A FISCAL THEORY OF GOVERNMENT’S ROLE IN MONEY

GEORGE SELGIN and LAWRENCE H. WHITE*

As an alternative to market failure explanations, we draw on theory and histor-

ical evidence to argue that fiscal considerations explain the roles governments
typically play in producing and regulating money. Public monopoly production of
coins and banknotes, substitution offiat for commodity standards, and restrictions
on substitutes for government money all generate revenue and especially provide
means for meeting fiscal emergencies. We argue that these arrangements do not
reflect conscious design so much as the evolutionary survival of the fiscally ad-
vantageous. (JEL E5, E6, H I , N1)

I. INTRODUCTION ernments shape monetary institutions to serve

“Economic policy has, up to the turn of money holders, by providing a more efficient
the century, been motivated primarily by and stable payments system than would exist
fiscal considerations... . [Fliscal measures under laissez faire. Thus private competition
have created and destroyed industries ... is not allowed in currency issue because mar-
even where this was not their intent, and
have in this manner contributed directly kets inherently would fail (or historically did
to the construction (and distortion) of the fail) there, and the legal restrictions we see on
edifice of the modem economy.” deposit banking are ones needed to prevent
[Schumpeter 1954, 71 market failures in that industry.
Recent research supplies three reasons for
Why do governments play the roles they doubting the adequacy of the market-failure
do in the monetary system? In particular, why approach for explaining monetary arrange-
have national governments almost universally ments. First, economic historians have found
taken over the business of issuing coins and that the actual forms taken by money and
paper currency, and replaced precious metals banking regulations, and the timing of their
with fiat money as the base supporting bank- adoption, often have little apparent connec-
issued money? Why have they not (in devel- tion to alleged market failures. Observed reg-
oped countries, at least) also nationalized the ulations (e.g. reserve requirements that freeze
production of checking accounts, choosing in- rather than enhance liquidity) are ill-designed
stead to tax and regulate private banks? to remedy the suboptimalities that are sup-
Standard answers to these questions refer posed to have motivated them. Second, mon-
to market failures (natural monopolies, exter- etary historians have found that systems close
nalities, or information asymmetries) that to laissez faire have (by and large) been at
might render unregulated private production least as successful as more restricted systems.
of money inefficient or unstable or infeasible. Finally, monetary theorists have pointed out
Market-failure explanations assume that gov- weaknesses in theoretical arguments for mar-
ket failure in money.’
If the market-failure explanation is doubt-
* We have received helpful comments from John Lott, ful, how else can one explain government’s
David Glasner, seminar participants at George Mason Uni- role in money? Charles Kindleberger [ 1994,
versity and North Carolina State University, and session
participants at the AEA and SEA meetings. xi] poses the challenge squarely: the econo-
Selpin: Associate Professor, Department of Economics, mist who doubts the market-failure approach
-University of Georgia, Athens, Phone 1-706-542-2734 “has to explain why there seems to be a strong
Fax 1-706-542-3376 revealed preference in history for a sole is-
E-mail selgin@rigel.econ.uga.edu
White: Professor, Department of Economics, University of
Georgia, Athens, Phone 1-706-542-3696
Fax 1-706-542-3376 I . Selgin and White [I9941 survey relevant historical
E-mail Iwhite@rigel.econ.uga.edu and theoretical literature.

154
Economic Inquiry
(ISSN 0095-2583)
Vol. 37, No. 1 , January 1999, 154-165 OWestern Economic Association International
 SELGIN & WHITE: GOVERNMENT’S ROLE IN MONEY 155

suer.” We propose a fiscal hypothesis: Gov- model? Puviani found a high degree of corre-
ernments have come to supply currency, and spondence between actual tax arrangements in
to restrict the private supply of currency and post-unification Italy and ones predicted by
deposits, not to remedy market failures, but to his modeL3 We argue that a fiscal approach
provide themselves with seigniorage and also accounts for monetary arrangements.
loans on favorable terms. Government cur- To avoid misunderstanding, we are not pro-
rency monopolies and bank regulations can posing that governments have consciously de-
thus be understood as part of the tax system. signed all monetary arrangements, from
The “strong revealed preference in history for scratch, to achieve purely fiscal ends. Such a
a sole issuer” is, fundamentally, the prefer- view would be at odds with the gradual and
ence of fiscal authorities, not of consumers.2 piecemeal historical development of
Economic historians have of course often governments’ monetary roles. Instead, as we
recognized fiscal motives behind specific discuss in more detail below, revenue-seeking
monetary arrangements, especially those of governments have opportunistically modified
ancient and medieval autocracies. Analysts of private-market arrangements as these devel-
developing countries today have recognized oped.4 Revenue-enhancing modifications tend
that policies of “financial repression” aim at to survive, while others are more likely to be
fostering “financial institutions and financial discarded. The resulting arrangements thus
instruments from which government can ex- look as ifthey were designed from scratch to
propriate significant seigniorage” (Fry [ 19881, generate government revenue. The “rational
14; Giovanni and de Melo [1993]). We go fur- dictator” model of monetary arrangements
ther in arguing that fiscal forces have typically should be understood in this “as-if” fashion.
shaped the industrial organization of money
production, throughout history and across II. SEIGNIORAGE-ENHANCING INSTITUTIONS
countries, and account for its major institu-
tional features even in advanced democracies An extensive literature analyzes the reve-
today. Such observed legal restrictions as stat- nue-raising device known as seigniorage or
utory reserve requirements, interest rate ceil- inflationary finance. The basic concept is
ings, foreign exchange controls, and monop- straightforward: a government reaps profit by
oly issue of currency impede efficiency but producing new base money at an expense less
raise revenue. than the value of the money produced. The
government finances expenditures by spend-
A Rational Dictator Model ing the new units of base money into circula-
tion5 Such expansion of the monetary base
To develop our hypothesis, we adopt a
method found in the writings of the Italian
fiscal theorists, especially Amilcare Puviani. 3. Levi [I9881 offers a modem exposition of Puviani’s
According to Buchanan [ 1960, 641, Puviani fiscal-predation thesis. An alternative “predatory” hypoth-
esis is private rent-seeking: government serves special in-
tried to account for overall government tax terests by restricting competition. This hypothesis certainly
arrangements by asking “two simple ques- helps explain a number of legal restrictions on private
tions.” First, what sort of tax system would a banks. But it can hardly explain the exclusion of private
mints and banks of issue from the market in favor of a
“rational dictator” put in place if his aim were state-run monopoly.
“to exploit the taxpaying public to the greatest 4. On the (conjectural) evolution of private-market
possible degree,” gaining the greatest revenue monetary arrangements in the absence of government in-
consistent with a given threshold of public re- tervention see Selgin and White [ 19871.
sistance? Second, to what extent do actual tax 5. The budget constraint for a government that issues
fiat money is G = T + A D + A H , where G is government
arrangements conform with those predicted by spending (including debt service), T is tax revenue, AD is
such a rational dictator (or “Leviathan”) the change in interest-bearing debt held by the non-gov-
ernment public, and AH is seigniorage, the change in non-
interest-bearing debt (fiat base money) held by the public.
2. A divergence between the preferences of fiscal au- For a textbook introduction to seigniorage see McCulloch
thorities and citizen-consumers implies that political “mar- [ 19821. McCulloch notes that the term “seigniorage,” from
kets” do not strongly tend to weed out policies that dimin- the French seigneur or lord, reflects the feudal lord’s prac-
ish aggregate wealth. On the contrast between “Chicago” tice of profiting from monopoly production of debased
and “Virginia” views on this broad question, see Lott coins that his subjects were compelled to accept at face
[ 19971. value.
156 ECONOMIC INQUIRY

implicitly taxes base money holders by dilut- making seigniorage an especially valuable fis-
ing the value of existing money balances. For cal resource during a emergency that threatens
the most part the literature treats the base- the state’s survival, such as an insurrection or
money expansion rate (or the associated price external military threat (Glasner [ 19971). Its
inflation rate) as the government’s choice unique revenue-raising speed helps to explain
variable, taking monetary institutions as why state monopoly of base money survives
given. The focus lies on the rate that maxi- into modern times, long after state monopolies
mizes seigniorage, or alternatively minimizes of other goods like salt have given way to
the deadweight burden of taxation subject to taxation of private producers. We later discuss
a revenue constraint. In contrast, we inquire surprise inflation and the time-consistency
here into what sorts of monetary institutions issue it poses.
enhance seigniorage.
Seigniorage from Commodity Money
Why Collect Seigniorage At All? What sort of outside-money regime would
Several features make seigniorage an at- a rational dictator prefer for fiscal purposes?
tractive option for raising revenue. First, a tax Precious metals offer the potential for sei-
on money balances might be consistent with gniorage extraction through debasement. By
the Ramsey rule for minimizing the dead- adding base metal, 100 silver coins can be
weight burden of raising a given amount of remade into 105 (or 150 or 200) apparently
overall government revenue. Several theorists similar coins. Coins entirely composed of
have argued, however, that when money is re- base metal, by contrast, cannot be further de-
garded as an intermediate good, any positive based. A cowrie shell or a peppercorn, being
inflation tax is inefficient, even given a posi- a naturally-occurring unit, cannot be easily re-
tive revenue constraint. The optimal inflation made or redenominated. Putting aside fiat
tax is then zero (Banaian et al. [ 19941; Correia money for now, fiscal considerations would
and Teles [ 19961). If so, the collection of sei- incline a rational dictator to favor the precious
gniorage, and the shaping of monetary insti- metals over other commodity monies.
tutions to that end, cannot be justified on the Although the earliest known coins appear
grounds of fiscal efficiency.6 to have been privately produced, ancient rul-
Second, seigniorage is a relatively “hid- ers seeking a new source of revenue (and pro-
den” tax. If the public blames inflation on paganda, by putting the ruler’s name or face
causes other than the government’s monetary on the coins) soon granted themselves legal
policy, the political resistance provoked by an monopolies in minting (Bums [1965]). A mo-
inflation tax may be lower, for a given amount nopoly mint extracts seigniorage from the
of revenue, than that of more obvious taxes. metal it coins, subject to the accounting iden-
A rational dictator concerned with maximiz- tity
ing his survival in power, extracting seignior-
age to the point where the marginal political
resistance incurred per dollar of revenue is M = PQ+ C + S,
equal to that of alternative taxes, will then
exploit the inflation tax even beyond the point where M is the nominal value assigned to a
where its marginal deadweight burden equals batch of coins (e,g. 100 “shillings”), P is the
that of other taxes. nominal price paid by the mint per ounce of
Finally, to the extent that changes in the precious metal, Q is the number of ounces of
nominal stock of base money can be made precious metal embodied in the batch of coins,
unexpectedly, they impose an ex-post capital C is the remaining average cost of minting,
levy on holders of the state’s unindexed nom- and S is the nominal seigniorage. Out of every
inal liabilities, including base money. Such a M’s worth of shillings coined, PQ is paid to
levy may yield substantial revenue rapidly, individuals who brought in precious metal, C
covers other mint expenses, and S is retained
6. Alcazar [I9941 briefly surveys empirical work on as profit for the mint-owner. Total seigniorage
whether actual seigniorage rates conform to the implica- per year depends on how many batches of
tions of the optimal-tax model. In most countries they do
not. coins are produced per year.
 SELGIN & WHITE: GOVERNMENT’S ROLE IN MONEY I57

Greater nominal seigniorage per batch is that coinage systems will have exclusive ter-
earned by debasement when Q is reduced for ritories that expand and contract with sover-
a given M. When reducing silver content, me- eign realms. The history of medieval coinage
dieval governments typically added base supports the fiscal hypothesis. European mon-
metal, reducing the fineness rather than size archs of the middle ages insisted that the right
of coins. Minting costs were lower because to mint coins belonged exclusively to the sov-
coin dies did not need to be resized, and the ereign (thus Bisson [ 19791 speaks of “the pro-
new coins would circulate more readily be- prietary coinage”), even when diseconomies
cause they closely resembled the old. The re- of plant scale led them to delegate actual coin
duction in metallic content might even go un- production to local moneyers. During the
detected for a time, enhancing short-run real early Middle Ages kings and princes had trou-
revenues. Alternately, each new shilling could ble enforcing their laws against independent
simply be declared to have a higher nominal coinages. This “fragmentation of monetary
value, increasing M for a given Q.’ Greater rights” was not due to changing economies of
seigniorage per batch can also be earned with- scale in coin production but “corresponded to
out debasement by reducing P, i.e. putting as the multiplication of territorial powers” (Bis-
much silver into each shilling but paying son [1979, 31). When kings regained power
fewer shillings per batch back to the provider over the nobility, one of their first objectives
of silver. “was to reclaim control over the coinage”
As an excess profit or rent in coin produc- (Glasner [ 1997, 271).
tion, seigniorage cannot persist without legal Many rulers also enforced legal restrictions
restrictions on entry. The fiscal motive thus that were designed to secure the profit from
accounts for state-enforced coinage monopo- issuing debased coins accepted at face value.
lies. In a competitive minting industry with Boyer-Xambeu et al. [1994, 49-59] note:
constant returns to scale, competition would “Until the sixteenth century princes in most
enforce the condition of price equal to mar- countries prohibited the weighing of coins and
ginal and average cost, M = PQ+ C. Every made people accept them all, even when used
mint, including the monarch’s, would earn up, simply in view of their imprints and in-
zero seigniorage if competing mints could be scriptions.” Even when weighing was later al-
established side by side, bullion owners were lowed (to encourage the return of worn coins
free to choose where to take their bullion to to the mint), the practice of valuing coins in
be coined, and no steps were taken to restrict exchange by bullion weight rather than by tale
the circulation of nongovernmental coins so was “expressly forbidden.” Payments in metal
that all coins were valued by precious metal other than the prince’s coin, and contracts
content. The few historical cases where com- specifying payments by bullion weight, were
peting private mints were allowed (e.g. gold- outlawed. The practice of culling good coin
rush California) do not exhibit the sort of mar- and passing on bad was a crime that “system-
ket failures-fraud, or lack of standardiza- atically carried the death sentence.” It is hard
tion-that are sometimes hypothesized to pro- to imagine an efficiency-enhancing rationale
vide an efficiency-enhancing role for the state for such restrictions.
in coinage.8 Two reasons consistent with the fiscal hy-
The efficiency theory of government coin- pothesis suggest why past monarchs preferred
age predicts that coinage systems will vary in owning monopoly mints to taxing private
geographic scope only in response to chang- mints. First, as the modern theory of vertical
ing economies of scale in coin production. integration suggests, monitoring and enforce-
The fiscal hypothesis, by contrast, predicts ment problems would likely be lower with
vertically integrated (state-owned) mints. Sec-
ond, increases in the seigniorage rate might
7. Note that medieval coins typically displayed no
numbers, only graphical identifying devices. be accomplished at lower cost than equivalent
8. The many brands of private coins minted from Cal- increases in the rate of mint taxation, in part
ifornia gold in the 1850s were all denominated in the es- because the incidence of an increased mint tax
tablished gold dollar unit. Assays “invariably found” that would be more transparent, more concen-
the coins’ gold content was very close to the precise legal
standard, and most coins apparently erred on the side of trated, and therefore likely to meet with more
exceeding the standard (Kagin [ 1981, 239-423). political resistance than a debasement. Both
 158 ECONOMIC INQUIRY

considerations become especially relevant great lengths to preserve the quality of their
during a fiscal emergency, when revenue “international” monies (monete grosse) even
needs to be raised immediately. Spufford’s while ruthlessly debasing the locally used
figures [I9881 indicate that, in times of war, petty coins. The Spanish government, for ex-
mint-owning medieval rulers raised as much ample, took pains to preserve the metallic
as 60?&92% of their total revenues through content of its silver coin, which by the late
debasement. 15th century had become Europe’s (and the
The value of the ability to meet a fiscal New World’s) most stable and coveted, while
emergency also explains why an insecure ra- actively debasing the petty copper coinage
tional dictator would prefer owning a monop- which it produced as a local monopoly
oly mint to the alternative of selling or leasing (Motomura [ 19941). The English government
monopoly franchises to private bidders. Fran- debased some small-denomination coins, but
chising substitutes fixed advance payments carefully protected the international reputa-
for what would otherwise be a variable flow, tion of larger coins, especially sterling (May-
but rules out recourse to surprise inflation and hew [1992]).
corresponding emergency capital levies. Ac-
cordingly we observe that central govern- Fiat versus Commodity Money
ments have typically retained operational con-
trol over mints. The seigniorage motive favors fiat over
commodity money in three respects. First,
Local versus International Coin government captures a one-shot profit from
A government that seeks seigniorage from replacing the existing stock of monetary metal
the monopoly production of coin may act as with fiat money.” Second, issuing fiat money
a discriminating monopolist when the elastic- is a cheaper way to capture an ongoing flow
ity of demand with respect to their deprecia- of seigniorage revenues each year. Finally, the
tion rates varies across coins: the revenue- demand for a fiat money is less elastic, be-
maximizing rate is lower for coins facing rel- cause users encounter greater costs in trying
atively elastic demand. During the early Mid- to employ any foreign money in its place. We
dle Ages in Europe, low-value or “petty” sil- elaborate these last two points in turn.
ver coin from local mints circulated almost Seigniorage flow is most profitably cap-
exclusively in local exchange. Higher-value tured with a money that can be produced (in
coin from the same mints was mainly used in nominal units) at zero resource cost, and
international markets (Cipolla [ 1956]), where whose nominal stock can be expanded at
whatever rate desired. In principle, nominal
it competed head-on with foreign coin.9 Be-
units of money can be created under a silver
cause the demand for high-value coins was
standard without incurring mining costs, and
much more elastic, a rational dictator would
subject high-value coins to lower rates of sei- the nominal money growth rate can be con-
gniorage (less frequent debasement). trolled, by continual debasement, i.e. by con-
tinually redefining the unit of account to equal
Medieval European governments accord-
ingly extracted less seigniorage from gold progressively fewer grams of pure silver. A
coin than from silver, and debasement of sil- mint that wants to earn a large annual profit
ver coins was much less frequent for large from debasement, however, must recoin a
large part of the outstanding money stock. In
denominations than for small.I0 Mints went to
practice, it is much more costly to expand the
9. Transportation costs, and penalties connected to nominal stock of coins 5% through recoinage
legal tender laws (which were most readily enforced in than it is to expand the nominal stock of a fiat
local transactions), made it normally uneconomical to im- money 5%, which requires only the expansion
port better small-denomination foreign coin for domestic
use. Legal tender laws compelling the acceptance of do-
mestic coin at face value were less readily evaded for small 1 1 . The government realizes a capital gain even if it
denomination coins, because weighing coins was less does not sell off its monetary gold stock. (The central
worthwhile for small transactions. bank’s balance sheet typically disguises this gain by count-
10. The fact that politically influential aristocrats were ing fiat money as “liabilities” of the central bank, as though
the principle users of high-value coins (Cipolla [I956 261) central bank notes were still redeemable.) Still, the failure
supplied a separate public-choice reason for rulers to re- of central banks to liquidate their gold stocks even decades
frain from debasing these coins as extensively as petty after the demise of the Bretton Woods system is puzzling
coins. from a purely fiscal perspective.
 SELGIN & WHITE: GOVERNMENT’S ROLE IN MONEY 159

of ledger entries and the printing of more usually has to become quite severe before
identical paper notes. “dollarization” of domestic transactions oc-
The process of debasement also invites curs.
substantial public resistance. Compulsion, no Because currency substitution and the elas-
lighter and no more popular than that neces- ticity of demand for domestic base money are
sary to exact ordinary taxes, is needed to pre- reduced under fiat currency, the fiscal hypoth-
vent market participants from exchanging and esis predicts higher inflation rates under fiat
valuing new (debased) coins by weight rather standards than under metallic standards
than face value, and thus to encourage them (which allow inflationary finance via debase-
to treat both old domestic coin and foreign ment). This prediction is borne out histori-
coin as mere raw material to be taken to the cally in a comparison of commodity-money
mint. The tax imposed by recoinage is fairly and fiat-money episodes after 1600 (Rolnick
obvious once the reduced precious metal con- and Weber [ 19941).
tent of the new coins becomes known. Sei-
gniorage flow can be extracted more easily Fiat-money Monopoly
and less obviously with a fiat money, whose Why does a revenue-seeking government
nominal quantity can be increased merely by itself issue fiat currency monopolistically, in-
spending new units into circulation that are stead of taxing private issuers? The reasons
identical to existing units, obviating the need for thinking that a seigniorage-seeking gov-
to recall or devalue the old currency. No one ernment would prefer a mint monopoly to tax-
objects to accepting the newly issued units at ation of private mints apply again. In the case
a value equal to old-since they are identical of fiat money, a more fundamental reason ex-
in (zero) commodity content and interchange- ists as well: open competition in the produc-
able-so no obvious compulsion is needed. tion of fiat currency is to date a purely hypo-
Fiat money also offers the public smaller thetical possibility, and one that might not be
opportunities for switching to alternative base sustainable in practice. If “competitive sup-
monies. Under a silver standard, alternative ply” of fiat money meant free entry into the
coins can always be evaluated (even if not production of fiat dollar notes--the equivalent
legally) by weight, making the substitution of of legalized counterfeiting-each counter-
foreign for domestic money a relatively sim- feiter would produce notes until even the
ple matter of measuring both in terms of silver highest denomination note was worth no more
content (measured in a fixed reference weight than the paper and ink it contained. If there
unit, a so-called “ghost money” unit). If do- were no upper bound on denominations, prof-
mestic money is being frequently debased, its from producing dollars would persist until
traders quoting prices in weight units would the dollar became worthless (Friedman
naturally favor more stable foreign coins, less [ 19601). Alternatively, with trademark protec-
frequently requiring weighing and assaying, tion, perfectly competing firms might issue
as their medium of exchange. By contrast, distinct irredeemable monies, bearing identi-
traders who consider switching from a domes- fiable brand names but perfect substitutes for
tic to an alternativefiat currency as a medium one another (Klein [ 19741; Taub [1985]). The
of exchange find that there is no simple com- result would again be an equilibrium without
mon metric. A network effect associated with economic profit, either (in the case where an
using the common unit of account protects the enforceable infinite-horizon pre-commitment
incumbent currency by imposing high trans- is feasible) with positive-valued money pay-
actions costs on those who would switch first ing a competitive rate of return, or (in the case
(Selgin [ 19971). Acceptance of an alternative where time-inconsistency or “cheating” can-
currency in transactions presupposes familiar- not be prevented) with the same worthless-
ity with its exchange value, but until its ac- money outcome as the legalized counterfeit-
ceptance is widespread, or at least until the ing case (Selgin and White [ 19941).
domestic unit has become thoroughly unreli- Monopoly revenues from the production of
able as a unit of account (as in a high infla- fiat money could in principle be obtained by
tion), there is scant individual incentive to a group of fiat-money issuing institutions
track the exchange rate between the incum- whose aggregate currency issue is set at the
bent and alternative currencies. Inflation thus monopolist’s revenue-maximizing level. The
 I60 ECONOMIC INQUIRY

principle drawback of this arrangement is that otherwise restricting their availability or at-
it requires costly monitoring to avoid cheating tractivene ss.
(issues in excess of allotments) by individual Alternatively, bank liabilities can simply
cartel members. Italy in the late 19th century be taxed-for example, by reserve require-
offers a case in which he cartel approach ments. Unlike competitive private issue of
proved unsustainable. Following the Risorgi- commodity or fiat base money, private bank-
mento, the new national Italian government, ing does not deprive the government of the
having failed in its early attempts to establish ability to manipulate the rate of inflation.
a single bank of issue, awarded legal tender When banknotes and deposits are redeemable
status to the (then irredeemable) notes of six claims to fiat money, their rate of expansion
established banks in return for their funding ultimately depends on the rate at which the
of government debt. The system broke down stock of fiat money expands. It follows that,
because one cartel member-the Bank of in allowing private firms to issue redeemable
Rome-was discovered to have cheated on the substitutes for (fiat) base money, a rational
cartel, secretly exceeding its note allotment dictator does not deprive himself of the ability
by issuing notes with duplicate serial numbers to increase short-run seigniorage via a sur-
(Sannucci [ 19891). prise inflation.
Dwyer and Saving [1986] show that, if
Restrictions on Substitutes bank deposits and currency are perfect substi-
The ability of a national fiat-money pro- tutes, and if government is as efficient as pri-
ducer to earn seigniorage is, like that of a na- vate firms in producing money, then govern-
tional mint, limited by the availability of sub- ment can obtain the same maximum steady-
stitutes for domestic base money. Potential state revenue by imposing a positive reserve
substitutes include foreign currencies. As ratio or other form of tax “licensing fee” on
noted above in the contrast between local and a private banking industry as it would by sup-
international coin in medieval Europe, oppor- pressing private banking altogether. Histori-
tunities for substitution into foreign currency cally, governments have typically chosen to
increase the elasticity of demand for domestic suppress private banknotes, while allowing
money. They thereby reduce the maximum checkable private bank deposits to coexist
steady-state real seigniorage, and raise the in- along with fiat money. A straightforward ex-
flation rate associated with achieving any tar- planation for this, consistent with the rational
get level of real seigniorage. A rational dicta- dictator model, is that the public treats repu-
tor will take steps to limit currency substitu- table banknotes as very close substitutes for
tion, and can do so using such means as ex- base money. In historical cases where private
change controls and legal tender laws (Nich- note issuance was relatively unrestricted, as
ols [ 19741). Nations threatened by loss of sei- in Scotland and Canada, commercial bank
gniorage due to currency substitution, because notes displaced coin (and, in Canada, govern-
they have for other reasons committed to dis- ment-issued “Dominion” notes) almost en-
mantle barriers to free capital flows, might try tirely where their denominations over-
to form a cartel-a multinational central 1apped.The government therefore enhances its
bank-and share its seigniorage. The move- seigniorage tax base by suppressing private
ment for a European central bank can thus be notes. l 2
given a fiscal interpretation.
A second set of close substitutes for domes-
tic base money consists of private financial 12. Federal Reserve oficials have recently been scm-
assets, including redeemable private tinizing electronic payments media that amount to the re-
banknotes and deposits, that function as ex- introduction of private currency. In an unusually candid
statement of the authorities’ concern for seigniorage, Alan
change media. Here again, a rational dictator Greenspan [1997, 49-50] worries that the result of new
would take steps to suppress the substitutes, electronic currency substitutes may be “simply a diversion
either by prohibiting them altogether (as has of seigniorage from the government to the private sector.”
White and Boudreaux [I9981 argue that (non-price) com-
been commonly done with private banknotes), petition to “divert” seigniorage is efficient, while nation-
or by capping their interest yield (as has some- alization of currency for the sake of seigniorage implies
inefficiently low quality.
times been done with bank deposits), or by
 SELGIN & WHITE: GOVERNMENT’S ROLE IN MONEY 161

Bank deposits, by contrast, are not such seigniorage revenue are positively correlated
close substitutes for base money, and compet- with political instability and polarization. In
ing private banks can typically produce depos- countries with more unstable and polarized
its and other banking services more efficiently political systems, established governments are
than government can.13 Taxes on private more willing to sacrifice their long-run infla-
banks are likely to bring in more revenue than tion tax base to remain in power, because such
a ban on private banking that enhances sei- a strategy will either preserve the particular
gniorage only slightly. In consequence, as government that resorts to it, or will at least
Glasner [1989, 331 notes, for fiscal reasons serve to “constrain the behavior of future gov-
“most governments have preferred allowing ernments ... with which t h e y disagree”
banks to operate and exploiting them as a (Cukierman et al. [1992, 5381). In general a
source of credit to suppressing them or to op- rational dictator cannot exclude the possibility
erating banks of their own.” of confronting a fiscal emergency at some fu-
Fiscal considerations can thus account for ture date, and so will value a monetary ar-
governments allowing competitive deposit- rangement that allows him to resort to an in-
taking (subject to statutory reserve require- flationary capital levy even if in ordinary
ments and other devices aimed at directly or times he collects little seigniorage (Glasner
indirectly taxing bank deposits) while sup- [ 19971).
pressing redeemable private banknotes. However, a capital levy strategy is time-in-
consistent: it yields more revenue (in present
111. MONETARY REPUDIATION AND THE
value terms) than steady inflation only if lev-
TIME-INCONSISTENCY PROBLEM ies are greater than expected. A capital levy
that appears “optimal” for each rational dic-
Governments, as we have noted above, tator, considered in isolation from his prede-
may sometimes seek revenue through a sur- cessors and successors, may be suboptimal for
prise inflation that acts as a “capital levy” on all successive rulers together. If the public
money. The capital levy is imposed by a de- fears that the government will expropriate
liberate short-run burst of money creation. much of their monetary wealth, they will hold
Holders of cash balances experience a loss of smaller real balances, reducing (to zero, in the
real wealth as the price level jumps more than limiting case where total expropriation is ex-
expected. Such a capital levy makes it possi- pected) the maximum yield to all successive
ble to generate more real revenue in the short governments from a steady-state inflation tax.
run, but at the cost of smaller steady-state sei- The time-inconsistency problem associated
gniorage once the public recognizes the risk with monetary repudiation supplies a rational
of a high-inflation period occurring and there- dictator with a motive for trying to convince
fore holds less real base money at any given the public that monetary policy will be based
non-peak inflation rate than it would hold if upon a long time horizon, beyond the term of
the inflation rate were viewed as stable. any particular ruler, In other words, the ratio-
The rational dictator will find inflationary nal dictator wants to be able to resort to sur-
capital levies most worthwhile during emer- prise inflation, but also wants the public to
gencies (especially wars) that put present rev- believe that he will probably not resort to it.
enues at a large premium over future revenues If the dictator is well-entrenched, faces few
by threatening his reign (Glasner [ 19891). A external military threats, and has credibly ar-
capital levy is attractive to a government that ranged for a line of successors who will main-
attaches a high discount rate to revenues ob- tain his policies indefinitely, then the public
tained in the future, or one that expects to be may recognize that he has more to lose than
short-lived without the levy. Consistent with to gain by repudiating the currency. On the
this view is the finding of Cukierman et al. other hand, short-lived dictators (and rulers in
[ 19921 that inflation rates and reliance upon democratic regimes) are typically unable to
make such arrangements, and so must seek a
different solution. One historical solution was
13. A number of American states experimented with retention of a fixed-parity metallic standard,
tate-run commercial banks in the decades before 1860.
4ost of the banks were so incompetently run as to incur modified to allow for the suspension of cen-
nancial losses despite their monopoly privileges. tral bank convertibility during fiscal emergen-
 162 ECONOMIC INQUIRY

cies.14 Drawbacks of this arrangement in- and predicts that independence is most likely
cluded its inability to yield much seigniorage to be withdrawn during periods of heavy fiscal
during non-crisis times, and the high cost of demand. Consistent with that prediction, the
sustaining (via post-crisis deflation) the Reichsbank lost its independence during
public’s confidence in the promise to preserve Hitler’s rearmament program, and the Federal
the ancient and honorable parity. Another so- Reserve System lost its during both world
lution, where rival parties are not severely po- wars (Sylla [ 19881).
larized (and so are willing to cooperate to at-
tain mutually desired ends) is the establish-
IV. THE EVOLUTION OF MONETARY
ment of an “independent” monetary authority ARRANGEMENTS
that is supposed, like a business corporation,
to operate with a time-horizon much longer On the face of it, present-day monetary in-
than its current directors’ terms. The decision stitutions display several striking similarities
to form an independent monetary authority is to those predicted by the rational dictator
most likely to be made when rival political model. Practically everywhere base money
parties have little to lose by cooperating to does take the form of fiat paper or deposit
restore a depressed inflation-tax base, such as credits, issued by a central bank.I6 These cen-
immediately following an inflation-based cap- tral banks enjoy exclusive monopoly privi-
ital levy that has greatly increased the public’s leges granted to them by their governments,
estimate of the likelihood of hture high infla- returning the bulk of their seigniorage reve-
tion. Historically, then, central banks are most nues to the sponsor governments. Currency
likely to be given independence by demo- areas correspond to national political bound-
cratic governments in the wake of relatively aries rather than to the criteria suggested by
severe inflations. The Reichsbank, for exam- the theory of optimal currency areas. Typi-
ple, gained independence at the end of the cally, no statute or rule limits the rate at which
German hyperinflation. In the United States, the central bank may expand the monetary
the “Accord” giving the Federal Reserve base. Private firms are typically prohibited
greater independence from the Treasury came from issuing redeemable banknotes. Banks
in the wake of the post-World-War-I1 infla- are, on the other hand, typically allowed to
tion. Cross-sectionally, our argument predicts supply checkable deposits, subject to reserve
that independent central banks--serving the requirements and other taxes.
need for a commitment device-should be Just how is it that monetary institutions
found more commonly in pluralistic democra- came to take a form so well suited for meeting
cies than in autocratic states where a ruling governments’ fiscal ends? An answer based on
lineage has secure tenure. This prediction is continuous seigniorage maximization, in
broadly consistent with evidence from the which governments are portrayed as designing
1980s. In Cukierman’s 1119921 ranking of 46 monetary arrangements from scratch purely to
central banks, the 14 most independent were achieve fiscal ends, would be far from ade-
found in liberal democracies, with the sole ex- quate. Fiscal motives, we have argued, do di-
ception of Hungary’s. Of the remaining 32 rectly explain why various rulers monopo-
less-independent banks, 16 were in countries lized coinage, providing a precedent for later
that were authoritarian for the entire 1980s, state monopolization of paper money. But fis-
and six more were in countries that were au- cal motives by themselves do not account for
thoritarian at the start of the decade.I5 the gradualness and seeming haphazardness
The fiscal hypothesis suggests that central with which revenue-enhancing reforms ar-
bank independence is unlikely to be absolute,

14. Glasner [1997, 391 notes that “recognition of the

16. Although Hong Kong (two private note issuers be-
time-consistency problem may help rationalize the seem- fore the mainland takeover), Scotland (two), and Northern
ingly irrational attachment to the gold standard” of pre- Ireland (four) retain private note issue even today, the base
1930 governments. money is fiat (dollars or sterling) in each of the three sys-
tems. We expect that private banknotes will be phased out
1 5 . Information on regime types is taken from in Hong Kong now that it has come under mainland Chi-
Derbyshire and Derbyshire [ 19961. Cukierman’s survey ex- nese rule. In Scotland and Northern Ireland, seigniorage is
cludes most communist countries, which had relatively de- extracted by a 100%marginal reserve requirement (in non-
pendent central banks. interest-bearing Bank of England liabilities) against notes.
 SELGIN & WHITE: GOVERNMENT’S ROLE IN MONEY 163

rived, culminating in monopoly issue of fiat ing redemption of government notes. As Ard-
money. ant [ 1975, 1921 puts it, “a developed economy
In modern times especially the govern- was the prerequisite. It was necessary that
ments of industrial democracies do not con- bank bills be common in all circles and that
tinuously act to maximize seigniorage. (Infla- the state could pay its soldiers, its functionar-
tion rates would be much higher if they did.) ies, even its peasants in paper money. ...
Yet monetary institutions capable of extract- France in the seventeenth century did not have
ing maximum seigniorage from the public the conditions for a successful state manipu-
have emerged and have persisted. Indeed, the lation of the money supply.” During the Re-
single most effective means for extracting sei- striction period of 1797-1 82 I , even while the
gniorage-monopoly issue of fiat money- rest of the United Kingdom operated on a
became a permanent feature of monetary sys- Bank-of-England-note standard, Northern
tems only during the twentieth century. Ireland’s continued adherence to a gold coin
Our explanation for the gradual and uneven standard indicated the fact that banknotes did
development of seigniorage-enhancing mone- not yet commonly circulate there. California
tary institutions consists of three parts. The likewise remained on a gold coin standard
first is that government monetary institutions during the American Civil War, accepting
represent to a large extent piecemeal and op- Greenback dollars only at a discount, and thus
portunistic modifications of private-market remained immune from seigniorage taxation
developments, including the growth of bank- through the issue of Greenbacks.
ing and substitution of paper notes and check- The second part is that governments, and
ing accounts for gold and silver coins. The democratic ones especially, are most anxious
more genuinely “Leviathan-like” govern- to obtain seigniorage revenues, and to alter
ments of pre-industrial times were simply un- monetary arrangements in ways that generate
able to take advantage of such technological more seigniorage, during fiscal emergencies,
developments, and so had to settle for rela- especially wars. Such emergencies act as fis-
tively limited seigniorage revenues obtainable cal catalysts for seigniorage-enhancing inno-
through mint monopolies. Eventually, as ex- vations that public resistance might otherwise
plained above, increased opportunities for for- preclude. Thus the fiscal hypothesis explains
eign currency substitution made the exploita- the observation that the move from commod-
tion of mint monopolies for revenue unprofit- ity to fiat money typically occurred in steps
able, causing governments to look elsewhere corresponding to fiscal emergencies.I* T h e
for sources of revenue, and emergency reve- first step away from the gold or silver stan-
nue especially. One such source was the bank- dard in many countries, as already noted, was
ing industry, originally perceived, not as de- the establishment of a government-sponsored
vice for earning seigniorage, but as a source bank, The Bank of England, the Bank of
of loans on favorable terms. Such loans were France, and the Swedish Riksbank are well-
typically obtained in exchange for awards of known examples of government-sponsored
monopoly privileges, especially in note issu- banks established to play the fiscal role of
ance (Smith [1990]). The harnessing of mo- lending the government funds on favorable
nopoly banks of issue-central banks-as terms. Over time, with the aid of further leg-
sources of substantial seigniorage came later, islation that granted it a note-issuing monop-
with the discovery that such banks (unlike
competing banks of issue) could suspend pay- 17. There were episodes of temporary suspension (e.g.
ments with relative impunity, opening the way the Napoleonic Wars in Britain, the Civil War in the United
States), but notes that were (correctly) expected to become
to the emergence of fiat money. redeemable in the future were not fiat money. Unlike a fiat
We hypothesize that the seigniorage motive note, a temporarily suspended note has a lower bound to
did not produce fiat money before the 20th its current value set by the discounted expected value of
its future redemption media. The government i s con-
century” because (redeemable) banknotes strained to (re-)accumulate a sufficient inventory of re-
had not yet become commonly accepted in demption media. If the “ancient and honorable parity” is
areas of lesser financial sophistication, so to be re-established, the government’s potential capital
gain is limited to the interest it might earn (or loan interest
those areas could not be subjected to a capital it might avoid) by lending out or spending from its reserves
levy by the government’s monopolizing the during the suspension.
issue of banknotes and permanently suspend- 18. Chown [1994, 2011 documents this observation.
I64 ECONOMIC INQUIRY

oly, the privileged bank’s liabilities became their evolution through the centuries. To say
high-powered money. In many European this is not to claim that the fiscal hypothesis
countries this step was reached by end of the accounts for every organizational detail of
19th century. Fiat money could then be estab- past or present arrangements, or that alterna-
lished by the suspensions of central bank lia- tive accounts are universally invalid, but
bilities prompted by the fiscal emergency of rather that the fiscal hypothesis provides a
the First World War.19 In the United States, useful “default rule.” It fits the overall histor-
where central bank liabilities achieved high- ical pattern of facts better than its leading
powered status somewhat later, the establish- competitor, the market failure hypothesis. Re-
ment of fiat money awaited the fiscal emer- searchers seeking to explain particular gov-
gency of the Great Depression. The leading ernment roles in the monetary system should
alternative to the fiscal hypothesis, the view therefore “follow the money”: they should not
that government’s purpose in establishing fiat fail to consider the fiscal implications.
money is to remedy a market failure to con-
verge to a more efficient monetary standard, REFERENCES
offers no explanation for the historical timing Alcazar, Lorena. “Political Constraints and the Use of
of the steps toward fiat money. Seigniorage: Empirical Evidence from a Cross-Coun-
try Analysis.” Manuscript, Washington University,
The third and final part of our explanation 1994.
is that, once a revenue-generating reform is in Ardant, Gabriel. “Financial Policy and Economic Infra-
place, it is more likely to survive than other structure of Modem States and Nations,” in The For-
arrangements even when it proves to be a mation of National States in Western Europe, edited
source of disorder. Glasner [ 1997, 361 argues by Charles Tilly. Princeton: Princeton University
Press, 1975, 164-242.
that early states with access to seigniorage Banaian, King, J. Harold McClure, and Thomas D. Willett.
“improved their chances of survival in mili- “The Inflation Tax is Likely to be Inefficient at any
tary competition.” During peacetime also, fis- Level.” Kredit und Kapital, January 1994, 30-4 1.
cally advantageous innovations prove espe- Bisson, Thomas. Conservation of Coinage. Oxford: The
cially durable, in part because they enjoy the Clarendon Press, 1979.
support of powerful interest groups: the recip- Boyer-Xambeu, Marie-Therese, Ghislain Deleplace, and
Lucien Gillard. Private Money and Public Curren-
ients of state spending, and the fiscal author- cies: the 16th Century Challenge. Armonk, New
ities and regulators themselves. The result is York: M. E. Sharpe, 1994.
a gradual accretion of revenue-enhancing Buchanan, James M. “‘La Scienza delle Finanze’: The
changes, culminating in arrangements that Italian Tradition in Fiscal Theory,” in Fiscal Theory
and Political Economy: Selected Essays. Chapel Hill:
look remarkably “as if’ they were designed University of North Carolina Press, 1960, 2k-74.
from scratch to maximize government reve- Bums, Arthur R. Money and Monetary Policy in Early
nue. Times. New York: Augustus M. Kelley, 1965.
Together these arguments imply that sei- Chown, John F. A History of Money from AD 800. New
gniorage-enhancing institutional arrange- York: Routledge, 1994.
ments will be observed emerging later in Cipolla, Carlo. Money, Prices and Civilization in the Med-
countries that face fewer fiscal crisis, and es- iterranean World. Princeton: Princeton University
Press, 1956.
pecially those facing fewer external military
Correia, Isabel, and Pedro Teles. “Is the Friedman Rule
threats. Thus central banking came early to Optimal When Money is an Intermediate Good?”
belligerent nations of Europe, and only later Journal of Monetary Economics, October 1996,223-
to Switzerland, North America, Australia, and 44.
New Zealand. Cukierman, Alex. Central Bank Strategy, Credibility, and
Independence: Theory and Evidence. Cambridge,
Mass.: MIT Press, 1992.
V. CONCLUSION
Cukierman, Alex, Sebastian Edwards, and Guido Tabellini.
Fiscal considerations explain the main con- “Seigniorage and Political Instability.” American
Economic Review, June 1992,537-55.
tours of government’s roles in money, and
Derbyshire, J. Denis, and Ian Derbyshire. Political Systems
of the World. New York: St. Martin’s, 1996.
19. Earlier suspensions-the Bank of England during Friedman, Milton. A Program f o r Monetary Stability. New
the Napoleonic wars, the U.S. national banks during the York: Fordman University Press, 1960.
Civil War-were also associated with fiscal emergencies, Dwyer, Gerald P., and Thomas R. Saving. “Government
but were eventually followed by full resumption at the Revenue from Money Creation with Government and
pre-war parity.
 SELGIN & WHITE: GOVERNMENT’S ROLE IN MONEY 165

Private Money.” Journal of Monetary Economics, Rolnick, Arthur J., and Warren Weber. “Inflation and
March 1986,23949. Money Growth Under Alternative Monetary Stan-
Fry, Maxwell J. Money, Interest, and Bankingin Economic dards.” Federal Reserve Bank of Minneapolis Re-
Development. Baltimore, Md.: Johns Hopkins, 1988. search Department Working Paper No. 528, 1994.
Giovanni, Alberto, and Martha de Melo. “Government Sannucci, Valeria. “The Establishment of a Central Bank:
Revenue from Financial Repression.” American Eco- Italy in the 19th Century,” in A European Central
nomic Review, September 1993,953-63. Bank? Perspectives on Monetary Unification Ajier
Ten Years of the EMS, edited by Marcello d e Cecco
Glasner, David. Free Banking and Monetary Reform. Cam- and Alberto Giovannini. New York: Cambridge Uni-
bridge, Mass.: Cambridge University Press, 1989. versity Press, 1989, 244-80.
~. “An Evolutionary Theory of the State Monopoly Schumpeter, Joseph A. “The Crisis of the Tax State
over Money,” in Money and the Nation State, edited [ 19181.” International Economic Papers 4, 1954.
by Kevin Dowd and Richard H. Timberlake, Jr. New London: Macmillan.
Brunswick: Transaction Publishers, 1997,2145.
Selgin, George. “Network Effects, Adaptive Learning, and
Greenspan, Alan. “Fostering Financial Innovation: The the Transition to Fiat Money.” Manuscript, Univer-
Role of Government,” in The Future ofMoney in the sity of Georgia, 1997.
Information Age, edited by James A. Dorn. Washing-
ton, D.C.: Cat0 Institute, 1997,45-50. Selgin, George, and Lawrence H . White. “The Evolution of
a Free Banking System.” Economic Inquiry, July
Kagin, Donald H. Private Cold Coins and Patterns of the 1987,439-57.
Unitedstates. New York: Arc0 Publishing, 198 I .
.- “How Would the Invisible Hand Handle Money?”
Kindleberger, Charles P. “Forward” to Boyer-Xambeu et Journal of Economic Literature, December 1994,
al. 1994, ix-xii. 1,718-49.
Levi, Margaret. Of Rule and Revenue. Berkeley, Calif.: Smith, Vera. The Rationale of Central Banking. Indianap-
University of California Press, 1988. olis: Liberty Fund, 1990.
Lott, John R., Jr. “Does Political Reform Increase Wealth?: Spufford, Peter. Money and its Use in Medieval Europe.
Or, Why the Difference Between the Chicago and Cambridge, Mass.: Cambridge University Press,
Virginia Schools is Really an Elasticity Question.” 1988.
Public Choice, June 1997,219-27. Sylla, Richard. “The Autonomy of Monetary Authoritfes:
Mayhew, N. J. “From Regional to Central Minting, 115% The Case of the U.S. Federal Reserve System, in
1464,” in A New History ofthe Royal Mint, edited by Central Banks ’Independence in Historical Perspec-
C. E. Challis. Cambridge, Mass.: Cambridge Univer- tive, edited by Gianni Toniolo. Berlin: Walter de
sity Press, 1992, 83-178. Gruyter, 1988, 17-38.
McCulloch, J. Huston. Money and Inflation, 2nd ed. New Taub, Bart. “Private Fiat Money With Many Supplies.”
York: Academic Press, 1982. Journal of Monetary Economics, September 1985,
Motomura, Akira. “The Best and Worst of Currencies: 195-208.
Seigniorage and Currency Policy in Spain, 1597- White, Lawrence H., and Donald J. Boudreaux. “Is Non-
1650.” Journal of Economic History, March 1994, price Competition in Currency Inefficient?” Journal
10&27. ofMoney, Credit, and Banking, May 1998, 252-60.
Nichols, Donald A. “Some Principles of Inflationary Fi-
nance.” Journal of Political Economy, March/April
1974,42330.