The Discount Window Is a Stigma

A central bank's oldest function is to lend when nobody else will. A solvent institution that cannot raise cash today fails for reasons that have nothing to do with its assets.

The authority to make those advances exists precisely so that a temporary shortage of liquidity does not become a permanent insolvency.

The design is sound and it has one dependency that no statute can supply, which is that an institution needing the facility must be willing to use it.

They are frequently not. Borrowing is read by counterparties as evidence of distress, and an institution suspected of distress finds its other funding withdrawn faster than the advance can replace it.

So the facility is least used at the moment it is most needed, which is a failure mode built into the signalling rather than into the mechanism.

The rate is the lever most often reached for and the least effective one, because the cost of borrowing is not what deters the borrowing.

The institution that operates it has worked at this problem for decades: renaming the facility, restructuring the rate, adjusting who may borrow and on what terms.

An institution facing a run will pay almost any rate for cash. What it will not do is accept being seen paying it, and no rate setting changes that calculation.

Each change helps at the margin and none removes the underlying inference, because the inference is correct. An institution borrowing at the window generally does need to.

The comparison with the deposit guarantee is instructive. That instrument works because it is automatic and requires no institution to admit anything.

The disclosure question is the sharpest version of the trade-off and it has no clean answer.

A guarantee that operates silently and a facility that requires an application are different technologies for the same problem, and only one of them is run-proof.

Publishing who borrowed makes the stigma worse and makes the facility less usable. Not publishing means a public backstop is extended without the public knowing to whom.

Both are real costs. The compromise that has emerged, disclosure after a delay, is defensible and satisfies neither concern fully.

The governing regulations set out the collateral and eligibility framework, and the framework is where most of the practical constraint sits.

Collateral is the part that decides whether the facility works for a given institution. An advance is secured, and assets that cannot be valued quickly cannot be pledged quickly.

An institution holding exactly the assets that have become hard to value is the institution most likely to need the window and least able to use it.

Pre-positioning collateral solves this and costs money in normal times, which means it is done by the institutions least likely to need it and skipped by the ones most likely to.

That inversion is the same one that runs through the deposit limit and through readiness spending generally: the preparation is cheapest for those who will never draw on it.

Supervisors have pushed institutions to pre-position and to test their access, on the reasoning that a facility nobody has ever used is a facility nobody knows how to use.

Operational readiness is a real and underrated part of this. Drawing an advance involves systems and paperwork, and an institution doing it for the first time in a crisis is slow.

The published statements record the eventual outturn, which in most periods is that the facility cost the public nothing, because advances are secured and short.

That record is the strongest argument for using it and does nothing to change the incentive of the institution deciding whether to be seen using it.

The proposal that would work is the least popular one, which is routine borrowing by everybody: if healthy institutions used the window regularly, using it would stop being a signal.

Supervisors have suggested as much. Institutions resist, because being early to normalise something still stigmatised means bearing the stigma first, alone, for the benefit of everyone later.

Which is a collective action problem with an obvious solution and no obvious way to reach it, and it is the reason a facility that works on paper keeps failing in practice.