Any attempt to strengthen the role of the Fed as lender of last resort is untenable in the current political environment. That is a fact of life. But apart from politics, what should one do to strengthen the Fed’s role? The beauty of the power of a strong lender of last resort is the power would never have to be used because runs would be deterred by the knowledge that the Fed would do what it took. Stopping runs before they occur avoids the serious problems of lending into a run. At the same time, a strong lender of last resort should not be one operating under an ill-defined framework. It is important to establish a detailed LLR framework, so that the legally permissible actions of the Fed are not unbounded. Further, a clear framework, which includes specification of actions that might be taken in a contagious panic could well forestall the panic in the first place—this is the essence of the “do what it takes” approach of the ECB.
It should also be noted that some have raised the question of whether the Federal Reserve’s large balance sheet may impede its lender-of-last-resort capabilities in the future. This is unlikely to be an issue since the Federal Reserve’s balance sheet expansion is temporary and it could offset any new LLR lending (which would further increase the balance sheet) by selling other assets, thereby keeping the LLR activities balance-sheet neutral.
The obvious first step for reform would be to restore the Dodd–Frank changes to Section 13(3), including the Fed’s independence from the Treasury. But as former Secretary Geithner has written, these powers were barely sufficient to deal with the crisis. What more could be done?
First, we should follow the lead of other central banks by dividing LLR powers between normal times through the discount window and special times through emergency procedures, rather than dividing our lending authority up between banks and nonbanks. In the modern financial world lending to banks and nonbanks should generally be treated the same, while recognizing the need for accurate solvency information will be more difficult to obtain for unsupervised nonbanks.
Second, there should be a more flexible and clearer policy about collateral for emergency assistance. The effectiveness of lender-of-last-resort liquidity in stemming risks of contagion hinges on two key factors: (1) financial institutions must have adequate collateral to access central bank liquidity and (2) short-term creditors must have confidence in the collateral adequacy of the financial institution and certainty that the central bank will indeed lend against it. With respect to collateral, it is important that the central bank accept a broad range of securities as adequate collateral. More risky and less liquid collateral can be subject to haircuts that reflect their riskiness. Requiring that banks publicly disclose their levels of adequate collateral would serve to increase the confidence of short-term creditors that their borrowers could get access to Fed liquidity. The Committee on Global Financial Stability advocates for a requirement that banks provide regular, standardized disclosures on the quality and amount of their collateral as well as the extent to which any assets are encumbered.1 Disclosures related to asset encumbrance and collateral adequacy should serve to improve market discipline, since unsecured short-term creditors would be in a better position to evaluate the riskiness of the bank’s short-term debt.2
A crucial issue, however, in collateral adequacy is the dynamic nature of the valuation of collateral; that is, an asset’s value or riskiness may change dramatically, thus compromising its usefulness as collateral. As noted by Federal Reserve Bank of Boston President Eric Rosengren, “widespread questions about the appropriate valuation of collateral during the crisis made it apparent that collateral solutions in and of itself was not sufficient to avoid runs.”3 Ultimately the valuation of collateral will be determined by the Fed, likely with help from an outside contractor (e.g., BlackRock’s role during the prior financial crisis). Given that Fed lending is likely to restore asset prices to their ex ante value, the Fed should make it clear in advance that collateral should be valued under “normal” market conditions without taking into account the effect of fire sales.
A third reform would be to abandon the ad hoc nature of decision making about when to lend by adopting an advance commitment approach. Solutions focused solely on the collateral of financial institutions do not address the ad hoc nature of central bank lending and the uncertainty of creditors as to whether the Fed will actually lend to their institution. As an alternative, the Federal Reserve could provide an explicit advanced commitment to lend against adequate collateral. Under an explicit contract, short-term creditors can have confidence in the availability of central bank lending in the face of liquidity problems. An advanced commitment by the Fed would specify the relevant details of lender-of-last-resort support, such as adequate collateral, penalty rates, and term, before the onset of a crisis. By removing most of the uncertainty of whether a financial institution will receive liquidity from the Fed (it might still fail to come up with agreed collateral), short-term creditors would be able to rely on advanced commitments. This is akin to the approach of the ECB in laying out the conditions for its assistance under its September 2012 Outright Monetary Transactions Policy, the “do what it takes” policy.4 The power of the Fed to stop contagion would thus be greatly strengthened. This approach would, of course, clash with the doctrine of “creative ambiguity.”5 The main idea behind the “creative ambiguity” concept is that to mitigate moral hazard concerns the central bank should be ambiguous in regards to its liquidity provisions, so that a financial institution cannot always expect that the central bank to provide liquidity when it is needed. However, while “creative ambiguity” may serve to limit moral hazard by introducing uncertainty, it is precisely the uncertainty that exacerbates the problem of contagion. A BIS workshop discussing the lender of last resort with representatives from the world’s major central banks concluded that “ambiguity had impaired the effectiveness of LOLR actions during the crisis,” by making it harder for central banks to respond and making banks and markets more pessimistic about the central bank’s ability to lend.6 However, some participants suggested that ambiguity in lender-of-last-resort policies was desirable for nonbanks, but not for banks. It is unclear why ambiguity would hinder a central bank’s actions for banks but not for nonbanks, particularly given that some major “banks” today are bank holding companies in name but broker-dealers in practice.
In general, academic theory suggests that while “creative ambiguity” does have merits in certain cases due to the reduction of moral hazard, the systemic risks associated with contagion of large financial institutions outweigh the systemic risks associated with moral hazard.7 While “creative ambiguity” may have a limited role during normal times, there is no place for such a policy during a crisis. However, given that regulators will likely not identify a crisis until it has already occurred, waiting to provide clarity may be too late to combat contagion. A summary of the September 2014 BIS conference on lender of last resort found that “many participants expressed the view that ambiguity with respect to provision of emergency liquidity to banks was not always constructive. Given the scale of liquidity stress, there were doubts that ambiguity had helped reduce moral hazard before the crisis.”8 Further it is hard to see how moral hazard is created by lending to victims of panic runs, as compared to financial institutions that have become insolvent due to their own bad decisions.
A program of advanced commitments to lend by the Federal Reserve does have historical precedent. In 1999, concerns of Y2K computer glitches led many financial institutions to limit exposure to these potential Y2K risks by planning to reduce trading volume. Foreseeing the liquidity problems that would be created by a reduction of trading activity the Federal Reserve created the Standby Financing Facility that “would provide securities dealers with a form of backup funding and ease market anxieties about year-end credit conditions.”9 This liquidity facility allowed primary dealers in government securities to buy options on temporary repos, which gave the contract holder the right to arrange a one-day repo with the Fed for $50 million at a price of 150 basis points over the federal funds target rate (a price determined through an auction).10 Secondary trading of the liquidity options was not permitted.11 The Fed’s auction of the options was successful as demand surpassed expectations. Even though a disruptive trading event never materialized as a result of Y2K and no dealer exercised their option, an argument can be made that the introduction of the Standby Financing Facility averted market disruptions.12 Not only did market repo rates decline substantially after the Fed announced the details of the facility and the strong results of the auction, but according to members of the New York Fed’s Markets Group, “many dealers indicated that the options program helped ease their anxieties about prospective market conditions around year-end.”13 This result again underscores a central point about lender of last resort—strong powers may never have to be used because they will deter runs in the first place.
While the Fed’s Standby Financing Facility does provide a framework for considering advanced commitments, most potential liquidity crises will not be as easily foreseen. Therefore it is crucial that a lending policy be outlined in advance. A policy paper released by the Center for Financial Stability proposes an implementation of advance commitments that is very similar to the liquidity options created through the Y2K Standby Financing Facility. In the paper, Bruce Tuckman introduces Federal Liquidity Options (“FLOs”) as an advanced commitment mechanism.14 Under this proposal, the Federal Reserve would auction off a limited number of FLOs to nonbank financial institutions at market-determined prices.15 The FLO would give the holder an option to borrow money from the Fed “on a secured basis at a predetermined rate and under prearranged collateral terms.”16 This proposal to use FLOs as the exclusive means of providing liquidity to nonbank financial institutions has several advantages. First, moral hazard should be minimized, since an effective FLO program would allow the Federal Reserve to credibly dismiss the possibility of further liquidity provisions or bailouts in the event of a crisis.17 In addition financial institutions that borrow from the Fed through an FLO would be required to pay for this privilege through market-determined prices, thus “forcing market participants to internalize the cost of lending of last resort ….”18
Three main issues arise regarding the potential effectiveness of the FLO proposal in stemming contagion. First, it is unclear why such a policy should only be adopted for nonbanks, given that contagion in the banking system is an increasing concern given the rising proportion of short-term uninsured liabilities, a point covered in more depth in the next chapter of the book on deposit insurance and guarantees. Second, an appropriate amount of FLOs may not be held by institutions that are experiencing a run, thus limiting their effectiveness in protecting against contagion. Purchase of the liquidity options would not be mandatory as they are sold at auction and completely discretionary. Seen in this light, such options should be additive to ad hoc powers, and not be the exclusive means of lending. The third issue of concern with FLOs is the collateral management. Since FLOs can only be exercised with adequate collateral, FLOs are only valuable, and hence effective in stemming contagion, if the holder of the FLO also has a sufficient amount of collateral. In terms of providing a solution to contagion, the question remains whether collateral management should be left to the individual FLO holders or if the Fed should impose collateral adequacy policies in conjunction with the FLOs. Proponents of the FLO program firmly believe that collateral adequacy will not be a major concern since financial institutions will likely exercise FLOs to replace lost funding that already required the same collateral. In fact the FLO proposal recommends that adequate collateral for the FLO should be the same as the collateral used in the tri-party repo market.19 The lack of adequate collateral could be further addressed generally by the enhanced collateral policies suggested above.
A fourth reform of LLR could be consideration of opening lender-of-last resort facilities to MMFs, given the dangers of MMF runs. Central bank injection of liquidity can, of course, be provided indirectly using banks as a third party to facilitate the lending. In this manner banks borrow directly from the Fed and pass the liquidity on to money market funds, using the MMF assets as collateral for the Fed loans, as was done in the 2008 crisis through AMLF, in which the Federal Reserve extended indirect access to the discount window to money market funds through a $150 billion facility.20 But money market funds could also potentially borrow directly from the Fed. Former Chairman Bernanke indicated a possible willingness to provide central bank loans to money market funds during a future crisis.21 However, any lending to money market funds faces an initial obstacle in the 1940 Investment Company Act, which limits the amount of leverage an MMF can incur.22 An SEC exception would be necessary to exempt loans from the central bank in regards to any leverage limits on money market funds for lender-of-last-resort liquidity to be a viable solution to MMF contagion. Another potential obstacle that must be overcome is the ownership structure of money market funds. Since MMF investors technically own the MMF assets, the MMF does not have ownership of the collateral that must be used to borrow from the Fed. A solution to this obstacle could be as simple as a requirement that investors agree to subordinate their claims to the Federal Reserve loans.
One reason that broadening lender-of-last-resort facilities to include MMFs may work is that MMFs are regulated by SEC, thus subjecting funds to regulatory oversight, much like the Federal Reserve regulates banks in conjunction with the discount window. In the unlikely case that a MMF needs liquidity support the reporting requirements of MMFs will allow regulators to determine whether the fund is insolvent; indeed, given the nature of money market fund assets, such determination will be easier than for banks. Insolvent funds would not be permitted access to lender-of-last-resort liquidity and would be allowed to fail, similar to insolvent banks. However, a concern arises with the SEC’s MMF reforms discussed later, in particular, the imposition of a floating NAV requirement, which could result in regulatory arbitrage. The potential flight of MMF funds to unregulated investment alternatives, including the “enhanced cash fund” market and separately managed accounts, or to private unregulated funds, as contemplated by BlackRock,23 would make it much more difficult to deploy lender-of-last-resort liquidity to where it is needed most.
This raises a fifth point of reform. A significant problem in the Fed lending responsibly to nonbanks is that it is not the supervisor of its potential borrowers and thus is not able to make an informed decision about their solvency. In Japan the BOJ has examination authority over any borrower. This ensures that it may examine potential borrowers that come for loans. While other central banks may not have the same statutory power, they could impose an examination requirement as a condition for making loans. But in Japan and elsewhere last minute examinations of unsupervised institutions would be problematic. This could be remedied by giving the central bank supervisory authority over any significant financial institution, but this is impractical, particularly with respect to firms that are generally unregulated for solvency, like hedge funds. At least the Fed should have free and complete access to supervisory information collected by other regulators that bears on solvency. Our fragmented regulatory structure contributes to Fed’s incomplete picture of the financial system.24 The lender-of-last-resort capabilities of the Fed would be strengthened by better regulatory coordination and collection of information, without necessarily increasing the regulatory authority of the Fed.
Additionally we should give the Treasury the power to direct the Fed to make loans to certain financial institutions, again whether banks or nonbanks, as is the case in the United Kingdom and Japan. Former Secretary Geithner does have a valid concern over whether the Fed will always want to loan in all cases it should. Consideration could be given to whether such direction should come with indemnity for losses as is the case in the United Kingdom but not in Japan.
Finally, it would be a good idea for the Fed to set forth a general policy of how it would conduct its lender-of-last-resort policy. The more the market knows about what its policy is—albeit that any policy can be altered under unanticipated circumstances—the less it will run out of ignorance. Elements of such a policy could well include setting forth the facilities that may be used to fight contagion (basically those used in the crisis), indicating that, in principle, it should only loan to solvent institutions, with due recognition that determination of insolvency is difficult, and that fire-sale prices of collateral will be disregarded in determining the value of collateral. Such policies would be similar to those contained in the current UK framework.
As is repeatedly pointed out in this book, none of these recommendations are politically feasible—all that can be done for now is to foster discussion.