This chapter sets two bodies of writing side by side. One treats a rule known in advance, followed by a single holder of the reserve, as what calms a panic. The other treats a common measure, acted on by many holders at once, as what makes one. What this chapter adds is a reading of Bagehot's Chapter VII as standing on both sides: his rule never asked the many holders of money to differ from one another, only that one holder, with a known rule, should act against them when the others all act alike.
2.1The holder who stands apart
Charles Kindleberger's Manias, Panics, and Crashes (1978) is a history of financial crises told as a recurring sequence: displacement, boom, euphoria, distress and panic. It is also a history of the lender of last resort (Kindleberger 1978). This study reads the pairing as giving a place to one actor whose conduct in a panic is meant to differ from everyone else's. That reading is the study's own addition to the book's sequence and history.
Bagehot's rule, as this study uses it, is that in a panic the holders of the final reserve should lend freely, at very high rates (Bagehot 1873, Chapter II and Chapter VII). Chapter II states the rule for the case where a foreign drain is added to a domestic drain, and gives the remedy in these words: "Very large loans at very high rates are the best remedy for the worst malady of the money market when a foreign drain is added to a domestic drain." A familiar later summary of the rule circulates in shorter form; that summary is not Bagehot's sentence, and this study does not treat it as his.
On this side of the debate, what steadies a panic is a rule that the public knows in advance. The last paragraph of Chapter VII says as much in its own terms. Confidence in the size of the Bank's reserve and in its efficient use is the best palliative to a panic, and until the public and the Bank share a clear understanding on the point, the liability to crises and the terror at them will be greater than they need be. The rule works through expectation: if the public expects the holder of the reserve to lend, no one need hoard against the day when lending stops.
2.2Measures that move what they measure
The second body of writing starts from measures rather than from rules of conduct. Jón Daníelsson's article (2002) argues that risk models such as value-at-risk perform worst in the crises they are meant for. When many firms use such models under regulation, their use changes the market. Risk, in his term, is endogenous. The models can therefore deepen the crises they measure (Daníelsson 2002).
Donald MacKenzie's An Engine, Not a Camera (2006) makes a broader claim about financial models, above all option pricing. Such models can change the markets they describe, not only record them. The name for this is performativity (MacKenzie 2006).
This study reads the two together as one claim about shared measures. A measure that many holders act on is part of the market it measures: when it rises, the holders who use it respond, and their responses are among the things it will record next. A measure that one holder uses alone may be a fair picture of risk. The same measure in every hand is a signal that moves every hand together.
Michael Power's Organized Uncertainty (2007) argues that risk management is a way for organisations to turn uncertainty into objects they can manage and account for (Power 2007). This study reads a common measure as one such object, one that makes a reserve countable; it does not take from Power any claim about why many holders read the same number alike.
2.3Liquidity that leaves together
Richard Bookstaber's A Demon of Our Own Design (2007) carries the argument from measures to markets. Markets are made complex and tightly coupled by their own innovations, and liquidity vanishes when it is most needed (Bookstaber 2007). This study draws from it a point about counting: funds that can be counted are not always funds that can be used. A holder may own enough on paper and still be unable to pay, because the means of turning holdings into payment has gone at the same moment for everyone. That gloss is the study's.
Charles Perrow's Normal Accidents (1984) argues that interactive complexity and tight coupling are two independent dimensions of a system, and that in systems high on both, accidents are normal in his sense: to be expected from the structure, not from a single error (Perrow 1984). Perrow wrote about high-risk technologies, not money; this study borrows only the idea of tight coupling, and applies it in Chapter 5 to a chain of payments.
2.4The debate: does a common rule steady a market or shake it?
On the first side, a rule known in advance, followed by one holder, steadies a panic, because it tells the public that one holder will lend. On the second, a common measure acted on by many holders makes the panic, because it tells every holder to pull back at the same moment. The table is this study's summary of the works as it uses them, not the authors' words.
| Side | Works | What the work argues |
|---|---|---|
| A rule for the holder who stands apart | Bagehot 1873 (Chapters II and VII); Kindleberger 1978 | Bagehot: lend freely at very high rates in a panic. Kindleberger: a crisis as a sequence, and the lender of last resort |
| Measures that move what they measure | Daníelsson 2002; MacKenzie 2006; Power 2007 | Daníelsson: models used by many firms under regulation change the market and can deepen crises. MacKenzie: models can change the markets they describe. Power: risk management turns uncertainty into objects to manage |
| Liquidity that leaves together | Bookstaber 2007; Perrow 1984 | Bookstaber: liquidity vanishes when it is most needed. Perrow: in tightly coupled, complex systems accidents are to be expected from the structure |
One side praises a known rule and the other blames a common measure. Are they speaking of the same kind of rule? A rule that tells one holder to lend when others refuse is shared knowledge about a single actor. A measure that tells every holder when to refuse is a shared instruction to the many. Both are common and known in advance, yet they may differ in the one respect that decides a panic.
2.5Where this study stands
This study holds that Bagehot's Chapter VII stands on both sides. It calls for a rule for the one holder of the reserve that the public can know. It also shows a board with a ratio and with the ordinary rules of its time, and no agreed account of what that ratio meant, as the board of that one holder.
What separates the two kinds of rule, on this reading, is the direction a rule points when its trigger comes. A rule may point with the crowd, telling its holder to hold back just as everyone else holds back. It may instead point against the crowd, telling its holder to lend when others will not. The second question is whether any holder stands apart at all.