Discount Window Stigma: What’s Design Got To Do With It?

Since the bank failures of 2023, Fed policymakers have consistently communicated the importance of bank readiness to use the discount window in times of liquidity stress. We saw that neither SVB nor Signature Bank had made preparations to use the window as a contingent source of funding before their depositors ran. While these banks had more fundamental problems that could not have been solved purely by discount window lending, a prompt decision to use the discount window could have limited the fallout for other regional banks.

However, readiness alone is not sufficient; for the discount window to be effective in promoting financial stability, banks must also be willing to use it. The stigma associated with discount window borrowing in the US is well-documented, and is a multifaceted phenomenon, as described in a prior YPFS article.

The Fed’s efforts to redesign and rebrand the discount window in 2003 were explicitly aimed at removing the stigma for borrowing by adequately capitalized[1] banks, while also retaining a source of central bank liquidity to facilitate the recovery and/or resolution of weaker banks. This post examines the performance of that rebranding in light of those goals.

Why is discount window stigma bad for financial stability?

The discount window, like any other central bank tool for liquidity provision, can mitigate systemic risk during periods of instability in the financial system, in three ways:

• By allowing a bank to borrow when needed to meet short-term, unexpected liquidity needs, the discount window creates confidence among depositors and investors that the bank will continue to have access to the funding it needs to meet its obligations—thus mitigating run risk.
• A bank’s access to a liquidity backstop like the discount window avoids the need for the bank to sell assets to generate liquidity—thus mitigating fire sale risk.
• Availability of this liquidity backstop can also mitigate contagion risk, by providing confidence to a broader set of bank depositors and investors to avoid bank runs— thereby averting the spread of stress through banking and capital markets.

The 2003 redesign of the discount window: an attempt to end stigma

It is useful to recall the origins of the current design of the discount window, which features two principal tiers of lending based on a bank’s financial condition.[2] Before the 2003 redesign, the Fed’s discount window also provided liquidity to banks through two tiers: adjustment credit, which provided reserves to adequately capitalized banks experiencing a temporary reserve deficiency they could not fund in the interbank market, and extended credit, which provided longer-term funding to support banks that had lost access to the funds market and needed time to identify alternative funding sources.

Importantly, both types of loans were made at below-market rates. This reflected the view of policymakers in the early days of the Fed’s existence that Bagehot’s principle of lending at an above-market rate would not be practical for the United States, given the wide variation in interest rates across regions and loan types at that time. Additionally, during the 1920s and 1930s, policymakers used the discount window as a tool to ensure that bank lending was supporting productive, not speculative activity.

Given the moral hazard accompanying central bank lending priced substantially below market rates, and the desire to discourage speculative activity by banks, discount window borrowing was accompanied by substantial administrative scrutiny and constraints for decades thereafter, such as the requirement for a borrower to demonstrate it had exhausted all other sources of credit, and a prohibition on selling funds obtained at the discount window credit into the market.

The Fed’s 2003 redesign of the discount window preserved the two-tiered construct, with primary credit lending to adequately capitalized banks, and a secondary credit program for weaker banks that is priced at a higher rate and with higher collateral haircuts than for primary credit. But it adopted a penalty rate approach to pricing credit that was more consistent with the classic Bagehot model of lender of last resort tools. And it featured two notable design changes aimed at destigmatizing primary credit borrowing:

• The primary credit facility was positioned as an integral part of the Fed’s monetary policy implementation framework. The primary credit rate was designed and communicated to the public as constituting a ceiling on the Fed’s monetary policy instrument—the fed funds rate—and was presented as an essential part of the corridor-type system the Fed had adopted to ensure rate control in its implementation of monetary policy.
• Primary credit borrowing involved no administrative scrutiny or constraints—adequately capitalized banks could borrow on a “no questions asked” basis. Indeed, banks were encouraged to borrow when primary credit was cheaper than the prevailing fed funds market rate and arbitrage the difference in rates by selling those funds into the market at the higher rate. In this way, primary credit was envisioned as a tool to contain spikes in the effective fed funds rate that would otherwise have complicated the Fed’s implementation of monetary policy in a “reserves scarcity” framework.[3]

Importantly, the redesign also aimed to restore the discount window’s power as a financial stability tool—by destigmatizing it. As explained in an article by staff at the Federal Reserve Board of Governors previewing the new design,

“Besides serving as a marginal source of aggregate reserves to the market and a backup source of liquidity to sound depository institutions, the discount window can also, at times, serve as a useful tool for promoting financial stability by providing temporary funding to depository institutions that are experiencing significant financial difficulties. The provision of central bank credit can help guard against the sudden collapse of depository institutions by addressing liquidity strains while an institution is making a transition to sounder footing.”

A destigmatized primary credit facility could serve not only as a safety valve to provide reserves to banks on days when open market operations provided fewer reserves than demanded, or when the interbank market was not effectively redistributing reserves to where they were needed, but also as a liquidity backstop that could mitigate the risk of runs, fire sales, and contagion.

Is the 2003 design working as envisioned?

The answer is clearly “No” with respect to primary credit—which remains stigmatized despite the efforts to normalize it. The original 100 basis points level of the penalty rate spread, widespread criticism and misunderstanding of the purpose of discount window lending during and since the Global Financial Crisis, a failure to align primary credit borrowing with supervisory and regulatory requirements governing banks’ liquidity risk management, the operational challenges involved in establishing access and pledging collateral, the presence of the FHLBs as a less-stigmatized source of financing, and Dodd-Frank requirements for the (lagged) public disclosure of loan details (including borrower name) have all played a part in perpetuating stigma. Secondary credit, on the other hand, seems to be playing the role that was envisioned for it as a tool to facilitate weak banks’ recovery and/or resolution, if only to a limited extent.

Secondary credit usage. As shown in Figure 1, secondary credit usage has been small and infrequent to date. Quarterly discount window borrowing transaction data available from Q3-2010 through Q4-2021 reveal that over this period, 288 depository institutions borrowed secondary credit. While the Federal Reserve’s transaction data do not indicate the borrower’s size or charter, it appears from a cursory review of the data that most secondary credit borrowing was done by relatively small depository institutions—mainly community banks and credit unions. The Federal Reserve’s transaction data do not distinguish between test and actual borrowings; using an assumption that loans below $100,000 were done for operational testing purposes, only 5 percent of these transactions appear to have been actual, not test, borrowings.[4]

Figure 1

Secondary credit as a tool for troubled bank recovery. The Fed’s quarterly transaction data for 2010-2021 show 132 instances in which a bank borrowed secondary credit, and subsequently borrowed primary credit at a later date. All but four of these borrowers were community banks (i.e., banks with assets under $10 billion) or credit unions; three were foreign banking organizations, and one was a regional bank. Secondary credit does appear in some cases to be serving as a source of funding for banks that are working their way back to better health, particularly for smaller institutions.

Table 1 — Instances of bank borrowing secondary, then primary credit — Q3-2010 – Q4-2021

Secondary credit as a tool for bridging to bank resolution. Table 2 maps FDIC bank failure data to the Fed’s quarterly transaction data for discount window borrowings, to see whether banks that borrowed before they failed were typically eligible for secondary credit. This mapping reveals 46 depository institutions that borrowed at the window and subsequently failed between Q3 2010 and Q4 2021. As expected, the majority (40) of these institutions were secondary credit borrowers, suggesting that secondary credit has at times played the role envisioned as a bridge to the least-cost resolution of a failing institution. The time span between secondary credit borrowing and failure over this period ranged from eleven days to four and a half years, averaging just short of one year.

Notably, there were two banks that borrowed primary credit before they failed. One, Washington Federal Bank for Savings, failed just a few months after borrowing primary credit; the other, Trust Company Bank, operated for another three and a half years before closure.

Table 2 — Banks that borrowed before failure, Q3-2010 – Q4-2021

What’s design got to do with it?

It is clear that the redesign of the discount window in 2003 was not successful in achieving its goal of destigmatizing discount window lending to adequately capitalized banks, with well-documented adverse outcomes for the stability of the banking system in the spring of 2023. As noted above, there are a number of factors that contribute to discount window stigma. Could the current design of the discount window be another such factor? There is clearly a need for a central bank facility to provide funding to support the recovery or orderly, least-cost resolution of weaker banks, as secondary credit appears to have done at times since the redesign in 2003. But is it possible that the combination of a standing facility for adequately capitalized banks with one that is for weaker banks has muddied the waters and created stigma by association? A future post will look at how liquidity provision is designed in several other central banks, to explore whether there are lessons we might draw to improve the effectiveness of the discount window as a financial stability policy tool.

Special thanks to Anmol Makhija for his excellent research support.

[1] Here I am using the term “adequately capitalized” as defined by the Federal Deposit Insurance Corporation Improvement Act.

[2] This post does not discuss a third type of discount window lending, seasonal credit, which has a very specific purpose of supporting the seasonal liquidity needs of small banks financing agricultural activity and is thus not stigmatized. Seasonal credit represents a very small share of overall discount window borrowing.

[3] “Reserves scarcity” refers to a framework for monetary policy implementation in which the central bank supplies a relatively low level of reserves to the banking system, and adjusts supply as needed to manage the policy rate to its target level. For more details see Afonso, Gara et. al, How the Federal Reserve’s Monetary Policy Implementation Framework Has Evolved, Liberty Street Economics blog, Federal Reserve Bank of New York (2022).

[4] Of the 826 secondary credit loans reported between Q3 2010 and Q4 2021, only 44 were for amounts greater than $100,000.