Chapter 7. Announced and Available: Headline Sums and What Could Be Drawn

7.1A figure is a decision too

When an authority announces money in an emergency, the number it announces is not a description of something that has already happened. It is part of what the authority decides. A stand-by arrangement, a ceiling on lending, a line of credit, a deferral of principal: each carries a sum, and each carries that sum with conditions about who may use it, against what stated basis, and over what period. The headline sum is therefore a promise with a term, and a promise with a term is exactly the kind of act this study asks its records to disclose.

Three words need setting out, because the five cases of this chapter turn on their difference. A stand-by arrangement is an agreement in which a lender, here the International Monetary Fund, approves a sum available to a borrowing country over a stated period, in tranches, that is, in instalments, each later tranche dependent on review and performance conditions. Emergency liquidity assistance is the term for a central bank's lending to banks that cannot raise funding in the market, advanced against collateral, and set as a ceiling: a maximum, not an amount already granted. A drawdown is the act of actually taking the money; an approval is not a drawdown, and a ceiling is not a drawdown either. Much of what a headline sum hides lies in the distance between these three things.

Two shapes recur in what these five records hold: one number with nothing to set it apart from, and an announced sum beside an amount separately named as available or drawable.

Among all eighty-four cases, the reading stands as follows: no figure, twenty-six; a single figure, twenty-six; an announced figure kept apart from what was available or drawable, seven; the rest undecided or carrying only an implied figure. Among the twenty-four cases this study reads closely: no figure, seventeen; announced and available kept apart, three; a single figure, one; the rest undecided. The three cases that keep announced and available apart are Korea's Stand-By Arrangement, The Credit Suisse Takeover and Mexico's Moratorium; they occupy sections 7.2, 7.3 and 7.4 below. The one case among the twenty-four that carries a single figure is The LTCM Recapitalisation, treated in 7.5. Greece's Emergency Liquidity, the fifth case of this chapter, carries no figure in its record at all, which is itself worth a close look.

The counts across the whole ledger are given in 7.6, together with the verdict. What this chapter reads, case by case, is what each record actually holds and what a reader can check against it later.

7.2Approved and available

The Korean record of 4 December 1997 states a headline sum and an immediately available amount in the same sentence, and the distance between the two is this section's subject. An approval is a promise with conditions, not money that arrived, and in this record the conditions that tie the rest of the sum are named only in general terms. That is what this section adds to the debate over what the Fund asked of borrowing countries after an emergency: from this record alone the debate cannot be settled, because the record gives the shape of the announcement and none of its schedule.

A synthetic scene, drawn only from the record's date and its contents:

In Washington on 4 December 1997, Fund staff and directors work through an approval statement for Korea while officials of the borrowing country read the same text and market observers wait for what it will say about the rest of the money. Hands move over a document that carries a total, a currency, a period of three years and an amount available at once. Eyes check that the currency of presentation is not the currency of receipt, and that the word for the money to come later is a condition rather than a sum already granted. What is waited for is the first review, when the conditions attached to the later tranches are tested and the next instalment either follows or does not.

On 4 December 1997 the International Monetary Fund approved a stand-by credit to Korea. The record's scenario and its answer to the concrete-fact question state the same two sums in the same terms:

SDR stands for Special Drawing Rights, an international reserve asset created by the Fund; the record gives the headline sum in SDR and the dollar figure as a conversion "at the time", which the record keeps as its own quantity. Read straight, the record says that of an approval of 15.5 billion SDR over three years, 4.1 billion SDR was available at once and the rest was tied to review and performance conditions. Approval, availability, condition and drawdown are four different things, and the record holds all four in one paragraph without merging them.

The record's third question asks directly whether the approved total may be treated as dollars that flowed in on that day in full, and whether SDR and dollars may be calculated using the same number. Its answer refuses both readings: presentation currency, conversion at the time, approval, immediately available amount, conditional subsequent disbursement and actual drawdown must each be distinguished. A later reader given only the headline approval in a single currency has one number and no structure; the record's own sentence gives the approved total, the currency of its presentation, the conversion, the immediately available amount, the condition attached to the rest, and the question of what was actually drawn.

The conditions that tie the rest are named in the record only in general terms: review and performance conditions, subsequent tranches. The record gives no schedule of the reviews and no list of the conditions. What it does give, in the answer to its fifth question, is what a later comparison would need:

That answer sets out three tests: liquidity by maturity, the borrowing entities, and what was actually implemented. It also forbids two things that a headline sum invites: counting expected inflows as available cash, and adding together the support of different legal entities. Both prohibitions bear on the sections below, where each of the five records of this chapter is read for what its figure separates and what it leaves together.

The word "conditions" in the record points to a standing framework, given by the Articles of Agreement of the International Monetary Fund, formulated at Bretton Woods in July 1944 and printed in the United States Statutes at Large. The text available to this study is a machine transcription of that printing, marked as uncorrected, so its sense is given here in this study's words rather than quoted. The Agreement's first article lists the Fund's purposes; its purpose (v) is to give confidence to members by making the Fund's resources available under adequate safeguards, so that they may correct maladjustments in their balance of payments without resorting to measures destructive of national or international prosperity. Article V, section 3 sets out the conditions governing a member's use of the Fund's resources: among them, that the purchase is needed for payments consistent with the Agreement, that the Fund has not declared the member's currency scarce, that the Fund's holdings of the member's currency not pass stated limits of its quota, and that the member have not been declared ineligible. Section 4 allows the Fund, in its discretion and on terms which safeguard its interests, to waive those conditions, and to require collateral as a condition of a waiver. The phrase in the record thus has its institutional setting in the Articles: resources available under safeguards, and the safeguards are the conditions. Chapter 8 treats the adoption of the Articles in full.

The notes for Boughton and Blustein give no sums; the record supplies them. This study reads those notes as evidence about the surrounding argument and takes no sum from them; the sum of 15.5 billion SDR comes from the record alone.

Across the ledger, seven of the eighty-four records set an announced figure apart from what was available or could be drawn; among the twenty-four cases this study reads closely, three do, and Korea's is one of them.

Chapter 2 reads this record in its demonstration of how the ledger reads a case; this section reads only the figure and its separation from availability.

7.3Two lines of 100 billion francs

On 19 March 2023 the Swiss National Bank announced support for Credit Suisse and UBS. The record keeps two separate statements:

Two lines, each of up to CHF 100 billion, and the record's word "separately" does the work that any single combined headline would destroy. The record's answer to the concrete-fact question repeats the same two sentences, and its boundary answer then sets out what must not be done with them:

The record's answer to the question is no; it then lists the five distinctions a reader must hold apart: borrowing entity, legal basis, collateral and guarantee, limit, drawdown. A sixth item, the acquisition price, is named as a thing to be kept separate from all five, and the record gives no figure for it.

Working through the record's own separation, the first line has a stated borrowing entity, UBS and Credit Suisse together, described as a total limit; its stated basis is that support could be received in that form, with the record naming one attribute of it, priority claims in the event of bankruptcy. A priority claim is a claim that ranks ahead of ordinary claims if a firm's assets are distributed after a bankruptcy; the record names this attribute and no other. The second line names one borrowing entity, Credit Suisse, and one additional attribute, that a loan of up to CHF 100 billion with a federal guarantee could be provided to it. A guarantee by the federal government is a promise by the state to meet the obligation if the borrower does not; the record names it and stops. Neither line gives collateral, and neither gives a drawdown. The record's fifth question states the condition of a later comparison in terms that mirror this structure: what must be observed is liquidity by maturity, the borrowing entity, and the actual execution record.

No Swiss government press release and no Swiss National Bank statement goes with this case in this part; the list of originals that could not be had is given in Chapter 2. Everything else in this section comes from Skeptron's record of what the Swiss National Bank's release said, in the English translation made once for this study. Chapter 4 reads this same record on the axes of who was protected and who was not; this chapter reads only the figures and what they can and cannot be added to.

The ordinance of 16 March 2023 entered into force at 20.00 on 16 March 2023 and gives itself a term in Article 15(2): "Sie gilt für die Dauer von sechs Monaten ab Inkrafttreten." In this study's English: it applies for a period of six months from its entry into force. In this study's English, the ordinance leaves the height of the additional liquidity assistance loans that the National Bank may pay out at most to the Federal Council, after consulting the National Bank, and the remaining conditions of those loans to the National Bank itself. The amendment of 19 March 2023 entered into force at 20.00 the same day. It adds, in Article 14a, a guarantee by the Confederation in favour of the acquiring bank in a merger between systemically important banks supervised by FINMA, covering losses on the acquired bank's assets to be wound down; Article 14a(2) states its cap in German: "Die Garantie zur Verlustabsicherung beträgt höchstens 9 Milliarden Schweizer Franken." In this study's English: the loss-covering guarantee amounts to at most 9 billion Swiss francs. Under Article 14a(3), in this study's English, the guarantee pays out only once all those assets have been definitively realised and the acquiring bank has itself borne definitive losses of 5 billion Swiss francs on them, as FINMA confirms. This study reads that order as placing the first losses on the acquiring bank and capping the Confederation's exposure at 9 billion Swiss francs. Article 10a of the same amendment, in this study's English, lets such a merger proceed without the shareholder resolutions ordinarily required and makes FINMA's decisions binding on the commercial registers.

Two observations belong to the record itself, and neither is a judgment about the act. First, the record holds a limit, not a sum drawn: "up to CHF 100 billion" is a maximum, and the record contains no figure for what was actually taken under either line. Second, the record names the acquisition price as something to keep apart and then does not give it. What this record shows is that a figure can be rich in limits and empty in receipts, and that the second emptiness is not the same as the first. A reader given only the two headline sums and nothing else could add them together into one total; the record's own text forbids that addition and gives no alternative total, because none exists in the record. This study reads that as the clearest illustration in the five cases of what a single figure would hide.

7.4A ceiling kept and a deferral counted

The Greek record of 28 June 2015 concerns a ceiling, and the first thing to notice is that the record gives no number for it:

The record gives no sum for the ceiling at all; what it names is a date, 26 June, two days before the announcement, on which the ceiling's basis had been decided. The ceiling is held at that level, and the record names what it is not: not a new withdrawal, not a further increase. The answer to the third question puts the distinctions in the record's own terms:

A ceiling on emergency liquidity assistance is a maximum that a central bank sets on the amount a bank or group of banks may obtain from it in an emergency, against collateral. Announcing that the ceiling will be maintained is therefore an announcement about a limit, not an announcement about lending. The answer to the fifth question supplies the record's own count of the interval between the two dated actions: two days, with the prior determination of the ceiling's basis and the later announcement of its maintenance kept as different stages, since the order of two dates alone does not prove a single cause.

This record gives no sum for the ceiling, and the ledger does not settle whether it presents the act as temporary or names the condition of its end. The record does name a condition of reconsideration, the readiness to re-judge the decision, in general words, and it names the referable events, the referendum decision and the non-extension of the adjustment programme. What it does not name is a date on which the ceiling must be reconsidered. Chapter 6 reads this record for who bore the cost while the ceiling was held; this chapter reads it for what the figure promised and for how long. The answer is that the promise was a limit held at an earlier level, and the record gives no period over which that promise was made.

The Mexican record of 20 August 1982 gives the opposite shape: a term in days, and a request kept apart from what was agreed:

A deferral of principal is a postponement of the repayment of the sum borrowed, as distinct from a reduction of that sum, and here the record states at once that interest was not deferred with it. The boundary answer refuses two readings of what the deferral was:

Four quantities are kept apart: the request period, which another IMF source gives as 120 days; the agreed period, which the adopted IMF history gives as 90 days; any extension, which the record does not give; and the scope of interest treatment, which the record states as interest paid in full as scheduled. The record's fifth question adds a further distinction: what is to be observed is the change in repayment burden and available resources after the reduction, and where the debt type, maturity and rules for treatment of principal differ, the effect of reduction is not compared as the same. That last clause keeps a deferral from being read as a write-down, which is exactly what the boundary answer forbids.

In the ledger, Mexico's Moratorium is read as presenting a stated term for what is temporary, as naming an observable condition for its end, and as keeping announced and available apart for its figure. The reason this record keeps its figure apart is not that it gives a headline sum and a drawn sum; it gives the term, ninety days, and sets it apart from the period requested, 120 days. A deferral's figure is its duration, and the record's separation of agreed from requested is the same analytical move that separates approved from available in the Korean record. This study reads both as one thing: a record that names what was offered and what was accepted under different terms. Chapter 6 treats the Mexican and Greek records for who bore the cost of each and for the critics' accounts of the programmes; this chapter does not repeat that treatment.

7.5A sum rounded

The LTCM record of 23 September 1998 carries no headline sum at all. Its figure is a rounded commitment of about 3.5 billion dollars, and the record holds it apart from a second figure, 3.6 billion, which it refuses to merge with the first. This section reads that record for one question the chapter asks of every figure: whether the number states a promise on terms, or a count of what happened. The record gives a third answer, and it turns on provenance rather than on drawdown. That is a case the chapter's question has not met in its other four records: a figure can be untestable for reasons that have nothing to do with what was available to draw.

The record's scenario states the treatment in one sentence:

The testimony the record cites is dated 1 October 1998, which the record keeps apart from the event of 23 September. The reading note on that testimony covers the Federal Reserve's role, the moral hazard question and the fire-sale risk, and it gives no sums; the figure of about 3.5 billion dollars is therefore this study's report of what Skeptron's record retains from the testimony it checked, not a number taken from the note. The record's boundary answer adds that the testimony distinguishes the Federal Reserve's funds as never provided or proposed, and that the coordination role, the actual source of funds, the moral hazard debate and liquidation risk must each be reviewed separately. Chapter 4 reads this record for the rescue, the testimony and the moral hazard dispute; this section reads only the two figures.

Rounding is part of a record, not a defect in one. "Approximately $3.5 billion" is a quantity written at a stated precision, and the record treats the word "approximately" as data to be retained rather than noise to be stripped. The other figure, $3.6 billion, is a quantity at a different precision, from a source the record does not name in its own text; the two differ, and the record gives the reason for holding them apart as provenance and precision rather than as an error to be corrected. What the record establishes is narrow and sufficient for its purpose: before a later reader compares two figures, that reader must know which source each came from and at what precision. A comparison between 3.5 and 3.6 billion would then tell that reader about the sources, not about the recapitalisation.

Across the eighty-four cases of this study only seven records keep the figure announced apart from the amount available, while twenty-six record a single figure with no second quantity to set it apart from; among the twenty-nine records of the credit-history group, five keep the two apart. This record is the single-figure case among the twenty-four read closely here, and no condition of the end of anything is named in it. What it adds to the study's hypothesis F4, that an announced figure and an available figure are usually not kept apart, is a fourth kind of unavailability: one figure, with its rounding retained, can leave a reader unable to test anything at all, because two sources give two quantities and no source settles one. The record itself does not say what became of the arrangement these figures describe; no later value is given for either number.

A single figure here is not a headline sum compared with a drawdown; it is one figure with its rounding retained, standing beside another figure from elsewhere that the record holds apart. This study reads the record's refusal to merge as the archive's own answer to that ambiguity, and reads it as an answer about provenance rather than about magnitude.

7.6Figures compared across the ledger

The counts across all eighty-four cases, from "The ledger's numbers", are:

The hypothesis goes against: of the fifty-seven cases in which a figure is present, twenty-six stand against the separation and only seven support it, with twenty-four undecided. The count leaves out the cases in which the record holds no figure or no act to carry one. Where the two readings of the ledger differ, the case counts as undecided rather than for or against.

The same verdicts, group by group, are:

Three features of that table deserve reading rather than skipping. First, the credit-history group, which contains Korea's Stand-By Arrangement and Mexico's Moratorium, is the only group that leans toward F4 without reaching it: five of its cases keep announced and available apart, four stand against and seven are undecided. Second, the price windows are twenty-nine cases of observed prices with no actor and no act, of which twenty-eight carry no act and one is not settled; in twenty of them one figure stands with nothing to set it apart from and nine are not settled, because there is no approval and no drawdown in such records to separate. A verdict reached by such records is a verdict about how a price record is written, and this chapter reads it that way rather than as evidence about emergency acts. Third, the money group carries the strongest against reading in its small count, with one of its five applicable cases supporting and three against.

What a figure set apart from its availability lets a reader check later is specific. If the record names an approved total and an immediately available amount, a later document can be compared with both: a disbursement report, a drawdown record, a list of tranches released after review. If the record names a ceiling at an earlier level, a later announcement can be read as a raise, a hold or a cut against that level. If the record names a deferral term in days and separates it from the period requested, a later record of the negotiation can be matched against both numbers. In each case the separation creates a second date-and-quantity against which a third document can be tested, and that third document is what would tell a later reader whether the announced sum was ever drawn, and by whom.

What a single figure hides is, in the same terms, everything the separation would have carried. It hides the period over which the sum was available, the entity that could draw it, the stated basis on which it could be drawn, the conditions attached to any part of it beyond the first, and the distance between what was announced and what was taken. Where the figure is a ceiling, a single number hides that it is a maximum; where the figure is a term, a single number hides whether the term was requested, agreed or extended; where the figure is a rounded commitment, the single number hides its source and its precision. The records of this chapter each show one of these hiding-places in their own words, and the ledger's counts show that the hiding-place is common rather than rare.

Seldom do cases name an observable condition, date or event on which their act would end or be revised; that count is reported in Chapter 12 beside the others, and this chapter does not pre-empt it. What F4 adds to it is a separate point: even where a record names a term or a ceiling, the figure may still stand alone, and a reader given only the headline sum cannot tell what could be drawn of it. A reader given the record's own sentence, on the other hand, has a promise with a term, and a promise with a term can be checked against the record of what happened inside its term.

Chapter 8 takes up the other long kind of emergency act, the founding, in which the figure is not a line to be drawn but a charter or a statute meant not to end.