Seigniorage, legal tender, and the demand notes of 1861.
Shortly after the onset of hostilities at Fort Sumter, the Lincoln administration asked Congress to authorize the first large-scale issue of federal paper currency. Officially labeled as United States Notes, they were popularly known as the Demand Notes of 1861. The administration's action is unremarkable. The extraordinary revenue requirements of long and costly wars are often met, in part, by issuing paper money. More surprising (and not widely recognized) is that this initial issue of paper money carried provisions that largely prevented it from yielding sustainable revenue (seigniorage).
These provisions were largely omitted for the successor currency to the 1861 Notes (also known as United States Notes, but popularly as Greenbacks). As a result, its seigniorage potential was greater (although seigniorage was limited by other provisions). In modern software jargon, the Greenbacks might be viewed as 1861 Notes version 2.0: a version better suited to sustain a flow of seigniorage. However, the government was unable to replace the latter with the former very quickly, resulting in an extended period during which two paper currencies with different properties circulated side by side both with each other and with gold coin.
This somewhat untidy situation provides an interesting (and heretofore underexploited) experiment for economists. These alternative monies circulated widely with market-determined exchange rates for a substantial period during which the amounts of both paper currencies in circulation greatly increased before the 1861 Notes were removed from circulation. We have collected daily data on these exchange rates and report them as differing gold premiums for the competing paper monies. Below, these data are exploited to calculate the seigniorage cost of the various provisions of the 1861 Notes and to provide evidence on alternative theories of expectations formation during this period.
Our paper has three sections. The first presents a more extensive discussion than has been previously available of the 1861 Notes. In the second, a simple analytical framework is used to demonstrate and, where possible, estimate the seigniorage costs of the provisions attached to these notes. The third section analyzes the market relationship between gold and the simultaneously circulating 1861 Notes and Greenbacks prior to the (all but complete) retirement of the former in 1863. The paper concludes with a summary.
II. THE DEMAND NOTES OF 1861
The major focus of this paper is on the Demand Notes of 1861, the first large-scale federal experiment with paper money. (1) We concentrate on these notes for two reasons. First, little has been written about them. Mitchell (1902), in his pioneering work on Civil War finance, devotes only 7 pages to them while devoting 17 pages to fractional currency and minor coins during the war. In addition, most of his discussion concerns the period prior to suspension. More recently, Calomiris (1988a, 1988b, 1992) makes reference to the 1861 Notes, but only as a part of his discussion of Greenbacks. Second, the paradox of issuing largely seigniorage-free paper money to help finance a war has not been well recognized and, hence, largely unexplained. This conflict between means and ends is interesting in its own right and illuminates some of what is puzzling in the sequence of monetary events over the first half of the Civil War.
As a part of our discussion of the 1861 Notes, we challenge or extend the existing literature in three ways. First, we demonstrate that the conclusion of Studenski and Krooss (1963, 141, 144) that these notes were only used by the government to pay salaries and, hence, in their words, were a "pseudomoney" is mistaken. Examination of the financial press of the day suggests that these notes were not only popular but also circulated widely as money. A convincing claim can be made that they were indeed a national currency. Second, we offer an explanation for why the Greenbacks came into being that may be regarded as an alternative or a supplement to that given by Hammond (1961, 1970). Third, we present a simple model of the gold premium in terms of the 1861 Notes after the United States suspended the gold standard that provides quantitative predictions and is at variance with Calomiris (1988a, 1998b).
A. The Origins of the Notes
The incoming Lincoln administration not only faced the prospect of imminent civil war but also inherited a precarious fiscal situation succinctly described by Bray Hammond as "the North's empty purse." After 4 yr of budget deficits during the Buchanan administration, in which the national debt rose from $28.7 to $64.8 million, the federal treasury was nearly empty. (2) From March through June 1861, the final quarter of the 1861 fiscal year, revenue from ordinary sources (primarily the tariff) amounted to about $5.8 million against expenditures in excess of $23.5 million. (3)
To deal with this fiscal crisis, President Lincoln summoned Congress to a special session beginning on July 4, 1861. The Congress acted rapidly on Secretary of Treasury Salmon Chase's financing recommendations and, on July 17, passed the first of many bills to finance the Civil War, authorizing the government to borrow $250 million. Of this sum, the Treasury could issue $50 million in non-interest-bearing notes, payable on demand in gold, and in denominations of at least $10 but less than $50. The act was amended on August 5 to allow the Secretary to fix the denominations (which became $5, $10, and $20) and to make them receivable for all public dues including customs duties on a par with coin. Thus, these notes were equivalent to a warehouse receipt for a fixed weight of gold.
The 1861 Notes were, then, an emergency measure to furnish a depleted Treasury with currency until the loans authorized could be floated to bridge the rapidly widening gap between expenditures and tax receipts. (4) On February 12, 1862, when the Treasury encountered a similar situation, Congress authorized an additional $10 million in notes. The total authorized, $60 million, was a substantial addition to the money stock. Exactly how substantial is unclear because precise estimates of the coin and the notes of state-chartered banks circulating in the Union have not been agreed upon. For example, the Director of the Mint (1861, 8) estimated that approximately $275 to $300 million in gold coin (including bank reserves) was outstanding in October 1861, of which $20 million was held in Confederate states. (5)
This appears to be a substantial underestimate of the amount held in the Confederate states. Godfrey (1978, 64) shows that in the same month, Southern banks alone held $24 million in specie not counting coin that either circulated or was hoarded in the South. Dewey (1909, 283) states that on January 1, 1862, banks in the loyal states held some $87 million in coin. A somewhat lower figure of $76.4 million is given in House of Representatives (U.S. Government 1862, 209). If we take Secretary Chase's (U.S. Treasury 1862, 13) estimate that $210 million in coin was in circulation in the loyal states (including that held by banks) on November 1, 1861, subtract from this the amount held as bank reserves (about $90 million), and add to this his estimate of the notes of state-chartered banks outstanding, $130 million (U.S. Treasury 1862, 7), we arrive at a total of $250 million for circulating money. (6)
The $34 million of 1861 Notes issued prior to suspension and the $60 million authorized ($26 million of which were put into circulation in the first quarter of 1862 as shown in Table 1) represented 14% and 24% increases in this measure of money. Secretary Chase stated that the emission of these notes was "not with the desire of furnishing a general currency, but for the purpose of making good the difference between the amounts obtained by loans and the sums required by the public service." Nevertheless, they were the first large-scale experience of the U.S. government in issuing paper money and they enjoyed wide circulation.
Notes were payable on demand in gold at four Sub-Treasury offices: New York City, Boston, Philadelphia, and St. Louis, and the Treasury depository in Cincinnati. (7) It was known early on that these notes would not be issued to buy gold, although they were authorized for that purpose. Rather, they were to be paid out to satisfy the immediate and pressing claims on the Treasury, and there was no free gold in the Treasury to serve as their backing. (8)
These notes presented formidable production problems. They were engraved and printed by the American Bank Note Company with the date August 10, 1861. Although this was the day that they were supposedly first issued to the public, evidence suggests that they actually entered circulation toward the end of August. (9) It was initially expected that the Treasurer of the United States and the Register of the Treasury would personally sign each note. This proved impractical for the millions of notes involved, and the plates were altered so that an army of clerks hired for the purpose could sign "for" the two officials. (10) Even an army takes time to move, however, and the last of the initial $50 million issued left the signers on February 6, 1862, just 6 days before Congress authorized an additional $10 million.
B. Legal Tender Status
The legal tender status of the 1861 Notes was complex. Since they were acceptable for all public dues including customs, they were, de jure, a full public legal tender. Since they could be exchanged at par on demand for gold coin (itself a legal tender), one might conclude that they were a de facto legal tender for all private debts as well (a view held by Secretary Chase). That this was not the case diminished their acceptability to banks prior to suspension. Not being "lawful money," the Notes could not serve as bank reserves on a par with coin. Hence, some banks refused to accept them on deposit or would only accept them as special deposits to be repaid in the same form should they be withdrawn. After suspension, banks were no longer obligated to redeem their notes in "lawful money" and no longer objected to receiving 1861 Notes on deposit. After the Greenbacks were introduced, Congress, on March 17, 1862, declared them to be a full legal tender for all debts public and private. However, the legal tender properties of the 1861 Notes remained superior to those of the Greenbacks in one important respect. The Greenbacks were a public legal tender for all public debts and dues except for customs duties, which were a major source of federal revenue. The 1861 Notes could be used for customs on a par with coin and, as we show below, this had major implications for both the market value and the seigniorage potential of the two classes of United States Notes.
C. General Acceptability
Despite his claim that these notes were not intended to be a general currency, Secretary Chase encouraged their use by having his salary and that of other Treasury officials paid in them. They were also used to pay troops in the Union army and to pay for supplies received from military contractors. (11) In fact, the notes quickly became genuinely popular as indicated by the following partial listing of quotations gleaned from newspaper articles from the months immediately after their introduction. And, contrary to the conclusion of Studenski and Krooss (1963), they circulated widely throughout most of the Union.
The increasing inquiry for United States Demand notes is an important feature of the financial situation. The Cashier of the Sub-Treasury finds it impossible to keep any on hand
....Some of the applications [for notes] are from parties going West [today's mid-west] ... and who find these notes the most convenient shape to carry money in.... $20,000,000 of these demand notes have gone West. (Herald September 14, 1861, 3).
At the West they are received with the greatest favor by all the people and by all moneyed institutions in Cincinnati, Chicago, etc..... At the North and East they are universally accepted among the people. (Times September 27, 1861, 4).
There was a brisk inquiry for the United States notes, which are rapidly entering into general circulation. All, or nearly all, the banks receive them on deposit as money, and they will undoubtedly become the general circulating medium of the North in the course of a few months. (Herald October 5, 1861, 8).
In the West, as in New England, they are heartily welcomed. We presume that we are not far wrong if we say that the commerce of the West has during the past ten years paid a tax of five millions a year in the shape of discounts and losses on its paper currency. (Herald October 14, 1861, 8).
Four-fifths of the currency at present in circulation here, in the other Central and in the Western States, consists of United States notes. Three fourths of the business which is giving employment to our manufacturers and merchants grow out of the war, and is liquidated in United States notes. (Herald February 5, 1862, 3)
The popularity of the United States Notes in every quarter of the country and their paramount convenience for army and other Government payments are already equal to the absorption of $60,000,000.... (Times March 27, 1862, 2).
Most of them have found a home in the West, where they are treasured as the best currency ever to circulate in that region, and they will be kept there until they are worn out. (Herald April 7, 1862, 5). (12)
D. The Value of 1861 Notes after Suspension
On December 30, 1861, the Associated Banks of the Atlantic Cities suspended convertibility of their notes for specie and most other banks soon followed. Because all transactions with the federal government had to be conducted in coin (or 1861 Notes), this forced a suspension of the gold standard on December 31. Although suspension meant that the government would no longer convert 1861 Notes into coin on demand, their full public legal tender status continued. In particular, customs duties (tariffs) could be paid either in coin or in 1861 Notes, even after suspension (the government continued to service the national debt in coin). (13) Below, we argue that this continuing legal tender provision ensured that 1861 Notes remained incapable of generating sustained seigniorage despite their inconvertibility. Before presenting this argument, we introduce the data upon which it is partially based.
Predictably, the values of a paper dollar and gold dollar diverged after suspension. We are able to quantify this divergence in the form of daily observations on the 1861 Note price of a gold dollar. The data are taken from the columns of the Times, Herald, and Commercial Advertiser (1861, 1862, 1863). (14) In most cases, the price of 1861 Notes and the price of gold were quoted in terms of one of the competing currencies (particularly Greenbacks after they were introduced) then converted into the exchange rate above. It was necessary to use particular care in determining which numeraire was being used for the various quotes. A detailed discussion of numeraire issues comprises our Appendix.
For dates after the gold market opened on January 13, 1862, the underlying price observation is the mean of all prices recorded for transactions on that day. A similar method was used for the price of 1861 Notes after they began trading on the stock exchange during August 1862. Observations prior to those respective dates are based on prices (either a single figure or a range) at which transactions involving private banking houses or brokers were reported. This method also applies to observations from October 25 to November 15, 1862, when trading in both assets on the exchange was suspended in an effort to curb "excessive speculation." An analogous Greenback price of gold was also collected for the period after Greenbacks were introduced. While we delay examination of this exchange rate until our discussion of the Greenbacks below, a graph of both gold premiums appears as shown in Figure 1.
E. The Disposition of the Notes
On February 25, 1862, President Lincoln signed the bill authorizing the first issue of $150 million in Greenbacks. Of this sum, $60 million were to replace the 1861 Notes as rapidly as practicable. However, since the printing of Greenbacks did not begin until March 13 and they did not begin to circulate until early April 1862, the Treasury continued to issue 1861 Notes until the full $60 million authorized by law was outstanding. (15) Thus, the costs of maintaining a full legal tender currency not only continued but increased until the more limited legal tender Greenbacks could be issued in sufficient numbers to replace them. In fact, there was a substantial amount of 1861 Notes outstanding in early 1863. However, the amount outstanding had been reduced to $3,822,610 by June 30 (as shown in Table 1) and their market price ceased to be reported. By June 1864, less than a million dollars of these notes remained in circulation and a small amount remained outstanding for many years. (16)
On February 25 and July 11, 1862, and January 17 (March 3), 1863, Congress authorized $150 million in United States Notes, popularly termed Greenbacks. (17) They were a legal tender for all debts, public and private, and receivable for all public dues except for import duties that were payable in gold or 1861 Notes. With the customs receivability provision omitted, the seigniorage potential of the Greenbacks in the new flexible exchange rate regime was enhanced. Nevertheless, their seigniorage potential was somewhat compromised before the first $150 million were issued. Their authorizing legislation contained two options by which they could be converted into interest-bearing debt. First, they were exchangeable at any time for 6%, 25-yr bonds whose interest was payable in coin (this option proved unattractive and was withdrawn in 1863). Second, they could be left on deposit at the subtreasuries in special certificates of deposit at 5% for a period of not less than 30 days provided that 10 days' notice was given prior to withdrawal. Initially, the amount that the Treasury could accept as certificates was fixed at $25 million. As these accounts proved to be popular, the ceiling was raised in several steps to $150 million. Concern, however, was expressed about the adequacy of reserves against a sudden runoff of the deposits. To address this, the second legal tender act directed the Secretary to retain as reserves a sum of not less than $50 million. Thus, of the first $300 million of Greenbacks issued, $50 million were retained as reserves against the certificates of deposit.
The Demand Notes of 1861, the first large-scale federal experience with paper money, were successful in the sense that some $60,000,000 of them were put into circulation expanding the money supply by nearly 25%. They were widely accepted by the public and, ultimately, the banks and exchanged at par with coin and, prior to suspension, the notes of the major banks (especially those of the associated banks of the Atlantic cities). They were replaced eventually by Greenbacks. However, an extended period passed during which the two paper currencies circulated side by side with gold coin at market-determined exchange rates.
Below, we formalize our claim that this switch involved a fundamental change in the seigniorage potential of United States Notes and, therefore, the delay in replacing the 1861 Notes involved a significant cost to the Treasury. The literature is silent on this aspect of the Notes and its relationship to the subsequent switch to Greenbacks.
[FIGURE 1 OMITTED]
III.SEIGNIORAGE IMPLICATIONS OF THE PROVISIONS ON THE 1861 NOTES
Between their inception and the end of the gold standard on December 31, 1861, the 1861 Notes were convertible at par into gold coin.
This provision undoubtedly enhanced their acceptability, but seriously hampered the generation of revenue, which was the government's purpose in issuing them. The notion that convertibility limits a currency's ability to generate seigniorage is intuitive. However, we wish to supplement this intuition with a very simple formal framework. Our purposes were to (a) analyze some specific questions regarding a convertible currency concretely and (b) lay the groundwork for similar analysis of the effect of the customs receivability provision after suspension.
A. Convertible Notes and Seigniorage
Let money consist of gold coin, G, and paper money, M (in this historical context, M includes bank notes). If the demand for money, L, is fixed in terms of goods and P is an index of the paper money price of goods, equilibrium requires that M/P + G = L. Suppose P = 1 and M = [M.sub.0] initially implying [M.sub.0] + G = L. Finally, suppose the government issues paper currency at rate dN/dt = n for a period of time At sufficient to add notes in the amount N (= n[DELTA]t) to M, then stops (n = 0). As the government issues currency and expends it on goods, the government's collection of seigniorage is measured by N the quantity of goods it purchases at initial prices. (18) M rises to [M.sub.0] + N and equilibrium is restored as prices rise to 1 + [DELTA]P = 1 + N/(L - G).
For a true fiat currency, the eventual rise in prices would complete the process. However, it is only an intermediate outcome for a convertible currency. For a convertible currency, the rate at which N changes is given by dN/ dt = n - x, where n is the instantaneous rate of note issue and x is the rate of redemption of notes for gold. After a general inflation, a gold dollar would be worth P = 1 + p paper dollars representing a "gold premium" of p = N/(L - G). Thus, starting from parity of paper dollars and gold dollars, a 10% increase in the amount of paper money leads to a 10% premium of the gold dollar compared to the paper dollar. This gold premium gives money holders an incentive to redeem their currency for gold. Acting on this incentive by presenting paper dollars to the government for redemption causes the paper currency in circulation to fall.
The response of note holders to such a gold premium could be formalized by: x = [alpha]Np, where a is a parameter measuring the percentage of notes outstanding redeemed per time period per unit gold premium. Thus:
(1) dN / dt = n - [alpha]Np.
The one-time issue of currency in the amount N described in the first paragraph of this section would correspond to a scenario in which the government issues and spends currency at rate n over time period At until N = n[DELTA]t is issued then reverts to n = 0. It is unnecessary to specify the timing of the price adjustment to conclude that P will eventually rise (P > 1). With p > 0, x is positive and dN/dt < 0. N falls until P returns to 1 and M returns to [M.sub.0] (N = 0). Every dollar the government spends while n > 0 will eventually be repurchased with gold coin obtained by borrowing, forgone spending, or additional taxes. Thus, with a convertible currency, the government can collect seigniorage temporarily but must return it eventually. The date of eventual return of seigniorage moves farther in the future as [alpha] falls and as P responds with a greater lag.
Although the above analysis of a one-time issue of notes illustrates the fundamental seigniorage problem of the convertibility provision, it fails as a characterization of the experience with 1861 Notes in two aspects. First, the amount of 1861 Notes outstanding increased steadily throughout the period of convertibility and reached its peak well into the period of inconvertibility. Hence, convertibility did not last long enough for the predicted pattern of inflation, redemption, and reduction in notes outstanding to be observed. Second, the one-time issue examined above does not allow for the recirculation of redeemed notes, which would undoubtedly occur when the government remained in need of revenue.
Thus, we examine a permanent issue of a convertible currency to (a) illustrate how the seigniorage problems of convertibility might have manifested themselves if convertibility had lasted longer and (b) establish the basis for our subsequent comparison of convertibility to the customs receivability provision of the 1861 Notes. Suppose the government maintains the amount of notes outstanding at N by continuously spending notes at rate [n.sup.*] as they are redeemed. Setting dN/dt = 0 above and solving for n yields the steady-state rate of note-financed spending:
(2) [n.sup.*] = [alpha]Np.
Assuming (a) no limit on n and (b) finite [alpha], convertibility is not an insuperable obstacle to maintaining a given amount of notes outstanding indefinitely. Although convertibility does not prevent the government from accomplishing a permanent increase in paper money, the initial collection of seigniorage tends to be gradually returned in the process of recirculation. For each paper dollar, the government accepts for conversion, it must return one gold dollar to the public and (potentially) forego the goods that the gold dollar could buy. It can recoup this loss of goods as it returns the paper dollar to circulation by spending it, but only in the amount I/P= 1/(1 + p) < 1, since a paper dollar buys 1/P goods. Thus, seigniorage is returned to the public on every dollar of redemption/ recirculation. The rate of loss per unit time is given by:
(3) Loss = [alpha][Np.sup.2]/(1 +p) (= [n.sup.*] - [n.sup.*]/P = [[alpha]N(P - 1).sup.2]/P).
It would be asking too much of this highly simplified framework to generate a realistic estimate of the drain on the Treasury caused by maintaining a convertible paper currency in competition with circulating gold (particularly with no guidance as to the value of [alpha]). However, it indicates that maintaining a fixed amount of notes outstanding sets up a system in which seigniorage is steadily returned to note holders. It also delivers the prediction that this yearly cost increases both with the size of the maintained note issue, N, and with the public's responsiveness to arbitrage opportunities, [alpha], and that it accrues as long as convertibility is kept in place.
B. Suspension and the Seigniorage of 1861 Notes
The suspension of convertibility might be expected to fundamentally alter the seigniorage potential of a paper currency. However, the continuing provision that allowed tariffs to be paid by 1861 Notes on a par with gold prevented the notes from generating significant seigniorage. Our attempt to support this claim begins with Figure 1, which displays the fact that 1861 Notes immediately sold at a discount to gold upon suspension and continued to do so throughout 1862. Given the alternative of discharging the same liability with gold coin or, readily available, but less valuable, 1861 Notes, it is hard to understand why savvy importers would ever pay tariffs in gold coin. From this perspective, the surviving customs receivability provision can be viewed as a form of indirect convertibility. After suspension, 1861 Notes could still be "redeemed" for more valuable gold coin at par by using them to pay one's tariff obligations. The significant change from explicit convertibility was that the rate of "redemption" was limited to the rate at which tariff obligations accrued. On the other side of the ledger redemption was less cumbersome than direct redemption since the Notes could be "redeemed" in the normal course of tariff collection and did not have to be transported to a subtreasury.
The framework used to analyze seigniorage under a convertible currency can be adapted to this case as well. Assuming that all tariffs are paid with 1861 Notes, the differential equation summarizing the evolution of N changes from Equation (2) to:
(4) dN/dt = n - T,
where T is the rate of tariff collection per time period. As was the case under convertibility, the government can maintain N at a target level by spending notes (in this case at rate [n.sup.*] = T) as they are received in payment of tariffs. As in the previous case, the government loses seigniorage in the process.
If the market value of a gold dollar is 1 + p paper dollars (p = gold premium), the government loses p gold dollars (the equivalent of p/(1 + p) paper dollars) for each paper dollar it accepts (and recirculates) in lieu of a gold dollar. Thus, the rate of loss (in gold dollars) per time period is given by:
(5) Loss = Tp (= [n.sup.*]p).
This rate of loss can be estimated from observed values for T and p. During the first postsuspension quarter ending on March 31, 1862, revenue from tariffs was $15 million (shown in the second column of figures in Table 1). The average of all daily values for the gold premium, p, for those 3 months was 2.61% (shown in the next column of the same table). Thus, the tariff provision cost the Treasury approximately .026 ($15 million) = $392,000 in the first quarter of 1862. Calculations for the remaining quarters of 1862 (also shown in Table 1) result in estimated losses of $493,000, $1,581,000, and $515,000: a total of approximately $3 million for the calendar year. To provide some perspective, using linear interpolations for the point values of N in Table 1, one can arrive at a rough estimate of average notes outstanding of approximately $45 million for all 1862. The seigniorage loss from the receivability provision expressed as a fraction of the average amount of notes outstanding is approximately 6.7% per year (=$3 million per year/$45 million). This is quite close to the yield on interest-bearing government debt (as measured by the "6's of 1861"). In Section III, we provide a theoretical argument suggesting that this correspondence between the rate of debt service on interest-bearing debt and the rate of Treasury losses on paper currency is not coincidental. Here, we limit ourselves to pointing out that this is a substantial carrying cost for maintaining a currency whose ostensible purpose was the generation of revenue for the Treasury.
C. Why Were These Provisions Attached to the 1861 Notes?
A paper currency issued with the stated purpose of generating revenue accompanied by provisions largely negating its ability to do so on a sustained basis presents the observer with a paradox. One interpretation of the fact that these provisions were removed from the successor Greenbacks is that Congress and/or the Treasury regarded them as mistakes in retrospect. While a detailed rendering of the economic and political discussion leading up to the various decisions is beyond the scope of our paper, we sketch three alternative (and overlapping) hypotheses regarding the decision to attach, then remove, these provisions.
First, the provisions attached to paper money in a gold coin environment must be devised with the need to make the currency attractive to the public and the countervailing need to maximize its revenue-generating potential. Policymakers were forced to make a judgment on the trade-off between these needs in designing the provisions of the 1861 Notes. Perhaps, in the light of experience and a changed environment, these provisions were later judged to have erred on the side of the former and were adjusted toward the latter in the case of Greenbacks. As the discussion of Greenbacks below indicates, some of the provisions applied to the Greenbacks were also designed to address the need for acceptability at the expense of revenue. This suggests that choosing the optimal place on the trade-off remained a vexing issue throughout the war.
Second, the political figures of the day may not have understood the seigniorage consequences of provisions like the acceptability of 1861 Notes for tariffs but arrived at a better understanding after noticing the Treasury's subsequent distress. The federal government had no experience with the intricacies of paper currency and the economic theories of the day had their limitations. The following long quote from Senator Fessenden regarding the legal tender provisions of the Greenbacks may serve to illustrate how perplexing the task seemed to the political leaders of the time.
Nobody knows much upon the question of finance, not even those who are most familiar with it; for sir, I declare today that, in the whole number of learned financial men that I have consulted, I never found any two of them who agree, and therefore, it is hardly worthwhile for us to plead any very remarkable degree of ignorance when nobody is competent to instruct us; and yet such is the fact. I can state to you, Mr. President, that on one day I was advised very strongly by a leading financial man, at all events to oppose this legal-tender clause; he exclaimed against it with all the bitterness in the world. On the very same day I received a note from a friend of his, telling me that we could not get along without it. I showed it to him, and he expressed his utter surprise. He went home, and next day telegraphed me that he had changed his mind and now thought it was absolutely necessary; and his friend who wrote me again that he had changed his, and they were two of the most eminent financial men in the country. (Bayley 1881, 80-81). (19)
Third, political leaders may have realized that these provisions would force the Treasury to eventually return any initial seigniorage but decided that outcome was acceptable because (a) the provisions were necessary to ensure the acceptability of the currency, which would raise urgently needed funds that could not be immediately acquired in any other way and (b) the war was initially expected to be short. (20) Under this hypothesis, seigniorage-limiting provisions were subsequently omitted from the Greenbacks as the true length and cost of the war were better appreciated by that time. Although these hypotheses may be viewed as reinforcing, rather than mutually exclusive, we tend to give greatest weight to the third. In particular, the fact that the Lincoln administration asked for little in the way of initial tax increases and called upon the states to furnish only 75,000 in militia suggests that it hoped for a relatively quick resolution of the conflict.
It is interesting to note that the bill authorizing the Greenbacks that originated in, and was passed by, the House of Representatives provided that they should be receivable for taxes, debts, and demands of all kinds due to the United States. That is, the Greenbacks would retain the customs receivability provision of the 1861 Notes but would be a private legal tender as well. The Senate amended the bill by requiring that duties on imports should continue to be paid in coin. It was this version of the bill that became law. The House/Senate inconsistency may be consistent with the hypothesis that policymakers were grappling with the proper choice along the acceptability/seigniorage trade-off or that there was confusion about the seigniorage implications of provisions aimed at making paper money more acceptable to the public.
It is a further puzzle that other research has not noticed the heavy cost to the Treasury of maintaining the 1861 Notes and has, for the most part, not listed a change in the seigniorage flow as a primary motivation for the switch from 1861 Notes to Greenbacks. For example, this motivation is not mentioned by Hammond (1961, 1970). Instead, he emphasizes a motivation originating within the banking community. The 1861 Notes presented the bankers with a dilemma. Not being a private legal tender, they could not legally be used by the banks as reserves. If the banks were to accept them on deposit, there was no guarantee that they could pass them along to customers. Thus, the banks might be burdened with an illiquid asset from which they derived no income. On the other hand, not to accept the federal government's money in a time of national emergency would make them appear unpatriotic.
According to this interpretation, the crucial innovation was that the Greenbacks were a full private legal tender, while the 1861 Notes were not. The fact that 1861 Notes were a full public legal tender, while the Greenbacks were not, is not presented as crucial.
IV. THE MARKET BEHAVIOR OF THE TWO CLASSES OF UNITED STATES NOTES
In this section, we attempt to evaluate alternative hypotheses regarding the determinants of the market exchange rate between gold and the competing paper currencies in light of the environment that existed between the end of 1861 and the (nearly completed) retirement of the 1861 Notes in mid-1863. There are two questions that may be asked in this regard. First, why did one "dollar" circulate at a discount to another dollar? Second, why did this discount fluctuate the way it did?
Since both 1861 Notes and gold could be used to discharge tariff obligations and Greenbacks could not, one would expect Greenbacks to trade at a discount to both. Returning again to Figure l, note that the data for the Greenbacks begin in April 1862, shortly after they began to circulate and exhibit a positive gold premium from the outset. The additional discount of Greenbacks to 1861 Notes was very small in the beginning but rose to more than 40% as the amount of Greenbacks continued to rise and the 1861 Notes (after several months at their peak circulation) fell as they were withdrawn from circulation. The fact that 1861 Notes exhibited a positive gold premium requires a more detailed examination.
A. Accounting for the Gold Premium of the 1861 Notes.
While a systematic daily time series on the gold premium of 1861 Notes has not been developed until now, earlier researchers mention the premium for selected days. Mitchell had little to say about the gold premium merely observing that because the 1861 Notes were a good substitute for gold in paying customs, anything that drove the price of gold also drove their price. Calomiris' (1988a, 210, Footnote 19) explanation is directly to the point. In his words: "Because demand note supply was less than the sum of future tariff payments, the parity of gold and demand notes in payment of duties created a market parity between the two through arbitrage. The discounting of demand notes probably reflects the possibility of future change in tariff parity." In a later paper (Calomiris 1988b, 113), the conditional is dropped and the discount is said to reflect: "... the risk ... that tariff receivability would be discontinued." The public's expectations of revocation are influenced by fiscal news and "war news."
To summarize this approach: since the present value of future tariff liabilities exceeded the supply of 1861 Notes, the Notes should have traded at a par with gold. That gold traded at a premium implies that there was a positive perceived probability that their receivability on a par with coin for customs would be altered. Variations in this perceived probability caused the fluctuations in the gold premium. This approach has a number of weaknesses.
First, the notion that the supply of notes fell short of the present value of expected future tariff obligations seems to be based on an erroneous premise. The present value versus fixed supply analysis is based on the notion that the 1861 Notes could not be reissued once they were returned in payment of the tariff. This is stated most dearly in Calomiris (1992) where he contends that the laws authorizing the creation of Greenbacks were novel in three respects. The one relevant to this discussion is that, unlike earlier Treasury issues (the Treasury Notes of 1814 and the 1861 Notes), the Treasury had authority to reissue any Greenback it received.
His basis for this conclusion is not clear. It also seems to be contradicted by the data. During the first quarter of 1862, the amount of 1861 Notes outstanding increased to the full $60 million authorized by Congress for the first time. However, the treasury collected $15 million in tariffs over the same period, presumably most in the form of 1861 Notes. This would be impossible if 1861 Notes received for tariffs could not be put back into circulation. Likewise, Bayley's (1881, 154) data show virtually no redemptions prior to July 1862. (21)
Second, the notion of a significant and fluctuating perceived probability of changes in the customs receivability provision seems questionable. There is an odd silence in the newspapers and in the Congressional debates of the day that revocation was ever considered, seriously or otherwise, as a policy option. There is no mention of or speculation on such a policy shift in the newspapers. Congress had at least three opportunities (February 25 and July 11, 1862, and March 3, 1863) when it authorized Greenbacks to revoke the customs receivability privilege from the 1861 Notes and it did not do so. Not only did it not do so but also on March 17, 1862, Congress declared the notes to be a full private and public legal tender. If revocation were seriously considered, that would have been the time to make their public legal tender status on a par with Greenbacks, but Congress did not do so. (22)
Moreover, during the period the gold premium on the 1861 Notes reached a peak (August to September 1862 as shown in Figure 1) indicating, according to Calomiris, that the public strongly anticipated revocation, the Treasury was actively withdrawing them from circulation in accordance with the Greenback legislation. Between June 30 and September 30, 1862, nearly half the 1861 Notes were withdrawn from circulation by means of the same tariff provision, which was supposed to be in jeopardy.
Third, the present value versus fixed supply comparison seems beside the point, even if 1861 Notes could not be reissued once received for tariffs. A premium should be expected in any case for a much simpler reason: 1861 Notes could only be indirectly "converted" into gold via tariff payment over time, not instantaneously. A simple thought experiment illustrates the implications of this fact. Suppose an importer anticipates tariff payments at the rate of $1 per day for the indefinite future, which can be paid either with gold coin or with Notes. Furthermore, suppose the importer holds $182 in Notes and some gold coin and the real interest rate, r, is 6% per year. What discount relative to gold would induce the importer to hold one more dollar in Notes? A logical answer is 3%.
The logic behind this claim is that the extra note can be converted into gold at par, but only after waiting 6 months during which the previous $182 are exhausted in paying the tariff. If the discount on notes is greater than 3%, the importer can purchase the note and discharge the tariff obligation 183 days hence at par with gold, thus earning an above average rate of return over this period. If the initial holding of notes were $364, an extra note could not be indirectly converted into gold until a year had passed. In that case, the incentive would be to sell notes if the discount/premium was at any level less than 6% until note holdings were low enough to make it unprofitable to sell more.
If we extend this reasoning to the economy as a whole, and assume a well-functioning gold/note market, the equilibrium note discount/gold premium should equal rNIT, where N is the level of notes outstanding and T is the rate of tariff collection. At that discount/premium, those with above average note/tariff ratios should be selling notes to those with below average ratios, thus equalizing those ratios across money holders and achieving market equilibrium. This instantaneous equilibrium should apply regardless of whether the government retires notes as they are paid for tariffs or recirculates them.
Based on this theory, the gold premium for 1861 Notes should have been positive even if no possibility of revoking the customs receivability provision was foreseen. In addition, the premium should increase as the amount of 1861 Notes outstanding increases (as it did prior to March 1862) and decrease as they are withdrawn from circulation (as began to occur after July of the same year). Using this formula and the data in the first two columns of Table 1, one can obtain crude calculations of the gold premium on the 1861 Notes predicted by this theory. We are limited by the length of the period for which we have observations on the gold premium, p (January 1862 through May 1863), and frequency of observations on N (quarterly only). This leaves us only six dates from this period for which the outstanding number of 1861 Notes is known beginning with December 31, 1862, and ending with March 31, 1863.
For each date, we take the outstanding notes N and calculate the number of days it would take going forward from that date for the entire stock to be exhausted by customs payments (dividing the quarterly values for T evenly between each of the days in a given quarter). This provides the operational value for N/T (in days) as shown in Table 1. Again taking the 6's of 1861 as representative of the real interest rate, the value r = .06 is used to arrive at the predicted premium, rN/T, also shown in Table 1. Finally, a corresponding value for the actual gold premium is calculated by taking the average of seven daily quotes centered on the date in question. (23)
There is some correspondence between these crude predictions and the actual values. The predicted and actual values correspond closely for the beginning and end points and both tend to be higher in the middle of the sample. (24) However, the fluctuations in the two series are only loosely related (simple correlation = .49). (25) In particular, this model is unable to account for the low actual premium on March 31, 1862, and the high actual premium on December 31 of the same year. In addition, even if higher frequency observations on N were available, it is unlikely that they would exhibit sufficient high-frequency fluctuations to mimic the volatility of the daily observations on p (our values of T and r are smooth by construction).
The rough correspondence between the fluctuations of the predicted and actual premiums notwithstanding, the fact that the magnitudes roughly match is suggestive because this theory generates an interesting steady-state result regarding the seigniorage costs of the customs receivability provision. As shown in the Greenbacks section above, the Treasury loses pT per period (where T is the rate of tariff collection) by accepting Notes with discount p at a par with gold. Expressed as a percentage of the maintained amount of notes outstanding, N, the cost is given by pT/N. According to this "waiting time" theory, p T/N = (rN/T) T/N = r. Thus, equilibrium in the gold/note market leads to a premium that compensates note holders for the time they need to wait, on average, to "redeem" their notes at par at the going interest rate. A premium of this magnitude causes the Treasury to lose revenue at the same rate on each dollar of notes outstanding as it loses on each dollar of interest bearing debt outstanding. Unless the notes supply some transactions services that allow the premium to be lower (doubtful when cheaper Greenback dollars could be used instead), notes are as costly as bonds. In this regard, the revenue incentive to replace 1861 Notes with Greenbacks was considerable.
B. Accounting for the Gold Premium of the Greenbacks
While our primary purpose was to attempt to understand the level and fluctuations of the gold premium on 1861 Notes, our results suggest a modest contribution to the discussion of the determinants of the gold premium of the Greenbacks, at least for the period during which the competing currencies overlapped.
A controversial, but inconclusive, literature has grown up over the cause of the Greenback's gold premium and of the inflation process in general. In the quantity theory approach, typified by Friedman and Schwartz (1963), Friedman (1992), and Lerner (1956), the commodity inflation during the Civil War was primarily money driven. In the Cagan (1956) adaptation of this model, expectations of inflation can become an important variable driving changes in the velocity of money and the subsequent price path taken by commodities. Commodity price changes can, in turn, influence the market price of gold and the exchange rate in a flexible regime.
In a series of innovative papers, Calomiris (1988a, 1988b, 1988c) has challenged this view. He shares the conclusion of Mitchell (1903, 188) that "... fluctuations in the premium on gold were so much more rapid and violent than changes in the volume of the circulating medium ..." as to discredit the "expected inflation" theory and that this premium is determined by the expectations of the public as to the date of and rate at which the government will ultimately restore gold convertibility to its currency (this might be dubbed the "eventual convertibility" theory). These expectations, in turn, are importantly influenced by "news," especially "war news." Mitchell goes through a catalogue of financial, political, diplomatic, and battle news to explain nearly every movement in the gold premium during the war much as current news commentators today explain each daily change in the average of stock prices. (26)
A weak point of this model is that it is based on, to use Calomiris' phrase, a "vague promise of future convertibility" (Calomiris 1991, 80). When the Greenbacks were first issued, several members of Congress expressed positive sentiments regarding eventual convertibility. It was not, however, the official view of the government.
In none of the three bills authorizing emissions of Greenbacks did Congress set a date at which resumption of specie payments would occur or set a postresumption exchange rate. It was not until 1875, 10 yr after the war was over, that Congress declared that on January 1, 1879, specie payments would be resumed at the prewar parity.
Nevertheless, this "vague promise" is the driving force of the "eventual convertibility" model, and war news is held to be the primary force explaining the expectations of individuals regarding the date and rate at which the Greenbacks would be made convertible into coin. (27) Major Union defeats, signaling a prolonged war with higher fiscal costs, are said to give rise to expectations that push the date of convertibility into the more distant future and, perhaps, lead to a lower gold conversion rate. This raises the current gold premium and vice versa for a significant Union victory.
Subsequent research on this theory is inconclusive. All studies use the period of the Civil War, but the vagaries of the data affect the actual sample periods. McCandless (1996), using the gold price of both the Union and the Confederate dollar, supports the war news hypothesis. Williard, Guinnane, and Rosen (1996), using structural break analysis to distinguish "turning points" in the data from mere "blips," offer more conditional support. Several of their major break or turning points are not the same as McCandless' and some are unrelated to any news, battle or otherwise, and others occur at insignificant battles.
Weidenmier's (2002) support is more limited in the sense that he finds that significant Union victories or defeats are not always reflected symmetrically in the exchange rate of Confederate currency for gold (McCandless argues that they should be). (28) In fact, Weidenmier finds instances when after a battle, both currencies move in the same direction. Burdekin and Langdana (1993) find that the gold price of the Confederate dollar responded more to war events than financial policy shifts. The former, they argue, are more likely to affect redemption than the former. Smith and Smith (1997) employ an exchange rate regime switching model that enables them to cumulate the effect of battle news on the dollar/sterling exchange rate and conclude that forces other than war news must have dominated the movement in the exchange rate as their calculations suggest that during the war, "the net effect of news was a fall in the price of gold, whereas the price actually rose by more than 30 percent during the war." (29) Finally, Pecquet, Davis, and Kanago (2004), using Confederate dollar/Southern bank note exchange rates, provide evidence in support of war news in that "Northern victories typically led to an appreciation of bank notes against the Confederate dollar, while Southern victories caused bank notes to depreciate."
It is possible to link the fluctuating premiums in Figure 1 to war news during 1862. In particular, the largest premiums for the 1861 Notes occurred in July and later in early September when the amount of 1861 Notes outstanding was at or near its maximum and battlefield news was disquieting. On April 2, a large Union army under McClellan began an advance on Richmond that ended in disaster. Through superior generalship, Lee forced McClellan to abandon this second effort to take Richmond in July. On July 22, the gold premium on 1861 Notes reached a maximum of 10.2%. By August 14, McClellan began the withdrawal of federal troops from Virginia and their embarkation back north. Soon another federal army, led by Pope, was routed at Second Manassas on August 29-30. Lee quickly followed up these developments with his first invasion of the North.
He led a large Confederate army across the Potomac and, on September 7, captured Frederick, Maryland, some 30 miles north of Washington, DC. For the first time in the war, a large Confederate army threatened the federal capital from north of the Potomac. From its high on July 22, the gold premium on the 1861 Notes declined. During August 8-31, it ranged between 6% and 7%. It began to rise on September 1, reaching a second high of 9.3% on September 3, and remained in excess of 8.5% through September 11. By this time, the public appeared to realize that the size of Lee's army had been overestimated and that the army was unlikely to attack either Baltimore or Philadelphia. The premium had fallen between 5% and 6% by the onset of Antietam on September 15. Once Lee withdrew, the premium dropped even more, ranging between 3.5% and 4% on September 19 and 20. There are, of course, no analogous observations for the second Confederate invasion of the North in June 1863 that ended at Gettysburg.
As many others have noted, bad news was associated with greater gold premiums and good news was associated with smaller ones. But what is the mechanism? One implication is that Calomiris' theory of war news, which emphasizes fluctuations in expectations regarding the date of "eventual convertibility" cannot be applied to the gold premium of the 1861 Notes. By September, these notes were already being withdrawn and would never be convertible (except by the "indirect" mechanism that was already taking them out of circulation). Yet, the fluctuations in the gold premium of the 1861 Notes mimics the fluctuations of the gold premium of the Greenbacks in the same figure suggesting a common mechanism. This could be seen as indirect evidence against the eventual convertibility explanation of Greenback fluctuations as well.
However, if these fluctuations cannot be explained by changes in the expected date of a return to convertibility, what is the explanation for the comovement of war news and two gold premiums? In this regard, it is worth noting that since the 1970s, the United States has used flexible exchange rates and the gold price of the dollar in this regime has fluctuated frequently and widely, often in response to world turmoil. Yet, no one today argues that this reflects some market consensus that the United States may some day restore a fixed gold value to the dollar. Rather, gold is recognized as a vehicle for capital flight to safety and liquidity.
It is plausible that when major cities seem threatened with invasion and their inhabitants anticipate possible flight (as they did in July to September of 1862), an increase in the demand for gold coin occurs. At least we offer this simple explanation that fits the facts of this limited period.
The first months of the Civil War yielded much experimentation with the unfamiliar expedient of paper money. In particular, the federal government issued notes with a number of provisions that limited their eventual usefulness as a source of revenue. In piecemeal fashion, these provisions were removed and/ or the notes were replaced by notes to which these provisions did not apply.
We have attempted to illuminate this underexamined experiment involving the "1861 Notes." In particular, we emphasize the theoretical and empirical evidence that their seigniorage-limiting provisions were important in the decision to replace them with the seigniorage-friendly Greenbacks, a fact that seems to have been overlooked or, at least, underemphasized by previous research.
In addition, we have constructed a time series of daily market values for these notes that we report and make available for other researchers. These observations are also used to (a) estimate the (significant) seigniorage cost of one of these provisions, (b) evaluate a simple theory of the factors that determined the value of these notes, and (c) offer some evidence on existing theories of the factors that determined the value of their successors, the Greenbacks.
The Sources and Puzzles in the Data
The data are subject to a numeraire question. Prior to suspension, prices on financial markets were quoted in terms of "current funds in New York City," which meant the notes of the Associated Banks. These notes exchanged at par for the gold dollar and the 1861 Note dollar. In the days immediately following suspension, the quoted dollar price of gold rose. But which dollar? There is clear evidence that gold was at a premium in terms of the notes of the Associated Banks. For example, from the Times (December 31, 1861, 3): "The quotation for Gold over current funds to-day is variously quoted from 100 1/2 to 101 1/2%." But did the dollar price of gold rise in terms of 1861 Notes? The answer would appear to be yes, especially after a gold market opened in New York City on January 13, 1862. Evidence comes from several sources. On January 9, the Times (4) reported: "The United States Treasury Notes are generally received and paid out at Bank, as the standard of current funds." The same edition carried the price quotation for gold and silver coins from Thompson and Brothers, a major dealer in currency and publisher of a bank note directory, who specifically state for the first time that U.S. currency is the standard for its price quotes (the quotations carried in the Times' February 13 edition from the same firm are expressed in terms of "United States Demand Notes"). On January 14, the same paper reported (3): " ... United States Treasury Demand Notes pass current at par." The Herald (January 13, 1862, 5) reported: "... the obligations of the same government, bearing no interest.... Are current everywhere as money--that is to say at par" and (Herald February 14, 1862, 3) "When we say that gold is worth 104 1/4, we mean that Treasury notes have declined 4 1/4% below par." Thus, the 1861 Notes began to depreciate relative to gold on the day of suspension.
This raises an interesting puzzle about the data. Since 1861 Notes could continue to be used after suspension to pay customs, one might have expected that they would have appreciated relative to bank notes. This does not happen to be the case, at least during the first 4 or 5 months after suspension. The lack of a premium can be deduced indirectly from the fact that papers of the day drew special attention to instances when the 1861 Notes did not trade at par. On February 13, the Herald (3) reported that "banks are refusing to receive demand notes on deposit and they are selling on the street at 1/4-5116 discount," and the Times (February 28, 1862, 2) commented that, on February 26 and 27, the days immediately following President Lincoln's signing the bill authorizing the first $150 million in Greenbacks, 1861 Notes were at a premium relative to "the bank currency of the City," reported to be 118% to 1/4%. On March 2, the Times reported (8): "The United States notes are no longer held up at Bank for a premium and their return to the ordinary channel of currency is rapidly restoring ease to the money market." The March 29 Commercial Advertiser (3) reported on special circumstances related to the need for customs duties that sent the notes to a premium of from 1/4% to 112%.
Finally, the data suggest that for several months after the Greenbacks were introduced, they also passed at par for the 1861 Notes. We found a quote from the May 15 edition of the Herald (2): "The old backs, which are extremely rare, are scattered throughout the country. They are worth 518 to 3/4 premium." This suggests that not only was the customs use of these notes finally reflected in a premium over Greenbacks but also the Greenback dollar had become the numeraire.
Bayley, R. A. The National Loans of the United States from July 4, 1776, to June 30, 1880. 2nd ed. New York: Burt Franklin, 1881.
Bomberger, W., and G. Makinen. "Indexation, Inflationary Finance, and Hyperinflation: The 1945-1946 Hungarian Experience." Journal of Political Economy, 1980, 550-60.
--. "The Hungarian Hyperinflation and Stabilization of 1945-1946." Journal of Political Economy, 1983, 801-24.
Buiter, W. H. "Seigniorage." NBER Working Paper No. 12919, 2007.
Burdekin, R. C. K., and F. K. Langdana. "War Finance in the Southern Confederacy, 1861-1865." Explorations in Economic History, 1993, 352-76.
Burdekin, R. C. K., and M. Weidenmier. "Inflation is Always and Everywhere a Monetary Phenomenon: Richmond vs. Houston in 1864." American Economic Review, 2001, 1621-30.
Bureau of the Census, U.S. Department of Commerce. Historical Statistics of the United States, Colonial Times to 1970, 1975. Washington, DC: Bureau of the Census.
Cagan, P. "The Monetary Dynamics of Hyperinflation," in Studies in the Quantity Theory of Money, edited by M. Friedman. Chicago: University of Chicago Press, 1956, 3-117.
Calomiris, C. W. "Price and Exchange Rate Determination During the Greenback Suspension." Oxford Economic Papers, 40, 1988a, 189-220.
--. "Greenback Resumption and Silver Risk: The Economics and Politics of Monetary Regime Change in the United States, 1862-1990," in Monetary Regimes in Transition, edited by M. S. Bordo and F. Capie. Cambridge University Press, 1988b, 86-132.
--. "Institutional Failure, Monetary Scarcity, and the Depreciation of the Continental." Journal of Economic History, 1988c, 47-68.
--. "The Motives of U.S. Debt-Management Policy, 1790-1880: Efficient Discrimination and Time Consistency." Research in Economic History, 13, 1991, 67-105.
--. "Greenbacks," in The New Palgrave Dictionary of Money and Finance, edited by J. Eatwell, M. Milgate and P. Newman. New York: Macmillan Press Limited, 1992, 281-85. Commercial Advertiser. 1861.
Dewey, D. R. Financial History of the United States. New York: Longmans, Green, and Company, 1909.
Friedberg, R. Paper Money of the United States. 10th ed. Fort Lee, NJ: The Coin and Currency Institute, Inc., 1978.
Friedman, M. Money Mischief" Episodes in Monetary History. San Diego, CA: Harcourt Brace Jovanovich, 1992.
Friedman, M., and A. J. Schwartz. A Monetary History of the United States 1867-1960. National Bureau of Economic Research. Princeton, NJ: Princeton University Press, 1963.
Gilbert, J. H. "Trade and Currency in Early Oregon." Ph.D. dissertation, Columbia University, 1907.
Godfrey, J. M. Monetary Expansion in the Confederacy. New York: Arno Press, 1978.
Hammond, B. "The North's Empty Purse, 1861-1862." American Historical Review, 67(1), 1961, 1-18.
--. Sovereignty and an Empty Purse: Banks and Politics in the Civil War. Princeton, NJ: Princeton University Press, 1970.
Hunt's Merchants' Magazine and Commercial Review. New York: William B. Dana, Publisher and Proprietor, 1862.
Kagin, D. H. "Monetary Aspects of the Treasury Notes of the War of 1812." Journal of Economic History, 1984, 69-88.
Lerner, E. M. "Inflation in the Confederacy, 1861-65," in Studies in the Quantity Theory of Money, edited by M. Friedman. Chicago: University of Chicago Press, 1956, 163-75.
Lester, R. A. "Retention of the Gold Standard in California and Oregon During the Greenback Inflation," in Monetary Experiments: Early American and Scandinavian, edited by R. A. Lester. Princeton, NJ: Princeton University Press, 1939.
McCandless, G. T. Jr. "Money, Expectations, and the U.S. Civil War." American Economic Review, 86(3), 1996, 661-71.
Mitchell, W. C. "The Circulating Medium During the Civil War." Journal of Political Economy, 1902, 283-97.
--. A History of the Greenbacks. Chicago: University of Chicago Press (1903).
Moses, B. "Legal Tender Notes in California during the Civil War." Quarterly Journal of Economics, 1892, 1-25.
New York Herald. 1861.
New York Times. 1861.
Pecquet, G., G. Davis, and B. Kanago. "The Emancipation Proclamation, Confederate Expectations, and the Price of Southern Bank Notes." Southern Economic Journal, 2004, 616-30.
Smith, G. W., and R. T. Smith. "Greenback-Gold Returns and Expectations of Resumption, 1862-1879." Journal of Economic History, 1997, 698-717.
Studenski, P., and H. E. Krooss. Financial History of the United States. New York: McGraw-Hill, 1963.
U.S. Government. "Annual Reports of the Condition of the Banks." House of Representatives Executive Document No. 25, 37th Congress, 3rd Session, Washington, DC, 1862.
U.S. Treasury. "Annual Report of the Director of the Mint for the Fiscal Year Ending June 30, 1861." Washington, DC, 1861.
--. "Annual Report of the Secretary of the Treasury." Washington, DC, 1862.
--. "Annual Report of the Secretary of the Treasury." Washington, DC, 1863.
--. "Annual Report of the Secretary of the Treasury." Washington, DC, 1864.
Weidenmier, M. D. "Turning Points in the U.S. Civil War: Views from the Grayback Market." Southern Economic Journal, 68, 2002, 875-89.
Williard, K. L., T. Guinnane, and H. S. Rosen. "Turning Points in the Civil War: Views from the Greenback Market." American Economic Review, 86(4), 1996, 23-98.
(1.) The notes of the first and second banks of the United States were not government issues. The U.S. Treasury did issue several small denomination notes during the War of 1812, which may have functioned as money. Some bore interest, others did not. All were receivable for all payments due to the United States at par with coin, although they were not convertible into specie at par. Some may have circulated as money and have had monetary implications. Those most likely to have circulated as money, being of very small denominations, aggregated to $3.4 million out of a total public debt of nearly $119 million (by September 1815). See Kagin (1984) and Calomiris (1991, 71). In addition, the Treasury issued a limited amount of small denomination interest-bearing notes in both 1837 and 1845 that may have circulated as hand-to-hand currency. Unlike money, these notes had a limited life, that is, they were payable after 1 yr. While the United States maintained a bimetallic standard, the existing mint ratio undervalued silver and, as a result, full-bodied silver coins did not circulate. The gold coin that circulated was a full legal tender.
(2.) Data are from Studenski and Krooss (1963, 125).
(3.) Data taken from Report of the Secretary of the Treasury, December 1861, 30-32. The fiscal year at this time ran from July 1 to June 30.
(4.) While the notes were an emergency measure, their production time was long. During August to September, some $15.3 million were issued followed by $18.5 million during October to December. Secretary Chase quickly negotiated a loan of $150 million with the banks, which furnished an initial payment of $5 million before the 1861 Notes entered circulation. In addition, Chase was given authority to pay to willing contractors the 6% Treasury notes of 1861. These notes, like the 1861 Notes, were "customs privileged." During July to September, some $26.9 million were issued. However, the gap between federal expenditures and revenue grew rapidly as events disproved the wishful thinking of July that all would be over by Christmas.
(5.) The Historical Statistics of the United States for 1861 lists gold coin outstanding at $266,400,000 and subsidiary silver at $16,000,000 for a total of $282,400,000 (Bureau of Census, U.S. Department of Commerce [September 1975, 995]).
(6.) Estimates of deposits exist for the banks in the states loyal to the Union. Unfortunately, both demand and saving deposits are combined into one figure. Using data from Dewey (1909, 283) and Secretary Chase's estimate of the amount of bank notes in circulation, $130 million, total deposits in January 1862 would be about $330 million.
(7.) New York Herald August 18, 1861, 3. Unless otherwise specified, all quotations are from New York newspapers. Thus, the words "New York" will be deleted from the references below.
(8.) The Times opined: "The suggestion that these Notes are to be represented, at all times, by a corresponding amount of specie actually on hand ... and that the Secretary will only employ them to the extent that Specie is first exchanged for them is simply a mistake. The whole purpose of this Circulation is the mutual one of facilitating the disbursement of the Government and the accommodation and convenience of the public" (Times July 15, 1861, 3). All the quotations reproduced in this study are exactly as they appear in the newspapers even if the capitalization and punctuation do not correspond with current usage and standards.
(9.) On August 19, it was reported that the notes would be ready before the end of the "present week" (Times August 19, 1861, 8). On August 28, it was reported that the American Bank Note Company was sending notes to Washington at the rate of $300,000 per day (Times August 28, 1861, 8).
(10.) Hand-signed circulating notes were not uncommon. All the millions of notes issued by the Confederacy during the Civil War and most National Bank Notes did not used engraved signatures. We thank Richard Doty, curator of the money museum at the Smithsonian Institution, for this information. The first $20 million of the 1861 Notes delivered by the American Bank Note Company consisted of 2.8 million bills.
(11.) As early as September 5, a Times headline read, "Treasury Notes to be Used in Paying Soldiers" (1).
(12.) This may be an overstatement. According to Moses (1892), Gilbert (1907), and Lester (1939), 1861 Notes and Greenbacks did not circulate in California and Oregon, where gold remained the major circulating medium during the Civil War.
(13.) The government also needed coin to make payments of less than $5 (the smallest denomination of the 1861 Notes). Secretary Chase requested that $25 million of the second issue of Greenbacks be in smaller denominations, stating: "Payments to public creditors, as especially to soldiers, now require large amounts of coin to satisfy fractional demands less than five dollars." (Hunt's Merchants" Magazine and Commercial Review July 1862, 71). An earlier report of the Sub-Treasury reads "New York office sent to Washington today $500,000 in small Gold Coin to make change in payments to the army of the Potomac" (Times November 7, 1861, 3).
(14.) By using three newspapers, we were able to detect any printing errors.
(15.) The first $4 million in Greenbacks, all in $1,000 denominations, were received by the New York Sub-Treasury on April 5 (Herald April 6, 1862, 5).
(16.) On June 30, 1897, the Treasury listed $54,347 as unredeemed. (The total is odd given that the minimum denomination of these notes was $5. Similar oddities occur with respect to the sums outstanding in Table 1.)
(17.) With the introduction of the Greenbacks, two classes of United States Notes existed and circulated. They were the same size and color, were issued in denominations of $5, $10, and $20 (Greenbacks also came in larger denominations as well as in $1 and $2 denominations), and had nearly identical faces. However, Greenback Notes bore the Treasury seal, did not bear the words "payable on demand," and bore the date February 25 (the day they were signed into law) rather than August 10. The reverse side of the two notes was distinctly different. As to the name "Greenback," Friedberg (1978, 11) claims that it originated with the 1861 Notes. Although this is true, it should be qualified. Only the Herald (beginning in early June 1862) referred to them as "Greenback Treasury notes" or "Greenback demand notes." Neither the Times nor the Commercial Advertiser used that term. In both papers, they were referred to as "United States Demand Notes," "U.S. Treasury notes," "old demand notes" (or demand Treasury notes), "notes of the Custom house issue" (or custom house issue or custom notes or custom house demand notes), "U.S. Treasury notes of circulation," "U.S. Notes of the July issue," and so forth, while the notes now commonly known as Greenbacks were called by the financial press of the day, including the Herald, "legal tender notes" or "Legal tender demand notes."
(18.) Definitions of seigniorage are not uniform in the literature. Buiter (2007) examines the relationship between three related concepts. The use of AN/P is consistent with his use of the term. A competing concept is the yearly interest saving that occurs when an amount, N, of fiat currency displaces an equal amount of interest-bearing debt in private portfolios. Buiter labels this "central bank revenue," although, in the context of the period, it might be better described as "treasury interest reduction." As we go through our examples of alternative provisions and their effect on seigniorage, the reader will see that the results are robust to variations in the definition.
(19.) There are ample historical examples of political leaders more experienced with paper money who devised seigniorage limiting monetary provisions with unexpected and disastrous results. Bomberger and Makinen (1980, 1983) report that during the Hungarian hyperinflation of 1945-1946, the government introduced indexed deposits and indexed currency, whose effects were to return the seigniorage from money creation to the public resulting in the highest rate of hyperinflation ever recorded.
(20.) Secretary Chase clearly understood the seigniorage aspects of using money as a means of finance. In his Annual Report dated December 9, 1861, he explains the concept on pages 2, 11, and 12 without using the word itself.
(21.) There is more compelling evidence that the 1861 Notes were reissued. Like the Greenbacks, they were eligible for the special 5% certificates of deposit. In the pressing financial situation that prevailed in early 1862, the Treasury began to accept them for deposit. By the end of March 1862, before the Greenbacks began to circulate, nearly $18.5 million had been exchanged for certificates. These Notes were promptly spent by the Treasury. They were not borrowed to withhold them from circulation.
(22.) The only discussion we could find in the newspapers related to this subject was the speculation that Secretary Chase would call the notes and fund them in long-term debt so as to force the payment of customs duties in gold (Times June 9, 1862, 3).
(23.) For December 31, 1861, the first seven daily quotes in the following January were used.
(24.) The negative value for the actual premium reported for March 31, 1863, may seem puzzling. We were able to find 19 daily price quotations during March 1863, 8 during April, and 5 during May ending on May 16. March 5 is an outlier in that the 1861 Notes appear to be at a 6.5% premium relative to gold. This occurs because the price of gold fell sharply during the course of that day. The price for the notes was recorded during the morning board, while the gold sales were heavily concentrated during the afternoon boards during which the price break magnified generating the apparent premium on the notes. A much smaller but similarly generated negative gold premium for April 4 generates the p = -.009 for March 31, 1863.
(25.) The respective means (and standard deviations) of rT/N and p are .030 (.018) and .024 (.019) with a simple correlation of .49.
(26.) It is interesting to note that from Mitchell's data, one can conclude that the market accorded nearly the same importance to the resignation of an assistant treasurer of the United States in New York City as it did to the assassination of President Lincoln (the Greenback price of gold rose 0.8% for the resignation as opposed to 1.1% rise for the assassination).
(27.) Calomiris' (1988a) view is also influenced by the behavior of the Continental during the Revolutionary War (and America's colonial experience prior to the war). For some 2 yr, 1781-1783, after the last Continental was issued, they continued to depreciate in terms of gold. Obviously, he contends, something other than emissions of money was driving their continued depreciation.
(28.) An advantage of Weidenmier's work is that it is based on high-frequency price quotes for the Confederate dollar collected from Southern newspapers. McCandless' data were lower frequency and taken "from a newspaper clipping pasted onto a blank page of the Virginia State Library's copy of the History of the Confederate Treasury by Ernest Ashton Smith" (source of data unknown).
(29.) While not directly relevant to the Greenback/gold exchange rate, Burdekin and Weidenmier (2001) cleverly exploit data from the eastern and western portions of the Confederacy to argue that a major currency reform carried out in 1864 provides support for the quantity theory view that Confederate inflation was primarily a monetary phenomenon as suggested by Lerner (1956).
WILLIAM A. BOMBERGER and GAIL E. MAKINEN *
* We gratefully acknowledge the very helpful comments of two anonymous referees as well those of Tom Woodward, Bob Anderson, Marc Labonte, Kurt Schuler, Ted Murphy, and Richard Dory. The technical assistance provided by Mike Kolakowski is also gratefully appreciated.
Bomberger: Associate Professor, Department of Economics, University of Florida, Gainesville, FL 32611. Phone 1-353-392-0135, Fax 1-352-392-7860, E-mail firstname.lastname@example.org
Makinen: Adjunct Professor, Georgetown Public Policy Institute, Georgetown University, The Car Barn, 3520 Prospect Street, NW, Washington, DC 20007. Phone 1-703-525-7170, Fax 202-687-5544, E-mail g.makinen@ hotmail.com
TABLE 1 Amount of 1861 Notes Outstanding and Tariff Receipts Notes Tariff Average Quarterly Outstanding Receipts Premium Loss Date (N) (T) (P) (PT) September 30, 1861 15.6 7.2 November 30, 1861 21.2 December 31, 1861 34.1 8.3 March 31, 1862 60.0 15.0 0.026 0.392 June 30, 1862 60.0 18.9 0.026 0.493 September 30, 1862 32.3 23.0 0.069 1.581 December 31, 1862 16.2 13.4 0.039 0.520 March 31, 1863 5.4 15.4 0.014 0.215 June 30, 1863 3.8 17.2 0.000 0.002 September 30, 1863 2.6 22.6 December 31, 1863 1.6 27.4 March 31, 1864 1.2 June 30, 1864 0.9 Waiting Predicted Actual Period Premium Premium Date (NIT) (rNIT) (P) September 30, 1861 November 30, 1861 December 31, 1861 182 0.030 0.029 March 31, 1862 302 0.047 0.016 June 30, 1862 318 0.052 0.042 September 30, 1862 201 0.032 0.026 December 31, 1862 94 0.016 0.042 March 31, 1863 29 0.005 -0.009 June 30, 1863 September 30, 1863 December 31, 1863 March 31, 1864 June 30, 1864 Notes: Notes outstanding, tariff receipts, and quarterly loss i Tariff receipts during the quarter ending on the date in questi 1861, 62, 63, 64. Source: Bayley (1881, 153-54).
|Printer friendly Cite/link Email Feedback|
|Author:||Bomberger, William A.; Makinen, Gail E.|
|Date:||Oct 1, 2010|
|Previous Article:||Political business cycles in the New Keynesian model.|
|Next Article:||Monetary policy regime shifts: new evidence from time-varying interest rate rules.|