Nobody Wants to Be Seen at the Discount Window
Central banks offer emergency funding against collateral. Banks avoid it because borrowing there signals distress, which defeats the purpose at exactly the wrong moment.
The Facility
The discount window allows banks to borrow from the central bank against eligible collateral, at a stated rate, typically overnight or short term.
It is the operational form of the lender of last resort function. A solvent institution facing a temporary funding shortfall can pledge assets and obtain cash rather than being forced to sell into a falling market.
The entire purpose is to prevent a liquidity problem from becoming a solvency problem through forced asset sales. The facility only works if institutions actually use it.
The Stigma Problem
Borrowing at the window has historically been read as a signal that an institution could not obtain funding in the market.
That inference makes counterparties cautious, which reduces available funding further, which increases the need to borrow. The signal is self fulfilling.
The result is that institutions exhaust every other option first and approach the window only when the situation is already severe, which is the opposite of what a preventive facility requires.
During 2008, banks reportedly paid higher rates in the interbank market rather than borrow at the window, which is only rational if the perceived cost of the signal exceeded the interest saving.
Attempts to Fix It
| Measure | Effect on stigma |
|---|---|
| Delayed disclosure of borrowers | Partial, disclosure still eventually occurs |
| Auction based facilities | Broad participation normalises use |
| Encouraging routine testing | Familiarity reduces signalling value |
| Narrowing the rate penalty | Reduces cost, not the signal |
The auction approach was the most effective response in 2008. By creating a facility where many institutions bid simultaneously, participation stopped identifying an individual bank as distressed. Broad use removed the signal.
Supervisors now encourage banks to pre position collateral and to test their access periodically, so that using the window is an operational routine rather than an emergency step. Uptake has been mixed.
The Collateral Question
Lending against good collateral is the Bagehot condition, and defining good collateral during a crisis is difficult.
Accepting only the safest assets limits the facility usefulness, since institutions holding those assets can usually sell them anyway. Accepting broader collateral makes the facility effective and exposes the central bank to valuation risk on assets nobody else wants.
Haircuts are the mechanism for managing this: lending less than the assessed value, with larger haircuts on riskier or less liquid collateral.
Pre positioning matters operationally. Collateral must be lodged and valued before it can be borrowed against, and an institution that has not done that work in advance cannot access the facility quickly when it needs to. Several 2023 failures involved institutions that had not adequately prepared their collateral.
The Alternative Facilities
Central banks have added standing facilities designed to be less stigmatised, including arrangements allowing borrowing against government securities at par regardless of market value, which addresses the specific problem of unrealised losses on high quality holdings.
These are effective and they raise the question of how far the safety net should extend, since a facility lending against par value on assets trading below par is absorbing interest rate risk the institution took deliberately.
The Bottom Line
The discount window provides emergency funding against collateral and is undermined by stigma, since borrowing signals the distress it is meant to relieve. Broad participation facilities reduce the signal, pre positioned collateral determines whether access is actually available in a hurry, and each new facility designed to avoid stigma extends the safety net a little further.