This article asked how the American monetary authorities wrote their power to lend and its limits in two public texts: Greenspan's testimony on LTCM of 1 October 1998 and the Board's release of 14 September 2008. The answer is that neither text states the legal power at all. The testimony writes the limit of the Federal Reserve's role as a list of things not done and as a threshold for involvement that, by its own account, was lowered in fragile markets; it names the bearers of the loss in detail and grants moral hazard while arguing against its weight. The release writes help to primary dealers and to markets, limited by widened collateral defined by the repo market's own practice and, for one facility, by amount; it writes one dated end, for the exception to section 23A, with the Board as the body that could extend it; and it names no institution and no bearer. The modern records of the two episodes keep the entities and the figures apart and refuse the generalisations that later shorthand invites, while dropping the threshold, the bearers, the dates and the end.
The contribution is to the debate between Ball and Bernanke, and to the older argument over what a lender of last resort should say in public. The public texts of the weekend of 14 September 2008 do not contain the terms in which the dispute over Lehman is now argued. That does not decide the dispute, but it marks where its evidence has to come from: not from what the authorities told the public on the day, but from what was written later.
Skeptron's record of Lehman holds two entities, keeping the holding company's bankruptcy of 15 September 2008 apart from its dealer subsidiary's use of the Federal Reserve's existing credit facilities, and this study does not say that Lehman as a whole received or was refused a loan.
The limits are plain. Two texts are a small sample. The statement of President McDonough, to which Greenspan's testimony refers, is not in the files read here, nor are any of the Board's later releases of September 2008, nor the internal records on which Ball and Bernanke draw. The modern records are summaries written for a public platform and read here in translation; the ledger reads those summaries, not the primary texts, and its two readings are one model run twice. Lombard Street is read in an uncorrected volunteer transcription and only as frame. No claim is made here about whether any act was right, or about what it caused.
Further research could extend the reading in three directions. The first is President McDonough's statement of 1998 and the Board's releases of the days after 14 September 2008, which would show whether the wording of the limit changed as events moved. The second is a comparison with other episodes in the parent study's ledger in which an authority lent or guaranteed, to ask whether the pattern of a limit written by negation, threshold or date recurs. The third is a reading of the later accounts themselves, the memoir, the critique and the inquiry, for the moment at which the question of legal authority to lend to Lehman first enters the written record.
Appendix: Ledger Figures Used
All figures are from the numbers file of Meridian's study of emergency acts in money and credit (content/books/skeptron-crises/numbers.md), in the ledger's own terms. The ledger reads Skeptron's records, not the primary texts. Its codings are one model run twice; a cell is settled where both codings give the same value and unsettled where they differ.
The ledger holds 84 cases; 24 are read closely. The study's own fields were registered with hypotheses F1 to F5 on 2026-10-06T09:55Z, before coding.
Case-by-case cells for the two cases read here:
| Case | act | act kind | temporary | exit named | outcome of term | dates apart | figure | bearer named |
|---|---|---|---|---|---|---|---|---|
| The LTCM Recapitalisation | unsettled | unsettled | unsettled | unsettled | unsettled | named | single-figure | unsettled |
| Lehman's Bankruptcy | yes | lending-guarantee-or-rescue | not-said | not-named | not-said | not-named | no-figure | not-named |
Across all 84 cases, the cell act: yes (stated) 38; no 21; no (implied) 17; yes (implied) 4; unsettled 2; no (stated) 2. The cell bearer named: not-applicable 38; not-named 34; unsettled 8; named (implied) 2; not-named (implied) 1; named (stated) 1.
Across all 84 cases, lenient readings: F1 (in most cases that show an act, the case's own text, apart from the studio's recurring sentences, does not name an observable condition, date or event on which the act would end or be revised) supports, with supports 30, against 6, unclear 8, not applicable 40, share 83%. F5 (in most cases that show an act, the case's own text does not name who bore the loss or the cost while the act was in force or not yet taken) supports, with supports 35, against 3, unclear 6, not applicable 40, share 92%.
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%); neither of the two cases read here is that one.