctions versus Fixed-Rate (Price) Standing Facility Securities borrowed from the SLLR could either be fixed in quantity with the rate set through an auction or fixed in rate with the quantity determined by the borrower. However, only a fixed-rate standing facility would ensure that the needed supply of Treasury securities would be available to all eligible borrowers. This construct seems to be most in line with the concept of “lender of last resort,” allowing market participants to borrow as much supply as needed to resolve acute and protracted settlement fails. • Rate, Maturity, Delivery and Reporting Options As noted above, these parameters should be set in a such a way that borrowing from the SLLR would be a viable option during rare periods of severe market stress but would be viewed as too expensive in normal market conditions. This could likely be achieved through an appropriate combination of the rate, maturity, delivery, and disclosure conventions. The SLLR could make securities available at an implied rate of zero percent.
The implied zero percent repo rate could be achieved by charging a lending fee equal to the appropriate term general collateral repo rate.12 The lending fee alone should limit borrowing from the SLLR to only those cases when the market repo rate for a particular security has fallen to zero. The use of the SLLR could be even more narrowly targeted by suitable specifications of the term of the loan and a delayed delivery convention. For example, all SLLR loans might be offered for a fixed term with a standard forward delivery.
Requiring that borrowers enter into a term securities loan with an implied zero percent repo rate and a forward settlement date would likely limit borrowing from the SLLR to periods of severe market disruption when the market repo rate was expected to remain at zero for some time and widespread settlement failures were expected to persist for an extended period. A forward settlement date would further discourage strategic use of the facility in implementing short-run trading strategies. It is possible that fairly lengthy term and settlement periods — perhaps a one-week term with a T+5 settlement convention — might be required to limit usage only to scenarios in which markets are severely disrupted. Alternatively, shorter-term loans with maturities of a day or two 12 The lending fee would need to be set so as to guarantee the absence of arbitrage in the case with an assumed specials rate of zero for a security borrowed from the SLLR.
For example, suppose the SLLR extended one-week term loans with a one-week forward start. In this case, a dealer could reverse in general collateral securities today for two weeks and earn the general collateral two-week term repo rate. The general collateral securities could then be financed for one week at the one-week general collateral rate.
After one week had passed, the general collateral securities would be returned to the dealer and they could then be pledged at the SLLR in return for scarce securities. The securities borrowed from the SLLR could then be financed, by assumption, for one week at zero percent. The lending fee in this case would need to be set equal to the one-week forward one-week general collateral rate to guarantee the absence of arbitrage profits. As an operational matter, the discussion here suggests that specifying the appropriate lending fee would likely require calculations based upon regular quotes of general collateral repo rates across a range of maturities. 8 .
and with next-day or skip-day settlement might be adequate. Input from market participants concerning appropriate settings for the term of SLLR loans and the forward delivery convention would be particularly useful. A final element under the terms of borrowing concerns reporting requirements. It may well be desirable to require borrowers to report their daily cash, repo, and futures positions, and fails to deliver and receive in the security borrowed over an interval bracketing the period of borrowing. Reporting of this type would be another factor that would discourage use of the SLLR during normal market conditions and could also be useful in guarding against possible inappropriate uses of the facility. • Collateral The SLLR would lend securities on a bond-for-bond basis, meaning that to borrow securities from the facility, a borrower would have to pledge other Treasury securities of equal market value, plus a margin, as collateral.
A bond-for-bond facility structure would not affect the Treasury’s cash position, which simplifies cash management for Treasury and open-market operations for the Federal Reserve. It likely would be desirable to allow institutions to substitute collateral while borrowing from the SLLR. If collateral substitution capabilities were especially important to market participants, the SLLR might include a tri-party arrangement in which a collateral custodian would handle the back office work in tracking frequent collateral substitutions over the term of a SLLR loan. • Available Securities The range of securities available through the SLLR could be defined in a number of ways. At one end of the spectrum, the SLLR could stand ready to lend additional supply for any outstanding CUSIP number.
That structure would tend to address the inherent difficulties in anticipating future problems that could arise. On the other hand, many of the market problems faced in the past have involved recently-issued nominal coupon securities. This might suggest that the program could be limited to on-the-run and once-off-the-run securities.
Input from market participants about the appropriate range of available securities would be quite valuable. • Borrowing Mechanics and Public Transparency All borrowing requests would be submitted to the Federal Reserve Bank of New York (FRBNY) in its capacity as fiscal agent for the United States Government. As with other Treasury and Federal Reserve operations, the aggregate daily volume of borrowing requests by CUSIP would be made public promptly and well before the loans are settled. 9 . Prompt disclosure would be critical to ensure that market participants with direct access to the facility do not gain a significant information advantage over those without direct access.13 In particular, market participants would need to know how the temporary supply of an outstanding security would change in order to make informed trading and investment decisions. In addition, prompt disclosure should help to dispel bond market rumors about potential borrowing from the SLLR that might otherwise add to financial market volatility. The names of individual borrowers would be kept strictly confidential. • Eligible Borrowers The complexity of collateralized bond-for-bond borrowing and the anticipated infrequent use of the SLLR suggest the need to limit the group of counterparties to a manageable number. For example, direct participation might be limited to primary dealers as designated by FRBNY. Primary dealers play a critical role in making markets for Treasury securities and maintain active trading relationships with FRBNY. Market participants who wished to obtain securities from the SLLR could place their order through a primary dealer.14 This should not represent a significant disadvantage to those entities lacking direct access to the facility. The SLLR borrowing rate would be known to the entire market and competition among primary dealers should ensure that other market participants would be able to tap the SLLR through a primary dealer at a minimal cost.
Moreover, details on the usage of the SLLR (the total amount of borrowing for each security) would be publicly available. • Collateral Margin and Valuation As noted above, one of the basic options for the SLLR involves the provision of term securities loans. In the interest of protecting the Treasury from credit risk, only Treasury securities would be accepted as collateral. The amount of Treasury collateral required from a borrower could also include a margin to protect the Treasury from the risk that the market value of the pledged securities might fall below the value of the borrowed securities. Protecting the Treasury could be enhanced by marking-to-market the value of the collateral each day.
If the market value of the collateral including the margin were to fall below the market value of the borrowed securities, a margin call could be made to the borrower to provide more collateral and reestablish the margin. Conversely, if the market value of the collateral were to change in the borrower’s favor, excess collateral could be released to the borrower. 13 Even with prompt disclosure, borrowers may have an information advantage. They will certainly know that aggregate quantity will rise before it is disclosed to the public.
Dealers submitting bids for others as well as themselves arguably would have the greatest information advantage. 14 This structure would be analogous to that employed during 2000-2001 when the Treasury conducted buyback operations. Non-primary dealers that wished to participate in such operations placed their bids through primary dealers. 10 . • Borrowing Limitations It may be prudent to place some limitations on the total amount of securities that any one participant could borrow. Policymakers might have some concern, for example, about the motivations and financial circumstances of a market participant wishing to borrow enormous amounts of a particular security. A per-issue limit could be set in such a way that the aggregate amount of securities available from the SLLR would be adequate to resolve or substantially mitigate any market disruption. • Rollovers/Loan Extensions Under conditions of severe market dislocations, borrowers may be unable to return borrowed securities to the Treasury on the closing leg of the lending transaction. In these circumstances, imposing harsh penalties for fails back to the Treasury would run counter to the intent of the program; market participants in this case would find it advantageous to fail to private counterparties in their efforts to avoid failing back to the Treasury, potentially exacerbating the fails situation that the SLLR would be intended to address.
For this reason, it might be reasonable to treat fails back to Treasury in the same manner that fails among private counterparties are treated. The original loan could be extended on a daily basis at a zero percent rate with the lending fee thus set equal to the overnight general collateral repo rate. 6. Legislative, Regulatory, and Implementation Issues Beyond determining the structure for the proposed SLLR, there are a number of issues that would need to be addressed prior to implementation, including statutory changes concerning the Treasury’s borrowing authority, debt limit accounting, and the tax treatment of borrowed securities. Each of these is considered in more detail below. • Authority to Issue Securities for the Purpose of Securities Lending Although this paper describes the proposed transactions of the SLLR as “lending,” Treasury would actually be issuing additional securities for a temporary period of time.
The Secretary of the Treasury (“Secretary”) is authorized under Chapter 31 of Title 31, United States Code, to issue Treasury securities and to prescribe terms and conditions for their issuance and sale. The Secretary is authorized to borrow amounts necessary for expenditures authorized by law and may issue securities for the amounts borrowed, and may also issue securities to buy, redeem or refund outstanding securities. These authorities do not appear to encompass the activities of the proposed SLLR.
As a result, Treasury would likely need to pursue new authority to issue securities for the purpose of securities lending in order to implement an SLLR. • Debt Limit Treatment Treasury would also need to consider the implications of issuing additional securities, even on a temporary basis, on the debt subject to limit. A bond-for-bond SLLR may not provide a one-for-one offset accounting treatment for debt limit purposes. Under the current debt limit 11 .
treatment, the par amount of the debt pledged as collateral to the facility could partially or fully offset the par amount of the securities that are lent. However, because the SLLR would likely use the market value of the collateral to determine the market value of borrowed and margined securities, to the degree that market values and par values differ, there would not be a one-forone debt limit accounting offset in a bond-for-bond SLLR structure. For example, if all securities trade close to their par values, borrowing at the SLLR would tend to reduce the debt subject to the limit because the par value of securities pledged as collateral (including the margin) would tend to exceed the par value of securities borrowed. However, if the market value of pledged securities were substantially above par value, borrowing from the SLLR would likely increase the debt subject to limit.
Given this uncertainty, Treasury might need to suspend the SLLR lending activity during the period leading up to debt-limit increases unless there is a legislative change to the current debt limit treatment. • Tax Treatment Some tax issues would need to be addressed. For example, to ensure that Treasury securities borrowed from the lending facility are fully fungible with the outstanding securities, both the outstanding securities and the securities borrowed from the facility would have to be treated for federal tax purposes as being part of the same issue. It may be necessary to seek legislation regarding this treatment. 7. Conclusion As noted at the outset, maintaining a safe, efficient, and liquid Treasury market is a critical public policy objective.
Treasury is seeking comments on whether a well constructed SLLR might provide low cost insurance against certain types of market disruptions during times of financial market crisis. An ideal facility would rarely be utilized, but would be available to mitigate strains in the Treasury market and in broader financial markets. As noted above, there are potential costs to be considered as well, including possible increases in moral hazard and the risk of significant gaming of the facility. Public input in evaluating and designing a SLLR is essential and we invite comment on any aspect of the proposed facility, including whether it should be established at all.
Treasury takes no position on whether a SLLR should be established or, if such a facility were established, how it should be structured. In this regard, comments focusing on potential benefits and costs associated with a SLLR together with an overall assessment of the desirability of establishing a SLLR would be particularly useful. In addition, comments on the various facets of the proposed structure, including various terms and conditions and other operational details, would also be most welcome. 12 .
All comments should be captioned with “Comments on Securities Lending Facility.” Please include your name, affiliation, address, e-mail address and telephone number(s) in your comment. All comments received will be available for public inspection by appointment only at the Reading Room of the Treasury Library. To make appointments, please call the number below. FOR FURTHER INFORMATION CONTACT: Jeff Huther, Director, Office of Debt Management, 202-622-2630 (not a toll-free number). 2 .
1. Introduction1 A safe, liquid and highly efficient U.S. Treasury securities market is an invaluable national asset. Treasury securities play a key role in financial markets as risk-free assets, and the extraordinary liquidity in the Treasury market has also led to a role for Treasury securities as pricing benchmarks for a broad array of private financial assets.
Moreover, market participants can execute and manage large positions in the Treasury market with relatively low costs, making Treasury securities the instruments of choice for many in managing interest rate risk. Market participants are willing to pay a premium price for these special attributes of Treasury securities, which in turn allows the U.S. government to borrow at the lowest possible cost over time. Confidence in the safety and liquidity of the Treasury market is supported by the efficient settlement and clearing of Treasury transactions.
This underscores the importance of safeguarding, and enhancing where possible, a well-functioning Treasury market. The Treasury market generally operates very well—but there have been a few instances in which market functioning has been impaired by forces such as attempted market manipulation, catastrophic operational disruptions, and complications associated with historically low short-term interest rates. Some of these episodes have been associated with elevated levels of settlement fails as outsized demands for particular Treasury securities have outstripped the available supply.2,3 Adverse market outcomes in these cases have included one or more of the following—distorted prices in the Treasury cash, derivative and collateral markets, and deterioration in dealers’ market-making activities.
Left unaddressed, such developments could pose risks to efficient Treasury market functioning and result in higher borrowing costs for the U.S. Treasury. In August of 2005, Treasury announced at its Quarterly Refunding that it had concluded that the concept of a backstop securities facility warranted further consideration, and indicated that further advice from market participants would be sought on this idea. At subsequent Quarterly Refundings, Treasury indicated that it was continuing to study the desirability of a 1 This notice was prepared by the staff of the Office of Debt Management, U.S.
Department of the Treasury, in consultation with the staff of the Markets Group, Federal Reserve Bank of New York. It has benefited greatly from comments provided by colleagues in the Division of Monetary Affairs at the Board of Governors of the Federal Reserve System. 2 Settlement failures occur when the party selling a security fails to deliver the security on the agreed upon settlement date. Settlement failures occur for a variety of reasons including errors and miscommunications.
These failures, often called frictional failures, are small and are generally resolved quickly. Larger, more chronic fails can occur due to wide-scale operational disruptions. In addition, under current market conventions, the costs incurred by market participants in failing to deliver securities fall with the level of the market repo rate.
The potential for chronic fails episodes thus increases in a very low interest rate environment such as that prevailing during the summer of 2003. 3 In the collateral market, market participants borrow securities by lending funds against Treasury collateral, typically through the use of repurchase agreements. At the inception of the transaction, the dealer “borrows” the security and lends funds at the repo rate. When the transaction matures, the security is returned and the loan is repaid with interest.
Although sometimes described in economic terms as a collateralized loan, a repurchase agreement consists of a purchase of securities, followed by a sale at the unwind of the transaction. 3 . standing, nondiscretionary securities lending facility. This concept was also discussed at meetings of the Treasury Borrowing Advisory Committee in August and November, 2005. To assist in further consideration of this issue, Treasury is publishing this notice to seek comment on the question of whether establishment by Treasury of an SLLR would be an appropriate response to the potential threat of financial market duress described above. Assuming that an SLLR was seen as an appropriate response, we further seek comment on how we could accomplish this goal. Treasury has not taken any position on whether a SLLR would be beneficial, or, if so, the way in which an SLLR should be structured or implemented. In order to focus the discussion, however, and to solicit meaningful reaction and comments, this notice also outlines one potential structure for an SLLR. To foster discussion and feedback on these basic questions, the notice identifies some important policy considerations underlying these questions.
Section 2 begins by discussing basic issues associated with chronic settlement fails and also notes some of the related history of past proposals to establish a securities lending facility. Section 3 contains some basic lender of last resort principles that might apply to the design of a securities lending facility. Section 4 summarizes some of the potential benefits and costs of a SLLR.
Section 5 then outlines one of several possible structures for a prototype SLLR and evaluates many critical design features, and section 6 addresses legislative changes that may be needed and other implementation issues. Section 7 concludes with a summary of critical questions for public comment. We invite comment on any and all aspects of this notice. 2. Chronic Settlement Fails When settlement fails become acute and protracted, the smooth functioning of the Treasury market is undermined. Such episodes can lead to increased and unintended credit exposures, and can also hamper efforts by investors to liquidate positions.
In these circumstances, resources are diverted from productive activities to the monitoring, controlling and clearing of unsettled trades.4 Protracted acute fails may also shake investors’ confidence in the safety and liquidity of U.S. Treasury securities at precisely those moments when bolstering public confidence is most needed. In such situations, the reliability and effectiveness of Treasuries in their benchmark and risk management roles could be compromised, with potential adverse spillover effects on the functioning of broader capital markets.
This was precisely the situation encountered in the second half of 2003, when persistent and chronic settlement fails plagued the May 2013 ten-year note.5 The risk of acute and protracted settlement fails could potentially be alleviated by a temporary increase in the supply of Treasury securities. While market participants may be able 4 Garbade, Kenneth, D. and John B Kambhu, “Why is the U.S.
Treasury Contemplating Becoming a Lender of Last Resort for Treasury Securities?,” Federal Reserve Bank of New York, Staff Reports, No. 223, October 2005, revised April 2006. 5 For a detailed discussion of this episode, see Fleming, Michael J. and Kenneth D.
Garbade, “Repurchase Agreements with Negative Interest Rates,” Federal Reserve Bank of New York, Current Issues in Economic and Finance, Volume 10, Number 5, April 2004. 4 . to implement changes in market conventions that improve the availability of securities in high demand, only the U.S. Treasury can increase the aggregate supply of securities.6 There are other options available to Treasury to address impaired Treasury market liquidity, including permanently increasing supply through a reopening or, in some cases, the issuance of a Large Position Report.7 However, these options may be limited in their effectiveness, disruptive to Treasury’s “regular and predictable” issuance patterns and costly to Treasury’s commitment to stability of supply.8 The 1992 Joint Report on the Government Securities Market identified a SLLR as a preferred option to reopenings in addressing acute supply shortages. The report states that “the securities lending approach has some significant advantages over auctions and taps. It would be a temporary measure to deal with a temporary market problem.
It provides for a better possibility for the Treasury to capture some of the pricing anomaly and thus in effect make money for the taxpayer. Finally, like a tap, it is a more flexible approach than auctions to ending a squeeze.”9 3. Objectives and Principles We anticipate that the structure and operation of a securities lender of last resort would embody many of the basic objectives and principles that underlie traditional lender of last resort facilities. Fundamentally, a well-designed SLLR would act as a form of “catastrophe” insurance in the Treasury market—in normal circumstances, its impact would be minimal, but it would play an important role in mitigating the impact of very rare but potentially very costly events that weaken investor confidence and threaten the overall functioning of the Treasury and broader financial markets.
Consistent with this broad objective, we anticipate that the design of a prototype SLLR could incorporate a few key principles listed below. 6 Garbade and Kambhu (2005/2006) posit that “forestalling chronic settlement fails by introducing a lender of last resort of Treasury securities is conceptually similar to forestalling systemic bank suspensions by introducing a lender of last resort of money.” Pg.2. 7 Under 15 U.S.C. § 78o-5(f), Treasury may require persons holding or controlling large positions in certain Treasury securities to report their positions for the purpose of monitoring the impact in the Treasury securities market of concentrations of positions. 8 Following the post 9/11/2001 reopening of the August 2011 ten-year note in October 2001, then-Treasury Under Secretary Peter Fisher made it clear that reopening securities on an ad hoc basis to address shortages was not something that would be utilized frequently to address shortages because of the impact on borrowing costs. In remarks to the Futures Industry Association on March 14th 2002, US Fisher stated “.
. . the unscheduled reopening of the 10-year note last October was undertaken because of concerns about the long-term consequences of systemic failure in our credit markets - even though the uncertainty it engendered may have added to our borrowing costs in the short run.
For that reason, unscheduled reopenings will remain the exception - the exceedingly rare exception.” 9 See Department of the Treasury, Securities Exchange Commission and Board of Governors of the Federal Reserve System, The Joint Report on the Government Securities Market (January 1992). The report also identified other options and stated that there were advantages and disadvantages of each option. 5 . • The SLLR would provide additional, temporary supply on rare occasions when market shortages threaten to impair the functioning of the Treasury and broader financial markets. The SLLR would be intended to act only as a backup source for Treasury securities during the rare episodes in which Treasury market liquidity and functioning has become impaired. The terms and conditions should ensure that program usage is confined only to those instances in which markets are not operating normally. • Usage of the SLLR would be determined by market forces rather than Treasury discretion. Crisis events can occur with little or no warning, and administrative discretion in determining whether the SLLR should be available could result in delayed access and in speculative uncertainty about its availability. The pricing of Treasury securities would be less certain in this environment and policymakers could be perceived as acting in an arbitrary or capricious manner or engaging in favoritism. (We note that such concerns led the U.K. Debt Management Office to establish a non-discretionary securities lending facility.10 ) In addition, a transparent program that is driven objectively by market forces would be in keeping with the Treasury’s commitment to a “regular and predictable” debt management policy. • The availability of the SLLR should strengthen investor confidence in the continued safety, liquidity and efficiency of Treasury markets. In many cases, the potential for a substantial decline in market liquidity can be selffulfilling—market participants fearing a deterioration in market conditions may pull back from market activities such as securities lending, thereby exacerbating the situation.
An effective SLLR should work to prevent this by bolstering confidence among market participants that an additional, transparent supply of highly sought after securities would be available. 4. Potential Benefits and Costs of Proposed SLLR Market analysts have observed a number of possible benefits and costs that might be associated with an SLLR. Among the potential benefits, an effective SLLR might bolster overall Treasury market liquidity, even in normal circumstances, by insuring against extreme shortages 10 The U.K. Debt Management Office obtained the authority to lend securities in the late 1990s.
However, market participants were unsure about the criteria that would inform the DMO’s decisions to influence the supply of securities in this way, and this uncertainty was a source of concern. To address such concerns, the DMO proposed a non-discretionary facility in 1999 that was implemented in June of the following year. Under the terms of the facility, eligible institutions could borrow securities at any time.
However, the securities were made available at a penalty rate that effectively discouraged borrowing except in those cases when market conditions were extremely tight or disrupted. Since its inception, the non-discretionary facility has been utilized quite infrequently and reportedly has had little, if any, adverse impact on the normal operations in the gilt cash, repo, and futures markets. 6 . of particular securities. Moreover, an SLLR could contribute to greater confidence during a financial crisis by assuring investors that additional supply of scarce Treasury securities will be available in periods of extreme market disruption. If this effect were significant, the SLLR could be an effective crisis management tool.11 Finally, by guarding against widespread settlement fails, a SLLR could substantially reduce expected operational and regulatory costs associated with settlement of Treasury transactions. Weighing against these possible benefits, some observers have pointed to the potential for significant adverse market effects. In particular, some have argued that a SLLR could contribute to moral hazard by effectively “bailing out” investors with short positions.
The increase in moral hazard, in turn, might contribute to excessive risk-taking in markets. In addition, some have pointed to the potential for a SLLR to be “gamed” by market participants in a way that would be detrimental to investor confidence and that could impair the overall functioning of the Treasury cash and repo markets. Such an outcome would ultimately feed back in higher borrowing costs for the U.S.
Treasury. An even broader conceptual question is whether there is a clearly-defined weakness in the market structure sufficient to warrant the involvement and intervention of the federal government in the market through a SLLR, and whether such an intervention would undermine or reduce private sector incentives to better (and perhaps more efficiently) resolve the issues that the SLLR is intended to address. Quantifying the potential benefits and costs associated with a SLLR is inherently difficult. Other countries have implemented securities lending facilities, apparently without significant adverse effects.
On the other hand, the level of activity in the U.S. Treasury market dwarfs that in other sovereign debt markets, so drawing inferences from the experience of other countries on this point may not be appropriate. 5. One Possible Structure--Terms, Conditions and Other Operational Details The critical design features for the SLLR are the basic distribution mechanism and the various terms and conditions of securities loans, including rate, maturity, and delivery conventions.
A number of other parameters, such as eligible borrowers, available securities, borrowing mechanics and transparency, collateral valuation, margins and rights of substitution, borrowing limits, and reporting and administrative criteria are also important. Each of these design features is discussed in greater detail below. The terms and conditions that are presented below are not being recommended by Treasury and are being provided solely as a vehicle for more focused comment and discussion. Treasury has found in conversations with market participants and the public that a “straw man” model is extremely useful in eliciting views that are ultimately applicable to any of the many possible models on which an SLLR could be structured. The substantial detail presented in this 11 When faced with unprecedented levels of settlement fails that persisted for weeks after the 9/11 terrorist attacks, the Treasury Borrowing Advisory Committee “overwhelmingly felt that Treasury should expand their ability to enhance liquidity in the Treasury market.
To accomplish this, they could set up a repo facility to help alleviate protracted shortages, in particular, large and persistent fails when for some reason emergency reopenings, large position reporting, and debt buybacks do not work.” Report of the Treasury Borrowing Advisory Committee (October 30, 2001). 7 . particular SLLR model should not be construed as an endorsement by Treasury of either the general concept of implementing an SLLR, or, of this model. • Distribution Mechanism: Auctions versus Fixed-Rate (Price) Standing Facility Securities borrowed from the SLLR could either be fixed in quantity with the rate set through an auction or fixed in rate with the quantity determined by the borrower. However, only a fixed-rate standing facility would ensure that the needed supply of Treasury securities would be available to all eligible borrowers. This construct seems to be most in line with the concept of “lender of last resort,” allowing market participants to borrow as much supply as needed to resolve acute and protracted settlement fails. • Rate, Maturity, Delivery and Reporting Options As noted above, these parameters should be set in a such a way that borrowing from the SLLR would be a viable option during rare periods of severe market stress but would be viewed as too expensive in normal market conditions. This could likely be achieved through an appropriate combination of the rate, maturity, delivery, and disclosure conventions. The SLLR could make securities available at an implied rate of zero percent.
The implied zero percent repo rate could be achieved by charging a lending fee equal to the appropriate term general collateral repo rate.12 The lending fee alone should limit borrowing from the SLLR to only those cases when the market repo rate for a particular security has fallen to zero. The use of the SLLR could be even more narrowly targeted by suitable specifications of the term of the loan and a delayed delivery convention. For example, all SLLR loans might be offered for a fixed term with a standard forward delivery.
Requiring that borrowers enter into a term securities loan with an implied zero percent repo rate and a forward settlement date would likely limit borrowing from the SLLR to periods of severe market disruption when the market repo rate was expected to remain at zero for some time and widespread settlement failures were expected to persist for an extended period. A forward settlement date would further discourage strategic use of the facility in implementing short-run trading strategies. It is possible that fairly lengthy term and settlement periods — perhaps a one-week term with a T+5 settlement convention — might be required to limit usage only to scenarios in which markets are severely disrupted. Alternatively, shorter-term loans with maturities of a day or two 12 The lending fee would need to be set so as to guarantee the absence of arbitrage in the case with an assumed specials rate of zero for a security borrowed from the SLLR.
For example, suppose the SLLR extended one-week term loans with a one-week forward start. In this case, a dealer could reverse in general collateral securities today for two weeks and earn the general collateral two-week term repo rate. The general collateral securities could then be financed for one week at the one-week general collateral rate.
After one week had passed, the general collateral securities would be returned to the dealer and they could then be pledged at the SLLR in return for scarce securities. The securities borrowed from the SLLR could then be financed, by assumption, for one week at zero percent. The lending fee in this case would need to be set equal to the one-week forward one-week general collateral rate to guarantee the absence of arbitrage profits. As an operational matter, the discussion here suggests that specifying the appropriate lending fee would likely require calculations based upon regular quotes of general collateral repo rates across a range of maturities. 8 .
and with next-day or skip-day settlement might be adequate. Input from market participants concerning appropriate settings for the term of SLLR loans and the forward delivery convention would be particularly useful. A final element under the terms of borrowing concerns reporting requirements. It may well be desirable to require borrowers to report their daily cash, repo, and futures positions, and fails to deliver and receive in the security borrowed over an interval bracketing the period of borrowing. Reporting of this type would be another factor that would discourage use of the SLLR during normal market conditions and could also be useful in guarding against possible inappropriate uses of the facility. • Collateral The SLLR would lend securities on a bond-for-bond basis, meaning that to borrow securities from the facility, a borrower would have to pledge other Treasury securities of equal market value, plus a margin, as collateral.
A bond-for-bond facility structure would not affect the Treasury’s cash position, which simplifies cash management for Treasury and open-market operations for the Federal Reserve. It likely would be desirable to allow institutions to substitute collateral while borrowing from the SLLR. If collateral substitution capabilities were especially important to market participants, the SLLR might include a tri-party arrangement in which a collateral custodian would handle the back office work in tracking frequent collateral substitutions over the term of a SLLR loan. • Available Securities The range of securities available through the SLLR could be defined in a number of ways. At one end of the spectrum, the SLLR could stand ready to lend additional supply for any outstanding CUSIP number.
That structure would tend to address the inherent difficulties in anticipating future problems that could arise. On the other hand, many of the market problems faced in the past have involved recently-issued nominal coupon securities. This might suggest that the program could be limited to on-the-run and once-off-the-run securities.
Input from market participants about the appropriate range of available securities would be quite valuable. • Borrowing Mechanics and Public Transparency All borrowing requests would be submitted to the Federal Reserve Bank of New York (FRBNY) in its capacity as fiscal agent for the United States Government. As with other Treasury and Federal Reserve operations, the aggregate daily volume of borrowing requests by CUSIP would be made public promptly and well before the loans are settled. 9 . Prompt disclosure would be critical to ensure that market participants with direct access to the facility do not gain a significant information advantage over those without direct access.13 In particular, market participants would need to know how the temporary supply of an outstanding security would change in order to make informed trading and investment decisions. In addition, prompt disclosure should help to dispel bond market rumors about potential borrowing from the SLLR that might otherwise add to financial market volatility. The names of individual borrowers would be kept strictly confidential. • Eligible Borrowers The complexity of collateralized bond-for-bond borrowing and the anticipated infrequent use of the SLLR suggest the need to limit the group of counterparties to a manageable number. For example, direct participation might be limited to primary dealers as designated by FRBNY. Primary dealers play a critical role in making markets for Treasury securities and maintain active trading relationships with FRBNY. Market participants who wished to obtain securities from the SLLR could place their order through a primary dealer.14 This should not represent a significant disadvantage to those entities lacking direct access to the facility. The SLLR borrowing rate would be known to the entire market and competition among primary dealers should ensure that other market participants would be able to tap the SLLR through a primary dealer at a minimal cost.
Moreover, details on the usage of the SLLR (the total amount of borrowing for each security) would be publicly available. • Collateral Margin and Valuation As noted above, one of the basic options for the SLLR involves the provision of term securities loans. In the interest of protecting the Treasury from credit risk, only Treasury securities would be accepted as collateral. The amount of Treasury collateral required from a borrower could also include a margin to protect the Treasury from the risk that the market value of the pledged securities might fall below the value of the borrowed securities. Protecting the Treasury could be enhanced by marking-to-market the value of the collateral each day.
If the market value of the collateral including the margin were to fall below the market value of the borrowed securities, a margin call could be made to the borrower to provide more collateral and reestablish the margin. Conversely, if the market value of the collateral were to change in the borrower’s favor, excess collateral could be released to the borrower. 13 Even with prompt disclosure, borrowers may have an information advantage. They will certainly know that aggregate quantity will rise before it is disclosed to the public.
Dealers submitting bids for others as well as themselves arguably would have the greatest information advantage. 14 This structure would be analogous to that employed during 2000-2001 when the Treasury conducted buyback operations. Non-primary dealers that wished to participate in such operations placed their bids through primary dealers. 10 . • Borrowing Limitations It may be prudent to place some limitations on the total amount of securities that any one participant could borrow. Policymakers might have some concern, for example, about the motivations and financial circumstances of a market participant wishing to borrow enormous amounts of a particular security. A per-issue limit could be set in such a way that the aggregate amount of securities available from the SLLR would be adequate to resolve or substantially mitigate any market disruption. • Rollovers/Loan Extensions Under conditions of severe market dislocations, borrowers may be unable to return borrowed securities to the Treasury on the closing leg of the lending transaction. In these circumstances, imposing harsh penalties for fails back to the Treasury would run counter to the intent of the program; market participants in this case would find it advantageous to fail to private counterparties in their efforts to avoid failing back to the Treasury, potentially exacerbating the fails situation that the SLLR would be intended to address.
For this reason, it might be reasonable to treat fails back to Treasury in the same manner that fails among private counterparties are treated. The original loan could be extended on a daily basis at a zero percent rate with the lending fee thus set equal to the overnight general collateral repo rate. 6. Legislative, Regulatory, and Implementation Issues Beyond determining the structure for the proposed SLLR, there are a number of issues that would need to be addressed prior to implementation, including statutory changes concerning the Treasury’s borrowing authority, debt limit accounting, and the tax treatment of borrowed securities. Each of these is considered in more detail below. • Authority to Issue Securities for the Purpose of Securities Lending Although this paper describes the proposed transactions of the SLLR as “lending,” Treasury would actually be issuing additional securities for a temporary period of time.
The Secretary of the Treasury (“Secretary”) is authorized under Chapter 31 of Title 31, United States Code, to issue Treasury securities and to prescribe terms and conditions for their issuance and sale. The Secretary is authorized to borrow amounts necessary for expenditures authorized by law and may issue securities for the amounts borrowed, and may also issue securities to buy, redeem or refund outstanding securities. These authorities do not appear to encompass the activities of the proposed SLLR.
As a result, Treasury would likely need to pursue new authority to issue securities for the purpose of securities lending in order to implement an SLLR. • Debt Limit Treatment Treasury would also need to consider the implications of issuing additional securities, even on a temporary basis, on the debt subject to limit. A bond-for-bond SLLR may not provide a one-for-one offset accounting treatment for debt limit purposes. Under the current debt limit 11 .
treatment, the par amount of the debt pledged as collateral to the facility could partially or fully offset the par amount of the securities that are lent. However, because the SLLR would likely use the market value of the collateral to determine the market value of borrowed and margined securities, to the degree that market values and par values differ, there would not be a one-forone debt limit accounting offset in a bond-for-bond SLLR structure. For example, if all securities trade close to their par values, borrowing at the SLLR would tend to reduce the debt subject to the limit because the par value of securities pledged as collateral (including the margin) would tend to exceed the par value of securities borrowed. However, if the market value of pledged securities were substantially above par value, borrowing from the SLLR would likely increase the debt subject to limit.
Given this uncertainty, Treasury might need to suspend the SLLR lending activity during the period leading up to debt-limit increases unless there is a legislative change to the current debt limit treatment. • Tax Treatment Some tax issues would need to be addressed. For example, to ensure that Treasury securities borrowed from the lending facility are fully fungible with the outstanding securities, both the outstanding securities and the securities borrowed from the facility would have to be treated for federal tax purposes as being part of the same issue. It may be necessary to seek legislation regarding this treatment. 7. Conclusion As noted at the outset, maintaining a safe, efficient, and liquid Treasury market is a critical public policy objective.
Treasury is seeking comments on whether a well constructed SLLR might provide low cost insurance against certain types of market disruptions during times of financial market crisis. An ideal facility would rarely be utilized, but would be available to mitigate strains in the Treasury market and in broader financial markets. As noted above, there are potential costs to be considered as well, including possible increases in moral hazard and the risk of significant gaming of the facility. Public input in evaluating and designing a SLLR is essential and we invite comment on any aspect of the proposed facility, including whether it should be established at all.
Treasury takes no position on whether a SLLR should be established or, if such a facility were established, how it should be structured. In this regard, comments focusing on potential benefits and costs associated with a SLLR together with an overall assessment of the desirability of establishing a SLLR would be particularly useful. In addition, comments on the various facets of the proposed structure, including various terms and conditions and other operational details, would also be most welcome. 12 .
