Chapter 11. Conditions for Ending: What Could Be Observed

11.1A condition of ending

One of the series' axes asks what would change a decision and whether that change could be observed. This chapter puts that axis to emergency acts in money and credit, and asks of each record whether it holds the thing a later reader would need: a condition of ending.

Such acts may suspend an obligation, close or reopen banks, lend, guarantee or rescue, relieve or defer debts, or change a rate, a peg or a monetary regime. A condition of ending is what a record names that would end, lift or revise any of them. The naming can take three forms. It can be a date: the holiday runs from Monday the sixth to Thursday the ninth. It can be a number or a threshold: a deferral of ninety days; a ceiling kept at the level of 26 June. It can be a named event or decision: a treaty of peace; a referendum; a decision to reconsider. Anything less than these three forms is either a general phrase about what may come later, or nothing at all.

Three examples set the shape of the chapter. The suspension of a convertibility stops a promise without cancelling it. Convertibility is the note-holder's right to exchange a banknote for a fixed weight of gold; a suspension ends that exchange, and the note remains in circulation. Lifting a ceiling is the reverse gesture: an authority that has been supplying emergency liquidity, or refusing to supply more, names the level at which it will hold and the decision that would move that level. A deferral of debt moves a payment across time and gives the move a length. In each of the three, the question for a later archive is the same: what did the record say, at the time, about when or on what this act would stop?

The study separates two things that are easy to run together. The first is whether a record presents its act as temporary. The second is whether it names the condition of that act's end. A text can call a suspension temporary without giving any term for it, and the ledger's numbers keep the two apart: of the eighty-four cases, nine present an act as temporary, seven of them stating a term, while eight name a condition of the act's end, six of them an observable one.

Three more axes stand behind this one and are treated elsewhere: Chapter 7 asks what became of a figure announced as a limit or a line, and Chapter 12 applies the registered rule to the counts that this chapter sets out and reports which of the study's predictions held. This chapter stays with one question: what could be observed of an end, from the record alone.

Two limits govern everything that follows. The study does not judge any act right or wrong, and it draws no conclusion that any act caused a recovery or a loss. It reports what each record holds, what its actor's text named, and what a later reader could count or check from those materials alone. The second limit is evidential: the counts come from "The ledger's numbers", produced by one model run twice, where a value counts only where both readings found it stated in the record's own words. Chapter 2 sets out the method, the frame and the source texts that could and could not be had.

11.2Conditions written into the act

Four emergency acts in this study name their own condition of ending, at least in part. Each is read here on the axis above, and each record is quoted once.

Hammurabi's relief in a failed harvest. The provision in question is a rule, not a case decided under it. Skeptron's record answers the question of what it prescribes as follows:

The law's own text, in the 1910 English translation of L. W. King held for this study, states its condition and its term in a single sentence: "If any one owe a debt for a loan, and a storm prostrates the grain, or the harvest fail, or the grain does not grow for lack of water; in that year he need not give his creditor any grain, he washes his debt-tablet in water and pays no rent for this year." The passage is quoted in King's 1910 English; where this chapter gives the sense of a passage that is not quoted, that sense is in this study's own English.

What a later reader can observe from that clause is the condition itself: a storm, a failed harvest, grain not growing for lack of water. These are named natural events, and the term attached to them is the calendar year — "in that year", "pays no rent for this year". What the reader cannot observe from the case is anything about its application. The record's own answer states the boundary plainly: the provision is not a document recording a court result actually applied to a particular farmer, and the case's Scenario adds that whether it was actually carried out requires separate contracts and court records. This study reads the tension as genuine rather than as an error; the provision names its own condition and its own term in the text of the law, while the record keeps the law collection apart from the enforcement of any individual debt, so that the condition is observable in the source and unobservable in the case as a case.

The green sprouts loans of 1069. The record of the Northern Song proposal gives its terms as follows:

The Song history that carries this proposal is Toqto'a et al., Song shi (1345), juan 176, the treatises on food and money; the excerpt held for this study is in classical Chinese, and the English of this study is this study's own. The record's answer states that the money lent before harvest would be repaid in two halves, one at the time of the summer taxes and one at the time of the autumn taxes, that the borrower might pay in grain or in coin, and that where disaster struck the payment might be postponed until the next harvest ripens. Song shi, juan 176, continues past the proposal with objections to the scheme — Su Zhe's objection and Han Qi's memorial of Xining 3 — and Chapter 5 treats them.

Here the condition is named in two forms at once: a natural event (disaster) and a fixed point in the agricultural calendar (the next harvest). A later reader could observe both in principle — the disaster is an event, and the harvest is a date in a season — but the case's own text keeps the proposal apart from its execution. Its Scenario warns that effects can only be verified where there are user costs, whether coercion was involved, and a repayment ledger, none of which the case holds. The proposal names a deferral to a named later harvest, and both readings found that condition stated in the record's own words; what became of the act is not said, and the two readings differ on whether the record gives a figure. This study reads the disagreement as recording what the proposal cannot carry into the case: an outcome and a precise figure.

Mexico's moratorium of 20 August 1982. The record's answer sets the agreed period apart from what was promised:

Ninety days is the clearest form a condition of ending can take: a number a reader can count from the record's own date. The deferral is narrow, and the record says so — principal on public-sector bank loans, with interest still to be paid in full as scheduled. The record's own Scenario field goes further in keeping stages apart, and the sentences below are quoted from that field:

No source text could be had for the Mexican documents of 1982, so this case is read from the record alone and from James M. Boughton's Silent Revolution: The International Monetary Fund 1979–1989 (2001). This study reads the ledger's numbers for this case as giving a numbered term: the act is presented as temporary with a term stated, the condition of its end is stated, and the announced figure is set apart from what was available. What became of the ninety days is not said. The condition is observable; its outcome is not in the record.

Greece's emergency liquidity of 28 June 2015. Emergency liquidity assistance, as the phrase is used here, is lending by a central bank to banks in its own currency area that cannot obtain funds in the market, against collateral, up to a ceiling that the central bank itself fixes. The record gives the European Central Bank's decision as follows:

This record names an end-condition of a different kind from the other three. There is no number and no date for the reconsideration. This study reads the case as naming a revisiting instrument without a time — the decision to reconsider is named, the moment it would come is not — which places it between the numbered term of Mexico and the absence of a term in the Restriction.

The four records can therefore be set beside one another on one axis. Hammurabi's provision names a natural condition and a year. The green sprouts proposal names a disaster and the next harvest. Mexico's agreement names ninety days. Greece's statement names an event, a ground and a decision to reconsider, and no time. In none of the four does the case's own text hold what became of the act after the moment it describes: what became of the act reads not said for all four; of the two read next, the Restriction's outcome is recorded as ended as stated and the bank holiday's is unsettled.

11.3Conditions not written

Two further cases in this chapter carry suspension rather than relief or lending, and they show what happens when the condition of a return is written in a document the case does not hold.

The Bank Restriction of 1797. Skeptron's record answers the question of what the Restriction suspended as follows:

The period 1797–1821 is in the case, taken from the official chronology cited by the record. It is worth being careful about what that period is. This study reads the span 1797–1821 as measured by an official chronology afterwards — the length the Restriction lasted — and not as a term the act itself named in 1797; the record does not say when that chronology was drawn up. The Bullion Report of 1810 is not quoted here; its sense, given in this study's own words, is this: the committee holds that the high price of gold and the depression of the exchanges showed an excess of paper currency, and recommends that cash payments be resumed. Under the law as it stood, the committee notes, the Bank would have to pay in cash six months after the ratification of a definitive treaty of peace. The committee proposes instead that two years be given, peace or war, for resumption, and that the manner of resumption be left largely to the Bank's own discretion.

The Report therefore names, in 1810, exactly what this chapter's axis asks for: a condition of ending (a definitive treaty of peace, an event), a period (two years), and an authority to whom the manner is committed. None of that is in the case as the case states it. The Restriction record holds the suspension and the official span; the condition of return is in a document thirteen years later than the act, written by a committee rather than by the authority that suspended payment. This study reads the gap as the point: a source text can hold a condition of ending that the act's own record does not, and a later reader must keep the two apart. The record also does not say who held the Bank's notes during the period, and this chapter does not add that question to the case; Chapter 6 treats who bears the cost of a suspension in force. Chapter 3 sets out the Restriction's record, Thornton's analysis of the paper credit system during and after the Restriction, and the readings of Fetter and Clapham.

The bank holiday of 1933. The record's answer separates three dated moments:

The proclamation of 6 March 1933 that closed the banks names its own term in its own words. It states that the President, "do hereby proclaim, order, direct and declare that from Monday, the Sixth day of March, to Thursday, the Ninth day of March, Nineteen Hundred and Thirty-three, both dates inclusive, there shall be maintained and observed by all banking institutions and all branches thereof located in the United States of America, including the territories and insular possessions, a bank holiday, and that during said period all banking transactions shall be suspended." The same proclamation lets the Secretary of the Treasury permit some banking functions during the holiday.

Here the divergence between record and source text must be stated plainly, and both given. The case's own text, read as the ledger reads it, presents the act as temporary without a term while giving three dated moments, and both readings find a condition of its end stated on those dates. The proclamation, which is a source text rather than the record, gives a term of four days inclusive and an explicit power to vary what banks may do inside it. A reader can count the proclamation's term; a reader of the record alone can only see three dates and a reopening that begins later than the order. Chapter 3 sets out the proclamation's full terms, the Emergency Banking Act of 9 March, and the readings of Silber on why the reopening restored confidence and of Wigmore on whether the closure reflected a run on the dollar.

What the record does not say is equally specific. Its Scenario keeps institutional measures apart from individual outcomes, and states this in its own words:

The two readings disagree on what became of the holiday after its stated term. Nothing in the case says when the holiday formally ended, what became of the banks that did not reopen, or what condition, if any, attached to their closure; the ledger records who bore the cost while the act lasted as named in this case, and this is the only case among the twenty-four read closely that names a bearer in its own record.

Across these two cases the pattern is consistent with the axis's question. The suspension of 1797 is described as temporary and is accompanied, in the record, by a period taken from an official chronology; its condition of return appears in a later report, not in the act. The bank holiday of 1933 is described in the record with three dated moments and, in the proclamation, with a term of four days; the reopening is a later event, and some banks never reopened. In neither case does the record hold what became of the act's term. Chapter 9 treats the wider question of rules with escape clauses — Bordo and Kydland's reading of convertibility as a contingent rule under which a country may suspend in an emergency and return at the old parity, set against Kydland and Prescott's warning that discretion can defeat a rule — and this chapter can say only what its two records hold: a suspension named as temporary, a period of twenty-four years on the record's own dates, and no reason for that length named in the record.

11.4Skeptron's own conditions

Every case in Skeptron's history series carries five questions, and the fifth is a different kind of question from the first four. It asks, in each case, what observations must be gathered for a comparison to be maintained or withdrawn. It is the archive's condition of a later comparison, and it appears in every case with only the names and dates changed.

The Bank Holiday of 1933 shows what the fifth question looks like when its answer is read:

Two things must be kept apart about this passage. The first is its role inside Skeptron's method. The recurring sentences of the fifth question are part of the frame that Chapter 2 explains: they are sentences that stand in case after case with names and dates changed, found by recurring across cases with only names and dates changed, and they are the archive's method rather than findings about any one case; a sentence that repeats across every case cannot, by repetition alone, be evidence that any particular authority named a condition of ending.

The second is what the fifth question is a condition for. It is not a statement about what would end the act. It is a statement about what would keep or withdraw Skeptron's own comparison. The answer above asks for restoration tests, alternative routes and residual deficiencies, and it gives a withdrawal rule — if only some functions returned, or if the alternative routes depend on the same cause, withdraw the full-restoration judgment. The record of 1933 does not hold restoration tests, alternative routes or residual deficiencies. Its own answer says what it does hold and what it cannot: the proportion of resources in reopened banks and the individual recovery of deposits are different metrics, and banks that were permanently closed remained.

The same structure appears in the other cases of this chapter. The green sprouts record names what must be observed — user costs, whether coercion was involved, a repayment ledger — and the case holds none of them. The Mexico record names the change in repayment burden and available resources, and adds that where the debt type, maturity or principal-handling provision differs, the effect is not compared as the same. The Greece record's own calculation states two days between the determination of the basis for the ceiling on 26 June and the decision of 28 June, keeps the two apart as different stages, and states that the order of the two does not by itself prove a single cause.

For the study's own counting, this matters in a precise way: the condition of a later comparison in every fifth question is not counted as the actors' condition of ending anywhere in the counts. It is dated after the event, it is written by the archive, and it concerns the archive's comparison rather than the act's term. This chapter's axis — what would end, lift or revise an act, and whether a later reader could observe it — is answered from the actors' own records and source texts, and from nothing else.

There is a further reason for keeping the two apart, and it is diagnostic rather than procedural. A fifth question can name what would make a comparison be withdrawn, and still leave the case unable to hold that evidence. The green sprouts record names a repayment ledger as what would verify the effects of the proposal; the case does not hold a repayment ledger, and its third answer says effects can only be verified where there are user costs and a ledger. The Bank Restriction's fifth question names what must be observed after a control — the change in volume, price and transaction possibility — and adds that if quoted figures were replaced with actual executed prices, certain computations are withdrawn; the case holds no executed prices. The condition is stated; the observation is named; the evidence is elsewhere. This study reads that sequence as the archive's own answer to the axis's question, and reads it as separate from what the actors wrote about the end of their acts.

11.5Counts across the ledger

The counts in this section come from "The ledger's numbers". Chapter 12 applies the registered rule to the study's hypotheses and reports which predictions held.

A condition of an act's end. Across all eighty-four cases the record holds the following:

The forty cases marked not applicable are the cases whose own text shows no emergency act at all: the twenty-nine price windows read in Chapter 10, plus crashes and other market movements with no act by an authority. Skeptron's price windows record observed prices up to one closing price; they hold no decision and no actor, and this study reads them only as a group.

Whether an act is presented as temporary sits beside it:

Nine cases present their act as temporary — seven of them stating a term and two without one — and only those seven give a term. The gap between the two questions is where this chapter's argument lives: presenting an act as temporary and naming what would end it are separate things, and the records hold them apart.

What became of the act. Where the case's own text says what became of an act, the record holds the following:

Of the cases in which an act appears, two say something about what became of it — one records that the act was reversed or ended early, and one that it ended as stated — while thirty-nine hold no statement of outcome at all and three are unsettled. This is the record behind the study's F2 hypothesis, which asks whether acts presented as temporary and accompanied by an outcome statement turned out to have lasted longer than first said or became permanent.

Both questions, by group. The ledger's five groups are Skeptron's crisis history, its credit history, its markets, its money, and the price windows taken out of the first and third. Skeptron's price windows record observed prices up to one closing price; they hold no decision and no actor, and this study reads them only as a group. The two questions by group stand as follows:

Skeptron's price windows are twenty-nine cases with no act, recording observed prices up to one closing price, with no decision and no actor, and this study reads them only as a group.

The price windows carry the same values in both questions — twenty-eight not applicable and one unsettled — which is what this study expects of records that hold no actor and no act. The credit history holds the largest number of observable conditions, five, and also the largest number of unnamed ones, fourteen. The crisis history holds nine unnamed conditions against one observable.

The four kinds of verdict. The ledger's counts for both, with all four kinds of verdict, are:

By group, counting only what both readings found stated, F1 gives nine supports, one against and six unclear in the crisis history; fifteen, three and ten in the credit history; one, zero and one in the markets; three, zero and two in the money group; and zero supports with ten unclear in the price windows. Counting also what the setting implies gives nine, one and two in the crisis history; sixteen, five and four in the credit history; one, zero and zero in the markets; four, zero and one in the money group; and zero with one unclear in the windows. In the credit history F2 is unclear where both readings must agree and goes against when the setting may imply; in the crisis history it is unclear in both counts; in the markets and the money groups it does not apply, and in the price windows it is unclear in both counts.

The series' own axis on observable conditions sits beside these counts. This study reads the one case among the twenty-four it reads closely as The Bank Holiday of 1933, the case whose record names three dated moments, whose act is presented as temporary without a term, whose condition of its end is recorded as observable on those dates, and whose Scenario names the individual recovery of deposits among the outcomes kept apart. Whether the proclamation's four-day term belongs in that reading is the divergence set out in section 11.3, and both readings are given there.

Where the two readings differed. On the condition of an act's end the two readings agree on seventy-six cases out of eighty-four, or ninety per cent; eight cases are unsettled. On what became of the act they agree on eighty-one of eighty-four, or ninety-six per cent, with three unsettled. Of the six cases this chapter reads, four leave the two readings in disagreement on this axis: Hammurabi's provision on both whether the act is presented as temporary and whether the record names a condition of its end; the green sprouts proposal on whether the act is presented as temporary; Greece's emergency liquidity on both of those; and the bank holiday on what became of the act. The green sprouts proposal's condition of its end is stated on both readings; Mexico's deferral is settled on a stated term and a stated condition of its end; the Restriction likewise on a stated term and a stated condition of its end.

The pattern in the unsettled cases is the same one that runs through this chapter. Where a record holds both an actor's act and a later archive's boundary — a law collection with no enforcement, a proposal with no implementation, an announcement with no follow-up decision — the two readings can disagree about whether the text names a condition of ending at all. Chapter 12 applies the registered rule to these numbers, sets the predictions written beside the hypotheses before any case was read against what the counts show, and reports which predictions the counts did not bear out.

The hand now goes to the verdicts themselves: what the rule does with these counts, which of the study's five registered hypotheses the counts support, and which of the registered predictions failed.