1
The Evolution of Market
Monetary Institutions
•In the beginning, goods were bartered directly for other goods. A theoretical
account of the evolution of monetary institutions naturally begins with
an attempt to explain how the earliest money emerged from a non-monetary
or barter economy. "Money" here, following standard usage among economists,
means a commonly accepted medium ofexchange. A "medium of
exchange" means a good that people acquire through trade with the intention
of trading away later (rather than consuming for its own sake or using
up in a production process). "Commonly accepted" means that the money
good is routinely offered and taken in trade for other goods, and so appears
on one side of nearly every transaction. The theory that follows aims to
explain why, and how, some good should acquire these characteristics.1
The Austrian economist, Carl Menger (1892), developed the classic explanation
of the origin of money. Menger showed how money can emerge from
barter without anyone inventing it, or to use Adam Smith's phrase, "as if by
an invisible hand." In his account, money emerges through a series of steps,
each based on self-seeking actions by individual traders, without the resulting
social order (monetary exchange) being part of anyone's intention. This
1
For a further teasing out of these defining terms see White (1989, ch. 11). There I used the
modifier "generally accepted"; here, following Wameryd (1990), I use "commonly accepted"
in the same sense. The present chapter draws on Selgin and White (1987) and White (1989,
ch. 9). For complementary accounts of these issues, see Glasner (1989, ch. 1) and Dowd
(1996, ch. 1).
2 THE EvoLUTION oF MARKET MoNETARY INSTITUTIONS
is a satisfying mode of explanation because it does not require heroic assumptions
about the knowledge possessed by any trader.
A number of writers before Menger expressed the idea that money was an
undesigned or spontaneously emerged institution. Among them are Adam Smith,
the French economists Etienne de Condillac and Destutt de Tracy, and the British
monetary pamphleteers, Thomas Hodgskin and Samuel Bailey. Menger
was certainly aware of Smith's writings, though he does not cite Smith in this
context. However, none of these earlier writers spelled out the emergence of
money step by step. The typical modem textbook discussion of the origin of
money is plainly inadequate.2
It lists the problems of barter exchange, and
shows that monetary exchange overcomes these problems. A prototype can be
found in Aristotle (Nichomachean Ethics, Book 5): "All the things which we
exchange need to be comparable. This need led to the invention of money to
serve as a medium giving value to every thing." Unfortunately, the simple contrast
between problems and solution does not explain how the solution (money)
was arrived at, any more than a list of the advantages of standard time zones
would explain how they came about. One is left with the impression that
barterers, one morning, suddenly became alert to the benefits of monetary exchange,
and, by that afternoon, were busy using some good as money. In one
version of the story, a wise head of state introduced the idea that a certain
commodity was to be sanctioned as a general medium of exchange.
'Taken seriously as a theory of the origin of money, this account would
suggest that the idea of money was fully grasped before money existed.
Money would be an invention, like the telephone, which existed in someone's
mind before a prototype was produced. In fact, money is not a product
of technological advance brought forth by a single mind or a research laboratory.
This is evident from the fact that gold dust or salt, used as money, is
not technologically different from gold dust or salt, not used as money. What
transforms gold dust or salt into a money is not some physical change, but
rather the development of a social convention concerning the use of that
good. The use of any particular item as money is a social convention, in the
same sense that the use of particular utterances or gestures to communicate
particular ideas is a social convention. Each of us (in an English-speaking
group) calls a certain fruit an "apple" because that is what everyone around
us calls it, and we wish to communicate with them. Likewise, each of us
uses item x as a medium of exchange because nearly all others in our society
do, and we wish to trade with them.
A money could not spring forth full-blown from barter unless people
throughout a society simultaneously arrived at the idea of using x as a medium
of exchange, and each person knew that he could count on others to
2 Notable exceptions are McCulloch (1982) and Goodhart (1989b).
MENGER's THEORY RESTATED 3
do so too. Such a scenario begs too many questions. It invokes the realization
of money in the attempt to explain how money was realized. It attributes
knowledge of the benefits of money to people who would not have
such knowledge in a barter economy.
Menger begins by emphasizing the "mystery" of money: why is everyone
willing to trade truly useful goods and services for mere tokens? In Menger's
day (a century ago), these tokens were otherwise practically useless disks of
gold and silver, or slips ofpaper (banknotes) representing claims to such disks.
Today, the mystery is even greater, as the tokens are otherwise completely
useless disks of cupro-nickel and slips of paper interchangeable with them.3
Menger's approach does not apply only to commodity money, though it
was originally framed to explain such money. It emphasizes that the use of
a commodity money has a "conventional" aspect, the convention being one
that develops through a historical process. By extension, the use of a fiat
money rests on the prior development of a commodity money convention,
because fiat money is launched by suspending the redeemability of claims
to a commodity money. However, we are getting ahead of the story.
It is worthwhile restating Menger's theory in detail for several reasons.
Our immediate interest, here, is its usefulness in explaining the origin of
money. Later in the book, we will return to the theory because it has implications
for the viability of projects to establish a new money, or a payments
system without money. The theory also draws out certain "essential features"
of money that have implications for the macroeconomic properties
of a monetary economy (Yeager 1968). Finally, the theory holds a general
interest to students ofthe social sciences because it provides a paradigmatic
example of an invisible-hand explanation of a social institution.4
A simple barter economy faces each trader with the problem of finding a
trading partner with preferences and endowments reciprocal to his own. (This
has come to be known as the problem of finding a "mutual coincidence of
3
In recent years, a number of monetary economists have offered non-evolutionary models
of money as solutions to the mystery of a positive value being accorded to "intrinsically
useless" and inconvertible fiat money: in particular, the overlapping generations model (Wallace
1980) and search-theoretic models (Kiyotaki and Wright 1989, Ritter 1995). Menger's solution
is different, and is less subject to the cogent criticisms made of the overlapping generations
model (Tobin 1980, McCallum 1983), of other general equilibrium models of money
(Bryant and Wallace 1980), and of search-theoretic models (Selgin 1997b). Of course, it is
subject to other criticisms.
4
It has been cited as such by Nozick (1974, p. 18), though Nozick actually cites a restatement
of Menger's theory by von Mises (1980).
4 THE EvoLUTION OF MARKET MoNETARY INSTITUTIONS
wants.") For Alanis to trade the asparagus she brings to market for the bacon
she prefers to take home and consume, via a direct pairwise exchange, she
must find some other trader ("B") who both has what Alanis wants (bacon),
and wants what Alanis has (asparagus). It may be difficult or even impossible
to find such a match, even when a unanimously preferred reallocation
of goods could be arranged in another way. McCulloch (1982, ch. 1) offers
a simple example: imagine three individuals and three indivisible goods,
where A has good 1 and prefers good 2 (but not good 3) to good 1, B has
good 2 and prefers only good 3, and C has good 3 and prefers only good 1.
Clearly, all are better off if good 1 goes to C, good 2 goes to A, and good 3
goes to B, but there is no pairwise exchange that makes both traders better
off. More generally, even where a pairwise trading partner could be found,
it may be difficult and time-consuming to find that trader among the many
in the marketplace.
A trader who is frustrated by her inability to find a rare or non-existent
matched trading partner need neither continue the effort fruitlessly nor give
up and go home. There is an alternative. Consider the three-agent case just
described. Suppose that each pair of individuals has met, and has quickly
discovered that only one party wants to make each possible pairwise trade.
Alanis has discovered that Bjork, who is selling the good that Alanis prefers
(bacon), does not prefer what Alanis has to offer (asparagus). In this situation,
it would not take too much cleverness on Alanis's part to ask Bjork
what good Bjork would prefer. Learning that Bjork prefers cabbage, which
Coolio has offered to Alanis in exchange for asparagus, Alanis will be led
by self-interest to trade with Coolio, even though Alanis does not want to
consume cabbage. Alanis will then be in a position to make an offer for
Bjork's bacon that Bjork will accept.
The general point illustrated by this example is that, potentially, a barterer
can economically achieve a preferred holding of goods by exchanging her
initial endowment for some good which can then be turned around and exchanged
for the good(s) she ultimately wants to consume. This practice is
known as indirect exchange, in contradistinction to the direct exchange of
simple barter. In the example, individual A has used good 3 as a vehicle for
indirect exchange or, as it is usually put, as a medium ofexchange.
Now consider a larger barter market, such as a trade fair, with anonymous
traders selling many goods. (As noted below, such a fair would
not historically have been found in a barter economy, because specialization
and trade could not develop far where trade remained so difficult to
accomplish.) To keep the discussion simple, assume that each trader still
arrives endowed with a single indivisible good and desires to take home
some one other good, though several traders may now be selling each good.
As before, every trader besides Alanis is trying to use direct exchange, and
will agree to trade only for the one good she wishes to consume. Alanis,
MENGER'S THEORY RESTATED 5
again coming to market with asparagus, will find that her trading problem is
now more difficult. With a larger number of traders, it may take more time
to discover a trader B who is selling the bacon that Alanis wants to buy. The
probability that this B wants to take home asparagus is smaller than before.
Once Alanis learns what good j this trader B will agree to accept in exchange,
it will take more time to find which (if any) among the sellers ofj
will accept asparagus in exchange. In this setting, it can easily be the case
that Alanis acquires the desired bacon most economically neither by direct
exchange, nor by waiting until she meets a seller of bacon to learn what
goods can potentially be used as a medium of exchange. Instead, Alanis's
best trading strategy is to exchange her asparagus for a good k which any
seller of bacon (and, for that matter, of good j) that Alanis may happen to
meet is relatively likely to accept.
Here Menger introduces the concept that different goods have different
degrees ofmarketability. Marketability is a "non-Walrasian" concept: in a
Walrasian general equilibrium model, with costlessly coordinated trade, and
with a single price at which a good may be either bought or sold, every
good is perfectly marketable. In a world ofcostly trade, it takes some amount
of time, effort, and expense to sell for a good price. (Anyone who has ever
tried to sell a used car knows this.) A more highly marketable good is a
good that is easier (less costly) to sell for a good price. A "good price" here
means a price close to the best price that could potentially be found with
full information on both sides of the market. (Menger calls this best price,
an "economic price.")5
A perceptive barterer will exchange her initial endowment goods for more
highly marketable goods, which can then easily be exchanged for the goods
he or she wants to consume. She wants to maximize her expected gains
from trade, which obviously depend on the prices at which trade takes place,
net ofthe costs of finding trading partners (costs of search), and of the costs
of consummating trade (costs of transportation, contracting, and the like).
Indirect exchange requires two trades, instead of one. It is therefore more
likely to be advantageous to the extent that:
1 the good to be used as a medium of exchange is more widely consumed,
and traded, than the endowment good, and hence trading partners
offering a good price in it, and for it, are easier to find; and
2 the costs of buying, holding, and reselling it (costs of contracting,
spoilage, and transportation) are relatively small.
5
Because marketability has at least three dimensions (time cost, other selling costs, and
percentage of economic price realized), it may not always be possible to rank the marketability
of different goods unambiguously.
6 THE EVOLUTION OF MARKET MONETARY INSTITUTIONS
Where indirect exchange is advantageous, it pays a trader to accumulate an
inventory of highly marketable items for use as media of exchange.6
Having
highly marketable items on hand allows a trader to find good buying
prices more easily for the things she wants to consume.
Other alert traders in the market, facing the same situation, will adopt the
same strategy of indirect exchange. Menger notes that some individuals
may not catch on to the advantages of indirect exchange immediately or on
their own. Eventually though, they will notice the success enjoyed by those
who are trading their produce for a medium of exchange rather than persisting
in attempts at direct exchange. They are then likely to imitate the practice
of using indirect exchange.
Once many individuals are using indirect exchange, the stage is set for
social convergence toward a common medium ofexchange. One perceptive
trader, say M, will learn from experience which commodities are most marketable,
and best suited for use as media of exchange. The knowledge that
he can unload them easily will lead him to accept these commodities all the
more readily, and in preference to other commodities. M's greater acceptance
of a good k incrementally reinforces its usefulness as a medium of
exchange for other traders, A ... L and N ... Z, because they can count on
one more place to spend it. Its marketability for them has increased. They
may learn of good k's improved suitability as a medium of exchange, either
through communication, or from trial-and-error experience, or as a last resort
by imitation of the successful traders. Traders N, 0, and the others will
then accept good k more readily, just as M did earlier. Again, each trader
who does so reinforces its usefulness for the others. With every trader preferring
more marketable to less marketable media of exchange, ultimately
one good (or at most a few, covering different sets of transactions) is elevated
to the status ofbeing commonly or generally or routinely accepted as
a medium of exchange. It becomes money.
This theory is not meant to suggest that extensive specialization and market
trade historically antedated the emergence of money. On the contrary, it helps
to explain why specialization and trade developed simultaneously with money,
a fact Leijonhufvud (1981, pp. 229-30) has emphasized. Pre-monetary communities
were basically autarkic (Dingle 1988). Direct exchange is so difficult
that the scope of specialized production "for the market" is limited by the
6 This was pointed out long ago by Adam Smith (1981, pp. 37-8): "In order to avoid the
inconveniency of such situations [in which the seller of a desired good does not want
the produce the would-be buyer has to offer], every prudent man in every period of society,
after the first establishment ofthe division of labor, must naturally have endeavoured to manage
his affairs in such a manner, as to have at all times by him, besides the peculiar produce of
his own industry, a certain quantity of some one commodity or other, such as he imagined few
people would be likely to refuse in exchange for the produce of their industry."
SOME IMPLICATIONS OF THE THEORY 7
scope of indirect exchange. Thus Adam Smith's dictum, that the division of
labor is limited by the extent of the market, may be supplemented by the observation
that the extent of the market is limited by the extent of money's use.
The Mengerian theory helps us to understand the important characteristics
of a monetized economy that are the result of these evolutionary origins of
money. Menger emphasized the following three.
Everyone in a monetary economy routinely accepts money, and routinely
attempts to trade output or endowment goods for money before
acquiring consumption goods.
2 The ability to purchase goods at "the going rate" with money is not at
all doubtful, even in anticipation ofdealing with an anonymous seller.
There is virtually no risk of meeting a seller who refuses to accept
money, or accepts it only at a discount.
3 Sellers are reluctant to accept goods oflesser marketability than money,
with the result that the marketability of the money good is discontinuously
greater than that of any other good. A buyer (with money)
has markedly less difficulty trading at close to economical prices than
does a seller (of a non-money good). As Menger notes, being forced
to sell on short notice imposes much more of a burden than being
forced to buy on short notice.
The theory also establishes that no collective decision or legislative act is
necessary for money to emerge. Menger emphasized this point with respect
to the defining characteristic of money: its general acceptance as a medium
of exchange. Money did not originate from, or fail to perform its mediumof-exchange
function fully until endorsed by, the legal decrees of rulers.
By extending Menger's theory, we can see that the role of money as a
"unit of account" also arises spontaneously. The "unit of account" means
some definite quantity of a good used as a pricing and accounting unit.
Strictly speaking, as Jurg Niehans (1978) has pointed out, it is not proper to
say that money is a unit of account, because money as such is not a unit.
Money is rather a medium of account. The unit of account is a specific
quantity of the good constituting the medium of account. For example, silver
may be the money and medium of account, while the "ducat" (defined
as so many grams of standard-fineness silver) is the unit of account.7
7
The ducato d'argento of Venice, 1201-1355, was 96.5% fine and weighed 2.18 grams.
8 THE EvoLUTION OF MARKET MoNETARY INSTITUTIONS
The medium of account in an economy will naturally coincide with its
commonly accepted medium of exchange or money. A seller pursues his
self-interest by posting prices in terms of the good or goods he is routinely
willing to accept in exchange. If this seller were to post prices in terms of
some other good, he would incur the extra trouble, for himself and for his
customers, of keeping track of and figuring in the current exchange rate
between the pricing medium and the payment medium he is willing to accept.
An accountant could not calculate profit and loss as easily, or as clearly,
were she to keep books in units of a commodity other than the commonly
accepted medium of exchange in which the income actually accrues, liabilities
come due, transactions balances are held, and for which other assets
can most readily be exchanged.
Accordingly, some common unit ofmoney naturally becomes the unit of
account. It may be an early popular coin (e.g. the Spanish dollar), a simple
bullion weight (e.g. the "pound sterling"), or a natural unit (one standardsized
cowry shell). No official proclamation is necessary to establish a standard
unit of account, any more than to establish a standard building brick.
Commercial practice can converge on a conventional unit, without any collective
decision being taken, in much the way that it converges on a money
commodity. Each seller will discover that he does best for himself by posting
prices in the unit most popular with his potential trading partners. Court
decisions can, and historically did, follow trade custom in deciding how
much of what fineness of what metal would legally satisfy a contractual
obligation to deliver a "ducat" (or whatever money unit). The courts did not
create customary units.
There is no denying, of course, that governments often have played a role
in pushing a new money or unit of account. The point is that no collective
deliberation or action is needed for money to emerge fully, or was historically
instrumental in its original emergence.8
Money only makes sense, given its origin in indirect exchange, in a world
with certain features. There must be three or more traders, and at least two
goods besides money, for otherwise only direct exchange is possible. There
must be varying degrees of marketability for these goods, or some physical
feature of other goods that makes them less suitable for shopping with;
otherwise, there is no advantage to indirect exchange. These features are
8 John Maynard Keynes (1935, pp. 4--5), even while asserting the relevance to the modem
world of "the doctrine that money is peculiarly a creation of the State," recognized that the
original establishment of a conventional unit of account preceded government involvement:
"Thus the Age of Money had succeeded to the Age of Barter as soon as men had adopted a
money-of-account. And the Age of ... State Money was reached when the State claimed the
right to declare what thing should answer as money to the current money-of-account- when
it claimed the right not only to enforce the dictionary but also to write the dictionary."
FROM SIMPLE CoMMODITY MoNEY TO CoiNs 9
not always present in abstract economic models purported to be models of
monetary economies. In a world where all goods are perfectly and costlessly
marketable (so that there are no economic barriers to barter), there is no
rationale for a distinctive money. Barter with interest-bearing financial assets
would dominate the use of any non-interest-bearing money.
An evolutionary or neo-Mengerian perspective can help to explain the emergence
ofgold and silver as the predominant commodity monies in the world,
and the later emergence of such monetary institutions as coinage and bankissued
paper money.
The earliest form of money, following Menger's account, must have been
a useful commodity. A good must have acceptability in barter before it can
acquire wider acceptability as a medium of exchange. It must have some
usefulness as a commodity to be accepted in barter. Anthropological evidence
indicates that the goods that became monies in several cultures originally
had ornamental uses (Melitz 1974).9
This was true of Pacific and
African shell monies, North American wampum, and, also, gold and silver.
Other primitive monies have been foodstuffs, like grain or salt.
The eventual predominance of gold and silver as money, over other commodities
which early on would have had equally wide acceptability, can be
explained by at least four (partly physical) characteristics that promoted
their ready marketability and convenience (low usage costs) as media of
exchange. These characteristics were a staple subject ofdiscussion in moneyand-banking
texts during the era of metallic monetary standards.
Goods like livestock or tobacco, whose quality is variable and difficult
to assess, are more troublesome to exchange than goods of uniform
and easily recognized quality.10
Pure gold and silver, as chemical
elements, are absolutely uniform. The purity (fineness) of a particular
piece of gold or silver can be tested at low cost by biting it, sounding
it, or (with a bit more trouble) by assaying it. Traders were
commonly experienced in these assessment methods in the past. As
9 I used to add "or ceremonial uses," until I discovered that "ceremonial use" is the anthropologist's
shorthand for "we have no idea what it was used for."
10 Armen Alchian's account (1977) of "Why Money?" relies exclusively on low authentication
costs for selecting which commodity will become money. See also King and Plosser
(1986). Other things equal, this characteristic can be decisive but, more generally, it is only
one characteristic among several that can play a part in promoting a commodity's use as a
medium of exchange.
10 THE EVOLUTION OF MARKET MONETARY INSTITUTIONS
will be discussed below, coinage arose to relieve the difficulties created
by the non-uniformity ofgold and silver in rawer forms (nuggets
or dust or ingots).
2 Gold and silver are durable, so that there are no extra carrying costs
due to spoilage. The deterioration of goods like grain and olive oil
makes them costly to hold in inventory. The possibility of deterioration
also creates the above-mentioned problem of exchange being
encumbered by the need for costly verification of the goods' current
quality.
3 The precious metals are easily divisible andfusible, so that payment
can be tailored to purchase size. Large pieces can easily be split into
small pieces, and small pieces can be united to form larger pieces.
This is not true of jewels or, certainly, of livestock.
4 Finally, gold and silver are portable, that is, have high ratios of value
to bulk. Portability means a low cost of taking the medium of exchange
from the site where it is acquired to the site where it is spent.
Commodities like salt lost their suitability as media ofexchange when
their value per pound became too low. The copper money of seventeenth-century
Sweden, a non-precious metallic money, was notoriously
cumbersome. Individual pieces of copper "plate money"
eventually weighed up to 20 kilograms (44 pounds). Strong young
men had to be employed to carry the copper necessary to make an
ordinary-sized commercial purchase. Finally, Swedes stopped using
copper, except in the smallest transactions. A similar process may
have promoted the historical dominance of gold over silver in international
payments of large sums: it was less costly to send one boat
laden with gold than to send fifteen laden with silver. 11
The displacement of one money by another can follow the general
Mengerian logic of a self-reinforcing convergence process. As individuals
from two regions with different commodity monies come into contact and
begin to trade, an entryway is created for the better of the two monies to
spread to the other region. Traders on the margin, not only those physically
adjacent to the border but also merchants who do a large fraction of their
trade with users of the foreign money, will favor the foreign money, if it is
markedly better in some of the four areas listed above. Merchants and border-dwellers
will accept the favored money on somewhat better terms, and
11 Fleming (1994) finds, however, that the general historical switch from silver to gold standards
was not market-driven. It was, in fact, mainly due to the legal overvaluation of gold
relative to silver by the governments of Britain and the US, which set Gresham's Law in
motion (the legally overvalued or "bad" money drove out the legally undervalued or "good"
money). Other nations deliberately followed suit in a sort of bandwagon effect.
BANK-ISSUED MONEY 11
can use it among themselves where, before, they used the local money. The
margin can then spread: those who deal substantially with these merchants,
and those who live adjacent to the areas adjacent to the border, can find it
advantageous to be paid in the foreign money. Its sphere of acceptance can
snowball, following the Mengerian logic, until a single money unites the
two regions.
Coinage, the practice of fashioning monetary metal into standardized
marked discs, though it involves technical advances and not merely the
formation of a social convention, also developed in step-by-step fashion.
Where nuggets or gold dust served as money, merchants had to assess weight
and quality when accepting payment. It made sense for a merchant to mark
a piece of assessed gold, so as to avoid the costs of re-assessment when
paying the piece out later. Other traders who trusted this merchant could
then also rely on his mark. To prevent the possibility of shaving off gold
around the marked area, the piece could be covered with marks. Punching,
stamping, and finally modem methods of minting developed as low-cost
methods of fashioning reliably marked pieces of gold. Historical examples
of these stages can be observed in the money of ancient Lydia (Burns 1927;
Cribb 1986).
Mints arose spontaneously, then, to meet the demand for authentication
services. With the development ofcoinage, the marketability ofcoined metal
became discontinuously greater than that of uncoined metal (in this context,
branded bars ofbullion may be thought of as large coins). Gold miners
found it much easier to spend coined than uncoined gold, and, therefore,
were willing to pay for the service of minting their raw gold into coins.
Numismatic publications indicate that more than twenty private gold and
silver mints operated during the gold and silver rushes in nineteenth-century
America (Kagin 1981), and one in Australia (McDonald 1987, p. 122).
In practice, governments have typically monopolized the coinage industry,
but there are no signs that coinage is a natural monopoly. There are
ample signs that governments have wanted to exercise monopoly over money
production so as to reap the monopoly profits known as seigniorage (Selgin
and White 1999). In a later chapter, we will consider in more detail both
seigniorage in its medieval form, and seigniorage in its modem form of the
profit from monopoly issue of fiat money.
The next step to consider, in the evolution of monetary institutions, is the
emergence of money issued by commercial banks. Full-bodied coins (and
other types of full-bodied commodity money, like shells) originate outside
of any commercial banking system. We may call them "outside" money,
12 THE EVOLUTION OF MARKET MONETARY INSTITUTIONS
whereas bank-issued money is "inside" money. Outside money is an asset
for its holder but not a liability of, or financial claim against, anyone else.
The media of exchange produced by a commercial bank, by contrast, are
claims against it. A large literature attempts to explain why banks exist as
intermediaries between savers and borrowers (Santomero 1984).12
Our object
here is, rather, to explain why banks participate in the payments system,
by offering a logical evolutionary account of why and how claims
against banks came to be used as money.
The earliest bank liabilities were claims to outside money deposited with
bankers. Historical records indicate that bankers in medieval Italy began as
money-changers, but by AD1200 had moved into accepting time and demand
deposits (de Roover 1974a, 1974b). In a region of numerous citystates,
each with its own distinct coinage, money-changers provided the
service of trading local coins for the less spendable foreign coins brought
by inbound merchants and other travelers, and of trading the reverse way
with outbound travelers. A simple explanation of why money-changers became
deposit-takers is that merchants found it easier to leave money with
them "on account," to be called for when needed, rather than to take away
domestic coin equal in value to the foreign coin tendered (or vice versa) on
every occasion. Essentially, this means that the money-changers' vaults were
being used for temporary safekeeping of coin. In this respect, the development
of deposit banking in Italy was similar to its development in England
where, according to numerous accounts, early deposits were taken by goldsmiths
whose vaults provided safekeeping.
Bank deposits began to play a monetary role when they became a medium
of exchange, that is, when transfer of deposit balances became an
accepted method of payment among bank customers. The practice of deposit
transfer evolved by steps. Where a bank provided safekeeping services,
depositors no doubt discovered cases in which party Alice planned to
withdraw coins from the vault and laboriously transfer them to party Bob,
who in tum planned to lug them back to the same vault and redeposit them.
At the end of the day, the coins were back where they started, Alice's deposit
balance had been reduced, and Bob's balance had been enlarged by
the same amount. Only a little imagination was needed for Alice and Bob to
recognize that an easier method of accomplishing this result would be for
them to meet in the banker's office (in the coin-lugging method, both had to
12 An intermediary is an institution that issues financial claims (debt or equity) against itself,
and uses the proceeds to acquire financial claims on other agents. Because it is irredeemable
and not a financial claim, fiat money is outside rather than inside money, and an institution
that issues it (typically a central "bank") is not, in that respect, acting as an intermediary. The
text's distinction between outside and inside money is different from the one used by Gurley
and Shaw (1960).
BANK-ISSUED MONEY 13
go to there anyway) and there persuade the banker simply to transfer the
desired amount of deposit balances on his books. Alice and Bob thereby
avoid physical lugging around of coins, which simply stay in the vault.
Early banking documents, studied by de Roover (1974a; 1974b), record
such three-way meetings among payer, payee, and banker to authorize deposit
transfers.
Later develdpffients made transfers still easier to accomplish. Written slips
for authorizing transfers made it unnecessary for both parties to travel to
the banker's office. (In a checking system, Alice hands Bob a check, and
only Bob goes to the bank, to deposit it; in a "giro" system, only Alice goes
to the bank, to authorize the transfer into Bob's account.) Today, we see the
growing use of electronic funds transfer, that is, methods of authorizing
deposit transfers using electronic messages (sent using a telephone, home
computer, automatic teller machine, or debit card and point-of-sale terminal)
in place of slips of paper. These methods do not change the nature
of the payment system as one of deposit transfer. The "front end" of the
deposit transfer is different from writing a check, but not the "back end"
(what happens on the bank's balance sheet). Nor- despite excited predictions
that the future holds "a world without money" - do they threaten the
definition, or real existence, of money. The depositor's bank balance, not
the transfer-authorization device (e.g. the check), is money.
In addition to deposits, bank-issued claims in currency form were important
historically, and may soon become important again. Banknotes are
bank-issued claims to outside money that are not in any customer's name,
but are payable to (redeemable by) whoever happens to be the bearer. Such
bearer claims are transferable without the bank's knowledge or involvement
- Alice simply hands them over to Bob - and can change hands repeatedly
before being redeemed. Today some versions of "smart card"
payments, namely those like Mondex which allow transfer of balances directly
from card to card without the bank's knowledge or involvement, amount
to the reintroduction of banknotes in digital form.
Banknotes may have evolved from the practice of making payment by
signing over a deposit receipt or cashier's check. When such payments are
foreseen, depositors could ask for deposit receipts in round denominations
for convenience, and in bearer form, to streamline and certify the payment.
Payment was streamlined because signing over is no longer necessary. It
was certified in the sense that the bearer note is a claim against the bank
only, and not against any account that might have insufficient funds, nor
against any subsequent endorser. No one who accepts a banknote- unlike a
deposit receipt that is successively signed over - needs to worry that the
goodness of the claim depends on the funds of those who have previously
held it, or that he or she might be called upon to make good on it for those
who hold it subsequently. A banker is happy to comply with requests to
14 THE EVOLUTION OF MARKET MONETARY INSTITUTIONS
issue such claims, as a way of increasing his circulation and profits. According
to several accounts, this was the path by which goldsmith's deposit
receipts historically evolved into banknotes (Usher 1943, Richards 1965).
The widespread use of banknotes historically preceded the widespread
use of checking accounts (Bagehot 1873). For most British banks, note circulation
exceeded deposits up to 1850. For banks in other countries, the
date at which deposits began to exceed notes in circulation came even later.
Ifbanknotes evolved from deposit receipts, however, deposits on some scale
must have preceded the use of banknotes.
Banknotes historically have paid no interest, even in competitive settings
where deposits have, because there seems to be no easy way to pay interest
on a bearer instrument whose convenience rests on its circulating at face
value. In smart card payment systems thus far test-marketed, card balances
similarly do not bear interest. The view that bank-issued bearer claims should
be expected to bear interest in a competitive banking system will concern
us in the last chapter of this book.
Suppose a payments system has a common standard money, arrived at in
the Mengerian way, with many banks issuing redeemable currency and deposit
liabilities, but each bank refuses to accept any other bank's liabilities
at par (face value). Bank-issued money then has limited marketability. This
section argues that the profit motive, without legal compulsion, will move
the banks toward par acceptance of one another's currency and deposits. An
important side effect is the formation of an institution for interbank clearing
and settlement of currency and deposit claims. The exposition refers explicitly
to banknotes, both for convenience and for historical applicability to
systems dominated by notes early on, but the argument applies just as well
to par acceptance of deposit claims and digital currency.
An individual who has come into possession of a sum of notes issued by
Bank X, and who wishes to deposit the sum into her account at Bank Y, has
two options when Bank Y does not accept X-notes. She may bear the expense
of taking the notes back to Bank X for redemption in outside money,
or pay a fee to a note-changer (in the form of a discount on the X-notes
and possibly a commission) who purchases the X-notes for outside money
or f-notes. Either option is naturally more expensive, the farther the
noteholder is from a redemption site for the notes in question. Given these
costs, X-notes are likely to circulate readily only in the vicinity of Bank X
offices. Coin or other brands ofnotes will be preferred for transactions else-
where.
In this situation, there are at least three logical scenarios whereby the
REGULAR PAR AccEPTANCE 15
pursuit of profit leads toward widespread par acceptance; Selgin and White
(1987, pp. 225-33) discuss these scenarios and offer historical illustrations:
1 Banks as note-changers
2 Note dueling
3 Mutual par-acceptance packs
Banks as note-changers: Banks can out-compete non-bank note-changers
because they have the advantage of being able to issue their own notes (or
deposit balances) to purchase other banks' notes. Where a non-issuing notechanger
must hold costly till money on the asset side of its balance sheet, an
issuing bank can hold interest-earning assets, giving it a profit from "float"
for as long as the notes issued remain in circulation. (The Suffolk Bank of
Boston in the 1830s succeeded famously at this business.) By swapping Ynotes
for X-notes, Bank Y can maintain a larger stock of its own notes in
circulation. Where the transactions and redemption costs of note changing
are low enough to be covered by the float profit from additional circulation,
competition will bring the issuing banks' note-changing fee down to zero.
That is, competition will bring the banks to practice par acceptance. (Bank
Y would never offer to buy X-notes at a price above par, because the notechanging
customer could make arbitrage profits by turning around and immediately
redeeming the Y-notes issued.) If all banks are thus drawn into
zero-fee note-changing, mutual par acceptance develops de facto.
Note dueling: Bank Ymay accept, or even aggressively purchase, X-notes,
and then, suddenly, return a large quantity to Bank X for redemption in
reserve money, hoping to gain a greater share of the banking market by
embarrassing its rival. The trouble with this tactic is that two can play it.
Bank X can collect and redeem Y-notes, both to return the damage, and to
replenish its own reserves. Vigorous pursuit ofthe tactic on both sides (known
historically as "note dueling," and practiced for example by the first and
second chartered banks in Scotland upon the entry of the second) may drive
note-changing commissions to zero. In a repeated game of this sort, however,
a non-aggressive "tit for tat" strategy (return an opponent's blows in
kind, but meet cooperation with cooperation) should evolve (Axelrod 1984).
When both sides practice note dueling, both find it ineffectual and expensive.
Greater non-earning reserves must be held at all times to meet a rival's
large redemption demands that may arrive at any time. If neither party can
win the duel, both should eventually recognize that a regular, and amicable,
exchange of collected notes would benefit both by allowing them to economize
on reserves. Fees will nonetheless remain zero as a defensive
measure, allowing each to collect enough rival notes to safeguard its re-
serves.
Mutual par-acceptance pacts: As the fee for buying Y-notes with X-notes
16 THE EvoLUTION OF MARKET MoNETARY INSTITUTIONS
falls, and a fortiori as it reaches zero (Bank Y offers commission-free par
acceptance ofX-notes), the cost of accepting X-notes falls and thus the circulation
of X-notes grows relative to the circulation of outside money. This
result is no part of the intention of Bank Yin the above scenarios, and may
not be anticipated before the fact. (The Suffolk Bank was evidently surprised
to find that by purchasing country bank notes at par, it was not vacuuming
them from circulation but, in fact, encouraging their wider
acceptance and use.) Bank Y aims only at an expanded circulation of its
own notes, which indeed is also a result. Banks that accept other banks'
notes at par improve the circulation both oftheir own notes, and ofthe notes
they accept. If two banks both recognize ex ante the availability of these
circulation gains from mutual par acceptance, they may explicitly enter a
pact to accept one another's notes at par. Par-acceptance pacts among pairs
ofprovincial Scottish banks provide historical examples of such agreements.
Acceptance at par in a wider area increases the marketability of each brand
of notes, and thereby the quantity willingly held by the public. The same
logic explains the recent spread in the USA and elsewhere, of agreements
among banks to form networks of mutual acceptance for cards giving access
to automatic teller machine services. By participating in an ATM network,
a bank improves the accessibility of its own deposits, and thereby
attracts more depositors.
Par acceptance, developed through any ofthese routes, is generally more
profitable, the wider its scope. The potential gains are not exhausted until
all reputable banks practice par acceptance toward all others. Thus every
bank's liabilities come to circulate at par throughout an economic region.
The boundaries of the region will lie where the circulation-enhancing benefits
of membership (presumably declining at the geographic margin as distance
from the financial center increases) become equal to the transaction,
transportation, and administrative costs of membership (presumably rising
at the margin). As transaction and transportation costs secularly fall, the
par-acceptance region expands. Par circulation of notes became nationwide
with the spread of railroads in the nineteenth century. ATM networks are
rapidly becoming global today.
Following any of these scenarios, Bank Y will be collecting X-notes, and
Bank X will be collecting Y-notes, during the course of the business week.
Each bank will want to redeem the collected notes, rather than to pay them
back out again (as the whole profitability of the arrangement comes from
placing and maintaining more of its own notes into circulation) or to accumulate
them indefinitely (reserve money is more useful, and no more costly
CLEARING ARRANGEMENTS 17
to hold). A regular meeting for bilateral redemption, where X-notes are traded
for Y-notes and the difference settled in reserve money, or some agreed substitute,
will be arranged when the banks find that it is cheaper than unilateral
or irregular redemption.
The practice ofregular bilateral redemption may emerge without the banks'
management planning it. If note-porters are sent from each bank to redeem
at the other bank's counter, and they happen to meet, it should readily occur
to them that an exchange of notes would save them time and the effort of
lugging a great deal of gold (all but the difference) back home. They will
arrange to meet regularly at a specified time and place to exchange notes.
Bank management will endorse the arrangement not only because it saves
transportation costs, but also because bilateral netting (using Bank Y's claims
against Bank X to offset Bank X's claims against Bank Y) allows smaller
reserves to be held.
The gains in going from unilateral to bilateral note-exchange are further
extended in going to multi-lateral exchange in a system of more than a few
issuers. Time and transportation costs are further economized by having one
all-encompassing meeting rather than numerous pairwise meetings, and the
holding of reserves can be further reduced with multilateral netting of claims
that in pairwise clearing would have to be settled in reserve money. Multilateral
exchange may evolve from bilateral exchange in the same way that bilateral
exchange evolves from unilateral exchange. The note-exchange agent
for Bank X, having concluded his or her regular exchange session at Bank Y,
may happen to meet the agent for Bank Z arriving for his or her meeting at
Bank Y. There are the economies just mentioned in combining the two meetings,
and absorbing as well the regular pairwise meeting between X and Z
agents. (The London note-exchange reportedly grew out of such noteporters'
meetings in pubs.) Unified computation, and settlement ofcombined
net clearing balances, can economically replace three bilateral exchanges.
Other banks may be invited to join the clearing sessions subsequently,
either individually or through combination with a similar multi-sided clearing
group. Eventually, all reputable banks within the par-acceptance region
will be linked through a single clearinghouse, or through a small number of
subregional clearinghouses that regularly clear against one another. The development
of clearing arrangements in Edinburgh, London, and New York
all conform to this general pattern. The final outcome - a unified clearing
system encompassing all banks- is not part of any bank's initial design.
Each aims only at increasing the market for its own liabilities, and at economizing
on redemption and reserve-holding costs. Systemwide par acceptance,
and its embodiment in the clearinghouse, in this sense, represent a
spontaneous institutional order.
The simplest and initial way ofsettling interbank clearing balances is through
the physical transfer ofoutside money at the end ofthe clearing session. Echoing
18 ThE EvoLUTION oF MARKET MoNETARY INSTITUTIONS
the original development of deposit-transfer banks to provide payments more
cheaply between bank customers like Alice and Bob, the banks may find it
economical to make interbank payments by means of a banker's bank. They
can settle up by transferring claims to outside money held in the clearinghouse
vault rather than by physically carting outside money around. Clearinghouse
association (CHA) banks in US cities in the nineteenth century issued claims
in the form of paper certificates to be used for interbank settlements. Other
CHAs have used clearinghouse deposits for the same purpose.
Historically, CHAs have been known to take on functions additional to their
core function of economically clearing and settling claims among banks. One
is the sharing ofinformation on loan defaulters, passers ofbad checks, and the
like. More significant are certain "hierarchical" functions associated with policing
the soundness of member banks. Chapter 4 discusses the possible connection
between such functions and the emergence of central banking.
At this point, we can take stock of the spontaneously or "naturally" developed
monetary system so far described. The definitive money is specie.
Except in interbank settlements, transactors commonly make payments using
bank-issued currency and transferable deposits. A specie unit is the unit
of account. Bank-issued money is denominated in the specie unit, and is
widely accepted at par. All banks are linked into a unified system by one or
more clearinghouses. These outcomes are not purely theoretical, but could
be seen historically in banking systems that were free of significant legal
restrictions.13
Is there a spontaneous or market-driven path from this system to the noncommodity,
or fiat, standards that prevails today? No. If any single bank in
the system were unilaterally to stop redeeming, it would have breached its
contracts with its customers. Ifit were to announce in advance that it would
stop redeeming next month, holders of its notes and deposits would redeem
them all before next month, and would take their business elsewhere. (Alternatively,
if the bank tried to replace ordinary open-ended notes and deposits
with new liabilities whose redeemability was scheduled to expire on
a specified date, nobody would take the liabilities as the date approached.)
The other banks and the public would reject the irredeemable liabilities
because without redeemability at par for specie, there would be no assurance
of continued par value in terms of the specie unit of account. (Chapter
12 considers the question of private irredeemable money in more detail.)
13 See Dowd (1992b) for case studies.
THE PATH TO FIAT MONEY 19
The forces that lead to convergence on a common monetary standard, as
in Menger's account, continue to operate once a standard is reached. Nobody
wants to make trading harder by offering or accepting only a nonstandard
money, different from that routinely accepted and offered by others.
Consequently, nobody would want to go first in switching to a completely
novel monetary standard, even if he were persuaded that, in theory, it would
work better supposing that everyone switched. (The policy implications of
this phenomenon are discussed in chapter 5). Ifnobody goes first, the switch
never occurs.
If all the banks, together, could coordinate a simultaneous switchover to
a fiat standard (a very big "if," but banks did coordinate the beginnings and
ends of temporary systemwide suspensions ofpayments during a few of the
nineteenth-century US banking panics) the new standard might stick. However,
it is not clear what market forces would compel banks to want to make
such a move. Also, if it meant breaching pre-existing redemption contracts
(as suspensions of payments historically did), it would not be a voluntary
switch by the users of money.
In historical practice, a nation's switch to fiat money was typically made
by the central government first granting a legal monopoly of note-issue to a
single institution, a central bank, whose liabilities became as widely accepted
as specie, and displaced specie as the reserves for other banks. The
government then suspended, permanently, the redemption of the central
bank's liabilities. With their permanent suspension, central bank notes and
deposits became a fiat base money. The fiat-money unit correspondingly
became the unit of account. Typically, the central bank for continuity's sake
retained the old specie unit name (e.g. "dollar"), which was printed on the
notes in circulation at the moment of suspension, while severing its specie
definition. The now-irredeemable notes can continue to circulate because
they are familiar, and the practice of continuing to accept them is selfreinforcing:
it is not in any one trader's self-interest to refuse them if she
expects others to continue accepting them.14
Thus, fiat money is possible where paper banknotes had previously gained
14
In addition, the government can reinforce their continued acceptance by making the nowirredeemable
central bank liabilities
1 publicly receivable- taking them for tax payments and for purchases from state enter-
prises;
2 legal tender for payment of old debts contracted in the unit of account;
3 forced tender in all domestic exchanges, including spot transactions that traders would
rather conduct in another currency and repayment of old debts specifically denominated
in metallic units.
As a final step, the government can, as the US government did, require the public to tum in
its specie.
20 THE EVOLUTION OF MARKET MONETARY INSTITUTIONS
common acceptance as redeemable notes.15
Likewise, to launch a new fiat
money today (for example, in former Soviet republics), it must at first be
made redeemable for the prevailing money (the ruble). Selgin (1994a)
likens initial redeemability to a "launching vehicle" that can fall away once
the new currency gets into orbit.
A few authors, in recent years, have argued that the story of market monetary
evolution should not end with banks operating on a silver or gold
standard. Absent legal restrictions, Kevin Dowd (1996, pp. 14-18) argues
that the next logical market-driven steps would be:
1 a discontinuation of direct redemption for precious metal, and
2 a switching over of the medium of account to a multi-commodity
standard.16
Dowd's scenario for the discontinuation of direct redemption runs as follows.
Once the stage is reached where gold coins no longer commonly circulate,
because everyone prefers to use bank-issued money for all
transactions, the public will no longer care about having the option to redeem
for gold as such. Because a bank can pay its depositors a higher return
by replacing all its non-interest-bearing gold reserves with interest-bearing
assets, competition would compel banks to make the switch, and, correspondingly,
to offer redemption in equivalent financial assets rather than gold
itself (call this "indirect redemption"). At this juncture, a weight of gold
would still define the unit of account, but gold would no longer serve as the
medium of redemption. If, for example, one "dollar" is one-twentieth of a
Troy ounce of gold, the holder of a $100 banknote could redeem it for financial
assets (perhaps blue-chip bonds or equity shares) equivalent at market
prices to five ounces of gold. Ounces-worth or "indirect" redeemability,
15 Two historical cases are instructive here. When the Bank of England suspended payments
from 1797 to 1819, Northern Ireland remained on a specie standard because banknotes did
not yet commonly circulate there. For the same reason, California remained on a specie standard
while the rest ofthe Union went on to an irredeemable "greenback" dollar standard during the
American Civil War.
16 Cowen and Kroszner (1994, pp. 38-44), using quite a different argument, imagine evolution
toward a variety of parallel standards. Greenfield and Yeager (1983) propose a multicommodity
standard, but do not claim that spontaneous market forces are enough to ensure
its adoption.
SPONTANEOUS SEPARATION BETWEEN THE MEDIA OF REDEMPTION AND ACCOUNT? 21
like direct redeemability, satisfies the customer's demand for contractual
assurance of the value of bank-issued money. Though the typical retail customer
might never exercise the new redemption option, just as he or she had
come never to exercise the old, the option maintains an avenue for arbitrage.
It keeps the market price of gold, quoted in any particular brand of dollardenominated
banknotes and deposits, tied to the definitional gold content
of the dollar.
Is this scenario convincing? There is historical evidence that, in a goldstandard
country with a sophisticated banking system, the use of gold coins
as a hand-to-hand medium of exchange has indeed tended to diminish substantially.
There is certainly evidence that banks take steps to economize on
their holdings of non-interest-bearing reserves. Dowd (1989, p. 155) cites
the case of the nineteenth-century Scottish banks whose customers were
often (which is not the same as always) satisfied to redeem their liabilities
for drafts on London correspondent banks (presumably because they had
payments to make in London) rather than gold. Those Scottish banks could
hold much (which is not the same as all) of their reserves in the form of
interest-bearing accounts in London. The case suggests that, if customers
and their banks sometimes mutually prefer redemption in something other
than gold coins, then a competitive bank would offer its customers that
option in addition to gold redemption.
There are, however, no known historical cases where competition led banks
to reduce their gold reserves literally to zero, and to remove entirely the option
of direct redemption in the medium of account from their banknote and
deposit contracts. Thus, we have little reason to be confident that the public
would, in fact, welcome the elimination of direct redemption. At the gold
reserve ratios observed historically in the most sophisticated banking systems
(2 percent and less), only very small increases in deposit yields are available
by reducing reserves the rest of the way to zero. Unless the option of demanding
gold rather than financial assets were completely valueless, it is doubtful
that the public would welcome the final reduction of reserves to zero, and the
complete elimination of the direct redemption option from bank liabilities.
The heaviest users of the redemption option are not retail bank customers,
but the banks themselves: every day, banks redeem huge volumes of claims
on one another at the clearinghouse. Whether direct redemption would spontaneously
disappear thus depends on whether banks themselves would welcome
the switch by their clearing partners, and would agree to accept settlement
in financial assets (bonds or equity shares) rather than in the medium of account
(e.g. gold or clearinghouse claims directly redeemable at par for gold).
As noted above, banks historically found it convenient to settle net clearing
balances by transferring claims to precious metal, kept in the clearinghouse
vault, rather than by physically carting bullion, or bags of coins, to and fro.
However, there are no known historical cases where member banks agreed to
22 THE EvoLUTION oF MARKET MoNETARY INSTITUTIONS
reduce the clearinghouse's vault cash to zero, and to settle in clearinghouse
claims that were not directly redeemable for outside money.17
A straightforward
explanation is that no historical clearinghouse in a gold-standard banking
system could do without gold reserves because its members had to be
prepared to pay gold to the rest of the world. No clearinghouse embraced all
the banks in the world, and gold remained the medium of settlement between
banking systems.18
A move to indirect redemption is conceivable only for the
clearinghouse of a closed economy, which in the world of an international
gold standard means a global clearinghouse.
Members of a clearinghouse might agree to have it hold fractional reserves,
so that their clearing account balances could bear interest. (The New
York Clearinghouse in the nineteenth century, however, typically held 100
percent reserves.) This does not imply an end to direct redeemability, or
settlement with financial assets. The convention of settling in one ofthe two
traditional ways, either by transferring physical units of outside money, or
by transferring directly redeemable (and hence par-valued) clearin
house claims to outside money, reduces transaction costs because it transfers
a single homogenous asset with an unambiguous unit-of-account
value. All parties can agree that a $10,000 clearing balance is settled with
$10,000 in full-bodied coins or a claim redeemable at the clearinghouse
for $10,000 in coins. An imagined system of settling with financial assets,
by contrast, appears to face banks with the problem of continually negotiating
agreements about which specific assets are acceptable, and for each asset
at what price within the interval bounded by the asset's current bid and
ask prices (which prices would have to be continuous tracked in a system of
continuous settlement). It does not appear that these agreements could simply
be reached once for all time, because the set of financial assets available
is continually changing. Even items within the set change in their
characteristics. In the bond market, for example, new bonds are being
issued, old bonds are being retired, extant bonds are shrinking in duration
and, hence, in interest-rate risk as they approach maturity, the default
17 Dowd (1989, p. 96) wishes to attribute the non-realization of his scenario to state intervention,
but it is not clear that private clearinghouse associations were everywhere prohibited
from taking such a step.
18 Describing the international payments system in the early part of the twentieth century,
Ludwig von Mises (1980, pp. 325-6) observed that although "the clearing system has without
difficulty transcursed political boundaries and created for itself a world-embracing organization
in the international bill and check system," there were still no bank-issued payment
media "that are recognized internationally and consequently able to take the place of [metallic]
money in international trade for settling the balances that remain over after the clearing
process." The absence of global branch banking and globally accepted bank liabilities was, of
course, at least partly due to legal restrictions.
SPONTANEOUS SEPARATION BETWEEN THE MEDIA OF REDEMPTION AND ACCOUNT? 23
risk associated with particular issuers is varying, and the liquidity of particular
securities (as indicated by the size of their bid-ask spreads) is changing.19
Dowd's scenario for the spontaneous mutation of the medium of account
runs as follows. So long as the unit of account is defined as a fixed weight of
gold, the price level is subject to disturbance by supply and demand shocks in
the market for gold. (This feature of a gold standard is analyzed in chapter 2).
"A time would therefore come when the banks would decide to reduce pricelevel
instability" by replacing gold with a medium of account "with a more
stable relative price," most likely consisting of "a basket of goods and services"
(Dowd 1996, p. 16).20
The banks would arrange to act in unison. The
public would go along because they prefer greater price level stability.
The question poses itselfin this case as well: ifthis unit-of-account switch
is a move worth making, why have banks and clearinghouses historically
avoided it? One explanation, parallel to the previous discussion, is that no
historical clearinghouse found it advantageous to switch its medium of account
while the rest of the world stayed on gold. If no single clearinghouse
(or its members' customers) would wish to "float" against the rest of the
world, then a move to the new regime- from the status quo of an international
gold standard- is again conceivable only for a global clearinghouse.
Chapter 5 below discusses in more detail the idea that no one finds it worthwhile
to switch the unit of account unless everyone switches together.
It is not necessary to switch the unit of account, however, for a bank or
clearinghouse to insulate the value of its liabilities from changes in the relative
price of gold: it could index its redemption rate. To allow continuous
adjustment of the redemption rate, the indexation could be to the price of a
basket of standardized commodities continuously traded on organized ex-
changes.21
The question becomes, then, why did banks and clearinghouses
on the gold standard avoid indexation of their liabilities? (White 1990, p.
19
Why wouldn't it be enough for the banks simply to agree once-for-all to accept settlement
only in default-risk-free securities, like present-day US Treasury bills? First, default risk is
not the only relevant risk for banks holding bonds. Second, under a commodity standard,
there simply are no default-risk-free bonds. Under a gold standard, for example, no government
can print up gold to repay its gold-denominated bonds. In this respect, the euro standard
acts like a commodity standard: no participating national government can issue default-riskfree
bonds denominated in euros, because no nation can print euros to redeem its bonds.
(Only the European Central Bank can print euros.)
20
This switch in the medium ofaccount could, in principle, be made whether the banks practiced
direct or indirect redemption. With the banks practicing direct redemption in gold, it amounts
to switching from an unindexed to an indexed redemption rate. With indirect redemption, the
banks would already be indexing the redemption rate in financial assets to the gold price of
the assets; now they would switch to indexing in the basket price.
21
This sort of indexation is discussed at greater length in chapter 12, in connection with the
reform proposal of Greenfield and Yeager (1983).
24 THE EVOLUTION OF MARKET MONETARY INSTITUTIONS
197) Perhaps the theory of indexation and the organization of commodity
exchanges were simply not developed enough by the time the gold standard
was ended by other means. Or, perhaps indexation would not have been
worth the bother because the instability of the relative price of gold (when
gold is money) is not much greater than that of any feasible index basket
that allows for continuous tracking.
1 Once traders begin using indirect exchange, why do they tend to converge
on a single good as the commonly accepted medium of exchange?
Could they converge on two or three goods?
2 If all goods were equally salable, would money still emerge out of
barter?
3 Why did commodity monies, rather than fiat monies, historically
emerge out of barter economies?
4 Consider a system where gold coins are the commonly accepted medium
of exchange. Absent legal restrictions, would private issuers
have incentives enough to establish a uniform monetary standard?
Or, would a variety of coins, of different weights and finesses, circu-
late?
5 Why have traders, historically, often preferred to use claims on banks
(like banknotes and checks) even though the claims might be
dishonored, rather than precious-metal coins?
6 Why has the unit of account typically been some quantity ofthe commonly
accepted medium of exchange?
7 J. Huston McCulloch (1982, pp. 6-7) has argued that "the development
ofmoney is very similar to the development oflanguage." There
is a tendency, in any society, for one verbal utterance to be singled
out as the spoken word for "fire," but it is "fundamentally arbitrary
which grunt or series of noises" people adopt for that purpose. Likewise,
there is a tendency for one commodity to be singled out as the
common medium of exchange, but "which commodity is singled out
is largely a historical accident." In what respects do you agree, in
what respects do you disagree, and why? Cite theoretical, and historical,
support for your position.
8 "As the supply of metals in the world has repeatedly been insufficient
to meet the increasing need for money, the use of paper [currency] has
spread." (Cribb 1986) Do you agree or disagree with Cribb's explanation
for the historical shift from coins to banknotes? Explain why.
9 John Browning and Spencer Reiss (1998) define "feedback, positive"
as "success that breeds success" and elaborate:
QUESTIONS
Positive-feedback loops create a winner-take-all world: whoever or
whatever starts ahead gets further and further ahead.... Incumbents
literally own their markets.... Once a positive-feedback loop
does kick in, the result can be awesome concentrations of economic
power. Bill Gates once explained it this way, talking about Windows:
"Momentum creates momentum. If you have volume, then
people write apps.lfpeople write apps, you have momentum." And
if it goes on long enough, you have $40 billion.
25
(a) Is Menger's theory of the origin of commodity money a story
about a positive-feedback loop? How or how not?
(b) Does convergence to a common commodity money standard lead
to an incumbent owning the market, or to a concentration of
economic power? How or how not?