Central-Bank Liquidity Is a Bridge, Not a Repair Shop for a Bad Balance Sheet.

The Bridge Has a Tollbooth

At some point in a funding event, a private institution discovers that the market has stopped asking what its assets will earn over time.

It has begun asking whether the next payment will be made before lunch.

This gets called a liquidity problem, which is useful, provided nobody then uses the phrase as a small handbag into which every unattractive balance-sheet fact may be quietly placed.

A central bank can do something unusually powerful here.

Its liabilities settle payments.

It can take eligible collateral and create a secured claim on itself: reserves, cash, the settlement balance that lets the borrower meet an obligation which has acquired such urgent manners.

In the Federal Reserve’s standing repo operations, eligible counterparties obtain liquidity against Treasury securities, agency debt and agency mortgage-backed securities.

That is a particular public balance sheet accepting particular assets for a particular purpose, which is limiting upward pressure in overnight funding rates.

It is not an announcement that all securities have become delightful.

The neat version is a one-day repo.

The counterparty sells securities today and agrees to repurchase them tomorrow, and the difference between the two prices implies the financing rate. Repayment reverses the entries, with interest, the following morning.

One may call that a bridge.

One should not call it a new kitchen.

That distinction is the first casualty when a funding run gathers speed. The public sees a central bank lend and assumes a rescue. The borrower sees cash arrive and would, in the circumstances, rather like to make the same assumption.

The operations people see a secured facility, an eligibility schedule, a valuation, a maturity, a legal agreement, a payment cut-off, and somebody holding an increasingly interesting telephone.

All three descriptions are correct.

Only one of them says anything about where the loss went.

Eight Questions Before the Key Is Handed Over

The lending window has eight tests, every one of which looks fussy until the first is omitted.

Eligible counterparty.

Eligible collateral.

Valuation haircut.

Rate.

Term.

Stigma.

Disclosure.

Solvency assessment.

That is the whole lending window.

It is not a buffet from which an emergency committee helps itself to the congenial bits.

Eligible counterparty asks who may come through the door.

Eligible collateral asks what may come with them.

The Federal Reserve requires discount-window loans to be fully collateralised to the satisfaction of the lending Reserve Bank, with haircuts set so that collateral value exceeds the loan.

That short sentence contains a great deal of adult supervision: valuation frequency, concentration limits, legal perfection, margin calls, and the uncomfortable fact that a nominal pile of securities is not a matching quantity of lending capacity.

The haircut is where market liquidity turns into a lending number. A security priced at $50 million with a 4% haircut supports a $48 million advance.

The retained $2 million is a cushion for a price fall, bid-ask cost, delay, and legal or operational uncertainty.

Raise the haircut to 10% and the same security supports $45 million. The borrower now has to find $3 million of cash or eligible collateral simply to stay where it was.

Its balance sheet has not changed character because a committee gave the haircut a sterner expression. Your usable funding has changed, which is generally the more immediate social fact.

Rate and term decide whether the bridge is usable, and for how long. A deposit rate can put a floor under overnight rates.

A lending or repo facility can restrain upward pressure, because an eligible counterparty may borrow against securities instead of paying whatever desperate price the private market has started to quote.

But borrowing for one week against an asset held for five years is repeated lender consent, dressed as a funding strategy.

The facility has bought time.

It has not taught the asset to mature sooner.

Then come stigma and disclosure, the two tests everyone treats as soft furniture until they find themselves sitting on them.

Using a facility tells depositors and counterparties something.

Disclosure tells them more, or tells them later.

The Bank of England releases policy-meeting transcripts and briefing materials after eight years, which is a serviceable reminder that public accountability and immediate market transparency are not the same appliance.

A lender-of-last-resort facility has to be legible after the event. A design that turns every use into an instant signal of infirmity will find its most useful borrowers taking great care not to be seen anywhere near it.

The final test is solvency.

It belongs first in your private conversation and last in the approval memorandum, that being the only arrangement likely to make anybody read it twice.

A liquidity facility is justified where a sound asset can survive a solvent valuation and temporary funding stops a funding spiral from choosing the outcome.

If the valuation does not survive, the facility has stopped being a loan against collateral. It is now deciding whether public money will absorb a loss.

Which is a different decision, and one usually followed by a meeting involving the fiscal authority.

The Market Is Not a Liquidation Value With Better Branding

The difficult admission is not that collateral has a price.

It is that the price on the screen may have very little to say about the price of a sale large enough to matter.

Market liquidity is scale and time.

It is not a compliment bestowed upon an asset class.

At a reference price of $100.00, buyers may want two thousand units at $99.98 and another eight thousand only at $99.80.

A small sale executes near the first level.

A sale five times the size consumes both, and prints a volume-weighted price well below the quote.

The last quote did not lie.

It simply had no plans to finance your whole afternoon.

Which is why a lender’s haircut is not an accounting markdown. It is a statement about expected liquidation value after price movement, bid-ask cost, delay and uncertainty.

When volatility rises or dealer capacity shrinks, demanding more cushion is the rational response. Dealers have capital, risk limits, hedging capacity and funding limits. They do not have a civic duty to warehouse every urgent seller’s inventory at yesterday’s mid.

So they ask for a lower price.

And the borrower discovers that an asset’s market liquidity and its funding liquidity have begun to converse.

The conversation is repetitive.

A loss or a funding withdrawal constrains dealers.

Quotes become less aggressive.

A customer sale has more price impact.

Prices become more volatile and less reliable.

Financiers raise haircuts.

The higher funding demand constrains dealers further.

That is the margin spiral.

The loss spiral adds mark-to-market losses on positions already held, then forced sales, then another decline.

Neither theory is mystical.

It is several balance sheets discovering simultaneously that they cannot all sell the same thing to one another.

The timing asymmetry does most of the mischief.

Lenders can change haircuts in an afternoon.

A borrower may need days or months to sell an asset, raise equity or replace financing.

Ten funds may each plan to sell the same bond within a week, each modelling a modest liquidation discount, while the same aggregate sale produces a punishing one once intermediary inventories and financing limits bind.

Your private spreadsheet can be entirely respectable.

Their common customer is the same finite market.

Ratios Describe a Calendar, Not a Character

Prudential liquidity ratios formalise a useful and deliberately bounded question. Coverage ratios ask whether unencumbered liquid assets cover modelled outflows over a defined stress horizon. Structural ratios ask whether the funding is stable enough over a much longer one.

Both tell a supervisor something material.

Neither tells the supervisor whether euros can be found at 08:00 when the stock is dollars, whether a concentrated counterparty will behave as modelled, or whether the collateral can be monetised before the settlement window closes.

That is not a defect in the ratio.

It is a defect in the meeting, at the moment somebody offers it as a character reference.

Central-bank liquidity belongs precisely where that calendar fails for a system reason rather than a private-business-model reason.

An individual firm can sell early, draw a line, or post collateral ahead of the crowd.

The system cannot do so all at once.

Every cash sale requires another balance sheet to supply the cash or the financing.

A public facility can reduce the need for a fire sale by accepting eligible collateral or supplying settlement balances. It cannot make every asset eligible, every valuation sound, or every funding structure clever in retrospect.

It can, however, prevent the fire brigade from arriving after the furniture has been auctioned.

Who Was Carried Across

Depositors and market counterparties benefit first, when a credible bridge stops a run from becoming the method by which value is discovered.

Payments continue.

Maturing obligations are met.

A counterparty that might otherwise hoard cash or liquidate collateral can wait while a solvent institution converts eligible assets into temporary funding.

That benefit is not a warm feeling about central banking. It is a forced-sale price being denied the chance to choose an institution’s fate before its assets have done their ordinary work.

Solvent institutions needing time benefit too, which is rather the point of the design. A dealer funded through repo may need to finance securities rather than dump them into a thin market.

A bank short of cash may need to borrow overnight rather than sell at a distressed price.

In an ample-reserves framework the interest paid on eligible reserve balances changes the opportunity cost of lending those balances overnight at a lower rate. So the facility works through a network of administered rates, collateral availability, counterparty limits and settlement frictions.

Less romantic than confidence.

Considerably more actionable.

The central bank benefits, in the narrower sense, from possessing a rulebook rather than an improvisation.

The operating framework states who may use a facility, at what rate, against which assets, in which currency, under what documentation, and with what consequence for failing to perform.

A decision record for a liquidity operation should name the statutory power, the maturity, the collateral terms, the risk controls and the reporting forum.

This is not bureaucracy attempting to spoil a crisis. It is how one distinguishes monetary support from a quiet fiscal choice while the coffee is still hot.

Who Paid the Toll

The central bank’s risk framework pays first, in the institutional sense. It must settle eligibility, valuation, haircut, legal perfection, concentration, margining and the terms of the advance.

Its balance sheet takes the secured claim, and with it the risk that the framework understood the collateral considerably better on paper than in a stressed sale.

A collateral schedule is therefore not a piece of administrative decoration. It is a distributional machine with very small print.

Other borrowers pay through policy effects and changed funding conditions. Central-bank rates and facilities alter the opportunity cost of reserves, overnight borrowing, dealer financing and a long chain of private decisions after that.

A move of 25 basis points is a quarter of a percentage point, not a quarter of your borrowing cost. Even so, short-term funding markets, floating-rate reset clauses, deposit offers, loans and refinancings all respond on their own contractual clocks.

A public bridge is not costless merely because its first entry is a repo rather than a grant.

Taxpayers pay only where the distinction is abandoned, and that is quite enough.

Lending against sound collateral at a haircut is one thing. Buying the asset at par when its realistic stress value is a fifth lower is another, and the difference has just migrated to the public balance sheet.

The valuation is hardest exactly when market liquidity is impaired. Which is why the solvency test is both essential and highly vulnerable to being called unhelpful by everybody who would prefer the money now.

Moral hazard is the price paid in advance.

If intermediaries expect an authority to take hard-to-trade assets at last week’s price, they will hold thinner private buffers and carry more liquidity risk. If authorities never intervene in a broad market-functioning failure, individually prudent deleveraging produces avoidable damage.

Which is why the arguments worth having are about eligibility, price, haircuts, duration, transparency and loss absorption. Support is the word people use for the whole subject while they are still avoiding the question of who bears the loss.

The bridge is necessary precisely because private funding can flee before a solvent valuation has time to prevail.

It is dangerous precisely because a bridge built across a hole in capital delivers the public balance sheet to the far side of the loss.

So the honest retrospective is neither that central banks rescue everybody, nor that lending is always a bailout wearing a false moustache.

The facility works when it buys time for value to be realised. It fails when it buys a story about value that is no longer true.

Lend against liquidity stress only after you have asked whether the asset can survive a solvent valuation.