This article asked of two sets of closure documents whether they said who was shut and for how long. For October 1907, Sprague's account shows closures without public decrees and without terms: a clearing arrangement withdrawn with notice given the day before, a trust company suspended after a run of hours, a restriction of payments decided by postponement, and a single private term, the savings banks' right to notice, held against their own depositors. The boundary that decided who was helped was membership of a private association, and an institution outside it that had met the association's reserve rule was not examined.
For March 1933, the documents resolve the tension between calendar terms and state conditions. Proclamation 2039 established a comprehensive scope and an explicit calendar term running to 9 March, which constituted an observable exit condition of time under hypothesis F1. The proclamation did not, however, formulate an objective state condition, such as a solvency test or reserve ratio. Congressional action subsequently replaced that designated calendar cutoff with an open-ended executive licensing framework under the Emergency Banking Relief Act, appointing conservators whose operating costs constituted a preferred claim on institutional resources while allowing ratable payouts under Section 206. The fireside address presented the staged reopening as an empirical examination, extending reassurance to sound institutions while withholding promises from the rest.
This investigation does not adjudicate between the view of panics as panic-driven runs on solvent enterprises and the perspective emphasizing institutional design and balance-sheet fragility. Rather, it reveals that the closure documents in each episode mark the institutional juncture where authorities deferred the distinction between illiquidity and insolvency, and show who bore the burden of that deferral. In 1907 the institution outside the clearing circle bore it. In 1933 depositors of institutions that failed to reopen bore it, as evidenced by the statutory ordering of liens and the presidential refusal of blanket guarantees, an outcome consistent with the modern summary's distinction between institutional resumption and depositor recovery.
The limits of this study are plain. Sprague is read from an uncorrected machine reading and only in paraphrase; the excerpt stops before his analysis of the New York banks' condition. The Act is read only in its first sections, and the provisions on currency and lending that the address describes are not among the texts at hand. No text read here says what became of the Knickerbocker's depositors, how long the 1933 holiday lasted after 9 March for any given bank, how many banks never reopened, or how much anyone lost. The two Skeptron records are modern summaries of modern essays and add nothing to the primary evidence except the dates they keep and the cautions they state. The ledger's reading of the 1933 record as the one close-read case whose record lets a later reader judge an emergency act by its own end does not hold when two further readers are added, as the Appendix records.
Further research could take three directions. The first is to read the instruments that carried the holiday past 9 March and the Reserve Bank provisions of the Act beside the address, to test whether the open term was closed by a date, a condition or a further proclamation. The second is to read the clearing-house committee's own minutes of October 1907, if they survive in a corrected text, beside Sprague's account and the record, to settle whether a request was made and refused. The third is to put the same two questions, who was shut and for how long, to the closures of other panics in Meridian's ledger, where most records with an act name neither a condition for its end nor who bore its cost.
Appendix: Ledger Figures Used
The statistics cited in this appendix derive from the registered ledger of emergency decisions (content/books/skeptron-crises/numbers.md), reflecting the direct codings of the catalogued cases. The ledger reads Skeptron's records, not the primary texts.
The ledger holds 84 cases; 24 are read closely by the study. The study's own fields were registered with hypotheses F1 to F5 on 2026-10-06T09:55Z, before coding.
Case-by-case cells (unsettled where the two codings differ):
| Case | act | act kind | temporary | exit named | outcome of term | dates apart | figure | bearer named |
|---|---|---|---|---|---|---|---|---|
| The Panic of 1907 | yes | unsettled | not-said | not-named | not-said | not-named | no-figure | unsettled |
| The Bank Holiday of 1933 | yes | suspension-or-closure | temporary-no-term | observable | unsettled | named | no-figure | named |
The record of 1933 is read as presenting the act as temporary without a term, although Proclamation 2039 itself gives a dated term; the ledger reads the record, and the record does not give the proclamation's term.
Across the eighty-four cases surveyed, the lenient evaluation profiles confirm the patterns detailed in Chapter 2: hypothesis F1 receives support in 83% of applicable instances (30 supporting against 6, with 8 unclear and 40 not applicable), hypothesis F3 is rejected with a 32% share (13 supporting against 28, with 3 unclear and 40 not applicable), and hypothesis F5 receives overwhelming support at 92% (35 supporting against 3, with 6 unclear and 40 not applicable). The dates-apart metric was explicitly recorded in 18 of the 84 cases.
Cases whose record lets a later reader judge an emergency act by its own end (an act presented as temporary; an observable condition of its end named; who bore its cost named; two dated moments named): 1 of 42 (2%); among the 24 cases the study reads closely, 1. On the case-by-case table, the cells carrying all four are those of The Bank Holiday of 1933. Where two of three readings agree, with two further readers on a sample of 30 cases, the count is 0 of 19.