7.1Bagehot on both sides of the debate
The debate set out in Section 2 places the danger in a panic either at a stage of a recurring sequence, to be met by a lender whose reserve stands outside the fall, or inside the system's own arrangements, where liquidity vanishes when it is most needed. Bagehot's rule, in the form the reading note gives it, belongs to the first side. The chapters read here show that his analysis of the reserve belongs, in large part, to the second.
Each chapter read here states that the buffer falls with the system it protects: unsellable securities and the Bank's own poverty in Chapter II, reserves called in from the brokers in Chapter XI, the twofold drain in Chapter XII. Bagehot's rule is his answer to that fact, and it works only because one holder is placed so that its lending can restore the market in which everyone else's securities have value.
Read this way, the two sides of the debate are less opposed than the table in Section 2 suggests. Stability that breeds instability, and liquidity that vanishes when it is most needed, are the conditions that call for a lender of last resort, and a history of such lenders records what was done when one was called for. Bagehot's chapters hold both: an analysis of how reserves fail together, and a rule for the one holder whose failure would end the system. What the second side adds is the claim that such fragility is produced in good times (Minsky 1986) and by the system's own innovations (Bookstaber 2007). Bagehot's version of that claim is local and institutional: the one-reserve system, he says, was made by privilege and history, and makes credit "more delicate, and more sensitive".
7.2A private system without a final holder
The records describe a proprietary trading system, not a banking system. They name no lender of last resort, and nothing in them suggests one. For such a system, Bagehot's analysis applies and his remedy does not. The analysis says that collateral falls with the position, that common collateral spreads a loss to every position that shares it, and that the defence takes time to work. The remedy says that the holder of the final reserve should lend. In an account, the only reserve is the account's own collateral, and the only lender is the one making the call.
Here the margin call and Bagehot's rule part most clearly. The rule asks the holder of the reserve to act against the instinct to hoard; a margin call is that instinct, sound for one lender in a calm market. What the sum of such sound acts does when the market is not calm, the records do not say. Bagehot's bill-brokers give one historical answer: the bankers who called in their reserves from the brokers, who under a different arrangement "would not aggravate the pressure on the bill-brokers, but relieve it", aggravated it under the one they had.
7.3Who bears the shortfall
Niklas Luhmann's distinction between risk and danger separates a possible loss attributed to one's own decision from a possible loss attributed to something outside one's control; the same loss is a risk to the one who decides and a danger to those who are affected but did not decide (Luhmann 1993). Bagehot's second chapter describes a striking case of the separation. The directors who decided the size of the reserve, he says, were men "who are named by and are agents for a proprietary which would have a greater income if it was diminished, who do not fear, and who need not fear, ruin, even if it were all gone and wasted." The loss from too small a reserve was a risk to them only in a weak sense; it was a danger to every bank and merchant who depended on the Bank.
The two records do not say who bears the cost of a collateral shortfall. In a proprietary system the bearer and the decider may be the same, but the records do not say whether the account belongs to the firm, a trader within it or a client, or who sets the limits they refer to. The parent study's two coders did not agree on the bearer in either record: in each, one read the cost as falling on the decider, by implication from the setting, and the other found no bearer. Across the study, the coders agreed on the bearer in 24 of 53 records (45%). The gap matters more here than in many records, because Bagehot's analysis of the reserve turns on the difference between those who hold it and those who depend on it.
7.4What the ledger shows, and what it cannot
Both records carry a condition for revisiting, as all 53 do, and neither condition contains a number, a threshold or a time that code can detect; across the study, 5 of 53 (9%) of conditions do. The pre-registered hypothesis that few decisions write in advance what would reverse them in words anyone could observe was supported on both readings, with 48 cases for and 5 against (91%), and the two records here are among the 48.
The ledger's test is narrow, and its narrowness is instructive. A condition that asks for "the case in which collateral prices and position prices move adversely together" names a computation, not an observation; code sees no number in it and counts it as not observable. Bagehot's rule of Chapter XII has the form the code would count: a level, a margin, a point at which to act, all in figures. But Bagehot also said that his figures were conjectures, offered for the judgment of others. A record that gave figures would pass the code's test and might fail Bagehot's. The ledger can say that these revisit lines contain no figures; it cannot say whether figures would have been better.
On the options, both coders agreed that none of the three in the cross-margin record reduces the immediate loss; in the co-deterioration record, both agreed that preparing alternative collateral does, and divided on keeping the collateral-quantity basis. The study's hypothesis that the options which cut the immediate loss most often give up evolvability and resilience was supported across the study, 24 cases for and 13 against (65%), but went against in the co-deterioration record, where the option that cuts the loss gives up nothing on either coder's reading. That fits Case one: alternative collateral adds to the buffer, and its cost is the idle money Bagehot's directors disliked.
7.5An alternative reading
One alternative reading takes the records as illustrations of Minsky's hypothesis: positions built in calm times that become fragile when the calm ends. The records are compatible with that reading, but they do not supply what it needs. Neither record describes how the positions came to be built, or what the calm before the loss looked like; both begin after a loss has occurred. Minsky's account, in its note, concerns the movement of finance through the good times. The records are set at the moment the good times end. They fit Bagehot's account of the panic better than Minsky's account of the boom.