Notes from the Vault
Larry D. Wall
Banks profit from providing liquidity and maturity transformation, that is, from using shorter-term deposits to fund longer-term loans. However, such transformation can lead to bank runs in which depositors demand repayment of their demandable and maturing deposits before the borrowers are required to repay the loans. Such risk played a prominent role in the 2008 financial crisis.
In response to these risks, bank regulators have agreed to new global standards to reduce banks' ability to engage in liquidity and maturity transformation, thereby reducing banks' exposure to runs. But these standards may be more effective in pushing liquidity and maturity transformation risk outside the banking system than in reducing the risks such transformation pose to the real economy.
This Notes from the Vault post discusses the economics underlying liquidity and maturity transformation, the new regulations, and some thoughts on the future.
Liquidity and maturity transformation
Liquidity and maturity transformation arises because bank creditors (hereafter called depositors) prefer to have ready access to their funds, but borrowers prefer longer-term funding that better matches their assets. People and businesses with excess funds like to have the ability to spend on short notice, in part to be able to pay for unexpected emergency expenditures such as an automobile or unscheduled maintenance at a manufacturing plant. People and businesses also like to remain liquid to exploit unexpected opportunities that may arise, such as a bargain on a new television or opportunity to acquire another firm.
Conversely, borrowers often prefer to fund their longer-term assets with a longer maturity loan rather than several shorter maturity loans. Relative to short-term loans, longer maturity loans reduce the transactions cost to the borrower of rolling over the loan. Moreover, if something happens and the borrower is not able to roll over a maturing short-term loan, she will default on the loan and likely end up in bankruptcy court.
Banks have historically stood between depositors and bank borrowers, giving each what they want and in return earning maturity and liquidity premiums. That is, the bank pays lower rates on demand and shorter-term deposits while receiving a higher rate from borrowers with longer maturity loans.
In order to earn these premia, one risk that banks take is that deposit rates will climb faster than asset yields when the yield curve shifts up. Banks manage this risk in a variety of ways, including setting limits on maturity mismatches, making longer-term loans that have interest rates that float with market rates (such as loans based on the prime rate or Libor), and hedging the risk with derivatives.
Another risk banks take to earn the liquidity premium is that on some days outflows will exceed inflows. This risk is manageable almost all of the time as banks can ordinarily predict their inflows and outflows with reasonable accuracy. Banks that anticipate net outflows borrow from other banks and the financial markets. Those banks anticipating net inflows lend to other banks.
The danger with this setup is that there will occasionally be days when a bank's outflows greatly exceed inflows. Moreover, the large outflow days are often a result of depositors' concerns about the financial stability of that bank. Such concerns could not only lead to deposit withdrawals but also to other banks cutting back their credit limits to a troubled bank.
One option for a bank facing such a run is to try to sell some of its more liquid assets. The bank will do so even if it thinks the assets are worth more than their current selling price, as failure to honor the withdrawals will result in immediate insolvency. Still, such asset sales can provide only a little time to reassure depositors and are limited by the liquidity of the bank's assets.
Whether a bank deposit run can threaten financial stability depends upon the circumstances. If only one, not too-large bank is being run due to perceived financial distress, that bank may fail or be put into resolution, and the banking system would otherwise continue normal operations. However, if a substantial fraction of the banking system is perceived to be in financial distress, the result could be that not only are the distressed banks at risk of failing but that their sale of assets at below long-run value (fire sales) could make other banks insolvent by decreasing the market value of the assets on their books.
Over time, the systemic problem of bank runs was recognized, and Rutgers University Professor Michael Bordo notes that lenders of last resort (LLR) arose in various countries to provide liquidity during banks runs. Typically, the LLR allows commercial banks to offer their illiquid assets as collateral for a loan that the bank can then use to honor deposit withdrawals. One of the primary reasons for the creation of the Federal Reserve was to create an LLR that would supply an elastic currency to mitigate damaging bank runs.1
The creation of an LLR provides a mechanism for preventing bank runs from causing widespread bank failures, but it does so at a cost. Banks are likely to become relatively more illiquid if there is an LLR to provide funding during runs. Moreover, prior to LLRs, bank runs had been the primary way of forcing insolvent (or severely undercapitalized) banks into resolution. This process could be slowed or prevented if the LLR were to lend to insolvent banks with a variety of adverse consequences. Ex ante, banks will take more risks if they believe that the LLR will lend to insolvent banks. Ex post, such lending may allow insolvent banks to continue to operate.2
The best-known guidance for the provision of credit by the LLR is known as Bagehot's rule. This rule calls on the central bank to lend freely on good collateral but only to solvent parties at a penalty rate.3 This solution discouraged unnecessarily lending and protected the central bank from losses. However, determining whether a bank is solvent and which of its assets are good can be problematic during a financial crisis.4
An alternative to relying on an LLR is to limit banks' ability to engage in liquidity and maturity transformation.5 The risks being taken by banks' liquidity and maturity transformation have long been supervised by the bank regulatory agencies.6 However, formal standards comparable to regulatory capital adequacy requirements are a recent development.
As a part of the Basel III revisions of the capital adequacy rules, supervisors from the developed countries also agreed to establish minimum standards for a liquidity coverage ratio (LCR) and net stable funding ratio (NSFR). The U.S. supervisors adopted a regulation implementing a version of the Basel III LCR in September 2014.7 When fully phased in, the LCR regulation requires large banking organizations to maintain high-quality liquid assets at least equal to their projected net cash outflows over a stressful 30-day period. The regulation implementing the LCR provides detailed rules on what items to include in the measurement of high-quality liquid assets, inflows and outflows, and how these items are to be weighted.
The final Basel III standards for the NSFR were issued in October 2014. The Basel III standards define available stable funding as "the portion of capital and liabilities expected to be reliable" over a one-year horizon. Required stable funding "is measured based on the broad characteristics of the liquidity risk profile of an institution's assets and OBS (off-balance sheet) exposures." The NSFR requires banks to have available stable funding at least equal to required stable funding. Similar to the LCR, the NSFR provides detailed rules for measuring available and required stable funding. The U.S. regulation spelling out the details for U.S. banks' implementation of the NSFR is still under development.
The intent of the LCR and NSFR is to constrain banks' ability to take liquidity and maturity risk in a manner similar to the way capital adequacy regulations limit banks' ability to take solvency risk. As is the case with the capital regulations, banks faced with binding limits on their LCR and NSFR are likely to respond along all of the possible margins, including holding a relatively larger proportion of liquid assets and obtaining longer maturity funding. However, the LCR and NSFR are also likely to be similar to the capital adequacy regulations in that over time banks will find ways of complying with the letter of the regulation while getting around its intent.8
Even to the extent the LCR and NSFR reduce banks' provision of liquidity and transformation services, they only change part of the underlying economics that motivate the provision of these services. That is, the LCR and NSFR will reduce the extent to which liquidity and maturity transformation are undertaken to exploit the underpriced provision of LLR services to banks. However, the regulations do not eliminate the other gains to depositors (or more generally, funds' providers) and to borrowers from such transformation. Hence, we might reasonably expect nonbank intermediaries and borrowers to find alternative ways of engaging in such transformation.
Money market funds (MMFs) are an example of a type of intermediary that already exists in part to provide liquidity and maturity transformation services. MMFs demonstrate that maturity and liquidity transformation can take place on a large scale at institutions with neither a history nor guarantee of direct access to an LLR prior to the crisis.9 The Securities and Exchange Commission recently tightened the regulations governing MMFs. However, experience shows that market participants will seek to develop new institutional arrangements to circumvent regulation.
To the extent that regulation successfully limits other intermediaries' ability to provide liquidity and maturity transformation, financial market engineering may provide a partial substitute. For example, prior to the crisis, municipalities issued longer dated claims called auction rate securities that were regularly remarketed, typically every seven to 35 days. As long as the auctions were successful, the municipality could meet its desire for longer dated debt while investors obtained the benefit of being able to liquidate their claim at a much shorter horizon. Yet while markets can supply liquidity during normal times, they can also break down during crisis. Federal Reserve Board economists Song Han and Dan Li have documented the breakdown of this market during the crisis and the adverse consequences for both borrowers and lenders.
Finally, to the extent that neither nonbank financial firms nor financial markets can meet the demand for liquidity and maturity transformation, depostors and borrowers will have to adjust their positions.
Depositors can supply longer-term funds to banks and receive a higher interest rate. But in doing so, they then take on the risk of being unable to meet emergency expenditure needs or take advantage of unexpected opportunities. Many of these costs would not be as visible as bank runs and corporate bankruptcies would be in a crisis, but in aggregate the costs may nevertheless be substantial.
Borrowers can accept shorter maturity loans at lower interest rates. However, in doing so they must take the risk that they will not be able to roll over their loans. The result could be that rather than banks being at risk of failing due to deposit runs, borrowers will be at risk of failing due to an inability to redeem the debt at maturity. Moreover, the extent to which depositors adjust their position versus the extent to which borrowers adjust will be endogenously determined by market participants. Thus, even if policymakers would prefer that depositors make most of the adjustment, in practice it may be that in normal times borrowers are more willing to adjust.
The recent financial crisis highlighted the riskiness of liquidity and maturity transformation to the banking system and the pressure this puts on central banks to act as LLRs. In response, U.S. bank supervisors have adopted an LCR requirement to limit liquidity risk and are in the process of developing a final NSFR requirement to limit maturity mismatches at a one-year horizon.
However, the LCR and NSFR can at most constrain maturity mismatches within the banking system. Many of the underlying economic reasons why funds' suppliers want liquid funds and borrowers want longer-term loans are not changed by the regulation. Thus, to the extent that the LCR and NSFR are binding, alternative methods of liquidity and maturity transformation will be developed.
When these alternatives encounter liquidity problems, those problems may not pose a systemic threat to the banking system. However, their liquidity problems can nevertheless have a substantial adverse impact on the real economy as occurred with the money market funds during the crisis. Thus, the effect of the LCR and NSFR may not be so much to reduce pressure on the Federal Reserve to act as a lender of last resort during a financial crisis but rather to transfer that pressure toward the establishment of a LLR function for nonbank firms and markets.
Whether this outcome is a net positive for financial stability is debatable. Nonbank firms and markets may engage in somewhat less transformation, as they cannot be as confident ex ante that they will have access to an LLR and may be constrained by other regulations. However, ex post when evaluating how to use its LLR powers, the Federal Reserve will likely have less information about these firms and markets than it would about the banking organizations it supervises.
Larry D. Wall is the executive director of the Center for Financial Innovation and Stability at the Atlanta Fed. The author thanks Paula Tkac for helpful comments on the paper. The views expressed here are the author's and not necessarily those of the Federal Reserve Bank of Atlanta or the Federal Reserve System. If you wish to comment on this post, please e-mail firstname.lastname@example.org.
1 Bank clearinghouses in the United States mitigated this risk prior to the creation of the Federal Reserve. However, banks became much less willing to provide this service after the Panic of 1907, which centered on New York City trust companies that acted like banks but refused to participate in the clearinghouses, according to University of Mississippi Professor Jon R. Moen and Oberlin College Professor Ellis W. Tallman.
2 Additionally, Stephen D. Smith and I noted that even if a bank eventually fails, the delay in its resolution will shift part of the losses from depositors who can run to those depositors and guarantors (including the deposit insurer) who cannot run.
3 See Federal Reserve Bank of Richmond economist Thomas M. Humphrey for a discussion of Bagehot's rule.
4 See a report by Bank for International Settlements economists Dietrich Domanski and Richhild Moessner and Federal Reserve Board economist William Nelson for a review of central banks around the world's experiences as lenders of last resort during the 2007–10 crisis and Columbia Professor Frederic S. Mishkin and Rutgers Professor Eugene N. White for an analysis putting the Federal Reserve's actions in an historical context.
6 The long-standing regulatory concern with liquidity is reflected in the U.S. bank regulatory rating system called CAMELS where the L stands for liquidity.
7 Both the Basel III LCR and NSFR requirements provide some level of discretion to national authorities to reflect conditions in their jurisdiction. Additionally, the standards set by Basel III are intended to be minimums with national authorities having the discretion to set even higher standards. Federal Reserve Board Governor Daniel K. Tarullo described the higher standards set by the U.S. LCR as being "super equivalent" to the Basel III requirements.
9 Admittedly, MMFs later did receive support from the Federal Reserve during the crisis. However, I view this as saying much more about LLRs coming under pressure to provide assistance whenever liquidity problems threaten the real economy, even if those problems arise outside the banking sector.