Abstract
The oldest act in this study is a Babylonian law that spares a debtor the grain and the interest owed in a year of storm or drought; the newest is the rescue of a Swiss bank by its rival in March 2023. Between them stand suspensions of payment, bank holidays, rescues, cancellations of debt, loans of grain, charters and abandoned pegs: acts taken in an emergency, many of them said at the time to be temporary. Of each such act this study asks three things of the record a later archive keeps. Does it present the act as temporary? Does it name the condition, the date or the event on which the act would end? And does it name the people who carried the cost while the act stood? It reads 24 cases closely, beside the texts their actors wrote where those survive and are free to reproduce, and codes 84 cases from Skeptron's crisis, credit, market and money histories on the same fields. Its answer is about records and the end they give an act. In most of these records the act is there and its end is not: the condition on which it would stop is seldom named, and the people who carried its cost are named less often still. Only one of the cases with an act holds all that a later reader would need to judge it by its own end.
What this study is
The study was made by RISCON Labs AI. It is the second of a series of studies that put the same question to different archives of decisions: what had to be decided, and what change should make that decision be revised. The archive read here is Skeptron's, a studio in the RISCON family that publishes history cases. Its cases are short dated summaries of public sources, made in Korean in 2026 and read here in English.
The study is not a history of financial crises. Better histories exist, and the first chapter names many of them. It is a reading of what a modern archive has kept of emergency acts in money and credit, set beside what the actors themselves wrote where their words survive: a proclamation that closed every bank in the United States for four days, an Act that suspended the gold standard "unless and until" a proclamation directed otherwise, a minute of the Privy Council that told the Bank of England to stop paying cash "until the sense of parliament can be taken", a central bank's release whose temporary exception expired on a stated date unless extended, and an emergency ordinance that gave itself six months.
Why read these records
An emergency act in money is usually announced as a bridge: a holiday, a suspension, a facility, a moratorium, a ceiling held for now. The question a reader most needs answered later is when the bridge was meant to end and who paid the toll while it stood. A record that keeps only the act, without its term and its bearers, lets a later reader see that something was done but not judge it by its own promise. The series asks of every archive whether its records let a reader know what would change a decision. Here the question takes its plainest form: what would end the act?
What the reader will find
An introduction sets out the question, the cases and the method. Two chapters review four literatures (the lender of last resort; manias and panics; debt, relief and the international lender; rules, discretion and the emergency that stays) and describe the ledger on which every case is coded. Part II takes the emergency act kind by kind, from the stopping of cash payments to the cancelling of debts, and asks who paid. Four more compare across the cases: headline sums and what could be drawn, charters founded in a crisis, pegs and regimes abandoned, and market records with no act at all, including twenty-nine windows of observed prices for crypto-assets. The last three chapters read the conditions of ending across the cases, judge the study's hypotheses, and set the results beside the earlier accounts and beside the first study of the series. A conclusion answers the question and says what would change the answer.
Every count in the study comes from the ledger, and every hypothesis was written down before any case was coded for it. Four of the seven predictions registered with them failed, and the study reports each one.
What this study does not do
It gives no financial, investment or legal advice. It forecasts no price, rate or crisis, and judges no institution's soundness today. It does not say whether any act was right or wrong, or whether it caused a recovery or a loss. It says what the records and the actors' texts state, what they leave unstated, and how far the counts reach.
Skeptron's records are its own readings of public sources, dated; the readings through the scholarly works are this study's; and the whole study is AI's expansion of Skeptron's records and the source texts. Where the study goes beyond what a work's reading note allows, it says so.