4.1The claim
Chapter II of Lombard Street separates two things that a balance sheet runs together: the quantity of assets a banker holds against liabilities, and the power of those assets to pay on the day payment is demanded. Bagehot shows that securities a banker reckons as reserve can pay in ordinary times and in a panic confined to that banker, and cannot pay in a general panic, because everyone else is selling the same thing. The point is made about Consols, the British government's perpetual debt, and about the Bank of England's own securities. The record Co-Deterioration of Collateral Assets makes the same separation for collateral posted in a margin account. The case adds to the second side of the debate. It shows that the dependence of a buffer's value on general liquidity, which Bookstaber's account places in modern innovation, was stated by Bagehot, on the first side, about the oldest and safest security in the London market.
4.2Promises to pay, and the time to pay them
Bagehot begins his general view with a definition that places time at the centre of credit:
The main point on which one system of credit differs from another is 'soundness.' Credit means that a certain confidence is given, and a certain trust reposed. Is that trust justified? and is that confidence wise? These are the cardinal questions. To put it more simply--credit is a set of promises to pay; will those promises be kept? Especially in banking, where the 'liabilities,' or promises to pay, are so large, and the time at which to pay them, if exacted, is so short, an instant capacity to meet engagements is the cardinal excellence.
The phrase "an instant capacity to meet engagements" already holds the distinction this case is about: a banker might own enough to cover every promise eventually and still lack the instant capacity. Bagehot draws the next line at once. Not all of a bank's cash is reserve:
It is to be observed that properly speaking we should not include in the 'reserve' of a bank 'legal tenders,' or cash, which the Bank keeps to transact its daily business. That is as much a part of its daily stock-in-trade as its desks or offices; or at any rate, whatever words we may choose to use, we must carefully distinguish between this cash in the till which is wanted every day, and the safety-fund, as we may call it, the special reserve held by the bank to meet extraordinary and unfrequent demands.
So there are three things in view: what the bank owns, the cash it uses every day, and the safety-fund kept for the extraordinary day. Only the third is the reserve. A margin account has the same three layers: what the borrower owns, the margin needed to keep positions open, and the headroom above that margin, which is all that stands between a bad day and a forced sale.
4.3Sound, and still unable to pay
The sharpest statement of the distinction comes where Bagehot explains why no one in London feared for the Bank of England, even though it had needed help three times since 1844:
But still there is a faith in the Bank, contrary to experience, and despising evidence. No doubt in every one of these years the condition of the Bank, divided or undivided, was in a certain sense most sound; it could ultimately have paid all its creditors all it owed, and returned to its shareholders all their own capital. But ultimate payment is not what the creditors of a bank want; they want present, not postponed, payment; they want to be repaid according to agreement; the contract was that they should be paid on demand, and if they are not paid on demand they may be ruined. And that instant payment, in the years I speak of, the Bank of England certainly could not have made.
Two senses of soundness separate here. In one, the Bank was sound: it could ultimately pay all it owed. In the other, it was not: it could not pay on the day, and the creditors' contract was written in the second sense. A holder can be solvent in the first sense and ruined in the second at the same moment.
4.4Consols on a Sunday
Bagehot then turns to the assets that bankers counted as reserve though they were not cash. His example is the most liquid security of the age:
The case is just the same if the banker wants to sell Consols, or to call in money lent on Consols. These he reckons as part of his reserve. And in ordinary times nothing can be better. According to the saying, you 'can sell Consols on a Sunday.' In a time of no alarm, or in any alarm affecting that particular banker only, he can rely on such reserve without misgiving. But not so in a general panic. Then, if he wants to sell 500,000 L. worth of Consols, he will not find 500,000 L. of fresh money ready to come into the market. All ordinary bankers are wanting to sell, or thinking they may have to sell. The only resource is the Bank of England. In a great panic, Consols cannot be sold unless the Bank of England will advance to the buyer, and no buyer can obtain advances on Consols at such a time unless the Bank of England will lend to him.
The passage sorts three situations. In ordinary times the security is as good as cash. In an alarm confined to one banker it is still as good as cash, because the rest of the market can buy. In a general panic it is not, because the rest of the market is trying to sell the same thing. Nothing about the security has changed between the second situation and the third. What has changed is who else needs to sell. The quantity of Consols in the banker's book is the same on Saturday and on Monday; the power of that quantity to pay depends on whether everyone else has the same need on the same day.
This is close to what Bookstaber means by liquidity vanishing when it is most needed (Bookstaber 2007). Bagehot is specific about the mechanism: the fresh money that would buy the security is held by people who are themselves "wanting to sell, or thinking they may have to sell".
4.5The directors' dream
The same reasoning applies, Bagehot argues, to the Bank of England itself. Its directors believed that a large holding of securities made the Banking Department safe even when its cash was nearly gone:
Both in 1847 and 1857 (I believe also in 1866, though there is no printed evidence of it) the Bank directors contended that the Banking Department was quite safe though its reserve was nearly all gone, and that it could strengthen itself by selling securities and by refusing to discount. But this is a complete dream. The Bank of England could not sell 'securities,' for in an extreme panic there is no one else to buy securities. The Bank cannot stay still and wait till its bills are paid, and so fill its coffers, for unless it discounts equivalent bills, the bills which it has already discounted will not be paid.
The dream had two parts. The first was that securities could be sold for cash in a panic. The second was that the Bank could wait for its bills to be paid. Bagehot rejects both on the same ground: the Bank's assets are claims on a market that will pay only if someone keeps lending into it. Dependent banks, he adds, count on the ultimate banks to discount more than usual at such moments; "ultimate banks have no similar rear-guard to rely upon." The ultimate bank's securities are worth what it can borrow against them, and it is the lender of last resort to everyone who might lend.
4.6The record: collateral that falls with the position
WSTiger's record Co-Deterioration of Collateral Assets is a synthetic setting in a proprietary trading system. Its scenario states one assumption, lists the evidence its judgment requires, and tags the setting:
Assumption: the value of the collateral in its reference currency or the recognized ratio declines together with losses on held positions. / Required evidence: reference currency and recognized ratio by collateral type · concurrent movement with position losses · actual availability of alternative collateral / Tags: prop system · collateral risk · co-deterioration
Two mechanisms are named in the assumption. The value of the collateral in its reference currency can fall, as when a security priced in one currency is counted in another. The recognised ratio can fall, which this article takes to mean the share of the collateral's market value that the lender agrees to count towards margin. Either way, the same quantity of collateral counts for less. The question the record asks is direct:
When the collateral assets pledged alongside a position also decline as the position incurs a loss, how does the risk change?
It offers three options: Maintain Based on Collateral Quantity; Stress Reflecting Co-Deterioration; and Prepare Alternative Collateral. Its example judgment, which WSTiger marks as its own dated reading, is this:
Reflect losses on positions and declines in collateral value at the same time. The fact that the collateral quantity is unchanged must not be interpreted as meaning that available margin is maintained.
The second sentence of the judgment is Bagehot's Consols passage turned into a rule of account. The banker's Consols were the same in quantity on the day of the general panic as on the day before; their power to pay had gone. The record's collateral is the same in quantity after the loss as before; the margin it supports has shrunk. The first option, Maintain Based on Collateral Quantity, is the position of the banker who "reckons as part of his reserve" what can be sold only while others are buying. The judgment rejects it in the same terms as Bagehot's rejection of the directors' dream: counting what is held is not the same as counting what it will pay.
4.7What the record adds and where it stops
The record adds one thing to Bagehot's analysis and leaves out another. What it adds is the coupling between the loss and the buffer within one account. In the Consols passage the security loses its usefulness because of what other bankers are doing; in the record the collateral falls "together with losses on held positions", so that one movement drives both. That is closer to Bagehot's account of the Bank of England in Chapter XII, read in Case three, where a single drain both empties the reserve and raises what must be lent.
What the record leaves out is the market beyond the account. Its third option, Prepare Alternative Collateral, assumes that alternative collateral can be found, and the record lists "actual availability of alternative collateral" as required evidence. It says nothing about where that collateral would come from if every account were short at once. Bagehot's answer, in the Consols passage, was that "The only resource is the Bank of England". The record names no such resource.
This case therefore adds to the second side of the debate evidence from Bagehot, a text placed here on the first: a reserve counted in securities is only as good as the market's ability to buy them, and that ability fails in a general panic. The record writes the same separation of quantity and value into a private margin account, and stops at the edge of the account.