An Essay on The Fed and the U.S. Treasury: Lender of Last Resort and Fiscal Policy – Hal S. ScottDownload PDF
An Essay on The Fed and the U.S. Treasury: Lender of Last Resort and Fiscal Policy
Hal S. Scott[*]
This essay explores the evolution of my thinking on risky emergency lending to non-banks. Like the famous 19th century British economist Ricardo, who recognized his views on machinery had undergone considerable change, the same can be said for my views on lender of last resort.[†] The Fed, in the pandemic, engaged in lending with potentially significant credit risk. While it appeared to the public that these were independent Fed programs, in fact the lending to non-banks was legally controlled, and was largely determined, by the Treasury due to their approval power. Treasury control of lending to non-banks is a fiscal decision and should be made by the elected government, not by an independent agency. And it should be the Treasury’s role, as advised by the Fed, to determine when there is significant credit risk. Once the Treasury determines there is not significant credit risk, the Fed should make the lending decision without control or approval of the Treasury, as part of its role as liquidity supplier and lender of last resort. When it comes to banks, the Fed should have more leeway. If the Fed determines loans to banks have significant credit risk, the Fed should be required to get Treasury approval, accompanied by an indemnity protecting the Fed against loss, resulting in the government taking on the credit risk. If the Fed determines that there is not significant credit risk, the Fed should have full control of the lending operation. This approach requires new legislation. In the shorter term, there should be much fuller disclosure of the actual role of the Treasury in the design and operation of emergency lending facilities.
As set forth in this essay, I have come to believe that risky emergency lending to both banks and non-banks should be done by the U.S. Treasury, with programs explicitly designed by and owned by the Treasury. These should not be Federal Reserve (Fed) programs, as they have been identified in the past. The role of the Federal Reserve should only be to advise the Treasury on needed programs and to execute the Treasury’s programs in accord with the direction of the Treasury. This is actually close to the situation today for emergency lending to non-banks, despite appearances to the contrary and the lack of disclosure as to how the Treasury does control such programs.
Lending with significant credit risk is a fiscal decision that should be made by the elected government, the Congress and the Administration, not by an independent agency like the Fed. And it should be the Treasury’s role, as advised by the Fed, in lending to non-banks to determine when there is significant credit risk. When there is no significant credit risk, the Fed should make the lending decision, without the need for control or approval of the Treasury, as part of its traditional role as liquidity supplier and lender of last resort. If there is disagreement as to whether there is significant credit risk the Treasury’s view should prevail. When it comes to lending to banks, the Fed should determine whether there is significant credit risk, and if so obtain Treasury approval and an indemnity against Fed losses. This essay explores the evolution of my thinking.
In my book Connectedness and Contagion, I took the view that the restrictions that Congress imposed on the Federal Reserve in the 2010 Dodd-Frank Act in lending to non-banks under section 13(3) of the Federal Reserve Act were ill-advised. The Fed had acted heroically and effectively in stopping the 2008 contagion but Congress wrongly, in my then view, believed it had overstepped its bounds. Indeed, Senator Dodd introduced a bill in 2009 that would have consolidated the federal regulatory structure for financial institutions, by removing the Federal Reserve’s supervisory authority over state-member banks and limiting its supervisory responsibilities to systemically important nonbank financial firms and holding companies with assets over $50 billion, as a result of its perceived failure to anticipate and prevent the financial crisis. While this broader initiative was dropped, the Dodd-Frank Act curtailed the Fed’s independence in its role as lender of last resort to non-banks. The most important of the congressional restrictions were that the Secretary of the Treasury had to approve the lending and that it had to be part of a broad program with collateral and could not be used for insolvent borrowers. Clearly, the Congress had in mind the loans made during the 2008 crisis to insolvent non-bank financial institutions like AIG. In the pandemic this framework was used to support the real economy, through loans to firms that were not only non-banks but also non-financial institutions.
I believed, at the time of writing my book, that the requirement of the Treasury’s approval was an undesirable infringement on the Fed’s independence as lender of last resort. Indeed, my concern was reinforced when it later came to light, in an article by Laurence Ball, that then Secretary of Treasury Hank Paulson, despite having no statutory authority to prevent Fed lending, pressured the then Fed Chairman Ben Bernanke not to loan to Lehman Brothers because, according to Ball, Paulson did not want to be known as “Mr. Bailout”. In the years since Dodd-Frank, without exception, every Treasury and Fed official with whom I conversed in private, agreed that the Dodd-Frank restrictions on lending to non-banks were undesirable, but that nothing could be done about them. Instead, the Fed concentrated its efforts on resisting further restrictions, publicly saying that the existing ones were acceptable. It was clear, of course, that any government official that questioned this new Dodd-Frank arrangement would be accused of wanting to again bail out Wall Street.
On the other hand, I was torn by the realization that section 13(3) loans could be made to non-creditworthy borrowers, partially due to the lack of Fed information about the credit risk they posed, so that there would be a significant likelihood that the Fed could lose money. This would not bankrupt the Fed since it can create money. Nonetheless, it might tarnish its credibility and reputation if the losses were severe enough. And the Congress and the Treasury would likely, therefore, feel compelled to recapitalize the Fed. Fed losses would also directly impact the taxpayer, since the losses would reduce the amount of profit the Fed annually remits to the Treasury. At their high in 2015, remittances constituted $117 billion, including a one-time capital surplus transfer of more than $19 billion, 3.6% of U.S. general revenue.
To some extent, these concerns with Fed losses were sufficiently alleviated by the Dodd-Frank section 13(3) restrictions that the borrower be solvent and that the loan be collateralized, without the necessity of going the further step of requiring the approval of the Secretary of the Treasury. One could argue that if the Fed adhered to these requirements it should still be independent in making loans to non-banks because there was no or very limited credit risk. However, a solvency determination is more art than science and very difficult to determine particularly in a crisis due to uncertain asset values. The Fed responded to pressure from Senator Warren and others to define solvency. Its 2015 regulation provided that the borrower could not be in bankruptcy or “generally” in default on undisputed debts in the previous ninety days. This, of course, leaves plenty of room for lending to borrowers with very substantial credit risk.
The second potential bulwark against credit loss is the collateral requirement. In 2008, the Fed bought unsecured commercial paper from highly rated issuers without real collateral. While it can be argued that there was little credit risk on these purchases, there obviously was some risk, and I thought that requiring collateral would prevent the Fed from doing what needed to be done in the future. But I was, nonetheless, sympathetic to the idea that collateral should generally be required to protect the Fed from taking on credit risk. But the uncertain value of collateral, in many instances, and therefore its ability to cover credit risk, particularly for loans to risky borrowers, would leave the Fed with significant exposure to credit risk.
3. The Fiscal Policy Concern
The basic problem, I came to appreciate, is the need to define when the Fed as lender of last resort wrongly crosses the line into making fiscal decisions which, in a democracy, rightly belong to elected government officials—the Congress and the Administration, not an independent agency like the Fed. The exercise of fiscal policy by the Fed does not just raise concerns about political accountability; it is also arguably inconsistent with the Constitution, which assigns the “power of the purse” to Congress.
Legitimate concern over the Fed’s involvement as a lender of last resort is not self-evident. After all, there is a broad consensus that the Fed’s independence from the Executive branch, and generally politics, is important for sound monetary policy and the stability of the financial system, and monetary policy has a profound influence on fiscal policy.
At the same time, in a democracy, fiscal authority should lie with political actors that are politically accountable to voters. The ultimate authority of elected officials over fiscal decisions is required by the Constitution, which provides that “No Money shall be drawn from the Treasury, but in Consequence of Appropriations made by Law.” That means that no federal agency or official, including the Fed, should engage in public spending without legislative authorization. Thus, this constitutional principle could operate to restrict Congress’s ability to delegate its fiscal authority to another federal agency, such as the Federal Reserve. Indeed, Congress used to adjudicate individual money claims against the United States, “on the ground that the Appropriations Clause forbade even a delegation of individual adjudicatory functions where payment of funds from the Treasury was involved.” There is no case law on point. The closest one can come is Synar v. United States, where a special three-judge panel of the U.S. District Court for the District of Columbia, in what it acknowledged was obiter dicta, held that the delegation to the Comptroller General of the power to estimate the federal deficit and impose mandatory program-by-program budget cuts, pursuant to the Gramm-Rudman-Hollings Act, did not violate the Appropriations Clause.
So, what is the line between (permitted) monetary policy and (forbidden) fiscal policy when it comes to the Fed’s role as legitimate lender of last resort? Basically, the line should be between relatively riskless liquidity provision versus riskier credit provision. If financial firms with strong balance sheets (i.e. strong credit profiles) need Fed lending simply because the financial system has pulled back private liquidity, even to solvent borrowers, then that is a liquidity problem that is more related to the traditional exercise of lender of last resort, and the Fed should remain independent. However, if the financial firms’ problems are a weak balance sheet (or weak credit profile), which contributes to their need for a Fed loan, then that is not only a liquidity problem but also a credit problem (even if there is also a concurrent pullback of private liquidity in general). In that case, the decision to provide risky credit to firms becomes a fiscal decision, given the risk of loss, and where that risk of loss is significant, the lending decision should be made by the elected government.
It is also the case that the Fed’s exercise of monetary policy can incur significant risk of loss, in the form of interest rate risk through its holdings of government obligations, principally Treasuries. However, interest rate risk is inextricably linked with the conduct of monetary policy, where there is a broad and long consensus of Fed independence. There is an important distinction between this interest rate risk, which applies to the purchase of any fixed-income securities, and credit risk. Credit risk deals with the risk of borrower default, of which there is virtually none in holding U.S. government obligations. Also, the Fed is protected against the full impact of interest rate risk since it does not mark-to-market its portfolio and only incurs gains or losses when it sells portfolio holdings.
This evolution in my thinking led me to join with Charles Calomiris, Glenn Hubbard, Douglas Holtz-Eakin, and the late Allan Meltzer, a group of conservative economists, in writing a 2017 article in the Journal of Financial Economic Policy, entitled “Establishing credible rules for Fed emergency lending.” The heart of the article was our recommendation to “[e]stablish specific, observable criteria, that will be used to determine whether emergency lending by the Federal Reserve becomes fiscal policy that should involve the Treasury, either exclusively or in conjunction with the Fed.” We believed that “loans to insolvent institutions or loans to institutions that have a substantial likelihood of becoming insolvent should be regarded as implicating fiscal policy.” We opined that not all loans to non-banks under section 13(3) would be fiscal, i.e. loans adequately secured by collateral or highly rated, so the necessity for the Treasury approval of all non-bank loans would not be required. If a loan were to be regarded as fiscal, however, it should be approved by the Treasury with indemnification (a guarantee), and possibly backed by a pre-established fund to make the indemnification good. We did not discuss whether the same approach should be applied to lending to banks. But a potential fiscal policy concern is triggered whether loans are made to banks or non-banks. However, as further discussed below, the adequacy of collateral is a much lesser concern for banks than non-banks.
I think a large part of my shift in thinking reflected the fact that the Fed’s willingness to take credit risk and, in effect, to engage in fiscal policy was a lasting problem even if one could justify the actions taken in the global financial crisis. The Fed may have acted in 2008 because the fiscal authorities were caught by surprise, and the unchecked contagion was a huge threat to the country, but this was not the right approach going forward. Moreover, that crisis was largely a liquidity event—lending to insolvent institutions like AIG was an exception rather than the rule.
The approach we recommended is close to the one currently deployed in the United Kingdom, one that emerged out of a confusion of roles between the Bank of England (BoE) and HM Treasury (Treasury) in the 2008 financial crisis, particularly as it concerned rescuing Northern Rock. The United Kingdom divides central bank lending into two parts. First, there is normal lending to banks and other borrowers specified by the BoE, which now include primary dealers, broker-dealers, and central counterparties. It appears that BoE, on its own, can further expand this category. Second, there is emergency lending to banks and non-banks. This is different than the U.S. approach of dividing lending authorities between banks, through the discount window, and non-banks through section 13(3), whatever the circumstances.
In the U.K., emergency lending is governed by a Memorandum of Understanding between BoE and the Treasury, established after the financial crisis, which permits the BoE to make loans to solvent but “at risk” firms, with the approval of the Treasury. Note that this approval is required for making loans to at risk banks as well as non-banks. In the United States, by contrast, the Fed is permitted by statute—through its discount window authority, on its own without the Treasury approval—to make loans to at risk, and even insolvent, banks at a premium rate, if it considers the collateral sufficient (although it rarely does so). Under applicable regulations, however, the Fed can only extend credit to undercapitalized banks if doing so is “consistent with a timely return to a reliance on market funding sources” or “would facilitate the orderly resolution of serious financial difficulties” of the bank. It appears, despite the absence of published guidance on the point, that the U.K. Treasury’s approval would come with the Treasury’s indemnity of BoE losses.
The United Kingdom’s arrangement largely leaves the judgment as to whether firms are at risk to the Bank of England. If the BoE judges they are not, the BoE is free to lend without the approval of the Treasury. That said, depending on the nature of the relationship between the BoE and the Treasury at a particular point in time, the BoE might seek informal approval from the Treasury even if it does not consider that approval strictly necessary from a legal perspective.
Quite unlike our system, the United Kingdom gives the Treasury the further authority, in exceptional circumstances, with parliamentary oversight, to direct the BoE to make risky loans to entities that the BoE does not judge to be solvent, on terms the Treasury dictates. In the event of such direction, the BoE is considered to be acting as the Treasury’s agent. The funds are placed in a special purpose vehicle (SPV) that is segmented from the BoE balance sheet, and the SPV and the BoE are indemnified by the Treasury for losses. This authority has been used in the COVID-19 pandemic. The COVID Corporate Financing Facility (CCFF), established by the Bank of England at the behest of the Treasury, directs the Bank to purchase eligible commercial paper in the primary and secondary market, including from middle-market firms that have not previously issued commercial paper, on terms comparable to those prior to the crisis.
In the U.S. system, a section 13(3) facility for non-banks can only be adopted if five members of the Federal Reserve Board (of the seven total) find that there are “exigent circumstances” justifying the facility. So even if the Treasury wanted the Fed to adopt a facility, it could not be done without the consent of five members of the Fed. If the Fed resists adoption, the Congress would have to authorize the facility, overriding the requirements of section 13(3).
The United States entered the pandemic crisis with the new Dodd-Frank framework, which I believe has been found greatly wanting and further underscores the need for revision.
As previously noted, Dodd-Frank requires the Treasury to approve all lending programs to non-banks. So when on March 17-18, 2020, the Fed announced its first three facilities to deal with the pandemic, modeled after similar facilities used in 2008, to buy highly-rated commercial paper (CPFF), to make loans to primary dealers (PDCF), and to make loans to banks to buy money market fund assets (MMLF) it was required to and did obtain the approval of the Secretary of the Treasury. As in 2008, the Fed used the SPV technique to buy commercial paper to circumvent asset purchase restrictions. 
There was an important difference in the 2020 deployment of the CPFF and MMLF as compared with 2008. Both facilities in 2020 were backed by the Treasury’s Exchange Stabilization Fund (ESF), the CPFF with a $10 billion equity contribution and the MMLF with $10 billion of credit protection. Because the MMLF did not use a special purpose vehicle that could accept an equity contribution—it involved direct loans to banks secured by assets purchased from money market mutual funds—the Treasury provided its backing in the form of credit protection. The Treasury provided no such backing for the equivalent facilities in 2008.
Why did the Treasury provide this backing in 2020? To a large extent, it was required to do so by the Dodd-Frank requirement of collateral. There was no real collateral provided in the CPFF, as discussed above. This was also the case for the MMLF since the Fed only had the assets purchased by the banks from the money market funds as security for their loans to the banks, for example, unsecured commercial paper. The facility further provided there was no recourse back to the banks if the obligors on the purchased paper defaulted. Thus, in 2020, Treasury backing served as the substitute for the collateral required by Dodd-Frank. Backing was not needed for PDCF, however, because the dealers provided real collateral to the Fed to secure their loans and there was recourse back to them in the case of default.
The amount of purchase or lending the Fed could do through the SPVs was, however, not capped, even though the riskiness of the facilities was dependent on their amounts. Such caps would come later for other facilities after enactment of the CARES Act.
For these three facilities, modeled after the 2008 facilities, it may have appeared to the public that the Fed designed the facilities and that the Treasury’s approval of their terms was merely nominal, except for the Treasury’s concern with losses signaled by its investments. So, if these facilities failed or were successful, blame or credit would go to the Fed, not the Treasury. But the reality is that the Treasury, due to its approval power, ultimately called the tune. The Treasury might have decided to defer entirely to the Fed’s design of the facilities, but it maintained the sole power to approve them, and consequently, the power to dictate their terms. I strongly believe that the Treasury did in fact dictate the important facets of the facilities.
The relationship between the Treasury and the Fed, with respect to Fed facilities, evolved significantly with the enactment of the CARES Act on March 27, 2020. Section 4003 of the Act appropriated $454 billion to the Secretary of the Treasury to make loans, investments, or guarantees to the Fed to support Fed lending to eligible businesses, states, or municipalities by purchasing obligations or making loans. The bill specifically ordered the Treasury to endeavor to seek Fed programs for such borrowers. Remarkably, the legislation further stated that if there was any doubt, the provisions of Section 13(3) should apply to any Fed facilities created under the CARES Act.
The Fed announced, with the required Treasury approval, four principal facilities: (1) to purchase asset-backed securities (TALF); (2) to purchase corporate bonds and ETFs in the primary and secondary markets (PMCCF and SMCCF); (3) to purchase bank loans to small- and medium-sized businesses (MSNLF, MSELF, and MSPLF) and nonprofits (NONLF and NOELF); and (4) to purchase state and municipal obligations (MLF). Each of these facilities had a similar structure, an SPV that was capitalized with a Treasury equity investment, and a cap on the amount it could lend. So, for example, the two corporate credit facilities operated through the same SPV, which was capitalized with a Treasury investment of up to $75 billion (though the Treasury actually only made a $37.5 billion equity contribution) and had a $750 billion cap on combined primary and secondary market purchases. The five Main Street facilities, for small and mid-size businesses and nonprofits, operated through a joint SPV capitalized with a Treasury investment of up to $75 billion (though the Treasury actually only made a $37.5 billion equity contribution to the Main Street SPV) and had a cap of $600 billion on aggregate loan purchases. The cap/equity investment ratio represents the leverage of the facility: ten times for the corporate credit facilities, and eight times for the Main Street facilities.
In addition, the facilities had detailed requirements with respect to the qualifications and riskiness of the assets purchased and the eligible borrowers. These requirements were not set by Congress; they were set by either the Treasury or the Fed, or some combination of the two. Thus, for example, the corporate credit facilities specified the minimum credit ratings of the issuers: issuers had to have been rated at least BBB-/Baa3, a relatively low risk rating, as of March 22, 2020, and an issuer that was subsequently downgraded had to be rated at least BB-/Ba3 as of the date of purchase. Detailed requirements were provided in the term sheets for Main Street’s three for-profit business facilities, which were revised several times, and for the last time on October 30, 2020, when the minimum loan size for new loans was lowered from $250,000 to $100,000.
The Main Street Expanded Loan Facility (MSELF), which permitted for-profit bank customers to increase the size of existing loans, can serve as an example of how the Main Street facilities worked. MSELF allowed a business borrower (1) with up to 15,000 employees or revenue of $5 billion or less in 2019 and (2) that had an existing loan from the bank that received an internal risk rating equivalent to a “pass” (the highest rating) in the supervisory rating system as of the end of 2019 (or upon origination or purchase, if the loan was originated or purchased in 2020), to increase the size of that loan by borrowing a minimum of $10 million and a maximum of the lesser of (i) $300 million or (ii) an amount that, together with the borrower’s existing debt, did not exceed six times the borrower’s 2019 EBITDA. The Fed’s Main Street SPV then bought 95% of the new tranche of the loan.
The interest rate on the new tranche of the loan was adjustable LIBOR plus 300 basis points and had a maturity of five years. The new loan might also have included up to 150 basis points of origination and transaction fees. The borrower could not use the proceeds of the new tranche of the loan to repay or reduce existing debt but could make mandatory principal and interest payments. The lender was specifically required to do a credit assessment of the borrower’s financial condition, although the lender would do so anyway since it would be on the hook for 5% of the loan. The borrower had to certify both that it would make commercially reasonable efforts to keep its employees and that it could pass a solvency test, meaning it had the ability (with the loan) to meet its financial obligations and did not anticipate going into bankruptcy in the next ninety days.
A major driving force behind the structure and terms of all the CARES Act facilities was the Treasury’s desire to minimize losses on its investments that were funded by Congress. To begin with, this can be seen clearly in the Treasury’s desire to not fully use the entire $454 billion Congress appropriated for the CARES Act facilities. Second, all the investments made by the Treasury, after CARES, were structured through SPVs with Treasury investment and a cap on lending. The leverage of each facility was set based on its perceived riskiness. Thus, as noted above, the corporate facilities were leveraged at ten times while the riskier small business facilities were leveraged at eight times. The limit on leverage permitted by the Treasury protected its investment.
When the CARES Act facilities stopped extending new credit on January 8, 2021, the Treasury had only invested $102.5 billion of the CARES Act $454 billion appropriation (although it had committed $195 billion), together with $11.5 billion from the ESF that it had invested before passage of the CARES Act to back the Commercial Paper Funding Facility ($10 billion) and the Money Market Mutual Fund Liquidity Facility ($1.5 billion). So, the Treasury had $351.5 billion of CARES Act funding that was never put at risk. Secondly, the actual amount of lending under the CARES Act facilities was very small. While combined lending under all the facilities was capped at about $2 trillion, actual loans as of January 8, 2021, when the last of the facilities stopped purchasing loans, were about $40.5 billion, approximately 2% of the cap. Indeed, the Congressional Budget Office (CBO) in its April 16, 2020, report on the CARES Act estimated that the income from the Fed facilities (in the form of interest payments and fees) would roughly offset any losses. The facilities were clearly designed to avoid credit risk.
In the Wall Street Journal op-ed titled “Main Street Needs More Fed Help,” Glenn Hubbard and I criticized an earlier version of the Main Street facilities, in which the credit protections for the Treasury were weaker, for prioritizing credit risk limits over getting funds to needy but risky borrowers. The desire to avoid loss became even stronger when the Main Street facilities were revised. The term sheets for the original Main Street facilities (the MSNLF and MSELF), for example, originally contained no “pass” rating requirement for existing loans and neither directed the bank to make creditworthiness judgements nor included any borrower solvency requirements. The Wall Street Journal’s Editorial Board, in its own editorial on the initial April 30 revisions entitled, “The Main Street Fakeout,” criticized the Fed and the Treasury for doubling down on credit risk limits. Glenn Hubbard and I subsequently published two additional op-eds in the Wall Street Journal, arguing that the Main Street facilities’ design would hamper their ability to save small and mid-size businesses. Specifically, we noted that the terms of the for-profit Main Street facilities were too onerous for borrowers and insufficiently attractive to lenders.
As already discussed, Treasury backing was in part necessitated by the Dodd-Frank collateral requirement. The actual Section 13(3) requirement is:[T]he Board shall establish . . . policies and procedures governing emergency lending . . . . Such policies and procedures shall be designed to ensure that . . . the security for emergency loans is sufficient to protect taxpayers from losses . . . . The policies and procedures established by the Board shall require that a Federal reserve bank assign, consistent with sound risk management practices and to ensure protection for the taxpayer, a lendable value to all collateral for a loan executed by a Federal reserve bank under this paragraph in determining whether the loan is secured satisfactorily for purposes of this paragraph.
Since many small business borrowers do not have any collateral, backing by the Treasury could serve as a substitute. The Congress recognized this problem in the CARES Act by giving the Treasury $454 billion to back Fed lending. The Treasury could use all the appropriated funds as collateral to back Fed loans, and was given leeway to judge the adequacy of collateral in relation to the permissible leverage, both matters of fiscal policy. While Congress gave the funds to the Treasury to back Fed loans, it did not actually order that all of the appropriated money be so used—this decision was left to the Treasury.
Section 13(3) also requires borrower solvency:
The Board shall establish procedures to prohibit borrowing from programs and facilities by borrowers that are insolvent. Such procedures may include a certification from the chief executive officer (or other authorized officer) of the borrower, at the time the borrower initially borrows under the program or facility (with a duty by the borrower to update the certification if the information in the certification materially changes), that the borrower is not insolvent. A borrower shall be considered insolvent for purposes of this subparagraph, if the borrower is in bankruptcy, resolution under title II of the Dodd-Frank Wall Street Reform and Consumer Protection Act, or any other Federal or State insolvency proceeding.
As previously discussed, the Fed issued a regulation in 2015 setting forth its solvency test: that the borrower is neither (A) the subject of any bankruptcy or insolvency proceeding nor (B) “generally not paying its undisputed debts as they become due during the 90 days preceding the date of borrowing under the program or facility.” The Fed regulation also allows the borrower to self-certify based on his or her reasonable belief that the solvency requirement as set forth in the regulation has been satisfied.
The revised Main Street facilities made this certification tougher than required by the Fed’s regulation. All three facilities required the borrower to certify that, after giving effect to the loan, the borrower reasonably believed “it ha[d] the ability to meet its financial obligations for at least the next 90 days and [did] not expect to file for bankruptcy during that time period.” The Fed’s pre-existing regulation looked to the past and only required the borrower to “generally” meet its obligations. Further, the borrower under the Fed’s regulation actually knows whether or not it is in bankruptcy, whereas under Main Street it was required to certify as to a more uncertain future. This appears to be the work of the Treasury. Why would the Fed otherwise insist –- indeed, can it even do so lawfully—on a tougher and different standard than set forth in its own regulation, which by its terms applies to any Section 13(3) lending?
Secretary Mnuchin revealed his desire to avoid losses when he told the Wall Street Journal and other reporters in late April, in response to a question about my first op-ed with Glenn Hubbard: “If Congress wanted me to lose all the money, that money would have been designed as subsidies and grants as opposed to credit support.” He further stated: “We’re looking at it in a base case scenario that we recover our money.” However, in response to criticism from lawmakers, Secretary Mnuchin stressed that that Treasury did “expect to take losses” on lending through the Main Street facility.
On November 19, 2020 Secretary Mnuchin sent a letter to Chairman of the Federal Reserve Jerome Powell ordering the Fed to close down, at the end of the year, all the facilities that had received Treasury backing with funds appropriated under the CARES Act. The fact that the Treasury Secretary resorted to unilateral action through a letter, ordering the Fed to comply, was an extraordinary and historic economic confrontation, given the Fed’s repeated statements that these facilities should remain in place past the end of the year. If the Fed had agreed with Treasury’s demand letter, this action would have been accomplished through a joint Treasury and Fed statement. Instead, the Fed immediately objected to the closure of these facilities.
The Treasury Secretary was given the authority, under the 2010 Dodd-Frank amendments to Section 13(3) of the Federal Reserve Act, to approve facilities to support the non-bank financial sector, and the CARES Act explicitly reinforced this authority. The power to approve encompasses the power to disapprove, so the Secretary was within his rights to terminate any and all of the facilities. If there was any doubt as to who called the shots on these facilities, the order to close them down removed it. However, for the record, the argument in the Secretary’s letter that he was compelled to close the facilities down, based on his being “personally involved in drafting the relevant part of the legislation,” was untrue. A plain reading of the CARES Act fails to support that claim. The CARES Act only set a deadline of the end of this year for the Secretary to make new investments in the Fed facilities. His decision to close the facilities was one of public policy, not one compelled by law.
The Secretary’s letter went further than just terminating new lending or purchases under the CARES Act facilities. It also requested that the Fed return $429 billion in “unused” Treasury funds. While the Treasury deposited at the Fed the entire CARES Act appropriation of $454 billion that might have been used to back Fed facilities, it had only actually invested $102.5 billion in the CARES Act facilities. And the Fed, pursuant to these facilities, had at the time less than $25 billion in loans outstanding. Thus, the Secretary, in requesting the return of only $429 billion, allowed the Fed to keep backing for its then-existing loans on a 1:1 basis. On the next day, November 20, 2020, the Fed acceded to the Treasury’s requests—noting that the Fed would work with the Treasury to return the unused portion of the CARES Act funds in connection with the facilities’ year-end termination—although the authority of the Secretary to order the return of his investments was unclear under the CARES Act.
Although the Secretary refused to renew the CARES Act facilities, the CARES Act itself left open the possibility that a new Treasury secretary could restart them, albeit without the funds that had already been returned by the Fed. The coronavirus relief legislation passed in late December, however, prevented the CARES Act facilities from making any new loans or purchasing new assets after December 31, 2020 (except for the Main Street facilities, which were allowed to continue to purchase certain loan participations until January 8, 2021). The new legislation also rescinded the CARES Act appropriation to the ESF, save for $42.5 billion that might be needed to absorb losses on the Treasury’s $2 billion in direct loans and the $40.5 billion in commitments of the CARES Act facilities.
Finally, the relief legislation prevents the Treasury from using the ESF to back any new Fed program or facility, except for the TALF, “that is the same as any such program or facility in which the [Treasury] Secretary made an investment pursuant to” the CARES Act. Concerned that the incoming Biden administration might reopen the CARES Act facilities, Senator Toomey proposed language that would not only have shut down the CARES Act facilities but also prevent the Fed from restarting those or any “similar” program or facility. Ultimately, the legislation that passed only prevents Treasury from using the ESF to back a facility that is the “same” as one of the CARES Act facilities. That provision therefore leaves open the possibility that Treasury could use the ESF’s remaining “core funds”—which stood at approximately $238 billion as of September 30, 2021—to back Fed lending facilities that are sufficiently different from the CARES Act facilities.
Currently, responsibility for the success and failure of Fed facilities designed and approved by the Treasury is murky. If these programs had failed and we had been plunged into a Depression (fortunately this did not occur), who would have been held responsible? And who will be held responsible for future programs? I fear that given that these are “Fed” programs, the Fed could unfairly take the blame for programs that are actually approved and dictated by the Treasury, with the result that its future independence, even for traditionally independent functions like monetary policy, could be threatened.
The lack of clarity about responsibility for the success and failure of the emergency lending facilities stems from the ambiguity over who was actually responsible for their design. On the one hand, the Fed may have actually agreed with the program design, or even suggested it, and been willing to step into fiscal territory as it had in the past. Vice-Chair Quarles testified before the Senate Banking Committee:
We are required by law that we structure these facilities so that they are loans to entities that we expect to be repaid and that the various measures and metrics that we have included in the Main Street facility are designed to try to balance as broad a reach as we can while maintaining fidelity to the statutory requirements.
One might indeed argue that if these Fed facilities were, in reality, those of the Treasury, and the Fed was fearful that that they would not work or were not consistent with the intent of Congress to take more risk, then the Fed could just refuse to adopt them, and that their failure to do so indicates their support. Reports indicate that the Fed sought to soften the terms of the facilities and take more credit risk. While some of the terms were, in fact, softened in the various revisions of the facilities, they were not softened enough to make them attractive to borrowers.
Expecting the Fed to say either do it my way or we will not adopt the program is unrealistic—one can just imagine the blowback by the Treasury and the President if the Fed resisted implementing the Administration’s program in the midst of a crisis. It is one thing to resist direction from the government on monetary policy, where the Fed is independent, it is another to resist direction where it is no longer independent due to Dodd Frank. Indeed, the shutdown order by the Treasury of the CARES Act facilities indicates the Treasury was in charge. Moreover, Chairman Powell’s accession to the Treasury’s demand for the return of its investments, though not required by law, further indicated the Fed’s subservience.
I have expressed my disagreement with the Secretary’s policy of not taking credit risk, and the Fed’s accession to this position, if they did in fact do so. I believe the Congress authorized the Secretary to risk all of the $454 billion to save the economy from even greater loss. But whether the Treasury was following the intent of Congress was a matter to be resolved between those two government branches. I want to make clear, however, that control of these programs, as between the Treasury and the Federal Reserve, should be the Treasury’s, given the inevitable losses from making many of the authorized loans, particularly to small businesses under the Main Street facility. Whatever the line between normal lender of last resort and fiscal policy, lending to small businesses, especially in a crisis involving their widespread collapse, is on the fiscal side.
This brings me to the punch line of this piece. Since all of these non-bank facilities were probably designed by the Treasury—and in any event could be solely designed due to the Treasury’s approval power—in the future the Treasury, and not the Fed, should be wholly responsible for them. And this should be made clear in legislation. These fiscal programs should be labeled as Treasury and not Federal Reserve facilities, and responsibility for their success or failure should lay squarely with the Treasury and the Congress that authorized them. The Fed should only be responsible for informing the Treasury of market conditions and the need for action, but then only serve as an agent of the programs devised by the Treasury.
Under the existing section 13(3) structure, where the Fed asks approval for a program, it looks like the Fed’s program even though the right of approval gives the Treasury the actual power to ultimately dictate the program’s terms. My proposed framework would dispense with the misleading requirement of the Fed to request approval for a program.
Instead, the Treasury should be responsible for adopting any lending program, whether for banks or non-banks, whenever there is a significant risk of loss (emergency lending) if such program is authorized by Congress and deemed necessary by the Treasury. Something very close to this procedure, as noted above, already exists in the United Kingdom.
A threshold question that arises in connection with the use of any lending facility, whether established by the Fed or the Treasury, is the standard for when it can be used. Under current law, the Fed can use the discount window to support banks without any triggering criteria. In contrast, the Fed (upon the vote of at least five of the seven board members) is only permitted to establish a section 13(3) facility for non-banks in “unusual and exigent circumstances.” The term “unusual and exigent circumstances” is not defined by statute or regulation and there is virtually no legal authority interpreting what conditions might meet that standard. But the history of the Fed’s invocation of section 13(3) indicates that it was never interpreted to require an emergency of potential systemic risk to the financial system.
Section 13(3) was enacted on July 21, 1932 to provide funding to businesses that were unable to get bank loans. Compared to other provisions of section 13(3), there is no legislative history to interpret the “unusual and exigent circumstances” language—the language was included in the original legislative proposal to broaden the Fed’s lending powers and remained unchanged afterward.
Once President Hoover and the Fed turned to implementing section 13(3), they began to discuss how the “unusual and exigent circumstances” requirement should be interpreted. Hoover, for example, argued that the condition was satisfied because of the number of borrowers who reported being refused for loans by banks, noting that “the unwillingness of the banks to take advantage of the facilities provided by the government” (referring to the discount window) gave rise to the kind of situation envisioned by section 13(3). The Fed Board appears to have taken as a given that “unusual and exigent circumstances” existed since the meeting minutes focus on the type of lending authorized and eligible counterparties, rather than whether the lending was authorized in the first place.
The Fed periodically authorized section 13(3) lending to nonmember banks (which at the time were not eligible for discount window loans) between 1936 and 2008, although no emergency existed. In the summer of 1966, the Board authorized Reserve Banks to lend to nonmember banks because of “the possibility that during the period ahead some nonmember depositary-type institutions … might be subjected to unusual withdrawals of funds,” a far cry from an emergency. The Board authorized lending to nonmember banks again in December 1969, on the ground that “the sharp further advance in market yields …, unusually large net savings withdrawals at depositary institutions …, and preliminary reports of rather poor savings experience in some areas … had all created some concern about the possibility of substantially enlarged savings attrition at such institutions.” In 1980, section 13(3) was activated again, but not used, to grant a loan to a Michigan nonmember bank to pay for cash letters presented to it. The Fed did not authorize section 13(3) lending again until the 2008 financial crisis. Clearly, the Fed has not seen the “unusual or exigent circumstances” as much of a limitation.
Under my approach, the appropriate standard governing when lending facilities could be established for either banks or non-banks would depend on the type of lending. For liquidity facilities with limited credit risk, which are designed and implemented solely by the Fed, the standard should be relatively forgiving. Like the Fed’s discount window authority, which is not subject to any threshold for market conditions, the Fed’s ability to provide liquidity support for non-banks should be very broad.
On the other hand, lending facilities that pose significant credit risk—and are under the control of Treasury, whether for banks or non-banks—should be subject to a higher threshold. Ultimately, that threshold would have to be set by Congress. However, that threshold should not require a credit crisis that threatens the stability of the U.S. financial system. Rather, Treasury should be empowered to act (and direct the Fed to act as its agent) in the event that economic conditions disruptions in lending markets severely impair access to credit by borrowers in general or by borrowers in a broad-based sector of the economy, even if short of an emergency. That standard is consistent with the Fed’s historical interpretation of the “unusual and exigent circumstances” threshold.
An important question concerns when a facility does or does not involve a significant risk of loss and thus should or should not be the responsibility of the Treasury. This is a difficult line to draw.
As a general matter, the first criterion should be whether the lending takes place in an economic turndown or crisis where there are serious questions about the ability of borrowers to repay. Possible objective criteria can be used to determine whether this is the case.
A second consideration is whether there is adequate security, while recognizing that the existence of collateral does not by itself necessarily ensure lack of credit risk. In all of the pandemic crisis facilities, only one—the Primary Dealer Credit Facility—met the collateral requirement. This was due to the fact that these dealers were able to post collateral and because there was recourse back against them in the event the collateral was insufficient to cover Fed losses. The lack of credit risk in this facility was also underscored by the fact that one must meet high standards to qualify as a primary dealer. The fact that there was no Treasury backing of this facility indicates that the Treasury believed the dealer collateral was adequate to protect against credit risk. Therefore, although this facility had been rolled out as part of the crisis response, this kind of facility could be part of the Fed’s independent responsibility in the future due to the lack of significant credit risk.
All the other pandemic facilities—the corporate credit facilities, the Main Street Facilities, CPFF, TALF, and MLF—lacked real collateral, since the loans to the SPV were merely backed by the assets purchased by the SPV and there was only recourse back to the SPV (which held the purchased assets) and no recourse back to the actual sellers of the assets. In the case of the MMLF, where no SPV was involved, and the Fed made loans directly to banks to purchase money market assets, including unsecured commercial paper, there was also effectively no collateral (since the underlying assets that served as collateral were unsecured) and no recourse back to the banks in the event the banks defaulted on these loans. Therefore, those loans to banks also lack collateral.
i. Emergency Lending to Banks
Currently, emergency lending to banks is governed not by Section 13(3) of the Federal Reserve Act, but by Section 10B, the so-called “discount window.” Under Section 10B, emergency lending to banks must be secured by collateral “to the satisfaction” of the Fed. Interestingly this same discretionary language was the standard in Section 13(3) for non-banks before Congress amended that provision in Dodd-Frank to require collateral that is sufficient to protect taxpayers from losses and is assigned a lendable value by the Fed.
I think the right approach to dealing with banks is twofold: (1) amend section 10B to specifically require collateral along the lines of the current section 13(3) requirement; and (2) as in the United Kingdom, require the Fed to seek Treasury approval of any loans to banks which the Fed believes involve significant credit risk, despite the collateral. Where Treasury approves such loans, it would indemnify the Fed for any losses. To ensure that Treasury exercises this authority responsibility, any decision by Treasury approving loans with such credit risk should be reported to Congress.
Within this framework, with discount window lending to banks, a facility where there is normally collateral and rarely significant credit risk, the Fed would remain independent and normally determine the terms of such loans, except where it believes individual loans to pose significant credit risk despite the collateral. In that case, the Fed would need Treasury approval. Unlike the case for non-banks, the Fed would have the final say as to whether the loans are risky, and thus would itself determine whether it needed to get Treasury approval. The same approach should apply to PDCF and other non-bank facilities that might be established in the future with collateral standards that are equivalent to those used by the Fed for discount window lending. This would require a further change of section 13(3) because all loans to non-banks, whether or not collateralized, are now currently controlled by the Treasury.
As with bank emergency lending, I think that non-bank emergency lending under Section 13(3) of the Federal Reserve Act should also be bifurcated into two categories, but different categories than apply to banks: (i) if the Treasury determines that non-bank emergency lending would pose significant credit risk, then the Treasury would determine whether to lend and the terms of such loans; and (ii) if the Treasury determines that non-bank emergency lending would not pose significant credit risk, then the Fed would have the sole authority to determine whether to lend and the terms of such lending.
Where a facility for non-banks does not provide for real collateral then the entire establishment and design of the facility should lie with the fiscal authorities, the Congress and the Treasury, and should be identified as a Treasury facility.
I am not suggesting, however, that the Treasury necessarily design and take responsibility for all facilities, without adequate collateral, in which it determines there could be credit risk. It might well decide that the credit risk is tolerable and let the Fed design and establish its own facility. For example, take the corporate credit facilities. The Treasury might well decide that that facility posed little credit risk even in the absence of real collateral since very few actual loans would have to be made (as was the case in the pandemic). So, the Treasury would determine that the Fed could establish and design this facility, subject only to a general lending cap and some Treasury backing. Or take the commercial paper funding facility, again a facility without real collateral. The Treasury could decide that since the facility was going to be used to buy AAA rated paper, there would be no significant credit risk, and thus again let the Fed establish and design the facility, subject only to a lending cap and backing.
However, where the Treasury believes that the facility poses significant credit risk for non-banks, the Treasury would establish its own facility, identified as belonging to, and designed and controlled by, the Treasury.
iii. Emergency Lending to Non-Financial Companies
If emergency lending authority is to be based on the presence of significant credit risk, then this has important implications for Fed lending to non-financial companies (as under the Main Street Lending facilities) or Fed purchasing the securities of non-financial companies (as under the CPFF, the Secondary Market Corporate Credit Facility, and the Municipal Liquidity Facility).
The assets of non-financial companies tend to be less liquid than those of non-financial companies; they are also less diversified than financial companies. As a result, they are more likely than financial companies to pose credit, rather than liquidity risk.
Consequently, I believe that lending to non-financial companies should be regarded per se as involving significant credit risk, and I recommend an outright prohibition on Fed lending to non-financial companies, which should be the sole purview of the Treasury.
Under my approach, the Treasury’s own identified facilities could be deployed through the Fed, permitting such leverage as the Treasury deems compatible with its willingness to take risk, as it has done with the pandemic facilities (however much I might disagree with its lack of risk appetite). But the Treasury would fully guarantee the Fed against all losses, either during the operation or the unwinding of its programs. In fiscal matters, all the risk should be with the elected government.
Any funds or guarantees provided by the Treasury would have to be authorized by Congress. The Congress would, therefore, be able to limit the appropriation or guarantee however it sees fit. From a budgetary perspective, there is little difference between either approach. Under the Federal Credit Reform Act of 1990, both direct loans and loan guarantees by the federal government must be accounted for in the budget on an accrual, net-present-value basis. That is how the $454 billion appropriation to the ESF under the CARES Act was scored; the CBO estimated that the budgetary impact of that appropriation would be zero, since income from the facilities would basically offset any losses. The treatment of a loan guarantee that, in the view of the CBO, exposed Treasury to an equivalent level of credit risk would receive the same treatment.
Whatever course of action Congress takes, it should be done in advance, as a crisis can quickly materialize requiring immediate action. This was more the case in 2008, with the widespread contagious run following the Lehman bankruptcy, than it was in the gradually evolving economic impact of the pandemic. If action is taken in advance, it is far more likely that it would be in the form of a guarantee, likely with a cap, rather than an appropriation that might never be used.
To the extent Congress caps the Treasury’s ability to guarantee Fed losses, for example, at a specific dollar figure, then the Treasury must be careful to limit guaranteed Fed lending to its best conservative estimate of losses that would be within the limits set by Congress. Nonetheless, if the losses did exceed the guarantee limit, there would likely still be an implicit guarantee, as there currently is under the deposit insurance system, to cover excess losses, with the clear expectation that Congress would cover the overrun.
Whether or not Fed leverage is used, the Fed will likely be needed to execute Treasury programs, due to its connection to the bank distribution channels. The Fed would only be responsible for its failure to execute in accord with Treasury directions. The Fed should rightly insist that it be operationally equipped to run any emergency lending program it is being required to administer. This, of course, was a problem for the Small Business Administration in operating the Paycheck Protection Program authorized by the CARES Act.
It is essential that the Treasury designs its crisis game plan so that it can be quickly rolled out when crisis strikes. The roll out time under a system of sole Treasury responsibility could actually be reduced compared to what it is today. In the Pandemic, unlike 2008, the Fed had to get Treasury approval, which probably caused some delay as the Fed formulated the programs which were reviewed by Treasury, with several iterations. The current system where the Fed requires Treasury approval slows down the response time.
The Treasury can act just as quickly as the Fed to implement a program, if so authorized by the Congress, as long as it prepares its programs in advance. Of course, the Treasury should fully utilize the expertise of the Fed in implementing the program, but again the responsibility of the program should be with the Treasury.
Under existing law, the Fed is only authorized to buy U.S. Treasury securities or government guaranteed debt, like mortgage-backed securities issued by government-sponsored enterprises. During the global financial crisis, the Fed used its Section 13(3) authority to circumvent these restrictions on direct purchases by lending to Fed-created SPVs that in turn purchased assets, like commercial paper, that the Fed could not purchase directly. The Fed repurposed this SPV structure for its pandemic lending facilities, albeit then explicitly authorized by the CARES Act appropriation which has now expired.
Pursuant to the proposed framework, Congress should clarify that the Fed’s emergency lending authority, when acting independently or as an agent of the Treasury, is not limited to lending but also extends to the purchase of debt or other interests directly from issuers or on the secondary market. In other words, the Fed’s ability to provide emergency support should depend on the substance of that support, not its form.
A significant motivation underlying my proposals is that responsibility for the structure and terms of emergency lending programs be made clear. Responsibility for the design choices made with respect to the CARES Act facilities was unclear, which made ultimate responsibility for their effectiveness hard to determine.
My approach to this problem is structural—define in law the relative responsibility of the Treasury and Fed. Where the Fed lends to banks or non-banks without significant credit risk, then the Fed is entirely responsible. Where the Fed asks for Treasury to approve risky loans to banks (and such approval is given), or the Treasury directs the Fed to make risky loans to non-banks, then the Treasury is responsible.
Pending the structural reforms I advocate in this piece, there should be full disclosure under the existing framework as to whom, as between the Treasury and the Fed, is responsible for the design of the facilities. How would such disclosure be made? This would be difficult to do during the actual crisis, but it could be done after the crisis has subsided. At that time, the Treasury and Fed could be required to issue a joint statement, laying out responsibility for major decisions about the structure and terms of the facility.
It is the case already that all Section 13(3) loans have to be disclosed within a week to the chairs of the Senate Banking and House Financial Services Committee. Putting aside whether the prospect of such prompt disclosure deters timely and needed borrowing, due to stigma concerns, disclosure of the borrowers does not address the concern with responsibility as between the Fed and Treasury for permitting such loans.
Apart from being blamed for the failure of a program that it did not design, there is a further potential risk to the Fed when it operates programs in which it exercises its discretion to choose specific borrowers or loans, rather than setting generally applicable terms and allowing borrowers to access liquidity or credit on those terms: the risk of picking winners and losers. Although the Fed will always exercise some discretion when determining program eligibility, the more it does so, the greater the concern that it has favored some prospective borrowers over others.
This was, to some extent, a problem with all of the pandemic facilities. Most of those facilities involved the Fed setting generally applicable terms for all willing borrowers. Thus, each of the PDCF, MMLF and TALF involved the Fed (or SPVs controlled by the Fed) lending to private entities—banks, primary dealers, and other financial institutions—that use the proceeds of those loans to purchase eligible assets. In the case of the CPFF, PMCCF, and MMLF, only eligible issuers could sell commercial paper or newly issued corporate or municipal debt to the Fed on general terms set by the Fed, rather than based on the Fed’s exercise of discretion. Still, whenever the Fed sets generally applicable terms, it must decide which types of borrowers or assets are eligible for participation and which are not. And wherever the Fed decides to draw the line, it can create the perception that the Fed is helping or penalizing certain sectors or industries.
Another potential issue of picking winners and losers arose in connection with the operation of the Fed’s secondary market corporate credit facility (the SMCCF), since the Fed exercised its discretion to determine which ETFs and bonds the SMCCF would buy and at what price. The Fed attempted to minimize this problem by initially only purchasing highly diversified ETFs, but it still had to decide which ones to buy. The Fed tried later to further eliminate this problem by constructing its own broad bond index composed of all secondary market issues that met the SMCCF’s eligibility criteria. But the Fed still set the eligibility requirements.
A similar concern might arise if demand for an emergency credit program exceeds supply and the Fed were forced to choose between different eligible participants—for example, if the demand exceeds a cap set by the Fed (under the current framework) or Treasury (under my proposed framework). Because of the limited take-up of the pandemic facilities, this particular concern did not actually arise. For instance, the CPFF, the PDCF and MMLF were not capped, and the amount lent under all three facilities has been relatively small, with the aggregate amount outstanding peaking at a total of about $86 billion in early April (as of year-end 2020, the three facilities had less than $4.7 billion still outstanding). The readiness of the Fed to supply funds through the CPFF and MMLF steadied the markets, so that funding could be obtained in private markets and actual large-scale Fed lending was unnecessary. This was also the case with the corporate credit facilities and TALF—their announcement steadied the market even before they became operational. Because the corporate credit facilities and TALF stabilized markets without large-scale purchases, and take-up of the Main Street and municipal facilities has been so limited, the capped CARES Act facilities have engaged in even less lending than the uncapped liquidity facilities. When they were shut down in January 2021, the CARES Act facilities had purchased only $40.5 billion worth of loans (the maximum amount outstanding over time), or approximately 2% of the $1.95 trillion cap established by the Fed and Treasury.
Some might argue that the Fed did not object strongly to the Treasury’s unwillingness to make the Main Street facilities more attractive because it believed that high credit standards would dampen demand and therefore protect it against the need to pick winners and losers. I do not think that is likely. First, the Fed has a lot more to lose in the future by failing to rescue the economy than it does from the winners-and-losers problem. Second, more importantly, the Fed does not really have to exercise discretion with respect to the Main Street facilities; it can just buy loans on a first-come, first-serve basis. The Paycheck Protection Program of the Small Business Administration (SBA) was vastly oversubscribed, but the SBA operated on first-come, first-serve basis and was not, like the banks making the loans, accused of favoritism. Obviously, SBA had operational issues, in terms of bad computer systems and poor guidance, but that is another matter.
Any Treasury program that delegates user eligibility to the Fed should require the Fed to make its terms clear and transparent. The Fed should then seek to operate the program to reduce the need to pick winners and losers by delegating authority to do so to a third party to pick qualified individual borrowers, or choose eligible borrowers on a strict first-come, first-serve basis.
As previously discussed, unlike the U.K. system, the Treasury cannot order the Fed to open a facility as lender of last resort. Congress should empower the Treasury to do so. As long as the Treasury has the ability to protect the Fed from loss—as it could do by use of the Exchange Stabilization Fund—the Fed should accede to the Treasury’s wishes. Such order would be public, and the Treasury would be held responsible for the success or failure of the facility.
One final point, albeit relatively minor in the big picture of this essay. The Fed is the most important bank regulator. In order to incentivize banks to use some of its facilities, the MMLF and its financing of bank loans to small businesses under the PPP Liquidity Facility, it specified that banks need not maintain risk-weighted capital or leverage capital to support assets the banks acquire pursuant to these programs. In addition, it exempted these assets from normal liquidity requirements under the Liquidity Coverage Ratio.
Theoretically, relaxation of these requirements increases exposure of the Fed to losses from riskier banks. While this may be unlikely in present circumstances, given what appears to be the strong capital and liquidity positions of the banks, together with the relatively small share of bank assets generated by these programs, there is the conceptual concern that such actions do increase Fed credit risk. However, these are actions clearly within the Fed’s general regulatory responsibility. While one could argue that the Treasury should call the shots on regulatory relief to accompany its own programs, I would leave regulatory decisions where they normally lie, with the Fed.
So, going forward when we have hopefully moved on from the pandemic, what would I recommend? In general, the Federal Reserve should have wide discretion to carry out its traditional function as the lender of last resort, providing liquidity to the financial system through lending that carries liquidity, but not credit risk. The Treasury, by contrast, should bear responsibility for designing and owning fiscal programs, including lending programs that carry significant credit risk. To the extent that the Fed is involved in these programs, it should be as an agent of Treasury.
The following summarizes my structural recommendations for different categories of emergency lending.
- The threshold governing when lending facilities could be established would depend on the type of lending. In general, the Fed’s ability to provide liquidity support for banks and non-banks should be very broad. By contrast, lending facilities to banks and non-banks that pose significant credit risk should be subject to a higher threshold. However, that threshold should not require a credit crisis that threatens the stability of the U.S. financial system.
- When determining whether loans pose significant credit risk, the relevant authorities should consider whether loans are adequately collateralized and whether the lending takes place in an economic turndown or crisis where there are serious questions about the ability of borrowers to repay.
- For emergency lending to banks under Section 10B of the Federal Reserve Act: (i) if the Fed determines that lending does not pose significant credit risk, then the Fed should remain independent in designing and implementing the lending program; and (ii) if the Fed determines that lending does pose significant credit risk despite the collateral provided, then the lending program should require Treasury approval together with indemnification for any losses incurred.
- For emergency lending to non-banks under Section 13(3) of the Federal Reserve Act: (i) if the Treasury determines that lending would not pose significant credit risk, then the Fed would have the sole authority to determine whether to lend and the terms of such lending; and (ii) if the Treasury determines that lending would pose significant credit risk, then the Treasury should determine whether to lend and the terms of such loans.
- Emergency lending to non-financial companies should be regarded per se as involving significant credit risk and be the sole purview of the Treasury.
- When lending to banks or non-banks has been deemed to pose significant credit risk, then programs should be clearly identified as Treasury and not Fed programs. The Fed’s role would then be merely advisory and operational.
- To allow the Treasury to serve this function, Congress should approve standing authority for the Treasury to engage in, or backstop, emergency lending. This authority could take the form of a standing emergency lending fund, funded with a fixed appropriation by Congress. Alternatively, it could take the form of congressional approval for the Treasury to guarantee loans made by the Fed, acting in its capacity as an agent of Treasury. Since the Fed’s role in such a program would be merely advisory and operational, Congress should also empower the Treasury to order the Fed to engage in Treasury-backed risky credit provision if the Treasury determines that such a program is necessary.
- As an initial step, the Treasury and Fed should be required to issue a joint statement, laying out responsibility for major decisions about the structure and terms of the emergency lending facility, if not during the actual crisis then promptly after the crisis has subsided.
I recognize that my proposal requires that the Treasury be responsible for risky lending. I reach this conclusion with a certain amount of regret. If I were to choose which party, as between Congress, the Treasury, and the Fed, would be likely to adopt the best policies in an emergency, it would be the Fed because they are independent and expert. Indeed, that is why I was critical in my book of the restrictions imposed on the Fed by Dodd-Frank. And there is a lot to complain about how the Treasury dealt with the CARES Act. But we do not have a government entrusting action to experts or philosopher kings. Nor should we.
[*] Emeritus Nomura Professor of International Financial Systems, Harvard Law School, Director, Committee on Capital Markets Regulation
[†] David Ricardo, On the Principles of Political Economy and Taxation, Chapter XXXI, On Machinery (3d ed. 1821).
 See Committee on Cap. Mkts. Reg., Revising the Legal Framework for Non-Bank Emergency Lending (Sept. 2021), https://www.capmktsreg.org/wp-content/uploads/2021/09/Revising-the-Legal-Framework-for-Non-Bank-Emergency-Lending-09.21.2021.pdf.
 Hal S. Scott, Connectedness and Contagion: Protecting the Financial System from Panics 94 (2016).
 Fed. Reserve Act § 13(3), 12 U.S.C § 343 (2012).
 See Walter W. Eubanks, Cong. Research Serv., R41176, Federal Financial Services Regulatory Consolidation: Structural Response to the 2007-2009 Financial Crisis 17–20 (2010).
 Matthew Karnitschnig, Deborah Solomon, Liam Pleven & Jon E. Hilsenrath, U.S. to Take Over AIG in $85 Billion Bailout; Central Banks Inject Cash as Credit Dries Up, Wall St. J. (Sep. 16, 2008), https://www.wsj.com/articles/SB122156561931242905 [https://perma.cc/L2EF-C4VJ].
 Laurence Ball, The Fed and Lehman Brothers 14 (2016), https://www.nber.org/data-appendix/w22410/The%20Fed%20and%20Lehman%20Brothers.pdf [https://perma.cc/7QTQ-6AGT].
 See Ben Bernanke, Final Keynote Address at the Brookings Institution Liquidity and the Role of the Lender of Last Resort Event (Apr. 30, 2014), https://www.brookings.edu/events/liquidity-and-the-role-of-the-lender-of-last-resort/ [https://perma.cc/35PB-XSMQ]; Janet Yellen, Chair, Fed. Reserve, Opening Statement on the draft final rule implementing amendments enacted by the Dodd-Frank Act to the Federal Reserve’s emergency lending authority under section 13(3) of the Federal Reserve Act (Nov. 30, 2015), https://www.federalreserve.gov/newsevents/press/bcreg/yellen-opening-statement-20151130.htm [https://perma.cc/663D-CR98].
 See Letter from Janet Yellen, Chair, Federal Reserve, to Elizabeth Warren, Senator (Nov. 30, 2015), https://www.federalreserve.gov/foia/files/warren-letter-20151130.pdf [https://perma.cc/8UEP-5343].
 Extensions of Credit by Federal Reserve Banks, 80 Fed. Reg. 78959, 78961–62 (Dec. 18, 2015) (to be codified at 12 C.F.R. 201).
 Since the Fed was not authorized by the Federal Reserve Act to buy commercial paper, it set up a SPV to do so. The Fed lent to the SPV and the SPV bought the unsecured commercial paper which was pledged, together with issuer fees, to the Fed. Of course, if issuers defaulted on the commercial paper, this “collateral” would be worthless, and the fees represented a small part of the Fed’s exposure. See U.S. Gov’t Accountability Office, GAO-11-696, Federal Reserve System: Opportunities Exist to Strengthen Policies and Processes for Managing Emergency Assistance 92–93, 192–98 (2011), https://www.gao.gov/assets/gao-11-696.pdf [https://perma.cc/ZH3L-B5SD]; Eric Posner, What Legal Authority Does the Fed Need During a Financial Crisis?, 101 Minn. L. Rev. 1529, 1552–53 (2017).
 U.S. Const. art. I, § 9, cl. 7.
 For an extended analysis of Congress’s constitutional authority over fiscal decisions, see Kate Stith, Congress’ Power of the Purse, 97 Yale L.J. 1343 (1988).
 See id. at 1381–86. Stith argues that the grant of spending authority by Congress to another agency is only consistent with the constitutional requirement that Congress control appropriations if Congress “clearly defines the activity being funded, provides a time limitation on the spending program, implicitly or explicitly decides the total amount of spending authority, and undertakes periodic legislative review.” Id. at 1383.
 See Office of Pers. Mgmt. v. Richmond, 496 U.S. 414, 430 (1990) (citing Wilson Cowen, Philip Nichols & Marion Bennett, The United States Court of Claims: A History (1978)).
 626 F. Supp. 1374, 1383 (D.D.C. 1986), aff’d on other grounds in Bowsher v. Synar, 478 U.S. 714, 721 (1986).
 Id. at 1385–86 (“The appropriations power is not functionally distinguishable from other powers successfully delegated by Congress.”).
 Bd. of Governors of the Fed. Reserve Sys., Financial Accounting Manual for Federal Reserve Banks 104 (2021).
 Charles W. Calomiris, Douglas Holtz-Eakin, R. Glenn Hubbard, Allan H. Meltzer & Hal S. Scott, Establishing credible rules for Fed emergency lending, 9 J.Fin. Econ. Pol’y 260 (2017), https://www.capmktsreg.org/wp-content/uploads/2017/08/JFEP-01-2017-00065b35d.pdf [https://perma.cc/2NCD-433U].
 Id. at 263.
 See Ian Plenderleith, Review of the Bank of England’s provision of emergency liquidity assistance in 2008–09 84 (2012), https://www.bankofengland.co.uk/-/media/boe/files/news/2012/november/the-provision-of-emergency-liquidity-assistance-in-2008-9 [https://perma.cc/BMS3-FS27] (“The purpose of this review is to learn lessons to inform the way the Bank conducts ELA operations for individual financial institutions. Such support operations will, in due course, be conducted under the new Crisis Management Memorandum of Understanding, which was published in January 2012. The review will build on the lessons learned in relation to the ELA provided to Northern Rock in 2007, as set out in the Treasury Committee’s report ‘The Run on the Rock’”); see also Press Release, Bank of Eng., Court of the Bank of England commissions a set of reviews to learn lessons (May 21, 2012), https://www.bankofengland.co.uk/-/media/boe/files/news/2012/may/court-of-the-boe-commissions-a-set-of-reviews-to-learn-lessons [https://perma.cc/J3EH-VZ7R].
 Financial Services Act of 2012, c. 21, Part 4, http://www.legislation.gov.uk/ukpga/2012/21/part/4/enacted [https://perma.cc/FVX8-6ZYD]; HM Treasury, Bank of England and Prudential Regulation Authority, Memorandum of understanding on financial crisis management § 1 (2012), https://www.bankofengland.co.uk/-/media/boe/files/memoranda-of-understanding/resolution-planning-and-financial-crisis-management.pdf?la=en&hash=57D8302D2AE09F004E 67BEF19A554547CAD2D47B [https://perma.cc/RT6M-XYDG].
 HM Treasury, Bank of England and Prudential Regulation Authority, Memorandum of understanding on financial crisis management, supra note 24.
 12 C.F.R. § 201.4(a) (“A Federal Reserve Bank may extend primary credit on a very short-term basis, usually overnight, as a backup source of funding to a depository institution that is in generally sound financial condition in the judgment of the Reserve Bank.”); § 201.4(b) (“A Federal Reserve Bank may extend secondary credit on a very short-term basis, usually overnight, as a backup source of funding to a depository institution that is not eligible for primary credit . . . .”); Id. §§ 201.104–110 (describing eligible collateral for discount window loans). In nearly two decades, the highest average weekly usage of the Fed’s secondary credit facility was less than $1 billion. See Assets: Liquidity and Credit Facilities: Loans: Secondary Credit: Week Average, https://fred.stlouisfed.org/series/WSC [https://perma.cc/7ZVT-KBJY](last visited Oct. 31, 2021).
 12 C.F.R. § 201.4(b).
 Hal S. Scott, Connectedness and Contagion: Protecting the Financial System from Panics 115 (2016).
 HM Treasury, supra note 24.
See Letter from Rishi Sunak, Chancellor, Exchequer, to Andrew Bailey, Governor, Bank of Eng. (Mar. 17, 2020), https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/873222/20200317_-_CX-Gov_Letter_re_CCFF_vF.pdf [https://perma.cc/Y2VJ-57ND]; see also Covid Corporate Financing Facility, Bank of England, https://www.bankofengland.co.uk/markets/covid-corporate-financing-facility [https://perma.cc/VY62-5P97](last visited Oct. 4, 2020).
 Fed. Reserve Act § 13(3)(A), 12 U.S.C § 343 (2012).
 Press Release, Bd. of Governors of the Fed. Reserve Sys., Federal Reserve Board announces establishment of a Commercial Paper Funding Facility (CPFF) to support the flow of credit to households and businesses (Mar. 17, 2020), https://www.federalreserve.gov/newsevents/pressreleases/monetary20200317a.htm [https://perma.cc/ZP38-TP5F]; Press Release, Bd. of Governors of the Fed. Reserve Sys., Federal Reserve Board announces establishment of a Primary Dealer Credit Facility (PDCF) to support the credit needs of households and businesses (Mar. 17, 2020), https://www.federalreserve.gov/newsevents/pressreleases/monetary20200317b.htm [https://perma.cc/MSP2-EMXR]; Press Release, Bd. of Governors of the Fed. Reserve Sys., Federal Reserve Board broadens program of support for the flow of credit to households and businesses by establishing a Money Market Mutual Fund Liquidity Facility (MMLF) (Mar. 18, 2020), https://www.federalreserve.gov/newsevents/pressreleases/monetary20200318a.htm [https://perma.cc/ME5L-YYJR].
 Press Release, Bd. of Governors of the Fed. Reserve Sys., Federal Reserve Board announces establishment of a Commercial Paper Funding Facility (CPFF) to support the flow of credit to households and businesses, supra note 35.
 Coronavirus Aid, Relief, and Economic Security Act (CARES Act), H.R. 748, 116th Cong. § 4003 (2020).
 See Press Release, Bd. of Governors of the Fed. Reserve Sys., Federal Reserve announces extensive new measures to support the economy (Mar. 23, 2020), https://www.federalreserve.gov/newsevents/pressreleases/monetary20200323b.htm [https://perma.cc/42A8-GUT4].
 See id.
 Main Street Lending Program, Bd. of Governors of the Fed. Reserve Sys., https://www.federalreserve.gov/monetarypolicy/mainstreetlending.htm [https://perma.cc/42ZF-9GTQ](last visited Dec. 7, 2020).
 See Main Street Lending Program, Bd. of Governors of the Fed. Reserve Sys., https://www.federalreserve.gov/monetarypolicy/mainstreetlending.htm [https://perma.cc/5DCK-EB4T] (Oct. 13, 2021).
 See Press Release, Bd. of Governors of the Fed. Reserve Sys., Federal Reserve takes additional actions to provide up to $2.3 trillion in loans to support the economy (Apr. 9, 2020), https://www.federalreserve.gov/newsevents/pressreleases/monetary20200409a.htm [https://perma.cc/9ZTD-AV6C].
 H.4.1 Factors Affecting Reserve Balances, Bd. of Governors of the Fed. Reserve Sys. (Dec. 28, 2020), https://www.federalreserve.gov/releases/h41/20201228/.
 Term Sheet for Secondary Market Corporate Credit Facility, Bd. of Governors of the Fed. Reserve Sys. (Apr. 9, 2020), https://www.federalreserve.gov/newsevents/pressreleases/files/monetary20200409a2.pdf [https://perma.cc/VT98-2FAH].
 Term Sheet for Main Street New Loan Facility, Bd. of Governors of the Fed. Reserve Sys. (Apr. 9, 2020), https://www.federalreserve.gov/newsevents/pressreleases/files/monetary20200409a7.pdf [https://perma.cc/5TCP-P94P].
 Primary Market Corporate Credit Facility, Bd. of Governors of the Fed. Reserve Sys. (July 28, 2020), https://www.federalreserve.gov/newsevents/pressreleases/files/monetary20200728a9.pdf [https://perma.cc/L2QP-BE8M]; Secondary Market Corporate Credit Facility, Bd. of Governors of the Fed. Reserve Sys. (July 28, 2020), https://www.federalreserve.gov/newsevents/pressreleases/files/monetary20200728a1.pdf [https://perma.cc/XUU4-QC6T].
 Press Release, Bd. of Governors of the Fed. Reserve Sys., Federal Reserve Board adjusts terms of Main Street Lending Program to better target support to smaller businesses that employ millions of workers and are facing continued revenue shortfalls due to the pandemic (Oct. 30, 2020), https://www.federalreserve.gov/newsevents/pressreleases/monetary20201030a.htm [https://perma.cc/GM53-S9S6].
 Term Sheet for Main Street Expanded Loan Facility, Bd. of Governors of the Fed. Reserve Sys. (June 8, 2020), https://www.federalreserve.gov/newsevents/pressreleases/files/monetary20200608a3.pdf [https://perma.cc/R4BA-4VSV].
 See H.4.1 Factors Affecting Reserve Balances, Fed. Reserve Statistical Release (Jan. 14, 2021), https://www.federalreserve.gov/releases/h41/20210114/[https://perma.cc/XVA4-EZJB]; Dept. of Treasury, Audit of the Exchange Stabilization Fund’s Financial Statements for Fiscal Years 2020 and 2019, 23 (March 20, 2021), https://home.treasury.gov/system/files/206/ESF-Annual-Audit-FY2020.pdf.
 Phillip L. Swagel, Cong. Budget Off., Preliminary Estimate of the Effects of H.R. 748, the CARES Act, Public Law 116-136, Revised, With Corrections to the Revenue Effect of
the Employee Retention Credit and to the Modification of a Limitation on Losses for Taxpayers Other Than Corporations, 2 (Apr. 27, 2020), https://www.cbo.gov/system/files/2020-04/hr748.pdf [https://perma.cc/WT3V-CBDT].
 See Glenn Hubbard & Hal Scott, Editorial, Main Street Needs More Fed Help, Wall St. J. (Apr. 16, 2020), https://www.wsj.com/articles/main-street-needs-more-fed-help-11587055459 [https://perma.cc/WSD4-75NA].
 Term Sheet for Main Street New Loan Facility, supra note 42 ; Term Sheet for Main Street Expanded Loan Facility, supra note 45.
 See Editorial, The Main Street Fakeout, Wall St. J. (Apr. 30, 2020), https://www.wsj.com/articles/the-main-street-fakeout-11588288799 [https://perma.cc/TUF9-28GS].
 See Glenn Hubbard & Hal Scott, Who’s Looking Out for Main Street?, Wall St. J. (May 17, 2020), https://www.wsj.com/articles/whos-looking-out-for-main-street-11589741411 [https://perma.cc/298D-J48S]; Glenn Hubbard & Hal Scott, ‘Main Street’ Program Is Too Stingy to Banks and Borrowers, Wall St. J. (July 20, 2020), https://www.wsj.com/articles/main-street-program-is-too-stingy-to-banks-and-borrowers-11595284266 [https://perma.cc/8T5R-HDDE].
 See ‘Main Street’ Program Is Too Stingy to Banks and Borrowers, supra note 55.
 Fed. Reserve Act, § 13(3)(B)(i).
 Fed. Reserve Act, § 13(3)(B)(ii).
 12 C.F.R. § 201.4(d)(5)(iii) (emphasis added).
 See 12 C.F.R. § 201.4(d)(5)(iv)(A).
 Term Sheet for Main Street New Loan Facility, Bd. of Governors of the Fed. Reserve Sys. (Dec. 29, 2020), https://www.federalreserve.gov/newsevents/pressreleases/files/monetary20201229a1.pdf [https://perma.cc/DJ6E-TV5Q].
 12 C.F.R. § 201.4(d)(5)(iii).
 Kate Davidson & Richard Rubin, Steven Mnuchin Says U.S. Aims to Get Back Its Money From Fed Programs, Wall St. J. (Apr. 29, 2020), https://www.wsj.com/articles/mnuchin-says-u-s-not-aiming-to-lose-money-on-fed-lending-facilities-11588178749 [https://perma.cc/8LV8-RPWB].
 Hybrid Hearing: Oversight of the Treasury Department’s and Federal Reserve’s Pandemic Response Before the H. Fin. Services Comm., 116th Cong. (2020) (statement of Steven Mnuchin, Secretary, U.S. Department of the Treasury), https://financialservices.house.gov/calendar/eventsingle.aspx?EventID=406869 [https://perma.cc/L7BZ-UAGM]; see also Victoria Guida & Zachary Warmbrodt, Mnuchin, Powell face new demands to rescue economy, Politico (May 19, 2020), https://www.politico.com/news/2020/05/19/mnuchin-powell-face-new-demands-to-rescue-economy-268331 [https://perma.cc/7R5M-LMGB] (“There are scenarios within Main Street where we could lose all of our capital, and we’re prepared to do that.”).
 Letter from Steven T. Mnuchin, Secretary, U.S. Dep’t of the Treasury, to Jerome Powell, Chair, Fed. Reserve (Nov. 19, 2020), https://home.treasury.gov/system/files/136/letter11192020.pdf [https://perma.cc/DL7V-65J5].
 James Politi & Colby Smith, US Treasury refuses to extend some of Fed’s crisis-fighting tools, Fin. Times (Nov. 19, 2020), https://www.ft.com/content/e4b3a063-db44-4e6c-b998-74a29d70b136 [https://perma.cc/4X89-KKNU] (“The Federal Reserve would prefer that the full suite of emergency facilities established during the coronavirus pandemic continue to serve their important role as a backstop for our still-strained and vulnerable economy.”).
 See supra note 66.
 See Memorandum from Jay B. Sykes on Section 4029 of the Coronavirus Aid, Relief, and Economic Security (CARES) Act and the Extension of the Federal Reserve’s Emergency-Lending Programs to the House Select Subcommittee on the Coronavirus Crisis (Dec. 17, 2020), https://coronavirus.house.gov/sites/democrats.coronavirus.house.gov/files/Memo%20re%20CARES%20Act%20Section%204029.pdf [https://perma.cc/6V4T-28BP].
 Id. at 5–7.
 See supra note 66.
 H.4.1 Factors Affecting Reserve Balances, Bd. of Governors of the Fed. Reserve Sys. (Nov. 19, 2020), https://www.federalreserve.gov/releases/h41/20201119/ [https://perma.cc/KU6K-JZZY].
 Letter from Jerome Powell, Chair, Fed. Reserve, to Steven Mnuchin, Secretary, Department of the Treasury, on emergency lending facilities (Nov. 20, 2020), https://www.federalreserve.gov/foia/files/mnuchin-letter-20201120.pdf [https://perma.cc/44GM-JY4F]; Sykes, supra note 69, at 6–7.
 Consolidated Appropriations Act, 2021, H.R. 133, 116th Cong., § 1005 (2020), https://rules.house.gov/sites/democrats.rules.house.gov/files/BILLS-116HR133SA-RCP-116-68.pdf [https://perma.cc/JB9F-67G3].
 Id. at § 1003.
 Id. at § 1005.
 See U.S. Dept. of Treasury, Exchange Stabilization Fund: Statement of Financial Position as of Sept. 30, 2021 (last accessed Nov. 4, 2021), https://home.treasury.gov/system/files/206/Trunc_Notes.pdf.
 See Hal Scott, Here’s How the Fed Can Do More To Support US Small Business, Fin. Times (Jan. 11, 2021), https://www.ft.com/content/0d0b0f48-df3c-4b2f-97b8-d2da0415d8d6 [https://perma.cc/7K4P-R8M6].
 There was also lack of clarity about legal authority over the duration of the lending facilities. The term sheets of the CPFF and the MMLF, both of which invoked the Fed’s section 13(3) authority, suggested that the duration of each facility could be extended unilaterally by the Fed—even though Section 13(3) requires the terms of any lending facility to be approved by both the Fed and the Treasury Secretary. See Commercial Paper Funding Facility 2020: Program Terms and Conditions (Mar. 17, 2020), https://www.federalreserve.gov/newsevents/pressreleases/files/monetary20200317a1.pdf [https://perma.cc/JV3F-C5U2]; Money Market Mutual Fund Liquidity Facility (Mar. 18, 2020), https://www.federalreserve.gov/newsevents/pressreleases/files/monetary20200318a1.pdf [https://perma.cc/XN2G-KCUE]. Subsequent revisions to the term sheets, issued several months later, corrected this mistake. Commercial Paper Funding Facility 2020: Program Terms and Conditions (Nov. 30, 2020), https://www.federalreserve.gov/newsevents/pressreleases/files/monetary20201130a1.pdf [https://perma.cc/3MBU-76PN]; Money Market Mutual Fund Liquidity Facility (Jul. 28, 2020), https://www.federalreserve.gov/newsevents/pressreleases/files/monetary20200728a4.pdf [https://perma.cc/P3CA-7SU7].
 Victoria Guida & Aubree Eliza Weaver, We’re In a Recession, By the Way, Politico (Jun. 9, 2020), https://www.politico.com/newsletters/morning-money/2020/06/09/were-in-a-recession-by-the-way-788358 [https://perma.cc/SU6Q-TXZW](“People familiar with the matter tell [Morning Money] that the Treasury Department has been more cautious on taking risk than the Fed. They’re the ones ponying up taxpayer dollars, so this makes some sense.”).
 Fed. Reserve Act, §13(3).
 Adam J. Levitin, In Defense of Bailouts, 99 Geo. L.J. 435, 496 (2011) (“[E]ven the ‘unusual and exigent circumstances’ requirement does not necessarily indicate systemic risk, just market disruption.”).
 Partinitha Sastry, The Political Origins of Section 13(3) of the Federal Reserve Act, FRBNY Econ. Policy Rev., Sept. 2018, at 23, https://www.newyorkfed.org/medialibrary/media/research/epr/2018/epr_2018_political-origins_sastry.pdf [https://perma.cc/LJ3C-TA9U].
 Id. at 19–23.
 Id. at 23–24.
 Id. at 24.
 Board of Governors of the Federal Reserve System, Fifty-Third Annual Report: Covering Operations for the Year 1966 92 (1967).
 Board of Governors of the Federal Reserve System, Fifty-Sixth Annual Report: Covering Operations for the Year 1969 92 (1970).
 Thomas C. Baxter, Jr., General Counsel and Executive Vice President, Fed. Reserve of N.Y., Presentation to Regulatory Response to the Financial Crisis on The Legal Position of the Central Bank, The Case of the Federal Reserve Bank of New York 6 (Jan. 19, 2009), https://web.archive.org/web/20130511222423/http://lse.ac.uk/fmg/documents/events/conferences/2009/regulatoryResponse/1160_Baxter.pdf.
 See Term Sheet for Primary Dealer Credit Facility, Bd. of Governors of the Fed. Reserve Sys. (Jul. 28, 2020), https://www.federalreserve.gov/newsevents/pressreleases/files/monetary20200728a8.pdf [https://perma.cc/YUY6-B43V]
 See Primary Dealers, Expectations and Requirements, Fed. Reserve Bank of NY, https://www.newyorkfed.org/markets/primarydealers#:~:text=Be%20either%20(1)%20a%20broker,2)%20a%20state%20or%20federally [https://perma.cc/48CW-HZRN ](last visited Oct. 4, 2020).
 Fed. Reserve Act, § 10B(a), 12 U.S.C. § 347b(a).
 See, e.g., Term Sheet for Primary Dealer Credit Facility (PDCF), Bd. of Governors of the Fed. Reserve Sys. (Nov. 30, 2020), https://www.federalreserve.gov/newsevents/pressreleases/files/monetary20201130a3.pdf [https://perma.cc/54JE-58TM ](“The pledged collateral will be valued by Bank of New York Mellon according to a schedule designed to be similar to the margin schedule for lending by the discount window, to the extent possible.”).
 The kind of lending the Fed engaged in through TALF would be more difficult to classify. On the one hand, the asset-backed securities to which the Fed was exposed through TALF included non-financial debt such as student and equipment loans; on the other hand, the loans made by TALF were to financial companies that were independent of the Fed.
 Fed. Credit Reform Act of 1990, Pub. L. No. 101-508, § 504 (1990). Previously, they were accounted for on a cash-flow basis. See Mindy R. Levit, Cong. Research Serv., R42632, Budgetary Treatment of Federal Credit (Direct Loans and Loan Guarantees): Concepts, History and Issues for Congress 4–5 (2014).
 See supra note 51.
 Stephanie Ruhle & Ben Popken, Thousands of applicants, zero loans: Trump’s small businesses lending program is a failure to launch, NBC News (Apr. 4, 2020), https://www.nbcnews.com/business/business-news/thousands-applicants-zero-loans-trump-s-small-businesses-lending-program-n1176766 [https://perma.cc/3TDG-DKW8]; Lauren Fox, Glitches hamper second round of small business loan funding, CNN (Apr. 27, 2020), https://www.cnn.com/2020/04/27/politics/state-of-play-paycheck-protection-program-small-business-administration/index.html [https://perma.cc/Z97J-5ZJG].
 Fed. Reserve Act § 13(2), 12 U.S.C § 343.
 See Federal Reserve, Commercial Paper Funding Facility: Frequently Asked Questions (October 27, 2008), https://www.newyorkfed.org/markets/cpff_faq_081027.html; Eric A. Posner, What Legal Authority Does the Fed Need During a Financial Crisis?, 101 Minn. L. Rev. 1529, 1551-52 (2017); Alexander Mehra, Legal Authority in Unusual and Exigent Circumstances: The Federal Reserve and the Financial Crisis, 13 U. Penn. J. Bus. L. 221, 235-36 (2011).
 See CARES Act, § 4003(b)(4); § 4029(b)(4). See also Letter from Treasury Secretary Steven Mnuchin to Federal Reserve Chairman Jerome Powell (Nov. 19, 2020), https://home.treasury.gov/system/files/136/letter11192020.pdf [https://perma.cc/K5YS-72KC ]
 Fed. Reserve Act, § 13(3)(C), 12 U.S.C. § 347d(3)(C).
 TALF also limited this problem by excluding single-asset, single-borrower commercial mortgage-backed securities as eligible collateral. See Term Sheet for Term Asset-Backed Securities Loan Facility, Bd. of Governors of the Fed. Reserve Sys. (Jul. 28, 2020), https://www.federalreserve.gov/newsevents/pressreleases/files/monetary20200728a6.pdf [https://perma.cc/8E8U-NBHY].
 See, e.g., Victoria Guida & Zack Colman, Fed’s expansion of lending program sparks oil bailout worries, Politico (April 30, 2020), https://www.politico.com/news/2020/04/30/feds-expansion-of-lending-program-sparks-oil-bailout-worries-227545 [https://perma.cc/RB45-8ZS5]; Robert Schmidt, Jesse Hamilton, & Sally Bakewell, Private Equity to Get Squeezed Out of Another Stimulus Program, Bloomberg News (April 23, 2020) (private equity owned companies concerned that they would not qualify for Main Street facilities), https://www.bloomberg.com/news/articles/2020-04-23/private-equity-to-get-squeezed-out-of-another-stimulus-program?sref=2lCQoM0A.
 Federal Reserve, Secondary Market Corporate Credit Facility (July 28, 2020), https://www.federalreserve.gov/newsevents/pressreleases/files/monetary20200728a1.pdf [https://perma.cc/2252-AX5C].
 Committee on Capital Markets Regulation, Treasury and Fed Lending Programs: An Assessment and Call for Continued Support for SMEs, 2 (Dec. 2020), https://www.capmktsreg.org/wp-content/uploads/2020/12/CCMR-Report-Treasury-Fed-Programs-Assessment-and-Recommendations-12.29.2020.pdf [https://perma.cc/VBL9-PJFR].
 Federal Reserve, Factors Affecting Reserve Balances, Federal Reserve Statistical Release H.4.1 (Jan. 14, 2021).
 Office of the Comptroller of the Currency, Treasury, the Board of Governors of the Federal Reserve System & Federal Deposit Insurance Corporation, Regulatory Capital Rule: Money Market Mutual Fund Liquidity Facility, 85 Fed. Reg. 16232 (Mar. 23, 2020); Office of the Comptroller of the Currency, Treasury, the Board of Governors of the Federal Reserve System & Federal Deposit Insurance Corporation, Regulatory Capital Rule: Paycheck Protection Program Lending Facility and Paycheck Protection Program Loans, 85 Fed. Reg. 20387 (Apr. 13, 2020).
 Joint Press Release, Board of Governors of the Federal Reserve System, Federal Deposit Insurance Corporation & Office of the Comptroller of the Currency, Federal bank regulatory agencies modify liquidity coverage ratio for banks participating in Money Market Mutual Fund Liquidity Facility and Paycheck Protection Program Liquidity Facility (May 5, 2020), https://www.federalreserve.gov/newsevents/pressreleases/bcreg20200505a.htm [https://perma.cc/X7UZ-CL25].