When it comes to responding to the coronavirus outbreak in the U.S., the Federal Reserve has emerged as one of the most active institutions at the national level. Its bold and timely interventions have halted a monetary breakdown that would have guaranteed a second Great Depression. And its continuing efforts to avert a vicious cycle of debt defaults are helping to address a sudden economic stop that has made a deep and lasting recession all but inevitable. Unfortunately, the Fed has repeatedly had to scramble and stretch its authority because it was not designed to address the current crisis. In this post, I examine what the Fed is doing and how its actions line up with existing law. I conclude with a few thoughts about how Congress might improve money and banking law in the future.
What is the Fed Doing?
In the last month, the Fed has launched 13 different emergency lending initiatives. Eight of these are “liquidity facilities” designed to stop a run in the money markets by making short-term loans to financial firms. The other five are “credit facilities” designed to invest in nonfinancial businesses and prevent a wave of business bankruptcies.
Through the liquidity facilities, the Fed is exchanging cash for relatively low-risk financial assets. These facilities function like a discount window for nonbank financial firms – securities dealers, money market mutual funds, finance companies, and foreign banks – firms that fund themselves like banks but do not have access to the Fed’s standing liquidity facility. The Fed’s standing facility is designed to backstop banks. The way it works is that the Fed lends them money against certain collateral by increasing their balances in bank accounts they have with the Fed. These balances are called reserves. In addition to borrowing reserves from the Fed, banks can also borrow reserves from each other in what is known as the “federal funds” or “fed funds” market.
The inability of nonbanks to borrow reserves is intentional. Congress designed the Fed to be a banker’s bank, the theory being that, in a credit crunch, the Fed would support banks, and the banking system would continue to lend to everyone else, including other financial firms. But the theory assumed that only banks would engage in banking: issuing demandable debt obligations like deposits that augment the supply of physical currency. The system was not designed to accommodate nonbanks issuing similar obligations.
And yet, practice moved away from theory. By 2007, a parallel “shadow banking” system had grown so large that it dwarfed the banking system. When the economy slowed, and asset prices fell, banks were either unable or unwilling to backstop the shadow banks. So, the Fed was forced to stretch its statutory powers to save the system from collapse. Why? Because by 2007 such a large fraction of economic activity depended on credit provided by shadow banks. Were the shadow banks to fail, price signals would go haywire and the economy would grind to a halt.
Unfortunately, in the aftermath, Congress did not restructure the financial system, nor did it redesign the Fed. (It actually rolled back some of the Fed’s powers.) So, when asset prices fell earlier this year, shadow banks needed help, and the Fed was forced again to improvise. It expanded its dealer repo facility and reopened four 2008-era programs designed to backstop different types of shadow banks – the PDCF, CPFF, MFFLF, and TALF. It also began lending dollars to foreign central banks through swap lines and a new program called FIMA. These programs are designed to on-lend dollars to financial institutions overseas. These institutions need dollars because they issue dollar denominated deposits and other short-term dollar debt.
Figure 1: The Fed’s 13 Initiatives
The Fed is underwriting each of these facilities itself, with the U.S. Treasury investing $10 billion of equity in three of them. Mostly, the Fed is working through known counterparties. But for the CPFF, which will buy commercial paper directly from non-banks, the Fed has hired Pimco and State Street to help.
It is worth dwelling briefly on the swap lines and FIMA facility. Swap lines are risky because there is no collateral. All the Fed gets is an account balance in foreign currency on the books of a foreign central bank. In the event of default, the Fed has little recourse. FIMA by contrast is safer because the Fed takes U.S. Treasury securities as collateral.
Bottom line: All eight of these programs extend the central bank’s classic lender of last resort function to shadow banks. They involve creating new reserves to backstop private sector money augmentation. Done properly, a relatively small amount of new reserves can prop up giant markets because, to be a bit reductive about it, once the Fed announces it will backstop a promise to pay dollars, it makes that promise as good as dollars.
The Fed’s new credit facilities are a whole different animal. These programs invest in nonfinancial businesses and municipalities. Whereas the liquidity programs backstop money markets, meaning they stabilize the value of deposits and deposit substitutes by ensuring that these private assets trade at par with actual dollars, the Fed’s credit facilities have nothing to do with money markets. These facilities are not designed to preserve existing credit arrangements by preventing fire sales – they are designed to expand credit.
Accordingly, these facilities raise all sorts of questions about underwriting and credit risk.
Here is what we know so far: The Fed is setting the general eligibility policies for each facility. For the PMCCF and SMCCF, which are designed to invest directly in big businesses, the Fed has hired BlackRock to apply these policies, and the Treasury will put up equity to absorb losses. For the PPPLF, MSNLF, and MSELF, which are designed to lend to small and medium-sized enterprises, the Fed has deputized banks to do the underwriting and servicing. For the PPP loans, the Fed will bear no risk of loss. For the Main Street loans, the banks will have skin in the game, and the Treasury will put up $75 billion of equity. The Fed is also going to help state and local governments by lending up to $500 billion, but it has not yet provided details on how it plans to do this. (The Treasury will contribute $35 billion in loss protection.)
Using What Authorities?
So, what are the legal foundations for this unprecedented balance sheet expansion? The Fed is relying primarily on two provisions of its enabling act: Section 13(3) and Section 14.
Section 13 Facilities
The third paragraph of Section 13 authorizes the Fed to lend to nonbanks in “unusual and exigent circumstances.” In 2010, Congress amended the law to add several significant new restrictions. One of these requires the Fed to ensure that “the security for emergency loans is sufficient to protect taxpayers from losses.” For three of the Fed’s facilities – the CPFF, MFFLF, and TALF – the Treasury is providing the necessary security by investing money from the Exchange Stabilization Fund, an account of around $100 billion. But the ESF is designed so that the Treasury secretary can stabilize the value of the dollar against foreign currencies and assist the U.S. in meeting its commitments to the IMF. The law provides that: “Consistent with . . . a stable system of exchange rates, the Secretary . . . may deal in gold, foreign exchange, and other instruments of credit and securities.” The law says nothing about Fed facilities.
When Treasury stretched this authority in 2008 to backstop money market funds, Congress wasn’t pleased, and it passed a law explicitly prohibiting the practice. Although Congress seems much more willing to look the other way this time around, it is hard to see how Treasury’s use of the ESF is any more legal. Treasury’s investment in these facilities is not related to maintaining an orderly system of exchange rates, the statutory predicate upon which the secretary is authorized to deal in securities. Nor is it clear that the secretary’s investment can be construed as dealing in securities – being that it is, at best, the purchase of a bespoke equity instrument that is not traded on secondary markets.
Another wrinkle involves the new requirement in 13(3)(B)(i) that the Board establish policies to permit lending only “for the purpose of providing liquidity to the financial system.” Under the original design of 13(3), the Fed could underwrite loans directly to individuals, partnerships, and corporations, which it did, albeit sparingly, in the 1930s. But Congress repealed this authority in 2010 and replaced it with something rather different: the power to create facilities to provide liquidity to nonbank financial firms. In other words, Congress chucked the part of 13(3) that authorized the Fed to lend to the real economy.
This would seem to prohibit all the Fed’s new credit facilities, but on March 27, Congress, in an amazing legislative sleight of hand, amended 13(3) sub silentio by instructing the Treasury secretary to invest $454 billion in Fed facilities that lend directly to the real economy. If the Fed could not create such facilities – if it could only provide liquidity to the financial system – then this appropriation would be a dead letter. To make the point clearer, Congress quoted the ostensibly restrictive language from 13(3)(B)(i) about facilities being for the purpose of providing liquidity to the financial system and used it to describe what the relevant real economy lending programs would do. Congress basically defined the text of its 2010 amendment down to nothing.
Section 14 Facilities
The Fed itself, not Congress or the Treasury, is doing the stretching when it comes to the Section 14 programs – the swap lines, repo operations, and FIMA facility.
In a swap, the Fed increases on its books the account balance of a foreign central bank. In exchange, the foreign central bank increases the Fed’s balance on its books denominated in whatever currency it issues. The arrangement is structured as a purchase of foreign currency, but it is really a loan. Sometime in the future, the purchase will be unwound and both balance sheets will shrink back down. One thing will have changed, however: The foreign central bank will have surrendered more dollars from its account than the Fed added to it to begin with. That amount is the interest.
The Fed has been doing this since 1961, when the Fed’s general counsel wrote an internal memo blessing the practice. His argument (although he acknowledged there could be criticism “on legal grounds”) was that Section 14 allows the Fed to purchase and sell cable transfers, which are foreign currency instruments, and to maintain accounts with foreign central banks. But he ignores, among other things, the words “in the open market.” Section 14 authorizes purchases and sales “in the open market,” and the swaps are not open market purchases or sales. The Fed is not buying pounds or yen in the foreign exchange market and paying a market price. It is lending dollars by using them to buy foreign currency at a non-market price.
This same problem plagues the Fed’s repurchase operations with dealer firms and the FIMA facility. Through these programs, the Fed lends cash against U.S. debt obligations, structuring the loans as sale-and-repurchase agreements. In the first leg, the Fed “buys” the debt security. In the second leg, the Fed “sells” it back. Neither leg transacts at a market price. The Fed buys the security for considerably less than it is worth – this is the haircut. Then the Fed sells it back for more than it paid for it – this is the interest payment.
The Fed claims that these facilities are authorized by the provision of Section 14 that permits buying and selling securities. But it ignores the language requiring it to buy and sell in the open market. The Fed’s buying and selling in a repo is not at a market price. Indeed, the Fed’s sale of securities in a repo is not a sale at all. It is the settlement of a forward transaction.
This “open market lending” is an area in need of further study. The Fed’s present leadership likely takes comfort in the fact that the Fed has been conducting swaps for 50 years and repurchase agreements for a century. Moreover, the legality of the Fed’s repo activities is buttressed by Section 13(13), which was added to the Federal Reserve Act in 1933. Section 13(13) explicitly empowers the Fed to lend to individuals, partnerships, and corporations against U.S. government securities for up to 90 days. Arguably, since the Fed has the power to make these loans, it also has the power to structure them as purchases and sales. Nevertheless, these initiatives reflect a fundamental mismatch between the design of the Federal Reserve as a bank for banks, created to manage a domestic monetary system, and the reality of the Federal Reserve as a monetary authority that acts primarily through dealers and backstops a global dollar system.
Given the ways in which the government has had to stretch to respond to the economic and financial dimensions of the current crisis, Congress ought to reform our statutory framework for money and banking. Since Treasury investments are required for the Fed to engage in most emergency lending, Congress probably should appropriate a standing fund for that, as Professor Jeffrey Gordon has recommended. Since the Fed is no longer explicitly authorized to lend to the real economy in its enabling act, Congress should consider updating the law or establishing a separate agency like the Reconstruction Finance Corporation. Since shadow banks will likely need a government backstop in every business cycle downturn, Congress ought to regulate shadow banks like banks. And since the dollar is a global currency, a framework for addressing runs overseas is also urgently needed. These are just a few steps that could help improve the government’s ability to respond to future crises predictably, efficiently, and equitably.
 See Federal Reserve Act, §§ 10(b), 13(2); Kathryn Judge, Three Discount Windows, 99 Cornell L. Rev. 785, 809-811 (2014).
 Memorandum from Walter Wyatt, General Counsel of the Federal Reserve Board, to Daniel Crissinger, Governor of the Federal Reserve Board 10 (Aug. 18, 1923) (on file with author) (“It was never contemplated by Congress that the Federal reserve banks should make direct loans to non-member banks nor to stock, bond and acceptance brokers or other individuals, partnerships or corporations which ordinarily would seek such accommodations from member banks.”).
 In 2007, there were $7.8 trillion of bank deposits and $14.8 trillion (gross) in private non-deposit money-claims. See Morgan Ricks, The Money Problem: Rethinking Financial Regulation (2016).
 See, e.g., Ben Bernanke, “The Real Effects of Disrupted Credit: Evidence from the Global Financial Crisis,” Brookings Papers on Economic Activity (2018).
 See generally Brad W. Setser, Addressing the Global Dollar Shortage: More Swap Lines? A New Repo Facility for Central Banks? More IMF Lending?, Follow the Money, Council on Foreign Relations (Mar. 17, 2020).
 Memorandum of Howard Hackley, General Counsel, to the Federal Open Market Committee (Nov. 22, 1961), reprinted in Hearings Before the Committee on Banking and Currency, House of Representatives, Eighty-Seventh Congress, Second Session 144 (Feb. 27 & 28, 1962).
 Such study is the subject of forthcoming work.
This post comes to us from Lev Menand, an academic fellow, lecturer in law, and postdoctoral research scholar at Columbia Law School.
Many thanks to Lev for this intelligent contribution to our public discourse. We read this piece with interest, and in the spirit of thoughtful engagement we wish to share some of the ideas that came out of a debate among the three of us. To put the bottom line up front: We don’t think it’s accurate to say that Congress stretched the law here; since Congress has the power to clarify its laws at will, and did just that by specific language in the CARES Act, we do not believe that the new credit facilities established by the Treasury and the Fed come anywhere close to a stretch of their statutory authority.
As to the Treasury – without expressing a view as to whether Treasury’s use of the Exchange Stabilization Fund in 2008 to backstop money market funds was on solid ground at the time, we believe that Lev is wrong to conclude that “this time around, it is hard to see how Treasury’s use of the ESF is any more legal.” It is not hard. Section 4027 of the CARES Act expressly states: “there is appropriated . . . to the fund established under section 5302(a)(1) of title 31, United States Code, $500,000,000,000 to carry out this subtitle.” The fund referenced is the Exchange Stabilization Fund. A major purpose of the subtitle was to fund Treasury’s participation in financial assistance programs carried out jointly with the Federal Reserve acting under Section 13(3) of the Federal Reserve Act. According to public statements made by Treasury Secretary Mnuchin before the CARES Act was passed, acting jointly with the Fed would allow the amount of assistance to be leveraged by many times the amount appropriated by Congress. Congress could not have spoken more clearly against the backdrop of an even clearer description of what the money would be used for and how it would be combined with the Fed’s authority under 13(3).
As to the Fed – it is strange to suggest that Congress “amended 13(3) sub silentio” by enacting the most intensely publicized statute in recent memory and by expressly indicating that the Treasury could use $454 billion to invest in facilities established by the Fed under Section 13(3). Consistent with that understanding, Lev states that Congress was “instructing the Treasury secretary to invest $454 billion in Fed facilities that lend directly to the real economy.” That should be where the argument ends. Lev goes on to suggest that Congress’ intent in 2020 is inconsistent with its intent in 2010, but we do not believe that is so. Moreover, even if it were, Congress can change its mind and the law, as long as the end is legitimate and the means are rationally related to that end. Lev refers to the requirement for the Fed’s emergency lending programs to be governed by policies and procedures “designed for the purpose of providing liquidity to the financial system” as if that means the Fed only has the statutory authority to support the banking system but not the real economy. But the central purpose of 13(3) is to authorize the Fed to use its lender-of-last-resort authority to provide secured loans to non-bank financial institutions, commercial companies and even individuals, if “unusual and exigent circumstances” exist, such as those associated with the COVID-19 pandemic. It reflects Bagehot’s dictum that during a financial crisis, central banks should “lend to merchants, to minor bankers, to ‘this man and that man,’ whenever the security is good.” Indeed, support for the real economy is the ultimate purpose of providing liquidity to the financial system, and of having a financial system in the first place. The 2008 Financial Crisis originated in the financial system, but the COVID-19 crisis is of a different nature. A public health emergency has halted business activity across the country, and the federal government is right to enlist the banking system to channel credit and other government financial support to the real economy through the Paycheck Protection Program, the Main Street lending facilities and other measures. The specific references to the Main Street Lending Program and a mid-size business program in the CARES Act make this very clear. The establishment of credit facilities by the Treasury and the Fed is clearly authorized by specific language in the CARES Act and by their statutory mandates.
Thank you Randy, Meg, and Andrew for your terrific comments.
A few further thoughts:
On the ESF, you are absolutely right to point out that Section 4207 of the CARES Act contemplates the use of the ESF to invest in Fed facilities — at least up to the amount appropriated to the ESF by the CARES Act for that purpose. Also cutting in favor of the Treasury’s MMFLF, CPFF, and TALF investments is Section 4015, which suspends the restrictions in 12 U.S.C. 5236 regarding the use of the ESF to provide guarantees to money market mutual funds, suggesting that making such guarantees is otherwise consistent with the statute. And yet, as with the restrictive language in section 13(3), Congress did not amend the core enabling act, which limits the Secretary to using the ESF to deal in securities, consistent with an orderly system of exchange rates. It appears to be another sort of sub silentio repeal of this earlier text — although it is not clear whether these words in effect no longer have any meaning, or whether they are trumped only in so far as they are inconsistent with implementation of the CARES Act. If the latter, it is not clear to me that Treasury can use the pre-CARES ESF balance to invest in Fed facilities.
Which brings me to a second issue. I also agree with you guys that Congress has the power to clarify its laws. My point is that the CARES Act does the opposite. It says enough to conclude that the Fed’s credit facilities are kosher — notwithstanding prior enactments — and that the Treasury investments, at least as they draw on CARES Act appropriations, are kosher. But it doesn’t tell us how to read the relevant enabling acts in the future: is the old version of 13(3) back whereby the Fed can lend to the real economy? Is the ESF an account Treasury can use for any sort of investment? I used the term “sub silentio” because Congress chose to ignore its prior enactments rather than explicitly repeal them.
Finally, I’m not sure I follow what you are saying about the purpose of 13(3). Bagehot’s LLR is a monetary role. Carter Glass’s bill was designed to authorize the Fed to go beyond LLR and serve as a state bank, a national investment authority. This was controversial at the time within the Fed, which is why the FRBs did very little 13(3) lending, and there are merits and demerits to idea of the Fed serving as both a monetary authority and a state investment authority. My point is that Title XI essentially repealed Glass’s bill in so far as it empowered the Fed to serve as a national investment authority and replaced it with a special discount window for nonbank financial firms. We agree that the CARES Act reverses course — it specifically contemplates Fed lending to the real economy — what’s odd is that it does so without removing the 2010 restriction that requires the Fed’s 13(3) lending to be “for the purpose of providing liquidity to the financial system.”
Needless to say, some of these interpretive questions are complex, and I would be very much interested in hearing more about other ways to read the statutes. I am currently finishing up a paper tackling some of these issues, which I will be sure to link in a future reply.