The Rescue of Fannie Mae and Freddie Mac–Module F: Federal Reserve’s Large-Scale Asset Purchase (LSAP) Program

The Fed had reduced the federal funds and discount rates since September 2007 (FRBG 2020). By November 2008, the crisis was far from over, and the federal funds rate was approaching the zero lower bound (FOMC 2008a, 22-23). The FOMC thus sought to develop and implement other nontraditional monetary policy measures to increase the availability of credit and reduce borrowing rates, thereby stimulating and bolstering the financial system (FOMC 2008a, 22-30; FRBG 2008).

On November 25, 2008, after discussions by FOMC participants, the Board of Governors announced the decision to purchase agency debt and MBS (FRBG 2008; English 2020). At the following FOMC meeting in mid-December 2008, officials discussed three nontraditional strategies—“simple” quantitative easing, purchasing longer-term securities, and creating or expanding special liquidity and lending facilities (FOMC 2008a, 17-18). The LSAP program fell into the second category, purchasing longer-term securities.

Nonetheless, the three strategies discussed in December 2008 may provide insight into the Fed’s implementation of the LSAPs as a nontraditional policy measure. Officials did not consider seriously purchasing private securities, as they believed this move would take them even further in the direction of credit allocation, which would result in more longer-term costs than benefits (FOMC 2008a, 34, 57, 71). The full list of benefits and drawbacks of all three approaches outlined in the December meeting can be found in Appendix C. While there was some uncertainty regarding the program’s size, FOMC members understood that the Fed would need to buy a substantial quantity of securities to give the intervention a chance of success (FOMC 2008a, 63-66). The FOMC’s directive gave FRBNY the authority to purchase up to $100 billion in housing-related GSE debt and up to $500 billion in agency MBS by the second quarter of 2009 (FOMC 2008b, 10). The directive also allowed the FRBNY’s Open Market Desk to determine the precise timing of the purchases (FOMC 2008b, 10).

Having already enacted the LSAP program, FOMC officials deliberated between purchasing agency debt and MBS alone or in conjunction with purchasing longer-term Treasury securities and decided to delay purchases of Treasuries (FOMC 2008a).

The FOMC saw advantages to purchasing agency debt and MBS over purchasing any Treasury securities for the following reasons:

• It would remove assets from the market (debt, MBS) that were lower in demand compared to Treasury securities (Gagnon and Holscher 2008, 140)
• Fed analysts concluded that debt and MBS purchases would result in a more rapid recovery of GDP growth than dollar-for-dollar purchases of Treasury securities (Gagnon and Holscher 2008, 140-141).
• It would complement housing refinance activities better than purchases of Treasury securities (Gagnon and Holscher 2008, 141).
• It was easier to explain the rationale for purchasing debt and MBS to the public (Gagnon and Holscher 2008, 140).
• Treasury spreads had already fallen, while private yields had not fallen (breaking a trend) (FOMC 2008a, 57). Several FOMC officials concluded that lowering Treasury yields, which were already low, likely would not have an effect on other yields (FOMC 2008a, 21, 57). By contrast, yields on agency MBS remained high.

The main drawback voiced about the purchase of agency debt and MBS was:

• The FOMC believed that purchasing of agency debt and MBS could be seen as credit allocation, or, in their words, “steering funds to the GSEs and to particular economic sectors” (FOMC 2008a, 21).

A secondary goal of the LSAP program may have been to augment the Treasury’s MBS purchase program for Fannie Mae and Freddie Mac, which some market analysts believed had been too limited and insufficiently transparent (Collins 2008, 19). The secondary goal of this program needs to be considered carefully, in light of the concern expressed by some Fed officials that purchases of longer-term securities through the LSAP program might be seen as credit allocation toward the GSEs and the secondary mortgage market (FOMC 2008a, 34, 57, 71).

Agency debt securities needed to be fixed-rate, noncallable, senior benchmark, at competitive prices, and off-the-run (FRBNY 2010b). In August 2010, the FRBNY also allowed on-the-run debt (on-the-run refers to a security that has been most recently issued) (FRBNY 2010b). The FRBNY allowed on-the-run securities in August 2010 because liquidity had improved and spreads relative to on-the-run securities had fallen (Gagnon et al. 2011, 46). The FRBNY also permitted on-the-run securities in order to reduce market dislocations that had resulted from the MBS purchase program (Gagnon et al. 2011, 46).

FRBNY concentrated MBS purchases on newly issued, 30-year securities (so-called “production” MBS) (Gagnon et al. 2011, 46). The Fed aimed to decrease primary mortgage rates by supporting the market for new loans (Gagnon et al. 2011, 46). Production MBS generally paid lower coupons than seasoned MBS because of then-low interest rates, so all of FRBNY’s MBS purchases paid coupon rates between 3.5% and 6.5% (Gagnon et al. 2011, 46; FRBG 2016). The FOMC also focused purchases on the to-be-announced (TBA) market, where MBS trade weeks or months before delivery, because it offered greater flexibility (FOMC 2009c, 26). Purchasing MBS in the TBA market allowed the FRBNY to respond to daily changes in the market as it could buy more, buy less, or use dollar rolls (Gagnon et al. 2011, 26).

Treasuries. The Fed’s purchases of Treasury securities were authorized under Section 14(2)(b)(1) of the Federal Reserve Act (FRA), which allows the Fed to buy and sell direct from Treasury (1) “bonds and notes of the Unites States with maturities not exceeding six months,” and (2) “bonds and notes, or other obligations which are direct obligations of the United States or which are fully guaranteed by the United States as to the principal and interest” of any maturity “but only in the open market.”

Agency MBS and debt. Section 14(2)(b)(2) of the FRA permits the Fed “[t]o buy and sell in the open market [ . . . ] any obligation which is a direct obligation of, or fully guaranteed as to principal and interest by, any agency of the United States.” Although the GSEs (Fannie Mae, Freddie Mac, and the Federal Home Loan Banks) are not government agencies but government instrumentalities, their Congressional charters and other federal regulations afford them certain special privileges (Reiss 2008, 1053-61). One of these privileges is that the Federal Reserve is required by statute to act as their fiscal agent, a role that the Fed primarily plays for the federal government (Reiss 2008, 1060-61). Another privilege is that many federal regulations permit banks and other entities to treat their obligations as similar to Treasuries and other direct government obligations (Reiss 2008, 1061). In similar fashion, the Federal Reserve has long treated agency debt and securities as “principal agency obligations eligible as collateral for advances” from its discount window and as “direct obligations of, and obligations fully guaranteed as to principal and interest by, any agency of the United States . . . eligible for purchase by Reserve Banks” pursuant to Section 14(b). This was the authority relied on for the agency debt and MBS purchases under the LSAP.^FNSee regulation 12 CFR 201.108, which interprets FRA Section 14(b) to encompass debt and securities of Fannie Mae, Freddie Mac, and the Federal Home Loan Banks, cited in Reiss 2008 at page 1062, footnote 197. See also Arthur L. Broida’s 1971 letter to the Federal Open Market Committee, which further explores agency issues in the context of open market operations. 

Reflecting on the LSAP program, Chairman Ben Bernanke noted:

Probably the most controversial form of unconventional policy adopted in recent years was what the Federal Reserve called large-scale asset purchases (LSAPs) but most of the rest of the world persisted in calling “quantitative easing,” or QE [ . . . ] By law, the Fed was able to purchase only Treasury securities and mortgage-related securities issued by government-sponsored enterprises. Other central banks, in contrast, have been able to buy a range of private securities, including corporate bonds and equities. The limits on the Fed did not seem to prevent its version of QE from being effective, although it was perhaps fortunate that, following a crisis centered on housing finance, the law did permit Fed purchases of mortgage-related securities (Bernanke 2017, 9).

Given the perceived advantages of purchasing agency debt and MBS, analysts at the Fed identified two approaches that the FRBNY could have taken related to the timing and size of its LSAP program (Gagnon and Holscher 2008, 142).

Approach A—Announcing a volume of purchases over a certain time period

Advantages included:

• The Fed could have better control over the size of its balance sheet.
• It resulted in less active trading of the Fed’s portfolio.
• It would be easier to achieve a balance of purchases across different market segments (possibly through an index replication strategy).
• The Fed did not have to assume full responsibility for the price of the securities.

Approach B—Establishing a ceiling for conventional fixed mortgage rates. For example, the Fed could announce that it would purchase all newly issued agency MBS with a certain coupon at par.

Advantages included:

• It would more clearly outline the Fed’s policy.
• The general public would understand it better.

The Fed’s decision to commit a substantial amount of reserves with a flexible timeline was a blend of these two strategies (Gagnon and Holscher 2008, 143). With respect to the first strategy, the FOMC announced an explicit commitment amount (Gagnon and Holscher 2008, 143). With respect to the second approach, the FOMC described this commitment as a ceiling, implying that it would not necessarily purchase the full amount of its commitment (Gagnon and Holscher 2008, 143). FOMC members seem to have recognized that establishing an upper limit could pose such a risk, given its debates over the limit and its decision to include the phrase “up to” in its agency debt and MBS announcements (FOMC 2008a, 79). 

The Fed published a press release on November 25, 2008, that communicated its decision to purchase up to $100 billion in GSE direct obligations and $500 billion in MBS in competitive auctions. In the same announcement, the Fed clarified its expectation for the purchases to occur “over several quarters” and did not describe the per-month path of asset purchases.

In March 2009, several members of the FOMC proposed an alternative to establishing an upper limit (FOMC 2009c, 78-80, 181). They argued that the Fed should announce its intention to increase the size of its balance sheet on a month-to-month basis, without specifying the total (FOMC 2009c, 78-80, 181). According to these members, a balance-sheet approach would allow the FRBNY to adjust its ongoing purchases of MBS, debt, and Treasury securities to reflect current economic conditions, since none of these assets would have a fixed limit (FOMC 2009c, 78-80). The balance sheet approach was not implemented.

Agency Securities. In March 2009, the FOMC decided to expand its purchases of debt and MBS because FOMC officials estimated that the hole in the demand for agency securities was still quite large (FOMC 2009f, 8). Some members argued that the market expected the Fed to extend purchases of the agency MBS at the same purchase pace of at least $500 billion every six months (FOMC 2009c, 73). The FOMC decided to increase MBS purchases by $750 billion (to an aggregate $1.25 trillion) to further support mortgage lending and housing markets. At least one FOMC member felt that the purchases of agency debt had not been effective but suggested that failing to increase the FOMC’s commitment to purchase debt could cause more problems than what it would solve (FOMC 2009c, 73).

Treasury Securities. Concurrent with the expansion in agency debt and MBS limits, the FOMC also decided to purchase $300 billion in longer-term Treasury securities (FOMC 2009h). FOMC officials noted the macroeconomics effects that resulted from the Bank of England’s (BoE) decision on March 5, 2009, to purchase longer-term Treasury securities (FOMC 2009c, 5-6). FOMC officials and staff observed the effects of BoE’s gilt purchases on public-sector and private-sector long-term yields (FOMC 2009c, 5-6; Gagnon et al. 2009, 5). Many officials also expressed less concern with the adverse effects of purchasing Treasury securities than in December 2008, such as creating the perception that the Federal Reserve was monetizing federal debt, which in turn could have adverse effects on term premiums and inflation (FOMC 2009c, 90, 92-94, 96, 176, 208). Some FOMC staff expressed concern that the continued purchases of agency debt and MBS could decrease the benefit of purchasing these securities to the point that Treasury securities would become more beneficial (Gagnon et al. 2009, 2). FOMC officials also suggested that the Treasury securities portfolio would be easier to wind down and sell off than the agency securities (FOMC 2009c, 90, 92, 170-171). It was also noted that given the “high degree of uncertainty” that had resulted from the Fed’s different policies since the beginning of the crisis “it [would] be prudent to consider including a significant share of Treasury securities in any further expansion of purchases” (Gagnon et al. 2009, 2).

The FOMC wound down the Treasury purchases months before it terminated its purchases of agency debt and MBS (FRBNY 2018; see Figure 2). The FOMC announced and gradually reduced its purchases of agency debt and MBS to avoid raising market interest rates (FOMC 2009j, 153-154). Gagnon et al. note that the FOMC was successful with its wind-down strategy, as the termination of the LSAP program did not raise interest rates by any noticeable amount (2011, 57).

The Fed’s designated primary dealers were the only institutions allowed to transact in any of the three securities of the LSAP program as broker dealers from whom the investment managers could purchase securities (FRBNY 2008d, 4, 28; FRBNY 2010b).

Given the size of the commitment and the FRBNY’s lack of experience in MBS purchases, FRBNY selected four investment managers and a custodian after the program’s announcement (FRBG 2008; FRBNY 2008a). The FRBNY chose BlackRock, Goldman Sachs Asset Management, PIMCO, and Wellington Management to serve as investment managers, based on their operational capacity, size, competitive fee structure, and knowledge of the MBS market (FRBNY 2008a). The FRBNY selected JPMorgan Chase (JPMC) to be the program’s custodian and also tasked it with providing fund accounting and administrative services (FRBNY 2008b, 1, 26; FRBNY 2010a).

The investment managers entered into nearly identical agreements with the FRBNY, and became responsible for purchasing agency MBS and exercising tangential rights (such as proxy rights or warrants) under the securities, in accordance with FRBNY investment guidance and objectives and on its behalf (FRBNY 2008d, 1-2; FRBNY 2008c, 1-2; FRBNY 2008f, 1-2; FRBNY 2008e, 1-2). FRBNY authorized investment managers to buy and sell agency MBS on the Fed’s behalf via the Fed’s System Open Market Account (SOMA) (FRBNY 2008d, 1-3). Transactions were to be communicated to and settled by the custodian (FRBNY 2008d, 3-4).

The investment managers had the sole right to determine the broker dealer for trades from a list approved by the FRBNY and to establish the rate for execution services (FRBNY 2008d, 4). Upon request, managers were required to both offer advice related to residential loan modification and to provide assistance to influence residential loan modification and policies of residential mortgage-backed loans tied to agency MBS (FRBNY 2008d, 2). In addition, managers were obligated to provide the FRBNY with monthly reports, submit weekly market updates to the FRBNY and the custodian, and meet with FRBNY representatives each month (FRBNY 2008d, 4-6).

Each manager was paid a fee calculated monthly and paid quarterly “based on the average quarterly notional value of the Agency MBS” in the LSAP portfolio from all investment managers (based on the records of the Custodian)^FNThe agreement provided that “notional amount shall mean the Current Face amount of the Agency MBS, including unsettled Trades and TBA's and not to be reduced by the unsettled Current Face sold as part of dollar roll transactions” (FRBNY-GS Agreement, Exhibit D). (FRBNY 2008d, Exhibit D). The quarterly fee rate was “1/16th of the annual rate of 1.25 basis points” (FRBNY 2008d, Exhibit D).

On August 17, 2009, the FRBNY announced that it would end its contracts with PIMCO and Goldman Sachs Asset Management (FRBNY 2009d; Gagnon et al. 2011, 45). Wellington Management Company would be retained as the sole investment manager and BlackRock would be retained to provide analytical support services (Gagnon et al. 2011, 45). The external investment managers were phased out because FRBNY developed its own “internal analytical and operational expertise” during the program and wished to reduce costs (FRBNY 2009f).

Under the new contract, Wellington retained the same rights and obligations to the FRBNY. From September 15, 2009, until the program’s termination in March 2010, Wellington was paid a flat fee of $1.3 million per month (FRBNY 2009c, Exhibit D; FRBNY 2018). As a provider of analytical services, BlackRock would provide to the custodian and the FRBNY: reports on the portfolio and compliance, client data, and access to analytical tools. BlackRock’s fee was $330,000 a month, paid quarterly (FRBNY 2009a, 13).

Beginning on March 2, 2010, the FRBNY began to use internal staff to partially execute MBS purchases and had assumed full trading responsibilities by the end of the program (FRBNY 2010a).

On December 31, 2008, the FRBNY entered into a contract with JPMorgan Chase (JPMC) whereby JPMC would serve as the LSAP program’s sole custodian (FRBNY 2008b, 1). Under the contract, JPMC was obligated to create accounts in the FRBNY’s name, one to hold assets and one to hold cash (FRBNY 2008b, 2-3). More accounts could be created at FRBNY’s request (FRBNY 2008b, 3). In accordance with instructions from the FRBNY, JPMC would credit or debit the accounts as needed to facilitate and settle trades conducted by FRBNY under the LSAP program (FRBNY 2008b, 3-4). Settlement was to be in accordance with market standards (FRBNY 2008b, 5).  

JPMC was also obligated to present assets to the FRBNY either at maturity or, at FRBNY’s request, before maturity, along with regular information and statements on the account (FRBNY 2008b, 4-5). JPMC was compensated per negotiated fees and reimbursement of any out-of-pocket expenses (legal fees, tax fees, etc.) related to the LSAP program (FRBNY 2008b, 7-8).

The FRBNY set the minimum debt offer at $1 million, with increasing $1 million increments (FRBNY 2010b). The auctions were conducted via FedTrade, which is the Fed’s trading system (FRBNY 2010b). Dealers were permitted to make up to three propositions each auction period, which typically lasted for 30 minutes (FRBNY 2010b). While the Fed aimed to purchase longer-term debt securities, most debt purchases were medium-term securities (with maturities of 2 to 5 years) because few longer-term securities were available (see Figure 3) (Gagnon et al. 2011, 46). The FRBNY created additional bank reserves to finance these purchases (FRBNY 2010b). The FRBNY generally held auctions to purchase GSE debt once a week, which it announced one day prior to the auction (FRBNY 2010b).

The FRBNY purchased agency debt and Treasuries through multi-price reverse auctions and bought MBS directly with the assistance of asset managers (Gagnon et al. 2011, 45). Multi-price reverse auctions of agency debt allowed primary dealer counterparties to indicate the quantities and prices at which they were willing to sell, so the FRBNY purchased the securities at the prices submitted by sellers (Gagnon et al. 2011, 45). This meant that potential investors competed for bids (Gagnon et al. 2011, 45). FRBNY also announced its Treasury and agency debt operations ahead of time by two weeks and one day, respectively (Gagnon et al. 2011, 45). Gagnon et al. also note that the announcements increased participation in the auctions, as they gave dealers time to appraise their inventories (Gagnon et al. 2011, 45).

By March 2009, settlement pressures emerged in the MBS market, so the FOMC used dollar rolls to mitigate temporary imbalances related to the supply and demand of specific coupon categories of MBS (FOMC 2009c; Gagnon et al. 2011, 46). A dollar roll is a repurchase agreement with a settlement date ranging from one calendar week to three calendar months for agency MBS purchases (FRBG 2016). FRBNY conducted dollar rolls with primary dealers by simultaneously purchasing/selling MBS for delivery in one month and selling/purchasing a similar security for delivery in the next month (FRBNY 2010a). After the Fed began using dollar rolls in the first week of March 2009, the forward financing rate for MBS dropped 100 bps compared with the MBS cash repo rate (FOMC 2009c, 6). The Fed began using dollar rolls to purchase high-coupon, seasoned MBS because their liquidity was the most affected (FOMC 2009d, 125). According to the FOMC, dollar roll transactions reduced the costs of managing mortgage inventory and helped lower forward financing spreads (FOMC 2009c, 6).

In June 2010, the FRBNY began using coupon swaps in addition to dollar rolls to settle transactions (FRBNY 2010a). FRBNY conducted coupon swaps with primary dealers by simultaneously purchasing/selling one MBS security and selling/purchasing another MBS security with a different coupon rate (FRBNY 2010a). Coupon swaps permitted FRBNY to exchange assets that were not ready for settlement for assets that were ready. Dollar rolls and coupon swaps did not change the aggregate purchase amounts but allowed the Fed to fine-tune the timing and composition of security settlement (FRBNY 2010a).

One FOMC official suggested that the Fed needed to be transparent about the LSAP program (and quantitative easing more generally) to reassure markets that the Fed had intentionally entered into a new monetary policy regime and that it still was in control of monetary policy (FOMC 2008a, 135-136, 191). The FOMC’s aim for more public transparency emerged in direct response to reports from several prominent media outlets in January 2009, which described the purchase programs as both unprecedented and unclear (FOMC 2009b, 37-38). Given these concerns, the FOMC established a Transparency Committee to consider ways to enhance the transparency of its policies. This workgroup was tasked with assessing the public information on all the FRBNY’s major rescue programs (FOMC 2009b, 37-38; FOMC 2009g, 9-10).

The FRBNY also attempted to make the LSAP programs transparent by announcing auction dates and changes to the programs beforehand. Gagnon et al. asserted that “the timely release of information was provided in order to reduce uncertainty and speculation about operational details. This information may also have helped to prevent erratic trading based on differential access to information or on rumors and misconceptions” (2011, 47). In particular, Gagnon et al. find that the FRBNY’s announcement of the program and subsequent announcements of changes to the program directly lowered longer-term interest rates (2011, 48-52).

In addition to the transparency measures directly related to the programs, the FRBNY continued to assure investors that it could raise short-term interest rates at any time (Gagnon et al. 2011, 42).