Liquidity is the second wing of regulatory reform, and one more aimed at preventing contagion than capital requirements. Minimum “private” liquidity requirements (as distinct from “public” liquidity supplied by the Fed) are supposed to assure banks’ uninterrupted holding of a pool of high-quality liquid assets that can be sold (or pledged as collateral) to accommodate a sudden surge of withdrawals by depositors and other short-term debt holders during a serious crisis involving contagion.1 In principle, maintaining sufficient high-quality assets should help financial institutions to withstand periodic instability created by the dependency on short-term funds. However, liquidity requirements are curiously in conflict with the Fed’s monetary policy efforts, insofar that quantitative easing is intended to get financial institutions to hold riskier assets to raise interest rates, whereas liquidity rules require banks to hold assets with low rates of interest reflecting their liquidity.
Initially, liquidity requirements seem to represent a more promising regulatory approach to contagion than capital requirements, since contagion originates in and propagates through runs that are fundamentally liquidity driven. Four primary objections to over reliance on private liquidity requirements should be considered.
First, like capital requirements, the liquidity proposals discussed below (with the exception of redemption restrictions and liquidity requirements for money market funds discussed later in this part) apply only to depository institutions, for short banks. In modern financial panics, as in 2008, contagion spread beyond the traditional banking sector—a fundamental point that is made throughout this book.
Second, the stock of high-quality assets that banks can hold to meet private liquidity requirements is limited by nature. Basel’s proposal, for instance, would require banks to retain sufficient liquid assets to match net cash outflows over 30 days.2 However, it is quite possible that persistent disruption to short-term borrowing markets leading to sustained investor outflows stretching over a longer period could eventually overrun even the strongest portfolio of liquid assets, making it difficult to liquidate even “liquid” assets and forcing financial institutions into liquidating long-term assets to meet incremental redemptions. Short-term creditors of a financial institution subject to such liquidity requirements would thus still have an incentive to exit sooner, while that portfolio was still intact, rather than later, after waves of outflow have exhausted it.
Third, holding assets suited to meeting the purposes of liquidity requirements entails costs to financial institutions and to the economy, since every dollar of capital allocated to low-yielding, liquid, short-term securities is unavailable to finance longer term lending to borrowers. This theoretically lowers the amount of new credit that financial institutions can create and raises the overall cost of capital to the real economy. Further, when combined with the leverage requirement, banks will tend to hold only the minimum amount of liquidity required by regulation. Higher levels of safe assets, with low returns, will be limited by the capping effect of a leverage ratio.
Fourth, securing emergency liquidity to the financial system through limited private liquidity is less efficient than traditional use of central bank lender-of-last-resort authority to provide unlimited liquidity to solvent institutions in emergencies. Indeed private liquidity requirements may even undermine the efficacy of the lender-of-last-resort system if, by selling privately held high-quality liquid assets to meet liquidity needs (before going to the lender of last resort), banks deplete the store of collateral available for pledging to the government in exchange for central bank loans.
For these reasons, private liquidity requirements are both under-inclusive and over-inclusive: under-inclusive because they provide coverage that is limited in amount, do not apply to nonbank financial institutions, and will not always forestall runs by short-term creditors; over-inclusive because they may unnecessarily raise the cost of real economic activities that depend on the intermediation of financial institutions but do not create systemic risk. We now turn to the details of the new liquidity requirements.
The Basel Committee has adopted a new liquidity standard for phase-in at the start of 2015, to be completed by 2019.3 Basel’s liquidity metric, known as the liquidity coverage ratio (“LCR”), requires banks to hold unencumbered high-quality assets sufficient to meet all outstanding 30-day-or-fewer liabilities.4 Financial institutions that achieve compliant LCRs must hold a “stock of high-quality assets” equal to 100 percent or more of their net cash outflows over a 30-day period.5 Maintaining a 100 percent LCR in principle is intended by Basel to enable an institution to use the sale of its own liquid assets to satisfy all potential net outflows during a full calendar month without impairing its capital by selling longer term assets at discounted prices, giving managers and regulators breathing room to devise a comprehensive response to a crisis or to wind down an institution, when necessary.6
Qualifying “high-quality assets” include liquid assets that can immediately be converted to cash equal to their carrying values.7 Among other restricting criteria, qualifying assets must be unencumbered securities with low credit- and market-risk and performance that is not correlated to riskier asset classes. Further they must be exchange-listed, trade in active and liquid markets, and easily be susceptible of valuation.8 Examples of high-quality assets satisfying Basel’s multifactor standard are cash, central bank reserves, marketable securities with 0 percent Basel II risk-weightings, and domestic currency government debt.9 Contractual committed liquidity facilities (“CLFs”) provided by a central bank can also be included as high-quality liquid assets.10 Originally the Basel Committee limited CLFs to jurisdictions that otherwise lacked sufficient liquid assets in the local currency, but it later expanded the permitted inclusion of CLFs more generally.11 The Basel Committee has also released guidance for using market-based factors of liquidity that could cause regulators to drop certain assets from an otherwise “high-quality” classification (e.g., Greek debt).12
The effectiveness of the LCR at meeting demand for liquidity during a crisis depends on making an accurate regulatory judgment ex ante about the required quantity and quality of assets. This judgment involves significant guesswork about the severity of future crises and assumes that assets thought to be of high-quality today will remain so during a period of market dislocation. Likewise, to be effective, the LCR must accurately estimate the 30-day net cash outflow that would arise from a “combined idiosyncratic and market-wide shock.”13 Regulators have promulgated minimum 30-day runoff rates for various liability classes, but have provided little empirical evidence to support these predictions.14 To some extent, these runoff rates are based on what happened in 2008, but the past is not always prologue.
Basel has also proposed a longer term metric called the net stable funding ratio (“NSFR”) designed to secure institutions with enough liquidity support for one year, to be implemented by January 2018.15 The components of “stable funding”16 are capital, preferred stock, other liabilities with maturities of more than one year, plus “stable” deposits.17 All components are discounted by weightings reflective of their relative stability.18 One hundred percent NSFR-compliant institutions must maintain stable funding levels in excess of total assets (both on- and off-balance sheet), weighted according to liquidity and resilience in a period of stress.19 In October 2014, the Basel Committee released its final NSFR standards.20 The Basel Committee included three additional requirements under the final NSFR: interbank loans with residual maturities of under six months must be backed by at least 10 percent of their value in “stable funding”; the initial margin on any derivative contract must be backed by 85 percent “stable funding”; and the ability to offset derivative assets by derivative liabilities was reduced.21 Under the NSFR, regulators are granted discretion in determining whether asset and liability items are sufficiently interdependent to be viewed as NSFR-neutral.22
Beyond LCR and NSFR, the Basel III proposal introduces other measurements oriented at facilitating supervisory monitoring of institution liquidity. Their focus is on maturity mismatching, wholesale funding dependency, and amount of available unencumbered assets. Finally, Basel endorses market-based liquidity monitoring using equity prices and CDS spreads.23 Like the LCR, the NSFR requires regulators to make accurate forecasts about the stability of funding and the quality and liquidity of a bank’s assets. It seems less justified than the NCR, since it is hard to imagine a year-long liquidity crisis for a bank.
The US implementation of the Basel III LCR was proposed by regulators in November 2013, and was finalized in October 2014.24 The US LCR applies to (1) large, internationally active banking organizations, (2) nonbank SIFIs regulated by the Federal Reserve that do not engage in substantial insurance activities, and (3) consolidated subsidiary depository institutions with total assets of greater than $10 billion.25
The US LCR requires all covered organizations to maintain a minimum LCR of 100 percent, calculated by dividing a bank’s high-quality liquid assets by its total net cash outflow amount over a 30-day period.26 It is substantially more severe than the Basel proposal. One of the most significant differences between the US LCR and the Basel III LCR lies in the assumed runoff rates for short-term creditors without a specific maturity date, including uninsured retail and wholesale depositors, the primary source of short-term funding for covered organizations.27 The US LCR assumes that these short-term creditors would withdraw their funding immediately on day 1 (essentially, a single day stress scenario), whereas the Basel III LCR implicitly assumes that these funds are withdrawn at a constant rate through day 30.28 The US LCR calculates the total net cash outflow amount based on the single day within a 30-day period with the highest amount of net cumulative outflows, while the Basel III LCR uses total cash outflows over a 30-day period.29
In order for an asset to qualify as a high-quality liquid asset under the Fed’s rules, it must be liquid and readily marketable, a reliable source of funding in repurchase agreement or sales markets, and not an obligation of a financial company.30 A standard stress scenario would assign specific outflow amounts to different categories of a bank’s funding.31 Importantly, during an idiosyncratic or systemic liquidity crisis, a covered organization would be permitted to convert its high-quality liquid assets into cash as necessary to meet withdrawals by short-term creditors, even if this required falling well below the minimum LCR.32
A Basel Committee Study assuming full implementation of the Basel III liquidity requirements as of December 31, 2013, found that for internationally active banks with over €3 billion in capital, the aggregate LCR shortfall at a minimum requirement of 100 percent is €353 billion, and is €158 billion at a minimum requirement of 60 percent.33 The aggregate shortfall for the NSFR is €817 billion.34 However, “the shortfalls in the LCR and the NSFR are not necessarily additive, as decreasing the shortfall in one standard may result in a similar decrease in the shortfall of the other standard, depending on the steps taken to decrease the shortfall.”35 The Basel III liquidity requirements, combined with a growing demand for safe collateral resulting from an increase in central clearing of derivatives,36 will put upward pressure on the prices of liquid assets and cause further increases in funding costs. The liquidity requirements are likely to be expensive and may ultimately have more impact on bank lending activity than capital requirements.
Whatever the virtue of “private liquidity” it can never be a complete substitute for public liquidity through a strong LLR. Indeed it was the shortcoming of the private liquidity system of the clearinghouses in 1907 that led to the creation of the Fed as a public lender of last resort in 1913. Governor Tarullo has stated that while private liquidity can never be considered a complete substitute, it bolsters the stability of the financial system by giving the Fed a breathing room of 30 days to make a determination of whether to provide liquidity but this heroically assumes that private liquidity would not be exhausted much faster, if runoff rates exceed expectations.37 Further any delay in lending to Bank X by the Fed will incentivize runs on other banks, which the Fed should seek to avoid. And, of course, there could be an immediate run on the shadow banking system that has no liquidity requirements. A recent Federal Reserve paper recognizes this problem by saying that only in the case of runs should the Fed act immediately. The paper thus acknowledges that private liquidity cannot be relied on in a run. Further, if the Fed were to wait 30 days to determine whether a run was taking place and discovered it was, the paper implicitly recognizes that this may be too late. Therefore the breathing space justification seems weak. Furthermore liquidity regulation may decrease lending between financial institutions, thus worsening a weak institution’s options during a crisis. A recent study found that banks have reduced lending to financial institutions as a response to liquidity regulation, but not to nonfinancial institutions.38 Critics of the LCR have also raised concerns that the rule locks up safe debt and increases more risky debt without any reduction of contagion risk.39 William Dudley, President of the Federal Bank of New York, has suggested a willingness to address liquidity concerns through adjustments to regulation.40
The strongest argument for liquidity requirements is a political one, that bank liquidity requirements allow the Fed to say it will only rescue banks from contagion that have exhausted their own resources first. This is probably smart politics but bad policy insofar as the markets may believe it will in fact delay the Fed’s response to a contagious run, thus causing the markets to run earlier and faster. Like the Dodd–Frank restrictions, it seems to put still another limit on the use of the Fed’s power. In the end, private liquidity is not a substitute for public liquidity.41