A closure order is a decision taken before the facts it depends on are known. When a bank, a group of banks or a whole national system stops paying, the people who order the stop cannot yet say which of the institutions they have closed are short of cash and which are short of value. The first kind can pay everyone in time if it is given time; the second cannot pay everyone at all. Much of the scholarly writing on banking panics is an argument about which kind a given panic was, and the argument is usually settled, where it is settled, by data assembled long afterwards: reserves, deposits, failures, gold flows.
This article puts a different question to the documents themselves. It asks what the closure and reopening texts of two American panics wrote down about two things: who was shut, or shut out, and for how long. The first is a question about the boundary of an emergency act. The second is a question about its term. Both can be answered from the wording of the documents without knowing how the panic ended, which is the reason for asking them. A reader in October 1907 or March 1933 had only the wording to go on.
The two episodes are the Panic of 1907, at the moment the Knickerbocker Trust Company in New York was refused support and suspended on 22 October, and the national bank holiday of March 1933, from the proclamation of 6 March through the Emergency Banking Relief Act of 9 March to the reopening of the first banks from 13 March. They are paired because one shows a private clearing house deciding whom to help, and the other a national government closing everyone and deciding whom to reopen. Between them stands the Federal Reserve Act, a statute that this article does not read.
The contribution is to the debate between those who read banking panics as runs that can strike sound institutions, and those who read them as the exposure of particular weak banks and of institutions built badly. The article does not take a side in that debate. It argues instead that the closure documents of 1907 and 1933 show where, in each episode, the authorities deferred the question that the debate asks, and what that deferral cost and to whom. In 1907, as Sprague tells it, the question whether the Knickerbocker could have been helped was never examined, because there was no time and no one whose duty it was to examine it. In 1933 the proclamation closed every banking institution in the country before anyone had answered the question for any of them, and the reopening became the examination. The article also has a smaller point to make about the modern records of these events: they keep the dates of the acts carefully, and they compress or alter the boundary.
Two further debates sit beside the first. One concerns the restriction of cash payments by banks. While Sprague evaluated the 1907 suspension of payments as an avoidable misjudgment, Friedman and Schwartz (1963) regarded cash rationing as a necessary shock absorber that averted a collapse of the national banking network. The other concerns 1933: whether the crisis that ended in the holiday was a run on the banks or also a run on the dollar (Wigmore 1987), and why the reopening restored confidence (Silber 2009). The documents read here bear on each of these, and on each they carry less than the debate would like.
The plan of the article is as follows. Section 2 lays out the debates and introduces the registered hypotheses. Section 3 describes the primary documentary sources, the case summaries, and the coding structure of the comparative crisis ledger. Sections 4 and 5 read the two episodes. Section 6 sets them side by side, and Section 7 says what the reading adds and what it cannot do.