Conclusion

8.1Contribution

This study has joined the critique of shared risk models (Daníelsson 2002; MacKenzie 2006) to Bagehot's account of the one banking reserve in Chapter VII of Lombard Street (Bagehot 1873, Chapter VII). Its finding is a shift of the danger. A shared measure does not destroy a variety that Bagehot's rule needs; it brings the sameness of a panic forward, with nobody placed to act the other way.

Put again, does a shared risk measure, with a figure of funds usable at once, remove the variety among holders that Bagehot's panic rule needs? On this study's reading it does not, since the rule never rested on that variety. It rests on one store of unused money held apart, and on a rule for its use that the public knows beforehand; the many holders already act alike when it applies. What a shared measure and a usable-funds figure change is the moment at which holders act as one. Each moves the panic's sameness earlier, to the measure's trigger and to every account's own standing prudence.

The contribution has two parts. For the reading of Bagehot, Chapter VII shows a system whose many holders act alike in a panic, a reserve whose proper size was left without an agreed meaning, and a rule whose force lies in its direction and in its being known in advance. For the debate over common rules, the shaking side's shared measure and the steadying side's known rule turn out to be different kinds of rule. They differ in direction, and in whether any holder stands apart from the rest.

The two WSTiger records add the words of a decision. One writes down the shared reaction; the other, the gap between funds owned and funds usable. Neither names who would stand apart.

8.2Limits

The records set the first limit. Both are synthetic settings written for a platform, modern summaries and not archival documents, and the one dated "1996 · synthetic setting" is not any real regulation. They show how the platform writes a decision, not what any bank did. Each judgment they carry is WSTiger's dated reading, not a finding.

The primary text sets the second. Chapter VII is not a neutral history. It is Bagehot's argued account of how the Bank of England kept and used its reserve, with his own figures and his own side in a controversy. It is read here in one public transcription, and this study adds no history from outside that text. Where Chapter VII records a panic, the account is Bagehot's, and a paraphrase of it can lose what the original keeps.

Scale sets the third. Two records cannot stand for a population of decisions, still less of banks, and the figures in the appendix describe a chosen set of synthetic records and the form in which they are written.

The readings of Bagehot through Daníelsson, MacKenzie and Kindleberger are this study's interpretation. Each work is used only as far as this study's account of it allows, so its argument appears in compressed form. Nothing here is advice. No option in either record is judged right, and no bank, desk or supervisor is told to do anything.

8.3Further research

Three lines of work follow. First, the parent study holds other records on shared models and common limits. They could be read for whether any of them names a holder who stands apart, and for whether their revisit lines name a level or a time rather than a general change of conditions. Second, other passages of Lombard Street on the size and the use of the reserve could be read beside Chapter VII's account of the ratio the Bank kept, keeping strictly to what the text says. Third, the question could be put directly to private systems that have no final holder. In such a system, what plays the part of the store that stays unused until the day everyone needs it, and does any decision record name it? A record that answered it would say who acts against the common move.

Appendix: ledger figures used

These figures are copied word for word from the parent study's numbers.md, computed by code from its decision ledger (2026-10-05), and from the ledger entries for the two records. The two coders are two models of different families. The count of 5 of 53 includes the first record only because its Korean text contains 초과; the condition itself holds no number, threshold or time, and the count is not a numeric threshold.

The cases

53 cases.

What the records themselves say

The hypothesis

The ledger entries for the two records

These are readings of single cells, copied from the ledger entries; they are not counts.

The same risk measure adopted by all banks (wstiger:wst:risk-ko-model-monoculture-past):

• A5 (condition for revisiting): one condition, in the field invalidation. The code marked it "observable" (with a number, a threshold or a time), so it is one of the 5 of 53. The code's test is a list of words; the Korean original of the condition contains 초과 ("exceeding"), which is part of 한도 초과 ("limit breach"), and that word is on the list. The condition itself holds no number, no threshold and no time.
• decider: unsettled; one coder read institution-or-state (implied: the supervisory authority recommends), the other team-or-organisation (implied: the banks choose whether to follow).
• bearer: unsettled; one coder read third-parties-or-the-vulnerable (implied: the market at large), the other unknown.
• reason recorded: no (settled, implied).
• Options as both coders read them: "Use the recommended measure as it is" does not reduce the immediate loss and gives up long-term stability and resilience; "Differentiate responses to breaches" reduces the immediate loss and gains long-term stability and resilience; on "Review in-house stress scenarios in parallel" the coders disagree.

Screen Balance and Funds Available for Immediate Use (wstiger:wst:wt-atlas-expansion-b02-p24-q):

• A5: one condition, in the field invalidation; not marked observable.
• decider: unsettled (team-or-organisation, implied; unknown).
• bearer: unsettled; one coder read the-decider (implied: the funds are the decider's own), the other those-nearest-the-stoppage (implied: the operator who needs the margin now).
• horizon: days-to-months (settled, implied). Reversal constraint: easy (settled, implied). Reason recorded: yes (settled, stated).
• Options as both coders read them: "Total-Asset-Basis Calculation" reduces the immediate loss, gains evolvability, and gives up resilience, long-term stability and ethical coherence; "Basis of Funds Available for Immediate Use" does not reduce the immediate loss and gives up evolvability; on "Reflect After Transfer Confirmation" the coders disagree on what it gains and gives up, and agree it does not reduce the immediate loss.