An Industry Perspective on JP Morgan's Discount Window Assumptions

Today, Alt-M features a guest blog post from Brian Nolan, a bank treasury professional and co-founder of Finteum in London, UK, which is creating an interbank market for intraday borrowing and lending.


In a speech on February 6 at the Money Marketeers Club of New York University, Federal Reserve Vice Chair Randal Quarles proposed that banks change their internal liquidity stress testing (ILST) assumptions—that is, their expected resilience to stress events—by assuming that they can obtain funds from the Fed's discount window on the first day of such events. On February 13, George Selgin wrote a post about Vice Chair Quarles' proposal, expressing doubts about the wisdom of Quarles' advice given the stigma attached to discount window borrowing.

Since then Jamie Dimon, Chairman and CEO of JP Morgan and Jennifer Piepszak, Morgan's CFO, have announced their decision to take Quarles' advice, and help shed some of the stigma surrounding discount window borrowing, by incorporating regular use of the discount window into Morgan's stress-planning scenarios going forward. They implied that JP Morgan also plans on taking advantage of Fed daylight overdrafts.

Here I offer a bank treasury perspective on Vice Chair Quarles' proposal and JP Morgan's decision to implement it, focusing on how the challenges of governance and mechanics relate to the discount window stigma and to the standing repo facility. (To be clear, the main challenge with Vice Chair Quarles' proposal is the stigma. The governance and mechanics challenges are secondary and complementary, but worth considering.)

Some Brief Reminders

Firstly, a reminder about the problem that Quarles is trying to solve: Quarles recognizes that the September repo market dislocation was caused partly by banks holding onto large reserve balances. They do so in part because they believe high balances are necessary for day one of their liquidity stress scenarios, particularly ILST. Of course, the more reserves they believe they need on hand, the fewer of them they lend through the repo market. Quarles believes this is a problem, particularly because the Fed is trying to incentivize banks to hold fewer reserves, which will help the Fed reduce its own balance sheet.

Quarles' proposed solution is for banks to assume they will not need high reserves on day one of a severe stress scenario, because they can borrow reserves from the Fed through the discount window instead. Instead of high reserves, banks would hold other assets on their balance sheets to cover their day one stress scenario liquidity needs. They could lend out their remaining reserves in the repo market. If each bank agrees to hold fewer reserves, the Fed can, in theory, reduce its balance sheet.

So, is Quarles' proposal credible? Who should decide—and on what basis?

ILST Requires Bank Boards to Decide

Ultimately, a bank's board of directors decides whether to assume central bank support as part of the ILST scenarios that it uses to determine the size of its liquid asset buffer. The board's decision is informed partly by regulation (especially Regulation YY) and guidance from supervisors, but also by the board's own judgment on sound risk management.

For example, JP Morgan's most recent Liquidity Coverage Ratio (LCR) disclosure says "… the Firm does not view this borrowing capacity at the Federal Reserve Bank discount window and the other central banks as a primary source of liquidity." This statement reflects the risk appetite of the board of JP Morgan, which uses the liquid assets on its balance sheet as its primary liquidity sources for ILST scenarios.

However, if the board of JP Morgan, or any other bank board, decides that the bank should be able to survive a severe stress event without using the discount window, but misjudges how severe that stress can get, they can request additional funding from the discount window as a backup.

While Vice Chair Quarles might believe that the discount window is a potential source of day one inflows for stress events, challenging the long-held discount window stigma will require each individual bank to change its discount window usage and revise its ILST assumptions, so it will take time to enact.

Boards Normally Decide to be Self-Sufficient

Usually, when bank boards decide how much they need in liquid assets to survive a severe stress scenario, they expect to survive as a self-sufficient company and not to rely on funding from the discount window. In fact, banks are expected to rely firstly on their own resources, and secondly on funding from private markets, before thirdly requesting central bank funding.

Vice Chair Quarles highlighted that global central banks and regulators "… expect firms to manage their liquidity risk prudently—to self-insure …". This is also reflected in the international BCBS Principles for Sound Liquidity Risk Management and Supervision: "… [A]n expectation that central banks will provide liquidity support, alongside the guarantees to depositors provided by deposit insurance, could diminish the incentives of the bank to manage its liquidity as conservatively as it should."

Vice Chair Quarles emphasis on banks maintaining their self-sufficiency, even while proposing that they assume discount window access on day one of a stress scenario, is an oxymoron. The idea that banks should be self-sufficient runs counter to the assumption that they will borrow from the Fed at the first sign of trouble. Vice Chair Quarles recognized but did not fully address this oxymoron in his speech. It is another matter for bank boards to consider before changing their ILST assumptions.

Committed Timing Mechanics Create a Credibility Problem

Let's change our focus from the process of changing ILST assumptions to the timing of borrowing from the discount window. According to its official description, the Fed's discount window does not provide immediate settlement of loans. Loans can be requested on the same day against pre-positioned collateral, but the Fed states that "proceeds are posted to the agreed-upon account (usually the borrowing institution's account or its correspondent's account), normally after the close of Fedwire."

Yet a bank U.S. dollar funding practitioner informed me that, when their bank recently tested their access to the Fed's discount window, it received credit immediately rather than after the close of Fedwire. There seems to be a discrepancy between how the Fed explains its discount window timing mechanics officially, and how these mechanics work in practice. Certainly, there is no official commitment from the Fed that borrowing institutions will receive funds immediately.

For banks to credibly assume they can obtain funding from the Fed's discount window on day one of stress events, they will require a commitment from the Fed that it will credit such funding that day, before a pre-determined cut-off time. Ideally, the cut-off should be early enough during the business day so that banks can use the reserves for their expected payment and settlement activity. It is not clear from JP Morgan's announcement whether the board of JP Morgan considered this when it approved the assumption change.

Why Spend Resources on a Discount Window Solution?

Dr. Selgin's earlier post describes the stigma associated with borrowing from the discount window. It is important to note that addressing the observations and recommendations above for governance and mechanics will not remove this stigma. The stigma remains the primary challenge with Quarles' proposal, whereas governance and mechanics are secondary challenges.

Providing assurances that the discount window can be used during stress will not change entrenched attitudes towards it. Discount window borrowing will probably remain an "option E", as a US bank treasurer recently described it. Even if the mechanics of the discount window were changed, and even if the Fed committed to settle discount window loans before an early cut-off time, banks might still not use it, except as a last-resort emergency lending facility, simply because of the stigma surrounding it.

Dr. Selgin proposes that a standing repo facility could provide a lower-stigma alternative to discount window borrowing, if designed correctly. Some of the governance and mechanics considerations mentioned above could inform the design of a standing repo facility—for example, funding should be available early in the day, and for same-day settlement.

Following Quarles' speech, some commentators suggested that the creation of a standing repo facility looks less likely. Creating a new facility requires Fed resources and Quarles stated a preference to use tools that are already at the disposal of the system.

However, enacting Quarles' proposal requires changing the discount window mechanics and changing banks' ILST assumptions. This, in turn, requires significant resources from supervisors and from banks' boards and treasury teams. Most likely, the Fed is already quantifying the resources required to achieve this, comparing them to the resources required to create a standing repo facility, and considering whether the latter represents better value for money.

It may be worthwhile for the Fed to consider a request for comment for specific and actionable proposals for the design and implementation plan of a standing repo facility, after the Covid-19 disruption has calmed again.

Vice Chair Quarles' proposal raises some other interesting questions about the minimum and maximum levels of reserves in individual banks: should banks be required to hold any reserves at all? What are the differences between LCR, ILST and Federal Regulation D? These are complex questions that deserve separate posts of their own.



Finteum is an early-stage company, based in London, creating an interbank market for intraday lending through FX swaps, repo, and securities lending. We have prepared guidance for banks' treasury teams and boards as an extension to this post, which can be obtained from covering (i) implications of Vice Chair Quarles' proposal for resolution planning, (ii) explanation and analysis of the USD system's minimum required reserves, (iii) analysis of the risks and early warning indicators today in US Treasurys, (iv) alternative proposals to reduce risk in the USD repo market and avoid a recurrence of the September dislocation, (v) alternative risk-free benchmarks to SOFR, and (vi) alternative proposals to reduce the Fed balance sheet.