Establishing a Monetary Union:
The United States Experience

Arthur J. Rolnick, Federal Reserve Bank of Minneapolis
Bruce D. Smith, Federal Reserve Bank of Minneapolis and Cornell University, and
Warren E. Weber, Federal Reserve Bank of Minneapolis.

The views expressed herein are those of the authors and not necessarily those of the Federal Reserve Bank of Minneapolis or the Federal Reserve System.

1. Introduction

In 1789, the Constitution prohibited the states from issuing their own coins or paper money, called bills of credit. Prior to that, each colony had issued its own paper money during the colonial period, and states had been free to issue bills of credit during the Revolutionary War and under the Articles of Confederation. Thus, the Constitutional prohibition of state-issued bills of credit constituted a radical departure from the previously existing monetary system. We have argued elsewhere (Rolnick-Smith-Weber 1993) that the newly formed government prohibited state issues of bills of credit because it desired to create a monetary union in order to reduce the costs of exchange rate variability and of converting between different currencies in inter-regional exchange.

The United States has, throughout much of its history, been committed to a multiple-issuer monetary system, and the monetary union created in 1789 was of this type. We previously demonstrated (RSW 1993) that the framers of the Constitution were well aware of a seigniorage incentive problem that exists in multiple-issuer monetary unions, as such a problem had been observed at various times before 1789. This problem derives from the fact that any issuer in a multiple-issuer monetary union can collect seigniorage from other members of the union, and the member with the fastest growing money supply will ultimately collect most of the seigniorage revenue in the system. Thus, one problem confronting the United States in 1789 was the construction of a multiple issuer monetary union in a way that addressed the seigniorage incentive problem.

We argue here that the solution arrived at was to allow each state to charter, if it so chose, note-issuing banks. We think it was generally held that these banks would redeem their notes for specie on demand, establishing a fixed exchange rate (indeed one with convertible currencies) system. Moreover, if banks were forced to redeem their notes on demand, then they had no discretion about the number of their notes outstanding. Thus, while it was possible to collect seigniorage revenue, no bank had any control over the amount of seigniorage it could collect.

The system established by the Constitution was not very successful in either eliminating exchange rate fluctuations between different kinds of notes or in avoiding seigniorage incentive problems. Until the National Banking System was established, the United States did not possess a uniform currency. Discounts on various kinds of notes fluctuated substantially through time and across locations. Nor was the seigniorage incentive problem avoided. Indeed, we will see that this problem arose even between different branches of the Second Bank, which was itself - as originally conceived - a multiple issuer system.

Why did the attempt at a monetary union represented by the Constitution fail in this way? We will argue that one reason was that the system that existed prior to the National Banking System left note-holders bearing the costs of note redemption. Thus discounts or premia could emerge between bank notes and specie reflecting the costs of note redemption. It was only in 1874, when note redemption costs were borne entirely by note-issuers rather than note-holders, that a uniform currency was achieved. Moreover, adequate mechanisms for preventing suspensions of convertibility (of notes) failed to exist until the National Banking system was established. The implication was that the methods for imposing discipline on the seigniorage incentive problem were flawed until about the same time.

2. Multiple-Issuer Monetary Unions

A monetary union is a collection of governments with some sovereignty that choose to eliminate exchange rate fluctuations between their currencies. This can be done in either of two ways. One is to have only a single currency issued by only one entity. We will refer to this as a single-issuer monetary union. The other is to allow some or all of the members of the monetary union to issue their own currency (or a common currency), so long as fixed exchange rates are maintained among the currencies. We will refer to such an arrangement as a multiple-issuer monetary union. We regard the monetary system established in the United States in 1789 to be a multiple-issuer monetary union. The currencies circulating were gold and silver (and copper) coins minted by the government and bank notes issued by banks chartered by the state governments.

Most authors take the view that, if the individual members of a multiple issuer monetary union are "small," the necessity of maintaining a fixed exchange rate among the currencies removes their discretion regarding how much money they issue. We have previously argued (RSW 1993), however, that if the currencies in a multiple issuer monetary union are perfectly substitutable - which we take to be the intent of forming such a union - this is not the case. If each issuer in such a system is not otherwise restrained, each issuer has complete discretion over how much they issue.

This discretion gives rise to what we term a seigniorage incentive problem. This problem derives from the fact that any issuer in a multiple issuer monetary union can collect seigniorage from other members of the union, and the member with the fastest growing money supply will ultimately collect most of the seigniorage revenue in the system. If one or more members of a monetary union takes advantage of this ability, thereby taxing the other members of the union, those members have an incentive to retaliate. Retaliation could take the form of all issuers in the union increasing their attempts to collect seigniorage, or of members in the union raising barriers against the use of the currencies of offending issuers. The former kind of action will tend to reduce the benefits of a monetary union because of excessive inflation, while the latter tends to defeat the purpose of monetary unification.

3. How the United States' Monetary Union Was Intended to Work

The currencies intended to circulate under the monetary union established by the Constitution were specie minted by the government and bank notes issued by banks chartered by the state governments. The federal government did not collect seigniorage under this system. There was no charge for minting bullion unless a person wanted to have the coins immediately. However, states could collect seigniorage through the note-issuing banks. The states were not prohibited from owning shares in these banks; indeed 100% state ownership of note-issuing, non-specie paying banks was held to be Constitutional in a variety of cases. Thus the states could collect seigniorage income through ownership of note-issuing banks, so long as these banks were chartered corporations. There were, of course, other methods of collecting seigniorage income indirectly, even if a state had no ownership of bank shares. States could, for example, tax bank profits, or could force loans from banks at below market interest rates in exchange for charters. Both were common practices, as documented by Dewey (1910) and Sylla, Legler, and Wallis (198X).

How did this system provide a fixed exchange rate between the notes of various banks, and how did it address the seigniorage incentive problem? If it had been the case - as we believe was conceived in 1789 - that all banks redeemed their notes for specie on demand then, subject to some qualifications we will state, there presumably would have existed a fixed exchange rate system (indeed one with convertible currencies). Moreover, if banks were forced to redeem their notes on demand, then they had no discretion about the number of their notes outstanding. Thus, while it was possible to collect seigniorage revenue, no bank had any control over the amount of seigniorage it could collect. The absence of any discretion on this dimension prevented the existence of a seigniorage incentive problem.

If this was the intention, why did the Constitution require states to charter note-issuing banks, while not allowing them to issue notes themselves? We believe that this was part of a mechanism to enforce the redeemability of notes. In particular, a note-holder of a private bank could sue for redemption if necessary, while the holder of a state-issued bill of credit could not sue the state for redemption without its consent. Thus the system of note issue through private banks left note-holders with recourse to the courts if redemption were ever refused, a recourse that would not have existed with states issuing notes directly. And indeed, Dewey (1910) argues that it was often necessary to resort to this recourse.

A. 1791-1812

Since states were empowered to charter and to regulate banks, as well as to enforce the regulations they imposed, enforcement of convertibility of bank notes was entirely left to the individual states over this period. Many states imposed no penalties for nonredemption of notes (Dewey 1910), and even when penalties existed, noteholders were often dissuaded from presenting notes for redemption. The First Bank of the United States also appears to have played no role in attempting to induce state banks to maintain convertibility (Holdsworth 1910).

In spite of this, the seigniorage incentive problem does not appear to have taken on a major significance over this period. Perhaps its absence accounts for the lack of concern with banks that were failing to redeem.

It is also true that, during this period, note-holders bore the entire cost of redeeming notes when notes were redeemable. As we have argued, this would permit the emergence of discounts even under full redeemability, and discounts on the notes of state banks were, in fact, observed. A uniform currency had not yet been obtained.

B. 1812-16

During the war of 1812 there was widespread suspension of convertibility of notes, along with a proliferation of banks. This suspension continued beyond the end of the war, and became a major concern of the Treasury, which was unable to collect revenue in a uniform currency. This led the Treasury to actively support the creation of the Second Bank of the United States (SBUS) as a mechanism for enforcing note redemption by the state banks, as well as for enforcing uniformity of the currency.

C. The Second Bank

The SBUS was created to "restore" uniformity of the currency, as well as to induce state banks to resume note redemption. While the SBUS had no direct power to force state banks to redeem their notes, it could refuse to accept the notes of nonredeeming banks. Such a stance would lead the Treasury to stop accepting the notes of such banks for taxes, which presumably would limit their circulation.

We have argued that having a system with multiple issuers of a uniform currency gives rise to a seigniorage incentive problem. The SBUS was itself subject to this problem in two forms. First it was, along with the state banks, part of a multiple issuer system, and it was therefore in a position to raise seigniorage at the expense of the rest of the system. (And indeed, a common charge against the SBUS was that it aggrandized its profits at the expense of state banks.) However, the SBUS was subject to the same control mechanism as the rest of the system, namely redemption of its notes.

More subtly, though, the SBUS was - in and of itself - a multiple issuer system. Each branch, at the inception of the Bank, could issue notes which were redeemed in Philadelphia. Moreover, many branches (including Baltimore and some southern and western branches) were not (early on) redeeming their own notes. As a result, a serious seigniorage incentive problem arose within the SBUS itself. Branches whose note issues were growing most rapidly gained capital at the expense of the rest of the SBUS.

In 1818 it was viewed as necessary to limit the seigniorage incentive problem operating within the SBUS. Two actions were taken: redemption of branch notes at other branches was limited, and many southern and western branches were ordered to cease note issues. There were two results. First, for many holders of SBUS notes, redemption costs rose. Accordingly, discounts on the notes of other branches appeared in Philadelphia. Second, in many areas only state bank notes circulated, and these were generally less uniform than SBUS notes. As a consequence, currency uniformity was not attained.

D. The Panic of 1819

In 1819 there was a systematic panic that caused a suspension of convertibility throughout most of the banking system. Under the multiple issuer system in place, such a suspension represented a major threat, as it eliminated the mechanism of choice for managing the seigniorage incentive problem.

This problem manifested itself almost immediately with general suspension. Several states already had wholly state owned banks. Once they were not subject to note redemption, their note issues could be expanded. Laws were passed to prevent these notes from being refused in payments, and strategic issues arose about forcing out-of-state-residents to accept notes issued (in effect) by the state. Such issues had been viewed as a serious problem under the Articles of Confederation (RSW 1993), as they were again during the panic.

This episode illustrates the limited enforcement mechanisms available against the seigniorage incentive problem in the absence of full redemption. Despite the Constitutional prohibition of state issues of bills of credit, note issues by wholly state owned (nonredeeming) banks were deemed Constitutional in a variety of cases even when the state had "guaranteed" the note issues.

E. Biddle's Policy

By 1823, note convertibility had been restored in most of the banking system, and in 1823 Nicholas Biddle ascended to the presidency of the SBUS. Prior to Biddle's presidency, the SBUS had followed the policy of "paying out" the notes of state banks whenever possible. Biddle reversed this practice, and it became SBUS policy to pay out its own notes, and to deliver (at its own expense) the state bank notes it received for redemption. As a result, agents could deposit state bank notes at the SBUS and withdraw SBUS notes. The result is that, to a significant extent, redemption costs for state banks could be transferred from individual note-holders to the SBUS. According to our previous arguments, this should reduce the discounts on state bank notes that could be observed. Such a reduction does, in fact, manifest itself in the data.

Biddle also adopted a new policy for addressing the seigniorage incentive problem that existed between branches of the SBUS. Branch notes were issued by buying inland bills of exchange. As these were presented in New York and Philadelphia they constituted claims on the branch of issue, and this prevented note issues from transferring capital between branches.

Biddle's policies reduced the general range of bank note discounts to their most limited level of the ante-bellum period. Nevertheless, it did not altogether eliminate discounts on notes, and indeed there were small discounts on the notes of SBUS branches. Such discounts presumably allowed Andrew Jackson to charge, in his message vetoing renewal of the SBUS charter, that the SBUS had failed to create a uniform currency.

F. Free Banking

Following the demise of the SBUS, many states experimented with free banking, in which bank charters were more generally available. As before 1812, the only mechanism for the control of seigniorage incentive problems was note redemption, and note-holders bore all redemption costs. This allowed discounts to be observed in excess of those that prevailed in the late 1820s.

G. The National Banking System

The formation of the National Banking System in 1863 saw an important change in the mechanism by which redemption costs were born: banks bore all costs of note redemption. This completely eliminated all costs of redemption for the typical note-holder, and no discounts were ever reported for National Bank notes in nonbank transactions.

Before 1874, however, redemption costs between banks were born by the redeeming bank. As a result, it was possible for discounts to be observed on national bank notes in inter-bank transactions (Friedman and Schwartz 1963). This problem was remedied in 1874: thereafter the costs of note redemption were born by the issuing bank, and discounts on national bank notes were eliminated completely.

It continued to be the case, of course, that note redemption was the means of controlling seigniorage incentive problems, and mechanisms for eliminating the suspension of note convertibility were put in place. Thus it was not until 1874 that the United States had a system in place for maintaining a completely uniform currency and for managing the seigniorage incentive problem associated with a multiple issuer system.

5. Summary

To summarize, in our view the U.S. was committed to having a multiple issuer monetary union. In order to address the implied seigniorage incentive problem, redemption of bank notes was conceived as an enforcement mechanism. Whenever this mechanism broke down, the monetary union was threatened. Thus suspension of convertibility by banks was a serious threat to the monetary system. Moreover, even if all banks were redeeming their notes, there were costs associated with note redemption. To the extent that these costs were borne by note-holders, discounts, or premia on notes were always possible. Both phenomena interfered repeatedly with the attainment of a monetary union.