Central Banking Theory and Practice before the Federal Reserve Act

Modern central banking theory began to develop in the eighteenth and nineteenth centuries under the gold standard. Because of the dominant position of England in trade, finance, and economic theory, much of the development took place either at the Bank of England or in response to its actions. The designers of the Federal Reserve System accepted a theory of central banking and a framework for policy operations that reflected the prevailing practices of European central banks, particularly the Bank of England. More important, the developers of the Federal Reserve System in the 1920s imported many of their aims and much of their understanding from the pre–World War I Bank of England. The blending of these imported elements with practices or principles from United States experience created the broad framework that guided Federal Reserve policy operations at its start and for many years after.¹

It would be comforting to find in the history of central banking a record of steady progress and orderly development from earliest antecedents to present knowledge. The facts are different. The discussion reached a high point very near its start in the first decades of the nineteenth century. Thereafter, the level of discussion drifted lower. Some of the subtle points were lost and, more important, the focus of the discussion shifted.

At the start of the nineteenth century Henry Thornton, building on his own earlier work and pieces of analysis taken from Smith, Locke, and Hume, developed some guiding principles for the conduct of monetary policy from an analysis of the relation of money, economic activity, prices, and balance of payments under fixed and flexible exchange rates. This framework was lost between David Ricardo’s emphasis on long-run comparative statics and the concern of men of affairs with short-term fluctuations in market variables. After two nineteenth-century experiments with what they regarded as the essential principles of Ricardian monetary theory, bankers and men of affairs became skeptical about the applicability of economic theory to their problems. Early work was ignored or lost.

At their best, as in Walter Bagehot’s Lombard Street, the discussions by men of affairs of the principles by which monetary policy should be conducted reached a very high level. Strict adherence to these principles would have avoided some of the worst errors of monetary policy in later years. Nevertheless, neither Bagehot nor those who followed his lead attempted to combine the theory of central banking or monetary policy with what is now called macroeconomic theory, as Thornton had done. Until Wicksell, Fisher, Marshall, Hawtrey, and later Keynes and Friedman reopened the discussion, very little was done to extend Thornton’s analysis or to develop an alternative framework connecting monetary policy to output, employment, prices, and balance of payments. Monetary policy—or bank rate policy as it came to be known in England—was assigned the task of regulating the gold flow.²

Why did Thornton’s rich and promising analysis degenerate first into a Bank of England policy of using bank rate mainly to protect the gold reserve and later into the Federal Reserve’s concern for short-term market interest rates and money market conditions? Three reasons appear to be important. First is the “automatic” gold standard. The gold standard gave monetary policy a clear and definite objective. Writers such as William Jevons and Alfred Marshall wanted to make improvements in the standard to eliminate or reduce procyclicality of money, but they paid little attention to implementation. Second, much of Thornton’s analysis considers an economy with an inconvertible currency. After the return to the gold standard, the continued relevance of other parts of his work was overlooked. Third, Ricardo and many of his followers not only failed to address the questions uppermost in the minds of the practitioners but failed to make clear that they were not addressing these questions. Ricardo’s analysis is almost entirely long-run comparative statics, and his policy recommendations consisted mainly of a set of rules for restoring and maintaining convertibility of pounds into gold at a fixed exchange rate. Important as is his work for economic theory, it gave very little guidance to the Bank of England on the issues of greatest concern to its directors. The governors and directors of the Bank of England were concerned with the profits of the bank, the avoidance of panic, and the appropriate response to short-term changes—for example, an increased demand for borrowing by the country banks or from banks abroad.

The wide gap between monetary theory and the practice of monetary policy, familiar to observers of the contemporary discussion of policy, had opened by the 1830s. The most able economists of the period participated in the discussion, and though they focused mainly on the longer-run consequences of policy actions and ignored short-term effects, they did not hesitate to recommend policy actions. Those bankers and economists whose writings show greatest interest in and knowledge of short-term operations and practices neglected, for the most part, the longer-run consequences of the policies and procedures they espoused. They tended to concentrate on the initial effects of policy actions and to ignore the longer-term consequences.³

In the history of economic thought, the participants in these discussions are grouped into schools known as bullionists and antibullionists for the first quarter of the nineteenth century and into currency and banking schools for the second quarter and into the third. While the groupings may be useful for certain purposes, they suggest more direct confrontation of ideas than appears to have taken place or than could have taken place given that one side was concerned much more with ultimate effects, the other mainly with initial effects.⁴ Indeed, the “disputants” most often failed to agree on the subject under discussion or even to mention whether they were concerned with short- or long-run consequences. One main result is that the link between short- and long-run effects of policy remained unanalyzed (Viner 1965, 139–40).

A second reason for the decline in the level of the discussion is related to the first. Throughout monetary history, the belief recurs that monetary policy has very limited effects on employment, expenditure, and output. Lacking an explicit theory of the transmission of policy changes, it was easy for the men who guided the Bank of England to mistake initial effects of a change in bank rate for the ultimate effect. Many of the writers in the socalled banking school, and many others in later generations, contributed to this belief by equating the effect of monetary policy with the change in the supply of “funds” in the money market. No reader of the discussions or interpretative accounts of nineteenth-century (or twentieth-century) monetary theory and policy can fail to be impressed by the frequency with which the idea reappears that any effect of monetary policy on the real economy is adventitious, the result of a particular and special conjuncture of forces that was either unlikely to be repeated or unlikely in the future to spread the effect far beyond the money market. Or if monetary actions had short-term consequences for the real economy, the effects were limited to specific sectors. Arguments about the “ineffectiveness” or noneffectiveness of monetary policy on the real economy became the official view of the working of policy.⁵

Ricardo’s dominant position and his failure to build on Thornton’s analysis of the ways in which the effects of monetary policy spread from the money market to economic activity, prices, and balance of payments meant that most of Thornton’s analysis was neglected. A century after Thornton’s promising start on a theory of money, his analysis leading to a statement of the principles by which monetary policy should be conducted to stabilize the economy had degenerated into the three main rules or principles for setting bank rate. These rules were accepted as basic at the start of the Federal Reserve System. First, there was the core principle of the gold standard: the central bank must raise or lower the discount rate as required to protect the gold stock and the exchange rate. Second, the central bank served as lender of last resort by offering to lend in a panic when markets did not function. Third, the central bank was to accommodate the needs of trade and agriculture by discounting only (or mainly) commercial paper, a principle known as the productive credit or real bills doctrine. This principle prevented purchases of government securities, mortgages, other long-term debt and the use of these instruments or equities as collateral for borrowing from the central bank.

The details of doctrinal history are less important than their consequences for the theory and practice of central banking. A number of excellent summaries of the literature of the period are available: Bagehot 1962, Clapham 1945, Hawtrey 1962, Keynes 1930, Rist 1940, Sayers 1957, Schumpeter 1955, Thornton 1965, Viner 1965, and Wood 1939. Since many of the issues that arose, and their solutions, reflect the economic events of the period, the chapter begins with a description of the background events. The rest discusses three major contributions to monetary and central banking theory that were ignored, at great cost, during most of the twentieth century. First is Henry Thornton’s analysis of the control of money and credit under either a fluctuating or a fixed exchange rate. Second is Walter Bagehot’s discussion of the responsibility of the central bank as lender of last resort. Third is Irving Fisher’s distinction between real and nominal interest rates. Thornton’s work was not well known. Bagehot’s work was well known at central banks, and Fisher was active until the middle of the twentieth century. Yet none of the three had a major influence on the conduct of policy. If they had, monetary history would have been much different.

The main issues in dispute during the period are familiar to contemporary economists. Can the monetary system be controlled? If so, which variables should be controlled, and how should this be done? What are the consequences of alternative systems of control? Did the central bank have an opportunity to exercise discretion, or is the real stock of money constant, so that central bank policy ultimately determined only the division of the real stock of money between gold (foreign exchange) or specie and paper? Should the central bank protect its own reserve, or is its main responsibility to protect the financial system in time of crisis? How could either or both of these ends be achieved? The answers to these and other questions given by central bankers and economists reveal the way the theory of central banking developed in the nineteenth century and the state of the art in the early twentieth century when the Federal Reserve was founded.

BACKGROUND EVENTS AND ARRANGEMENTS

During most of the eighteenth century the main policy actions involved the choice of standards, the establishment of de facto or de jure rules, and the provision of currency. By the end of the century England was on a de facto bimetallic standard at a ratio that undervalued silver, so full-weight silver coins did not circulate, and the currency consisted of gold, underweight silver coins of small denomination, and note issues of the Bank of England and other banks.⁶ Bills of exchange had come into use, and in some areas endorsed bills served as a medium of exchange. Usury laws restricted interest rates, including the rate on advances from the Bank of England (bank rate), to a 5 percent maximum. Bank rate for inland bills was put at 5 percent in 1746. The rate on foreign bills rose from 4 to 5 percent in 1773 and did not change again until 1822.

In 1697 the Bank of England was granted a charter to operate as a joint stock bank, but until 1826 other banks could not have more than six partners. The restriction of joint stock banking meant that partners were required to pledge their personal fortunes in periods of crisis. Consequently there were numerous small individually owned banks and very few branch banks.⁷

The main business of a banker consisted in issuing notes and discounting bills of exchange.⁸ Since a large number of country banks accepted bills and issued banknotes, and since many of the bills were drawn by small local merchants in payment for merchandise, it became common for country bankers to develop correspondent relationships with London bankers to provide information and clearing arrangements. This tendency strengthened because the supply of bills of exchange did not grow at a uniform rate throughout the country. London became the financial center through which deficit areas were able to sell bills to surplus areas. Country bankers held deposits with their London correspondents and purchased or sold bills. The continued increase in the number of country banks during the early nineteenth century made it increasingly difficult for London bankers to clear the bill market by operations between correspondents. A new institution, the bill broker, arose in London to perform part of the market clearing function.

The growing importance of the bill broker resulted also from the arrangements prevailing at the time. First, the usury laws prevented the London banks from changing rates to attract a larger volume of Bank of England notes and gold from country banks and to reduce the supply of bills. Second, the Bank of England adopted rules designed to reduce the demand for discounts from country bankers. To be eligible for discount at the Bank of England, bills had to be endorsed by two London names, one of which was the merchant or manufacturer accepting the bill. Many of the bills originating in the country did not meet this requirement.

As the system functioned at the start of the 1790s, the country banks maintained deposits and bought or sold bills from correspondent bankers in London. Bill brokers operated in the market in much the same way that a federal funds broker operates in the present New York money market. When the quantity of money demanded in London (bills supplied to London) exceeded the quantity of money supplied, bill brokers searched for buyers in the sections known to have surplus reserves. Just as the presentday federal funds market redistributes reserves from surplus to deficit banks, the bill brokers and correspondent banking system of the time drew bills and money to and through the London money market. The Bank of England participated in the market process as a banker. In addition, the bank absorbed gold and its own note issues as the market required and, without formally committing itself to do so, functioned as lender of last resort by advancing to banks on eligible paper.⁹

The system had an obvious flaw. With the rate of discount set at the maximum permitted under the usury law of 1714, the bank could not keep the market price of gold equal to the mint price of gold, maintain convertibility, and discount all of the eligible paper offered in periods of expansion. The reason is that the bank had only one means, and that a very ineffective means, of limiting or reducing the rate of monetary expansion: using qualitative controls or eligibility requirements to reduce the amount of discounts. After 1793 the government chose to finance the budget deficit incurred to wage the Napoleonic Wars by borrowing from the bank, so the bank’s notes and deposits increased. The monetary expansion and deficit spending generated an increase in private expenditure. Under the prevailing payments system, this meant an increase in the number of bills of exchange drawn, including bills eligible for discount at the bank.

From 1790 to 1795 the bank saw total securities (private and public) rise from approximately £8 million to £16 million and bullion reserves decline from £8 million to £4 million. The price index (base 100 in 1821–25) started to rise in 1793. Between 1792 and 1795, prices increased by 30 percent, a 9.9 percent compound annual rate of increase (Gayer, Rostow, and Schwartz 1978). To stem the gold outflow, the bank attempted to reduce the size, or perhaps the portfolio’s growth rate, by restricting the banks’ right to discount. Any step of this kind, however, raised fears that the resources of the financial system would prove inadequate to redeem outstanding bills at the fixed rate of interest. The policy of controlling the quantity of discounts by rationing, exhortation, and eligibility requirements failed on this occasion, as on many subsequent occasions. With the gold reserve reduced to less than £1.5 million, the bank asked the government to order an end to convertibility. From 1797 to 1821, the pound was an inconvertible currency.

LESSONS FROM RESTRICTION AND RESUMPTION

The events of the next twenty-five years and the analysis they engendered make the period known as Restriction and Resumption a remarkable epoch in the histories of money, monetary theory, and of particular interest here, the theory and practice of central banking. A key figure in the early discussion is Henry Thornton, whose contributions to monetary and banking theory reveal an understanding of monetary process and policy that is far better than can be found in much of the professional writing a century or more after his death.

Thornton’s contributions fall into five main areas. First, he provided a thorough discussion of the way the monetary arrangements of his day worked in practice and discussed some of the main implications of alternative arrangements and alternative monetary standards, including an inconvertible paper currency. He recognized that money produced by the banking system, paper credit, was part of the (circulating medium) means of payment. The effect of bank deposits on prices was the same as an increase in currency or gold. Second, he analyzed the monetary aspects of international exchange. David Hume had developed the basic flow analysis of monetary changes acting on home prices relative to foreign prices, thus on gold flows. Thornton for the first time used this analysis to explain the effects of actual price changes on international currency movements and the domestic economy. His discussion of currency movements is superior to the work on the same subject for the next century. Third, he saw clearly the difference between nominal and real interest rates, distinguished expected from actual rates, and offered an explanation of the rise in interest rates during an unanticipated inflation that is superior to the discussion in many later textbooks.¹⁰ In his testimony of 1797 and in his book, he attacked both the usury law and a fixed rate of discount on the grounds that by fixing the rate the bank relinquished control of money (the circulation). He argued that the discount rate had to be changed to raise or lower the cost of borrowing when the (anticipated) rate of return to real assets changed (Thornton 1965, 253–54). Thornton saw that the absolute level of the rate was not a proper criterion. The nominal rate had to be judged relative to the nominal rate of profit or, in modern usage, the return to capital. Fourth, his contributions do not appear as vague suggestions or dimly perceived truths occurring in the midst of an otherwise flawed argument. They are part of a carefully articulated explanation of the relation of money, prices, output, interest rates, credit, and balance of payments.

An important key to Thornton’s analysis is the distinction between the demand for bank credit (indebtedness to the banking system) and the demand for money. At the start of an (unanticipated) inflation, the demand for bank credit and the demand for money move in opposite directions. These movements reflect a common cause, changed anticipations of the return to real assets and the rate of change of prices. Increases in the stock of money, resulting from an issue of paper or a gold inflow, increase the demand for goods, raising the prices of the goods demanded and encouraging borrowing by businessmen, whose sales and prospective profits rise. Velocity increases—the demand for money falls—not only because (some) businessmen are for a time more optimistic and velocity depends on “confidence” (ibid., 96), and thus on anticipations of the future, but also because inventories decline (237). These are short-term cyclical effects, but for a time they persist and generate additional increases in the demand for credit and in velocity. One reason the demands persist is that not all prices adjust at the same rate. Some are fixed by contract in nominal terms and rise or fall more slowly than others. Thornton used money wages as an example of a price that was fixed in nominal amount and argued that, as a result, real profits rise and real wages fall in periods of (unanticipated) inflation. Once real balances adjust to their desired level, total wealth is “nearly the same,” but there has been a once and for all redistribution from workers and other creditors to debtors (189–90).¹¹

Thornton saw that short-term monetary disturbances had no lasting real effect. Money is neutral in the long run. One of his main reasons for short-run nonneutrality is that it is difficult to distinguish between permanent and transitory changes when they occur.¹²

In contrast to Ricardo, Thornton argued that replacing a convertible currency with inconvertible paper causes the market price of gold to rise above the mint price even if the nominal amount of paper money remains unchanged. His reasoning is that if money holders anticipate a decline in the purchasing power of money, they attempt to shift out of money. This argument makes the demand for money and short-run price changes depend on the anticipated rate of price change.¹³

Thornton’s fifth contribution to the theory of central banking is a part of his theory of money and in this respect also stands in marked contrast to much of the literature on monetary theory and policy that followed. By combining short-run and long-run adjustment, he was able to deal with issues that Ricardo neglected or dismissed. In all important respects, his analysis of the long-run consequences fully anticipated Ricardo’s.¹⁴

Neglect of Thornton’s work and reliance on Ricardo’s meant that the directors of the Bank of England, after periodically facing large changes in gold stocks and several threats to convertibility at the fixed exchange rate, concluded that Ricardian theory was inapplicable or useless. They therefore abandoned monetary theory as a basis for monetary policy and substituted ad hoc notions about money markets. These notions are at their best in the brilliant essays of Walter Bagehot and perhaps at their worst in the writings of central bankers during and after the depression of the 1930s. But either at their best or at their worst, the principles and practices of monetary policy became divorced from any analysis of the mechanism linking changes in money to short- and long-run changes in output and employment.

As part of his development of the price-specie flow mechanism, Thornton analyzed the effects of price changes on the gold stock of the Bank of England under convertible and inconvertible paper standards. Although he recognized the short-term effects of anticipations on the demand for money, he placed responsibility on the Bank of England for long-run inflation and any permanent divergence of the market from the mint price of gold. In a lengthy discussion of the relation of the country banks to the Bank of England, he argued persuasively that the expansion of country banknotes depended on expansion by the Bank of England and that the expansion of money was a necessary condition for inflation. But unlike the currency school, he emphasized that neither the total stock of notes in circulation nor the price level rose and fell in direct proportion to the note issue of the Bank of England (Thornton 1965, 219–29). If some resources are idle, the Bank of England can increase their employment by increasing the note issue, but “the increase of industry will by no means keep pace with the augmentation of paper,” and inflation results (239).

The three duties of the Bank of England were to protect the gold reserve, function as lender of last resort, and control the note issue. These duties were best performed, Thornton thought, by keeping the market price of gold equal to the mint price, limiting the note issue by discount rate policy, except in periods of crisis when the bank must expand the note issue and lend more freely. Any attempt to limit the note issue by rules controlling the quality of credit as proposed in the real bills doctrine was to lend “countenance to the error …of imagining that a proper limitation of bank notes may be sufficiently secured by attending merely to the nature of the security for which they are given” (ibid., 244 and elsewhere in chap. 10).¹⁵ The appropriate policy for the bank was to change the discount rate so as to control the quantity of money. In Thornton’s words, the policy should be

to limit the total amount of paper issued, and to resort for this purpose, whenever the temptation to borrow is strong, to some effectual principle of restriction; in no case, however, materially to diminish the sum in circulation, but to let it vibrate only within certain limits; to allow a slow and cautious extension of it, as the general trade of the kingdom enlarges itself; to allow of some special, though temporary, increase in the event of an extraordinary alarm or difficulty, as the best means of preventing a great demand at home for guineas; and to lean on the side of diminution, in the case of gold going abroad, and of the general exchanges continuing long unfavorable; this seems to be the true policy of the directors of an institution circumstanced like that of the Bank of England. To suffer the solicitations of the merchants, or the wishes of government, to determine the measure of the bank issues, is unquestionably to adopt a very false principle of conduct. (259; italics added)¹⁶

The bank did not accept Thornton’s prescriptions. From 1797 to 1815, the securities portfolio of the Bank of England increased threefold, and in the next seven years it declined as much as it had risen in the previous fifteen.¹⁷ At the end of the period as at the beginning, the Bank of England’s portfolio was approximately £15 million.¹⁸ Most of the increase in money during 1800 to 1810 resulted from the increase of commercial paper at the Bank of England. At its maximum of approximately £23 million in 1810, the bank’s portfolio of “private securities” was larger than at any time in the next hundred years. From 1810 to 1815, by far the largest part of the increase in money resulted from the bank’s acquisition of government securities to finance wars with France and the United States. The Gayer, Rostow, and Schwartz (1978) price index shows that the price level (December 1790 = 100) follows a similar course. Prices rose to 198 in May 1813, then fell to 94 in mid-1822.

There are several important developments for central banking theory and practice during the period.¹⁹ One is that bankers found, as Thornton had insisted, that the Bank of England had no effective means of limiting its portfolio and the rate of monetary expansion in a period of inflation. With prices doubling from 1790 to 1812, the average rate of inflation is 5 percent, but at times it was much higher.²⁰ Since the usury law fixed the discount rate at 5 percent, the realized cost of borrowing was zero on average. There was no previous experience with managing a paper currency and, even neglecting the usury law, no tradition of limiting the volume of discounts and allowing the market to determine the rate on bills of exchange. The bank, in an early application of the “real bills” approach, attempted to control the quantity by controlling the “quality” and restricted commercial discounts to short and “sound” bills. The policy failed on this occasion, as on many future occasions.

At the end of the Napoleonic Wars, the government deficit declined from £35 million in 1814 and 1815 to £2 million in 1817. After 1816 the Treasury retired debt, and the bank’s holdings of public securities declined. The monetary base measured as the sum of gold and total securities appears to have fallen after 1815. At first the bank’s holdings of private securities continued to rise, presumably because the anticipated cost of borrowing remained far below the anticipated rate of return on real assets after fifteen years of inflation. But the anticipations probably changed quickly. From mid-1814 to late 1815, the price index fell about one-third, and the bank’s holdings of private securities dropped to the low levels of the early 1790s. By March 1816 the Treasury was able to issue exchequer bills below the 5 percent usury rate. In the severe postwar deflation, the bank accumulated gold and lost earning assets. Thus it came to recognize a second problem of monetary management, a problem that was in fact the mirror image of the first. As long as the discount rate remained above the market rate, the bank could not take action to expand its portfolio. It eventually resolved this problem. Under pressure from the prime minister, it lowered the discount rate to 4 percent in 1822, the first change in fifty years. For a time the bank’s holdings of private securities, the note issue, and the bank’s deposit liabilities expanded.²¹

The third main problem of monetary management that the bank faced during these years was the restoration of convertibility, or Resumption. As early as 1810, the Committee on the High Price of Bullion, under the influence of Thornton, who was a member of the committee, and of Ricardo, who was not, urged a resumption of cash payments at the price of gold that had prevailed in 1797. Since prices had increased, the market price of specie was above the former mint price. To resume specie payments at the old mint price, the Bank of England had to engineer a deflation. On May 1, 1821, Britain returned to the gold standard at the historical mint price.

A century later, the Bank of England faced a similar problem and made a similar decision. Both decisions were followed within a few years by severe and prolonged depressions. The decision to resume specie payments (1819) allowed for a four-year delay and came after a deflationary policy had been in effect for six years. After the decision to resume cash payments was announced, gold flowed into the Bank of England. Much of the gold inflow occurred because the deflationary policy had pushed the price of silver and the exchange rate for the paper pound close to the mint price.²² Between 1821 and 1824, the bank’s holding of gold never fell below £10 million.

The years from 1820–24 are one of the more interesting episodes in early monetary history. The Bank of England’s holdings of bullion tripled in the brief span of seventeen months and reached £12 million, the highest level attained to that time. Prices continued to fall until 1822, then rose, on average, 8 percent a year for the next three years. Part of the rise was the result of an agreement between the bank and the Treasury calling for the bank to advance £13 million to the Treasury in exchange for a forty-four-year annuity, known as the Dead Weight debt. These and other special advances combined with the gold inflow to increase the bank’s deposit liabilities and note issue. Private borrowing expanded, and with bank rate reduced to 4 percent in 1822, the bank acquired bills and issued money.

Throughout the period, the government ran an almost constant budget surplus of £3 million to £4 million per year and used the surplus to retire debt. The net effect of the Treasury’s debt retirements and the special issues to the Bank of England was the same as would have occurred had the bank engaged in open market purchases. The expansive effect of the open market operation on the monetary base and the economy was not entirely unexpected. The prime minister, Lord Liverpool, informed the bank in 1822 that he wanted to increase the circulation, and it is likely that the bank’s purchases of Dead Weight debt were part of a plan to expand the stock of money and slow or stop the fall in (agricultural) prices. There is additional evidence that the idea of using open market operations to expand the note or monetary liabilities of the bank was understood, though the term open market operation was not used. At about the same time the bank purchased exchequer bills in the market at the request of the Treasury.²³

Judging from the increase in the bank’s holdings of private securities from 1822 to 1824 and particularly in the latter year, the economy expanded. The data are not sufficiently accurate to conclude that the expansion of the economy and the return of inflation can be attributed solely to the rise in the monetary base and the reduced discount rate at the Bank of England. However, the timing and direction of changes are consistent with the hypothesis that the deflationary policy before 1820 (1) induced a subsequent inflow of gold that increased spending, (2) thereby raising realized returns in agriculture and trade above previously anticipated rates, (3) stimulating additional borrowing, and (4) resulting in a further expansion of the monetary base. The reduction in bank rate and the open market purchases added to this process by increasing the growth rate of the base.

Ricardo had urged that paper money be kept in circulation. He recommended that bullion be held in ingots at the bank and the mint or held by private owners when demanded. The directors of the bank preferred bullion to paper for coins and notes of small denomination. English notes of less than five pounds, issued during the Restriction, were withdrawn from circulation after the Resumption. The effect was to raise the demand for gold in England and increase both the resource cost of maintaining the money stock and the rate of deflation required to restore convertibility at the previous fixed rate. By the winter of 1823–24, the bank’s gold stock reached a maximum and started to decline. The decline continued throughout 1824, accelerated in 1825, and reached a trough early in 1826. At the trough, the bank held only £2 million pounds after suffering a drain of £12 million in two years.

To stem the decline and protect the gold reserve, the bank refused to discount eligible paper, but it did not at first raise bank rate. Hawtrey (1962, 14–15) argues that raising the rate would not have been effective because short-term interest rates rose above 70 percent per annum. He overlooks the fact that before this occurred the crisis had intensified for several months and had become a panic after the bank restricted its loans. Bagehot, in a graphic passage, describes the money market in December 1825.

In the panic of 1825, the Bank of England at first acted as unwisely as it was possible to act. By every means it tried to restrict its advances. The reserve being very small, it endeavored to protect the reserve by lending as little as possible. The result was a period of frantic and almost inconceivable violence; scarcely anyone knew whom to trust; credit was almost suspended; the country was, as Mr. Huskisson expressed it, within twenty-four hours of a state of barter. Applications for assistance were made to the Government, but …the Government refused to act. (1962, 98)²⁴

In previous crises, such as 1793 and 1811, the government had issued exchequer bills to the merchants. Sir Robert Peel believed that issuing bills would help only if the bank agreed to purchase them from the market. Since “intervention would be chiefly useful by the effect which it would have in increasing the circulating medium, we [Peel] advised the Bank to take the whole affair into their own hands at once, to issue their notes on the security of goods, instead of issuing them on Exchequer Bills, such bills being themselves issued on that security” (ibid., 99). With the government’s guarantee in hand, the bank raised the discount rate to 5 percent and resumed lending. Bagehot describes the turnaround:²⁵

“We lent it,” said Mr. Harman, on behalf of the Bank of England, “by every possible means and in modes we had never adopted before; we took in stock on security, we purchased Exchequer bills, we made advances on Exchequer bills, we not only discounted outright, but we made advances on the deposit of bills of exchange to an immense amount, in short, by every possible means consistent with the safety of the Bank, and we were not on some occasions over-nice. Seeing the dreadful state in which the public were, we rendered every assistance in our power.” After a day or two of this treatment, the entire panic subsided, and the “City” was quite calm. (Ibid., 25)

The crisis, coming at the end of a period of alternating inflation, deflation, expansion, and depression that characterized the first quarter of the century, provided the impetus for an examination of monetary arrangements and produced some major changes in banking and central banking practices. Repeal of the usury law (1833) permitted the bank to raise the discount rate on short-term bills above the 5 percent limit, a very important step for the future development of central bank policy. Other changes made to improve the functioning of the banking system included the opening of branches of the Bank of England, the extension of the joint stock form of organization to commercial banks, and the granting of legal tender status to Bank of England notes. The latter changes show the tendency of governments (repeated on many subsequent occasions) to adopt new arrangements after a crisis even if there is little reason to believe that the previous arrangement was a major contributing cause of the crisis.

As is often the case, changes in informal arrangements were far more important than the new laws. Of particular interest here are the changes in central banking practices, since they reveal the attitudes and understanding of the directors. But there were also important changes in the practices of bankers and bill brokers.

The Bank of England was forced to accept, or at least to share, the responsibility for maintaining the payments mechanism and to function as lender of last resort to the economy. The bank did not publicly acknowledge the responsibility, and during the crisis the government had been forced to prod the bank and offer guarantees. A tradition was established thereby, and the bank was able to demand and get similar guarantees in later crises. The bank’s caution was partly a consequence of private ownership, as was alleged at the time, and perhaps partly of bureaucratic slowness resulting from its monopoly position, but it was partly lack of understanding of the responsibility of a central bank to serve as lender of last resort.

The basic cause of the crisis, however, was the bank’s inability or failure to slow the rate of monetary expansion in 1823–24. Many of the bank’s directors believed that the expansion had been partly the result of “speculation” and the panic a result of “overspeculation.” Influenced by the real bills doctrine, some directors attributed the start of speculation to the bank’s purchase of government securities, that is, the purchases of Dead Weight debt and other issues. But most of the directors recognized that the problem arose because of the expansion of the circulation or, as some of them put it, the reduction in interest rates.²⁶

The experience led many of the directors to conclude that the bank had been too ambitious when it agreed in 1822 to assist the government in a policy of reversing the price level decline. The bullion reserve had been at one of the highest points in the bank’s history when the policy started, but even so large a reserve had proved insufficient to satisfy the demand for bullion during the crisis. The experience seemed to support the extreme bullionist view that the combined circulation of gold and paper currency had to be kept equal to the amount of gold that would otherwise have circulated. Any excess would raise the price level in proportion to the excess issue, causing a fall in the exchange rate and a loss of bullion. This point was well known because Ricardo had stressed it in his writings and testimony, and it had become a main point of emphasis for the writers in what came to be known as the currency school. By accepting this point from Ricardo and the members of the currency school, the directors in effect rejected Thornton’s earlier stress on the effect of business conditions on confidence and of confidence on the demand for money. Neglect of Thornton’s promising start on an analysis that combined short- and long-run consequences of a change in gold or money closed off one of the few opportunities in a century to develop a general equilibrium analysis of money, bank credit, output, prices, and balance of payments.

The wave of bank failures profoundly affected London bankers and bill brokers, just as a similar experience was to affect their American counterparts a century later. Innovation and changes in practices followed. The risks inherent in the previous practice of holding very low ratios of reserves to deposits and relying on the sale of bills, or in some instances advances from the Bank of England, had proved larger than anticipated. The banks were partnerships at the time, so failure often meant the loss of a personal fortune. For a time the banks increased their reserve ratios by increasing their holding of bullion and deposits at the Bank of England relative to their deposits. They no longer relied on the purchase and sale of bills of exchange to adjust portfolios but held bills to maturity and adjusted portfolios by making or calling loans to bill brokers. As the banks withdrew from the bill market, some of the larger brokers accepted many of the functions the banks had performed. They bought and sold bills from their portfolios instead of acting as brokers (Scammell 1968, 133). Others continued a brokerage operation as before the panic.

Just as the United States banks in the 1930s virtually stopped all borrowing from the Federal Reserve, after 1825 English banks no longer relied on advances from the Bank of England. To increase cash, banks reduced call loans to the bill brokers. After 1830 the brokers were allowed to discount at the Bank of England, and they did so when the banks reduced call loans.

Some may find in these developments the origin of a “tradition against borrowing.” Wood (1939, 90–104) points out that the facts do not support that hypothesis. Generally the bank’s discount rate was a penalty rate, above the rate on bills of highest quality, the only type eligible for discount. London banks did not borrow even in periods of crisis but relied on call loans to adjust their cash position. When the banks’ demand for cash assets increased, the country banks sent more bills to the bill brokers and surplus areas purchased fewer. London banks reduced call loans, and the bill brokers borrowed from the Bank of England. The so-called tradition against borrowing by banks should be seen, therefore, as a tradition of borrowing by the bill brokers and dealers who supplied reserves to the banks.

Furthermore, there were other ways the bank’s rate policy affected the market. There were seasonal swings in the volume of exchequer bills. If the market refused to absorb the bills at existing rates, the bank was asked to lend to the Treasury or purchase bills in the open market. By raising bank rate, the bank induced the market to hold more bills. With the usury law repealed, the bank experimented with the use of bank rate as a means of controlling base money.

FROM PALMER’S RULE TO PEEL’S ACT

By the 1830s the main features of the monetary system were in place.²⁷ A money market had developed, and the principal institutions had accepted distinctive roles. The Bank of England had a set of social objectives, some partial understanding of the steps required to achieve these objectives, and glimmerings of an understanding of the short-run consequences of its actions. Both the market and the bank realized that the bank’s responsibilities went beyond those of an ordinary bank to include the role of lender of last resort. Moreover, the bank accepted responsibility for maintaining specie payments at a fixed pound price of gold and had become familiar with the traditional central banking control techniques—discount rate changes, qualitative restrictions, and in a limited sense, open market operations.

In 1827 the bank added a rule of procedure to guide policy actions, known as Palmer’s rule after the governor who announced it at the parliamentary hearings of 1832. John Horsley Palmer saw the rule as a means of reducing the variability of the quantity of money in circulation and the exchange rate, and he apparently regarded such smoothing operations as part of the responsibility of a central bank.

Palmer’s rule attempted to tie the liabilities of the Bank of England to the stock of bullion. When the exchange rate was at par, the sources of the monetary base were to consist of one-third bullion and two-thirds securities. Except for seasonal adjustments, discussed below, the security portfolio would be kept constant, and the bank would increase or decrease the note issue as gold flowed in or out.

Every monetary rule is based on a theory of the monetary process, Palmer’s rule no less than those that came later. The theory behind the rule was the Hume-Thornton-Ricardo theory of the long-run consequences for prices, gold stock, and the exchange rate of changes in money. The rule accepts two propositions from that analysis. One, emphasized during the Restriction, is that depreciation of the exchange rate is evidence of an excessive issue of notes. The second is that the gold reserve is held against the bank’s notes and deposits, not just notes as the currency school proposed (Mints 1945, 83).²⁸

The main defects of Palmer’s rule as a guide to operating policy bring out some differences between the monetary theories of Thornton and Ricardo. First, Thornton accepted proposition one as a long-run proposition, but he argued at length that, in the short-run, changes in the demand for money (or monetary velocity) cannot be neglected. Such changes occur when new substitutes for money appear or their use expands. Thornton was clear that “paper credit,” which is to say bank deposits, differs from gold but that both are part of the “circulating medium” and both affect the price level. Second, Thornton urged the Bank of England to expand the monetary base with the long-run growth of trade. Third, he stressed the effects on money, output, and prices of temporary changes in the demand for currency. Under Palmer’s rule, expansion and contraction of money (currency and deposits) were tied to gold flows. However, the rule made no provision for changes in the amount of currency produced by country banks and no provision for changes in the distribution of the liabilities of the Bank of England between government deposits and base money.

The members of the currency school attacked Palmer’s rule on two grounds, both familiar. The rule allowed the bank discretion, not only because it had been adopted voluntarily but because in practice the rule was sufficiently complex that the bank could abandon it or make exceptions whenever it wished. A second, and more frequent, criticism concerned the definition of money. Palmer’s statement of the rule allowed the bank’s total liabilities—deposits and currency—to rise and fall with gold movements. Since the bank offset the effect of quarterly fluctuations in Treasury deposits on the base, it had to raise or lower the monetary base as gold flowed in and out. The currency school defined money as the sum of currency (notes) and bullion but excluded deposits. It argued that gold movements would have their expected effect on the price level and exchange rate only if changes in currency matched the changes in gold, and it urged the Bank of England to operate as if the source of the monetary base consisted entirely of bullion.

Some of the issues raised by the currency school have had a long life. The issues resurface periodically when there are changes in the types of financial institutions or their activities. One part of the currency school position is that the price level depends on the type of liabilities or assets issued and repurchased by the central bank. These writers understood that money was neutral in the long run, but they emphasized a narrow definition of money. To maintain stable prices they believed the central bank should limit currency issues. The modern statement of the proposition usually assigns importance to a broader definition of money that includes checkable deposits and often time and saving deposits as well as some additional items.

From 1827 to 1836 bank rate remained constant (at 4 percent), and the base fluctuated with market forces. The stock of bullion varied between £4 million and £12 million and the security portfolio between £20 million and £34 million, both narrower ranges than in the previous decade. For this reason the rule might be regarded as a partial success. Some small portion of the fluctuations consisted of seasonal movements resulting from continuation of the bank’s practice of offsetting seasonal fluctuations in market interest rates caused by differences in the timing of Treasury payments and receipts. Apparently Palmer’s rule was never intended to apply to short-term portfolio changes of this kind, because neither Palmer nor most of the other directors believed at the time that short-term fluctuations in money (however defined) had a permanent effect on the exchange rate. The bank had not yet accepted the state of the money market as an indicator of bank policy, but it had started a move that would bring it to that position within a decade (Wood 1939, 45).

Wood (1939, 102–3) argues that Palmer adopted the rule because he did not believe discount rate changes provided an effective means of controlling the bank’s portfolio, a point that Hawtrey repeats (1962, 14–15). These writers neglect that the rule was promulgated in 1827 and announced in 1832 when the bank still operated under the usury law.²⁹ Whatever Palmer’s earlier views on the effectiveness of rate changes may have been, the bank under his leadership changed the discount rate seven times between the summer of 1836 and the winter of 1839–40 in response to a series of crises, first in the United States and later in Belgium. Equally important, the bank raised the discount rate above 5 percent, first to 5.5 and then to 6 percent, to stem the outflow of gold in 1839. By changing the rate and borrowing abroad to increase its bullion holdings, the bank was able to maintain specie payments throughout the period despite the loss of two-thirds of its bullion in 1839.

The experience convinced most observers that monetary problems had been exacerbated by the defects of Palmer’s rule. The bank lost gold when deposits were withdrawn, and although currency remained unchanged, it had acted to increase security holdings so as to restore deposits despite the loss of gold. The currency school and the bankers drew very different conclusions, however. Anticipating a dispute that continued for the next century and beyond, Palmer and the group of writers known collectively as the banking school, generalizing from their experience with the rule and the events of the previous half century, concluded that no set of rules could guide the bank adequately. To the extent that they proposed solutions, they favored changes in the discount rate, to be made at the discretion of the bank, as required to protect the bullion stock and the exchange rate. The currency school, on the other hand, blamed the failures of the rule on the exercise of discretion by the bank and particularly the failure of the bank to keep a constant stock of gold and notes.

The proponents of rules triumphed over the advocates of discretion. With the passage of Peel’s Act in 1844, the currency principle was established as the governing principle of the monetary system.

NEW LESSONS FROM RENEWED CRISES

The Bank Charter Act of 1844 (Peel’s Act) accepted a main point in Ricardo’s plan for a central bank: separation of monetary operations (the control of the note issue) from banking operations (control of deposits and lending). The bank was to have two departments, an Issue Department and a Banking Department. The Issue Department carried out the monetary operations under a formula that tied the note issue to the stock of gold, as the currency principle required. The bank’s maximum note issue was set at £14 million plus the stock of gold coin and bullion held by the Issue Department. The bank gained a monopoly of the note issue.³⁰

The Banking Department was expected to function in much the same way as any other private bank. Its reserves consisted of the notes of the Issue Department and a small stock of gold held to facilitate exchange of notes and deposits into gold. Whenever the Banking Department accumulated more notes than it wished to hold, the Issue Department redeemed them by paying out gold. The proponents of the act expected the bank to compete for discounts and to hold deposits for other London banks, and they saw no conflict of interest in these functions. The custom of country banks’ keeping deposits in London and of London banks’ keeping deposits at the Bank of England was well established. More important, the proponents of the act denied that deposits were money.

The bank apparently welcomed its new freedom of action. Prices had fallen 20 percent in four years, and the gold stock had increased by £11 million to £12 million. It reduced bank rate from 4 percent to 2.5 percent within the month that the act passed, and the bank aggressively competed for discounts in London and at its branches. From the low level of £2.6 million in 1844, the discount portfolio jumped to £18.5 million in 1845, £34 million in 1846, and £38 million in 1847 while the bank’s income from discounts rose from £80,000 to £380,000 (Wood 1939, table 9, 137; Scammell 1968, 145).

With the decline in British prices, gold had come to England at a steady rate during 1842 and 1843, increasing the bank’s gold holdings and expanding the monetary base. The 1844 act required the bank to follow the currency principle by issuing or withdrawing currency (notes) as gold held by the Issue Department rose and fell. The Banking Department could use all the gold it acquired to expand deposits. As a result the directors, for a time, took no action to control deposit expansion. Bank rate remained at the market rate of discount on bills of highest quality, whereas before it had served as a penalty rate. Prices started to rise in the second half of 1845. The gold flow reversed, so the bank raised the discount rate to 3.5 percent. The following year the rate was reduced to 3 percent, where it remained during the fall while gold flowed out.³¹ Between 1844 and 1847, prices rose by more than 20 percent, with much of the increase in 1847.

The panic of 1847 is in many respects similar to the panic of 1825.³² The bank raised the discount rate, first by steps of 0.5 percent and in April by 1 percent. When gold continued to flow out, the bank limited discounts and called advances. The brief panic that ensued ended when the gold flow reversed. During May and June the bank accumulated about £1 million of gold, and the reserves of the Banking Department increased by £3.5 million. In July the gold drain resumed. The bank met the new threat by raising bank rate to 5.5 percent and again placing restrictions on the type of discounts it would accept. A series of bank failures called forth new restrictions, and although the bank did not refuse to lend, it raised substantial doubt about its intentions by announcing that after October 15 bills could be discounted only at rates to be decided at the time of application. The currency drain intensified and produced a new wave of bank failures. The internal and external drain of £1 million pounds in the following week reduced the reserves of the Banking Department to £2 million and raised the fear that the bank would soon be unable to issue notes or accept deposits.³³

The bank looked to the government for assistance, just as it had in 1793, 1811, and 1825. This time assistance took the form of a letter indemnifying the bank for damages arising from violation of the 1844 act and empowering it to expand its discounts provided it raised the minimum discount rate to 8 percent. The bank quickly adopted the policy later known as “lend freely at a high rate.” Within a few days, the panic ended; bank rate was lowered to 7 percent within a month, and by late December it was back to 5 percent.

The panic was mainly a monetary crisis, as that term is now understood, brought on by the tardy and hesitant actions of the Bank of England during the period of rising prices after 1844, followed by a very restrictive policy in 1846–47. The bank seems to have recognized that its policy had either caused or contributed to the crisis. During the next few years the securities portfolio fluctuated much less than in the past, and the bank was able to reduce the discount rate in a series of steps to 2.5 percent by 1849. At the time the bank had £17 million in gold, and the Banking Department had £12 million in reserves and only £25 million in securities.

The panic helped to resolve some disputes about the role of a central bank. The currency school argued that before the act there was no effective limit on the note issue. The Bank of England was not required to maintain a fixed gold reserve ratio, and the regional banks could issue notes in response to demand. The currency school claimed that, as a consequence, money growth was procyclical in the early stages of an expansion. The increase in currency raised the price level, causing a loss of gold and a crisis.

The so-called banking school differed about the importance of currency and did not rely on any of the monetary aggregates. Its proponents wanted the bank to discount only real bills, and they claimed that a real bills policy would prevent over- and under-issue of money and credit. To the extent that they had a uniform view, they emphasized real events at home or abroad as the principal cause of crises. In their view, the role of the Bank of England was to serve as lender of last resort to the financial system. This distinguished the bank from other banks. (See Laidler 1988, 100–102.)³⁴

The crisis showed that the system had not worked as the currency school predicted. The Banking Department had not been able to operate as an ordinary bank. At the beginning of the expansion, as in 1824–25, the bank held a stock of gold that was much larger than the stock usually held. Yet the gold stock and the reserve of the Banking Department had proved insufficient, and the bank had been forced to appeal to the government to suspend provisions of Peel’s Act. The currency school interpreted the crisis of 1825 as the consequence of the bank’s failure to keep note issue tied to gold. Prominent members of the banking school, who had opposed the act of 1844, interpreted the crisis of 1847 as evidence of the failure of the currency principle. These writers now urged the bank to adopt a more effective means of maintaining the exchange rate, protecting the gold reserve, and avoiding crises.

Evidence of the change in policy and in the approach to policy is shown by the variability of the discount rate. Between 1793 and 1844 the bank changed the discount rate only eleven times, and except for a brief period in 1839, the rate was never less than 4 percent or more than 5 percent. Between 1844 and 1849 the rate changed sixteen times and varied between 2.5 percent and 8 percent, eight of the changes occurring in the crisis year 1847. During the quarter century beginning in 1850, bank rate changed more than 225 times, an average of once each five or six weeks. There are only three years from 1844 to 1914 in which the discount rate did not change. In two of them, 1895 and 1896, the rate remained constant at 2 percent, a rate that tradition had by that time established as a minimum.

If the act of 1844 was a victory for the currency school, the victory was short-lived. We do not know the precise date on which the bank’s policy changed from reliance on the currency rule to reliance on discretion, but there can be no doubt that it pursued a less aggressive lending policy and a more variable bank rate policy after the panic. Bank rate remained above the market rate, and the bank’s stated policy was “to follow the market” (Wood 1939, 138). Attention shifted away from the theoretical issues raised by Thornton or Ricardo and toward the solution of so-called practical, or managerial, questions about how best to operate under the gold standard and how to avoid suspension and inconvertibility. The bank learned to use the discount rate to attract balances to London from country banks and from abroad or, when required, to send balances to the country or abroad. Gradually, the bank accepted responsibility for maintaining convertibility at the fixed mint price of gold and relied on changes in bank rate to attract and repel balances.³⁵

Thornton had described the relation between capital movements, exchange rates, and demand for gold and pounds. He recognized also that changes in the quantity of gold affected the monetary base, the volume of currency and deposits, expenditure, and prices. Most writers accepted the general framework, but many failed to emphasize the effects of gold movements on the prices of goods, on output, and on the price level that Thornton had stressed.

By the 1850s most discussions of central bank policy that mentioned long-run (or general equilibrium) effects acknowledged that increases or decreases in money (however defined) changed prices in the same direction. However, many policy discussions ignored long-run effects and emphasized short-run changes in the money market and short-term capital flows. Effects of relative price changes on the balance of trade were said to operate slowly and as a consequence received much less attention than they had earlier. Critics of these orthodox views raised many of the objections that have been repeated ever since. Announcement effects and destabilizing speculation are common in the writing of the period. Some claimed that an increase in bank rate encouraged a withdrawal of gold from England because speculators anticipated a further increase; others claimed that monetary policy was counterproductive because exports had to be financed at a higher interest rate, whereas imports were not reduced until later, so gold was withdrawn (Wood 1939, 125–26; Viner 1965, 278–79).

Hawtrey (1962) notes that bank rate policy was more variable from 1860 to 1880 than after 1880. He explains the greater variability during the earlier period as a consequence of the interaction between the bank and the market and argues that during this period a change in bank rate had a substantial effect on the balances country bankers held in London. A rise in bank rate restored the bank’s reserve by drawing balances from the country and permitted it to lower the rate. Once the bank lowered the rate, the balances went back to the country, forcing it to raise the rate again. The process continued until the bank’s reserve was restored. According to Hawtrey (xii), London banks acquired many of the country banks as branches after 1880 and centralized reserves at the head office, so a rise in bank rate drew fewer balances to London and a fall drew fewer balances out.

The argument is of interest for two reasons. First, Hawtrey’s argument that institutional changes weaken the effectiveness of monetary policy returns many times. Second, his argument is opposite to the argument, frequently made in the United States, that a decentralized banking system is less responsive to interest rate changes. Most arguments of this kind confuse levels and changes. If banks holding negative excess reserves are penalized or incur a loss of utility greater than the cost of adjusting (by borrowing or by other means), on average each bank will hold positive excess reserves. Centralizing the banking system reduces the average level of reserves. Neither centralization nor decentralization, however, implies that the response to a change in the discount rate is larger before or after the change, only that variations take place around a new level. Moreover, the evidence for the period does not support Hawtrey. His argument requires the time series on bank rate to show advances followed by declines more frequently in the earlier years than in the later years. Most of the changes are unidirectional movements in both periods, contrary to Hawtrey’s hypothesis.

A more plausible partial explanation of the higher variability in the ten to fifteen years before 1880 than after 1880 is that trade expanded and in the 1870s several countries accumulated gold to prepare for a resumption of specie payments. Gold production was considerably smaller and more variable than it had been in the decade following the discoveries of gold in California and Australia. The growth in world demand for gold meant that the Bank of England had to either pursue a more deflationary policy than in earlier years or let its reserve ratio of gold to monetary liabilities decline. The bank chose the latter course. The monetary base, deposits and notes at the Bank of England, rose relative to gold during the sixties and seventies. As a result, a given change in gold both permitted a larger expansion and required a larger contraction in the monetary base and the stock of money.

The variability of the bank’s policy meant that the economy had to adjust frequently to large swings in monetary policy. Table 2.1 gives the data on the size and frequency of changes in bank rate. When reading these data, it is important to keep in mind that the changes were always made in a series of steps; most often the rate increased in steps of one percentage point and decreased in steps of one-half point. The data understate, to a considerable extent, the frequency with which the economy had to adjust to changes in bank rate.

A key difference between the earlier and later policies appears in the data at the end of table 2.1. After 1873 bank rate remained within a narrower range than before. More often than not, the rate was between 2 and 6 percent, and until 1907 it did not exceed 6 percent again, and then only for a short time. There is little doubt that the Bank of England’s discount policy became less variable. The reason is that between 1876 and 1886 the price level fell at a compound rate of nearly 4 percent a year. Since the bank did not set bank rate below 2 percent, the economy had to undergo enough deflation to equilibrate the balance of payments and maintain the bank’s reserve position. When economic expansion took place and gold was withdrawn, the bank raised the rate by steps of 1 percent just as it had done during the period of inflation. The bank did not recognize that a 5 or 6 percent bank rate was a much more restrictive policy with 4 percent average deflation than during years with positive average inflation. A short period with bank rate above 5 percent brought contraction, renewed deflation, and an inflow of gold. During the last quarter of the nineteenth century as a whole, bank rate remained above 5 percent a combined total of only twenty-six weeks.

The bank failed to recognize the effect of inflation on interest rates or to distinguish between market rates and real rates of interest. Thornton’s recognition of this distinction was lost. For the next century, during wars, depressions, inflations, and deflations, central bankers used the absolute value of market rates to judge whether rates were high or low and monetary policy tight or easy. Many of their most serious mistakes resulted from this error.³⁶

We can only speculate on other reasons for reduced variability, but one plausible reason is that the variable policy was followed by a series of disturbances and some crises. In 1857, 1866, 1873, 1878, and 1890 the Bank of England was forced to respond to an internal, external, or combined internal and external demand for gold. These disturbances to financial stability were by no means wholly a result of the bank’s policy. In 1857 the expansion of the economy was brought to an end by an external drain to the United States to satisfy the demand for gold during the United States panic of 1857. The disturbance was made worse by the failure of a large discount house, Sanderson and Company. In 1866 a crisis occurred, intensified by the failure of several banks and discount houses—including the largest discount house, Overend, Gurney and Company—a failure often attributed to poor management but no doubt intensified by the balance of payments deficit of 1865. In 1873 there were drains of gold to Germany, Austria, and the United States owing to crises in those countries, and in 1878 the failure of a large Scottish bank—the City of Glasgow Bank—induced an internal demand for gold and Bank of England notes as well as a series of bank failures. The problem in 1890 involved one of the largest and most prestigious merchant banks, Baring Brothers (see Schwartz 1987b, 272–74).

During the crises of 1857 and 1866, the bank relied entirely on bank rate until the government announced that it would indemnify the bank against issuance of currency in excess of its gold holdings. As table 2.1 shows, the rate was raised from 3 percent or 5.5 percent to 10 percent. During the Baring disturbance the maximum rate was 6 percent, and this rate remained in effect only four weeks. A considerable part of the difference in the discount policies of the two periods is explained by the alternative policies the bank developed. On receipt of the news that Baring would have to suspend payments, the bank increased its gold reserves by borrowing abroad, as it had done in 1839, and by purchasing gold. The governor of the bank then organized a syndicate of leading London banks to guarantee Baring’s liabilities. The disruption was short, and the sizable reduction in output that occurred several years later cannot be attributed to the Baring crisis.

Several important changes in the bank’s policies took place after the panic of 1847. In the panics of 1825 and 1847 the bank attempted to restrict the volume of loans. In 1857 and again in 1866 it made large loans at ever higher rates and made no attempt to restrict the volume. Within a few weeks of the start of the 1857 panic, the bank’s discounts doubled. The loss of gold was so great that it was forced to make use of the temporary power to issue notes without gold backing, that is, to temporarily suspend the provisions of the act of 1844. At the start of the 1866 panic in January, the bank had a larger reserve and did not have to suspend the act until March.

A second important change occurred between 1857 and 1866. The bank sold government securities during the crisis of 1857 in an attempt to reduce the growth of its portfolio. By 1866 the bank recognized the role of lender of last resort more clearly; it increased “private securities held” by a larger amount than in 1857 and made no attempt to sell government securities. Bagehot, a major critic of the bank’s directors, congratulated them for at last recognizing that the Banking Department was not an ordinary bank but the protector of the country’s reserve and the lender of last resort for the financial system.³⁷

The series of panics from 1847 to 1866 also contributed a classic to the banking literature: Walter Bagehot’s book Lombard Street: A Description of the Money Market (1873). The book gives a clear description of the institutional arrangements of the time and proposes rules for the conduct of monetary policy. Bagehot did not criticize the bank for failing to respond to crises. Long before economists emphasized anticipations and policy credibility, Bagehot criticized the bank for failing to announce its policy in advance. A main point of the book is that directors of a central bank must publicly acknowledge their responsibility as lender of last resort and prepare for future crises under a commodity standard by holding a larger reserve than ordinary banks. Failure to do so creates and intensifies panics. In the course of the argument, Bagehot reveals a clear understanding of fractional reserve banking.

Like Thornton, Bagehot distinguished between the appropriate response to an internal drain and an external drain. If an internal drain increased the demand for gold or Bank of England notes and there was no reason to be concerned about the exchange rate, the drain should be met by substantial loans from the bank. But if large amounts of gold went abroad, the crisis was external and should be met by raising the lending rate of the Bank of England. A rise in bank rate encouraged foreign lenders to buy bills in London and, by reducing internal demand and the price level in England, encouraged exports and reduced imports: “The rise in the rate of discount acts immediately on the trade of this country. Prices fall here; in consequence imports are diminished, exports are increased, and, therefore, there is more likelihood of a balance in bullion coming to this country after the rise in the rate than there was before” (Bagehot 1962, 23).³⁸ Later the gold attracted by the higher rate reversed the decline in domestic demand and the price level.

Bagehot recognized that if the bank allowed an external drain to persist, it would face an internal drain as well. The domestic public, seeing the decline in the gold reserve, would exchange deposits and the note issues of country banks for gold and Bank of England notes. If the two drains occurred together, the Bank of England must discount willingly at a high rate.

Before we had much specific experience, it was not easy to prescribe for this compound disease; but now we know how to deal with it. We must look first to the foreign drain, and raise the rate of interest as high as may be necessary. Unless you can stop the foreign export, you cannot allay the domestic alarm…. And at the rate of interest so raised, the holders—one or more— of the final Bank reserve must lend freely. Very large loans at very high rates are the best remedy for the worst malady of the money market when a foreign drain is added to a domestic drain. Any notion that money is not to be had, or that it may not be had at any price, only raises alarm to panic, and enhances panic to madness. (Ibid., 27–28)³⁹

Throughout his book, Bagehot (1962, 31–32, 79–82) made it clear that by “very large loans” he meant the absolute volume and not the volume of loans relative to the gold reserve.

Lombard Street is a clear and definite statement of some important principles of central banking. Consistent application of these principles would have avoided many of the worst consequences of monetary policy in the century that followed its publication. The book is a high point in the statement of the banking school view of the role of a central bank and its responsibilities.

The weakness of the book is the weakness of the bankers’ approach. Bagehot’s analysis of the relation between the operations on the money market and their consequences for the economy never reaches the level of Thornton’s. Throughout his book he shows an awareness of the feedback and response between monetary policy and the economy, but his astute observations never produce an incisive analysis of these effects to stand beside and supplement his analysis of the money market. Nowhere in the book does he attempt an analysis of the relation between discount rate and market rate that would be required to carry out his policy of increasing the gold reserve of the Bank of England. Nowhere does he mention that his policy of increasing the gold reserve at times requires deflation, although he was aware of this possibility. The appearance of the book before the start of a long period of deflation helps to explain this lack of attention, but examination of the many proposals to increase the bank’s reserve ratio, produced in response to his book, shows no recognition of the deflationary effect of the policies required to increase the bank’s gold reserve.

Bagehot’s book makes it clear that bankers and the banking school did not regard the international gold standard as a completely self-regulating system, and the data in table 2.1 show that their actions were consistent with their views in this respect. The bank was expected to regulate the money market. Indeed, the views of the period erred far more on the side of minimizing or ignoring the consequences of variability for the economy than of overlooking the role of a central bank in keeping the exchange rate between the gold points.

By automaticity of the gold standard, bankers most often meant that in the long run a central bank could not both keep the exchange rate fixed and prevent prices from rising or falling. Virtually every English nineteenth-century writer on banking understood that inflation or deflation must be accepted as the cost of keeping the market price of gold equal to the mint price. The statements of United States bankers and world central bankers during the 1960s on the importance of “discipline” and “confidence” could just as well have been made by English writers a century earlier.

The series of panics and disturbances after 1857 silenced any remaining adherents of the currency school view. Although the act of 1844 remained on the statute books, the Bank of England accepted de facto responsibility for controlling the reserves of the Banking Department and the broader responsibility for the functioning of the international monetary system. After 1870, world demand for gold rose as other countries joined England on the gold standard. The Bank of England accepted the resulting deflation as the price of maintaining convertibility, just as it accepted deflation after 1815 and 1919 as the price of restoring convertibility at the previous mint price.

The actions of the Bank of England after 1844 make it clear that the bank gradually accepted money market rates as the principal indicator of current monetary policy. Bank rate was set in relation to money market rates, and without any distinction between nominal and real rates. In both the inflation of the 1850s and 1860s and the deflation of the 1870s, 1880s, and early 1890s, the bank kept 2 percent as the minimum rate of discount. Neither the bank’s management nor the economists of the period recognized that under the gold standard, the bank’s refusal to lower the discount rate below 2 percent meant that the only equilibrium was at an anticipated rate of deflation equal to the actual rate. Alfred Marshall’s testimony before the Gold and Silver Commission of 1887–88 shows him to be struggling toward such an explanation without reaching it.⁴⁰

One aspect of the act of 1844 that should not be overlooked is that the Bank of England was forced to respond to internal as well as external drains. The bank did not—and could not—ignore the series of bank failures and the internal drains as the Federal Reserve did in the 1930s. Any failure or hesitation to assist the banks by refusing to lend or restricting discounts increased the demand for loans, gold, and Bank of England notes. Wood (1939) summarizes the period up to 1858: “The principle was thoroughly engrained in the minds of the business community that good bills were convertible into Bank funds and, regardless of the state of the reserve, there were only two occasions (1825 and 1847) when the principle was called into question. On these occasions the Bank’s action was quickly reversed after an understanding with the Government.”

The bank’s understanding of its responsibilities during crises improved so much after 1858 that it organized the banking community and participated in the guarantee of Baring’s assets in 1890 in full knowledge that these were illiquid. “Good bills” were, of course, preferred, but the bank showed by this action that convertibility could be maintained and the crisis ended if it acted promptly and did not refuse to lend.

FISHER ON REAL AND NOMINAL RATES

At the end of the nineteenth century, Irving Fisher (1896) analyzed the relations between real and nominal interest rates using a number of examples to illustrate his argument. He repeated and extended the argument in subsequent work. His economic writings and policy campaign for a stable standard of value recognized that stable purchasing power of money would remove most of the problem.

Fisher’s discussion (1930b) is much clearer than Thornton’s.⁴¹ He distinguishes between perfect and imperfect foresight. Under perfect foresight, expected appreciation or depreciation of purchasing power is fully reflected in the market rates (39, 41–42). People do not anticipate correctly: “When the cost of living is not stable, the rate of interest takes the appreciation and depreciation into account to some extent, but only slightly and, in general, indirectly.” Fisher does not fully explain why the public always underestimates the rate of price change.⁴²

Fisher’s explanation of the relation (1930b, 439) is similar to Thornton’s: “Rising prices increase profits both actual and prospective, and so the profit taker expands his business. His expanding or rising income stream requires financing and increases the demand for loans.”

Fisher was not an obscure author of unread economic tracts. He was the leading American academic economist and an active participant in policy discussions. He worked hard to get his ideas about money and monetary standards adopted. In the 1920s a citizens’ league promoted his ideas. Congress considered legislation to mandate his monetary standard. Yet I have found no mention of the distinction between real and nominal interest rates in Federal Reserve minutes during the deflation from 1929 to 1933 or until late in the inflation of the 1960s and 1970s. In both periods, and in many others, the Federal Reserve (and other central banks) used an absolute standard to judge whether interest rates were high or low and associated high and low market rates with tight and easy money.

Why was the distinction between nominal and real interest rates lost? Central bankers seem generally to have regarded Fisher as a bright but annoying crank. The Federal Reserve Board was dominated throughout the 1920s and early 1930s by advocates of the real bills doctrines who, like their predecessors, denied any relation between their actions and inflation or output. They ignored Fisher’s emphasis on the role of money, much as the banking school dismissed the arguments of the currency school without meeting them.

Further, Fisher often minimized the empirical relevance of the distinction between real and nominal rates. He viewed foresight as typically poor, so that interest rates did not reflect anticipations very accurately, if at all. But lack of foresight does not eliminate the importance of the effect of inflation on interest rates. The more myopic the public is, the larger are the losses and gains, and the effects on realized returns, when inflation or deflation occurs.

Fisher’s writings are also exemplary for the clear distinction he makes between permanent effects and temporary, transitional changes. Chapter 4 of his Purchasing Power of Money (1920) is concerned entirely with transitional effects. The same is true of his paper later in that decade relating inflation to unemployment. This insightful work had no influence on or meaning for adherents of the real bills doctrine, so it had no influence on policy decisions in the first twenty-five or thirty years of the Federal Reserve’s history. Later much of the staff and many of the policymakers adopted a type of Keynesian analysis that emphasized short-term or transitional effects and ignored long-term, permanent effects.

MONEY, CREDIT, VELOCITY, AND INTEREST RATES

Perhaps the most disconcerting aspect of the nineteenth-century discussion is that as central bankers improved their understanding of the effects of their actions, the techniques of central banking, and the responsibilities of the central bank during crises, their understanding of how their actions affected economic activity declined. The distinction between nominal and real magnitudes is more carefully observed at the beginning of the century than at the end. There is a clearer analysis at the start than at the end of the effect of substituting one means of payment for another. The distinction between money and credit blurred during the century, and most of the now familiar arguments about the “ineffectiveness of monetary policy” appeared. Although these issues returned to the academic literature at the end of the century, there is no evidence that academic writing had much influence on central banking. The gold standard and the real bills doctrine dominated policy action.

During the course of the century, the Bank of England (and others) learned to offset panics by serving as lender of last resort, to prevent large inflations or deflations by adopting the gold standard, and to manage short-term demands for credit by adjusting the discount rate to limit or increase the amount of discounts. Twentieth-century concerns about employment and economic growth were heard but had little effect.

Much of the academic writing failed the operational test that central bankers required. Definitions were not tight. There was little agreement and, with a few exceptions, no attention to the distinctions between temporary and persistent or transitional and permanent effects.

Ricardo, a leader of the bullionists, argued that the depreciation of the paper pound during the early years of the century could not be explained by the actual increase in paper money. The relevant change was the increase or decrease in paper money over the amount that would have circulated had there been convertibility. For this reason he tended to deny the validity of any simple comparisons between changes in paper money and changes in exchange rates. His argument, of course, allows for an effect of changes in the demand for money arising, as in Thornton, from changes in confidence and anticipated changes in the volume of transactions. Ricardo denied that these factors had operated between 1797 and 1810, however, and he offered no other explanation of changes in the demand for money. Further, like Thornton, Ricardo recognized that a change in the stock of money does not immediately affect prices and exchange rates, and he criticized his opponents for expecting immediate effects (Viner 1965, 135–42). Some of the other bullionists ignored these subtleties and wrote as if they too expected a very prompt and close adjustment between changes in money and changes in prices and exchange rates. A century later, rational expectationists also erred by understating the time required for the price level to respond to money.

By extension, the name bullionists is given to another group, those willing to extend the definition of money to include paper currency issued in fixed proportion to the stock of gold or specie. The Ricardians and their intellectual descendants in the currency school presented both views, and they inspired Sir Robert Peel, the author of the act of 1844. When writers in this tradition referred to deposits at the Bank of England or at other banks, they did not use the term money; they talked about means of payment, bills, credit, and sometimes the circulating media.

The group known as antibullionists generally denied that there could be an excess issue of paper money if banks restricted issues to the amount issued as part of the process of discounting commercial paper—at the time, bills of exchange. The core of their argument is that as long as the sources of the monetary base consist of gold and commercial paper, money cannot be overissued. The explanation they gave is that “no one would borrow at interest funds he did not need” (ibid., 148). If by chance a bank overexpanded notes or deposits, the excess issue would return to the bank either to reduce loans or, under convertibility, to acquire specie.⁴³

The central argument of the antibullionists reappears periodically and had a powerful influence in the early days of the Federal Reserve System. The influence has become a less important cause of errors in policy, but it still survives in two distinct ways. One is the belief that some increases in “credit” are productive while others are “speculative,” a belief that in its milder form generates periodic concern about the “quality” of credit as an independent factor. The second is the notion that the monetary base is demand determined, an argument that has been used at times to absolve central banks of responsibility for their errors and even for their policies.

The antibullionists, and all later adherents of the real bills doctrine, failed to distinguish between propositions that superficially appear similar. The first is the proposition that in the long run there can be no inflation if the stock of money grows at the growth rate of real output. The second is that changes in output induce changes in the demand for bank credit and the stock of money but that if the credit is limited to the change in output, expenditures cannot increase more than output and therefore inflation cannot result.

Thornton was the first to distinguish these two arguments and to recognize that the fallacy in the second argument resulted from the failure to differentiate nominal quantities and rates of interest from real quantities and rates of interest. The error, said Thornton, was the error of John Law, who “considered security as everything and quantity as nothing.”⁴⁴ Under an inconvertible paper currency, real bills provide no effective limitation of the currency and no defense against depreciation of the exchange rate. With a convertible currency the situation is no better, because the bank could not maintain convertibility if it allowed the base to be determined by the demands of the merchants. Thornton’s argument against the usury laws, discussed above, is a trenchant criticism of the real bills doctrine for the failure to distinguish between nominal and real rates. The antibullionists never replied to this argument.

Neither the antibullionists nor other proponents of the real bills doctrine recognized that it is the total quantity of notes, not their backing, that affects the price level. Commodities are sold and resold; each sale gives rise to a real bill. In the limit, there may be one increase in output backing many real bills.

As long as the payments system remained relatively simple, there was very little discussion of the definitions of money and the monetary base. Disputes about the definition start after the development of a market for bills of exchange and their use as a medium of exchange, the growth of deposit banking, and the increased use of banknotes in place of specie. Many of these disputes came to the fore during the Restriction period. Under convertibility, the requirement to pay gold on demand limited the quantity of money. The use of inconvertible paper raised questions about the effect of paper money on prices and exchange rates. In fact several different, but related, questions arose. One was whether the deposits at the Bank of England were money, and if they were money why they differed, if at all, from deposits at any other bank. To some writers if seemed obvious that because one liability is a substitute for another, there is no reason to draw fine distinctions between types of liabilities. A related issue is the possibility of controlling the stock. Those who emphasized substitutability generally concluded that efforts to control the stock of money, however defined, were a waste of time. Others focused on the gold stock and currency, items they believed to be money. Other items they regarded as part of the “circulating medium.” Related to these issues were disputes about the appropriate indicator of Bank of England policy—exchange rates, gold stock, gold flows, interest rates, or balance of payments—and about the variables that determined prices and exchange rates.

One long-lasting source of confusion in the monetary literature can be traced to the absence of accepted definitions. Since money originally meant gold or bullion, an increase in paper money was described as an increase in velocity. The reasoning was that the gold was held as a reserve by the issuer of notes, who thereby increased the “circulation” of the reserves he held (see Schumpeter 1955, 319). The source of this confusion lies in the origins of the banking system, particularly the “goldsmith” principle, under which goldsmiths could hold gold but could not increase its “circulation” by issuing claims in excess of the amount held. When the definition of money broadened to include notes as well as specie, the definition of monetary velocity changed also. Velocity included the “turnover” of notes, including the notes that might have been issued had bankers not elected to hold deposits at the Bank of England. The spread of deposit banking was often described, therefore, as an increase in monetary velocity.

Adding to the confusion was the practice of referring to an increase in deposits as an increase in credit, or sometimes as an increase in bank credit, and the related practice of referring to the stock of deposits as a stock of credit. For example, when Hawtrey discusses currency and credit, he means what is now called currency and deposits. Many of the writers who wrote that velocity declined secularly meant the same thing that others meant when they wrote that in the long run the ratio of currency to deposits declined with habits of payment.

A difficulty with any attempt to interpret parts of the discussion is that writers who denied that some asset or group of assets should be labeled money often did not make it clear whether they meant there was no point to defining the assets that serve as medium of exchange or that prices and the exchange rate did not depend on the quantity of money, however defined. Thomas Tooke, an early and prolific writer of the banking school, appears to have believed that the bank could affect the price level and exchange rate by changing market interest rates. But he also believed there was no need for interference by the Bank of England if bankers discounted real bills and notes and deposits remained convertible. For him the world price level was determined by the world’s gold stock, but he offered no explanation of English prices and denied that money, credit, or base money bore any consistent relation to prices. Most Federal Reserve officials remained in this tradition in the 1920s. They denied that their actions affected prices. A modern version of Tooke’s argument is found in Sayers 1957, 5, which argues that “to label something as ‘money’ …is to build on shifting sand.”⁴⁵

Later writers in the banking school tradition (Bagehot, for example) believed that the exchange rate and gold flows could and should be regulated by the bank. Bagehot emphasized that the bank had to regulate reserves, but a main reason for his emphasis is that the point was often denied. Under the act of 1844, currency issues were tied closely to gold movements, and it seems likely that if the currency issues were not regulated, Bagehot would have argued for control of both uses of the base—reserves and currency. Viner (1965, 243–44) points out that many in the banking school thought of banknotes and bank deposits as money, meaning mediums of exchange, even if they were vague about the effect of money on prices.⁴⁶

All of this was a far cry from Thornton, who recognized that the “possession of a right to draw [deposit] obtained in the one case, is exactly equivalent to the possession of the note [banknote] obtained in the other” (1965, 134). But few later writers saw that both reserves and currency affected market rates, money, and prices; most did not or, if they did, were inclined to emphasize one type of money rather than another.

The two points that the banking school emphasized most were the determination of interest rates in the money market and the determination of the exchange rate by the demand for and supply of pounds. If they saw beyond these points to the effect of changes in the base on money and of money on prices, market interest rates, and exchange rates, they did not stress the latter relations, and some denied them vigorously. As early as the 1830s John Horsley Palmer, a governor of the bank, had testified that the bank affected market rates by changing the “circulation” and suggested that during periods of crisis, market indicators are useful indicators of the state of the money market (Wood 1939, 45–47). With the passage of time and a series of crises and disturbances, this view gained adherents. The avoidance of panics seemed a more attainable goal if bank failures and internal drains could be avoided. Panics appeared to be money market responses, so avoiding panics required stability of the money market. Moreover, with the increased size and growing emphasis on short-term capital movements under the gold standard, market interest rates and other money market variables gained acceptance as indicators of near-term gold flows. As maintenance of the exchange rate and the bank’s reserve became the principal goal of bank policy, stability of the money market became its primary concern, the best means of ensuring exchange rate stability and avoiding domestic crises.

Some of the writers in the banking school deserve praise, however, for recognizing that the private sector is able to produce a variety of substitute means of payment. They described the development of branch banking, the pooling and centralization of bank reserves, and the use of deposits and bills of exchange as innovations that increased credit by finding more efficient uses of a given monetary base. They failed to see that the introduction and use of deposits and other substitutes for base money are limited by the return from producing substitutes. Nor did they see the related point that more efficient use of a given stock of base money does not imply loss of control of the stock of means of payment by the central bank.⁴⁷

The failures of the banking school included a failure to analyze the monetary system as part of the economy and often to define terms. Its adherents made no attempt to distinguish substitution in demand from substitution in supply, the use of substitutes from the production of substitutes. As late as 1867 Thomas Hankey, a director of the Bank of England, denied that either the bank or the banking system could create or destroy means of payment (Viner 1965, 255 n. 3). His argument repeats the central notion on which the act of 1844 rested and shows that Hankey was unable to distinguish partial substitutes from perfect substitutes. Bagehot’s criticism of Hankey and those who shared his view stresses the difference between bank reserves and bank deposits but not the difference between partial and perfect substitutes or between substitutes in demand and substitutes in supply.

The fundamental relation governing substitution on the demand side is that for each type of “money” the sum of the marginal product per dollar (or other unit) and the anticipated rate of deflation must equal the return per dollar (unit) in services and income. If two means of payment are relatively poor substitutes in supply, equilibrium is reestablished mainly by relative price changes, with little change in relative supplies. At the opposite pole, if the two are relatively close substitutes in supply, equilibrium is reestablished by changes in the respective outputs, with little change in relative returns to the two assets.

In the British monetary system, in our own, and in most others, currency and checking deposits are close substitutes in supply at a fixed supply price. Anyone who wishes to exchange one means of payment for the other does so at a fixed exchange rate. Technological change or other change in the relative cost or relative return from using one means of payment rather than the other affects relative demands. The return received by holders and users of these assets and the (real) demand for the sum changes. However, under a fractional reserve banking system, changes in relative demand for currency and deposits mainly change the combined nominal stock of the two and the price level at which the combined stocks are held.

The writers of the banking school, and many of those who repeated their arguments, observed the use of substitute means of payment and concluded, incorrectly, that the use of substitutes meant that each substitute means of payment had the same effect on the price level or the exchange rate as any other. Or if they were more sophisticated, they regarded changes in “credit” as changes in the “velocity” of the existing means of payment. They argued that the changes in velocity permanently changed the price level, the exchange rate, or the gold stock. For them money was a weighted average with the weights equal to the respective, but ever changing, velocities. Hence, they concluded, there was no point to defining money and no prospect of controlling money. These notions survive in discussions of “unstable velocity,” the impossibility of defining money, or the importance of controlling total “liquid assets,” credit, or the total liabilities of all financial institutions.

The currency school did not respond to the banking school arguments correctly or even uniformly. Some argued that the currency to deposit ratio was approximately constant, so control of currency issues meant control of money. Others argued that deposits could not be controlled because of substitution. Still others contended that the velocity of deposits was much lower than the velocity of currency and that the velocity of bills of exchange was lowest of all.

Improved understanding of the reasons for the failure of the currency principle had to await Fisher and Keynes. Fisher argued that the currency principle worked badly because currency was a poor indicator of monetary expansion and inflation. According to Fisher (1920, chap. 4), bank loans and bank deposits increased much more than currency during the early stages of the expansion. In his notation, M’ (deposits) rose relative to M (currency) following an inflow of gold or other source of base money. Later in the expansion, consumer expenditures increased and the ratio of currency to deposits rose. This rise forced fractional reserve banks, operating under gold standard rules, to surrender reserves. Faced with a loss of reserves, the banks raised loan rates, called loans, and reduced deposits. By redistributing the monetary base between reserves and currency, the rise in the currency ratio reduced the circulating medium, M + M′, or in current terminology, reduced money, currency and deposits. The reduction in money brought the expansion to an end and started the contraction.

Fisher’s explanation links short-term changes to the long-term value. The key element in his explanation of cycles is that businesses, banks, and households failed to anticipate promptly the inflation caused by monetary expansion. A central bank operating on the currency principle (or gold standard rules) would always react too slowly to prevent inflation. Since currency increased for a time by less than the inflow of gold, the central bank did not increase the discount rate or take action to slow the monetary expansion. As a result, commercial banks expanded deposits and loans by more than the amount consistent with the given gold stock and the longrun average ratio of currency to deposits. Inflation was under way before currency increased relative to gold. Changes in money and credit were therefore procyclical.

Fisher was one of the first to emphasize that differences in the timing of cyclical changes in the stocks of currency and deposits caused the currency ratio to fluctuate around its long-run trend. Others commented on the fluctuations but failed to see that they were systematic, not random events. However, Fisher did not draw the implications for central bank policy of adherence to the currency principle. This was done by Keynes, in rather picturesque language.⁴⁸

Most nineteenth-century writers not only failed to analyze the timing and proximate causes of changes in money but did not consider the failures of monetary policy as a main cause of fluctuations in output. Neither the fluctuations under Palmer’s rule nor the series of crises under the act of 1844 stimulated anyone to analyze the relation of the Bank of England’s policy to interest rates, output, prices, and specie flows. A few writers in the currency school accused the bank of overexpanding its discounts in 1844, but they did not follow the charge far enough to see the conflict with the currency principle. Had they considered the implications of their argument, they would have been forced to recognize that a central bank should keep control of both uses of the monetary base—reserves and currency— not currency alone. Almost certainly, some would have recognized that the way to control the uses was to control the total sources. Again, this point was not recognized until John Maynard Keynes’s Treatise on Money (1930, 2:225–26).

Thomas Tooke, testifying before Parliament in 1832, anticipated Marriner Eccles’s “pushing on a string.”⁴⁹ Wood (1939, 48) summarizes a part of his statement: “An increased issue of notes might only swell the note reserves of the London bankers or be deposited by them in the Bank.” Samuel Gurney, a leading banker of the period, testified that if there was an abundance of notes, extra notes would remain in the tills of the bankers, the “natural depository” of surplus notes. If there were ample means for speculation, mere idle funds would not encourage speculation (54).

William Blake argued in 1823 that inflation was caused by fiscal policy (government purchases) financed by new debt issues. The debt issues activated previously idle balances, that is, increased velocity. And Lord Lauderdale blamed the contraction before 1820 on the government’s budget surplus. A surplus reduced “effectual demand,” so production declined. Lauderdale wanted the government to replace the war expenditures that ended in 1815 with expenditures on public works (Viner 1965, 192–94).⁵⁰

CONCLUSION

Every complete theory of the monetary system must provide answers to a number of related questions. What is the monetary standard, and what are the source components of the monetary base? Why do the source components expand and contract? Which items are included as uses of the monetary base? If the uses of the base consist of more than one item, what effect does the substitution of one item for another produce on the monetary system? By what means and to what end should the government or a central bank seek to control the base? What are the short- and longer-term consequences of a change in the base on the stock of means of payment? What are the short- and longer-term consequences of changes in the means of payment on prices, output, employment, and balance of payments? What, if any, is the feedback from the changes in prices and real variables to the source components of the base?

In practice, there have been three distinct types of answers. One, following Thornton, stresses the relation of the base to the stock of money, the effects of money on economic activity, prices, balance of payments, and specie flows or exchange rates. A second approach, following Ricardo, puts aside questions of the relation of the base to the stock of money or means of payment and avoids analyzing the effects of substituting one means of payment for another. For the purposes of analysis, money and the base are identical or proportional. In both of these approaches, the monetary base stands at the bottom of a pyramid. Substitution of one type of credit instrument for another is of secondary importance, or no importance at all, once the determinants of the base and the stock of money have been specified and the relation between the two analyzed or dismissed. Corresponding to each set of tastes, state of technology, and anticipation of the future, the economy has a real rate of return at which the public willingly holds the stocks of money and real capital. The variety of claims and debts cancel each other out and affect the solution only to the extent that they represent changes in taste or technology, and then only as much as any other change in taste or technology. Corresponding to the real rate of interest, there is a market rate that differs from the real rate by the anticipated rate of price change. When anticipated and actual rates of price change remain equal, the economy reaches and remains in long-run equilibrium.

A third approach looks at the credit system as one that issues a variety of claims and debts that substitute for money or more generally for means of payment. In this approach, money lacks fundamental importance in the explanation of price and output changes or specie flows. Credit, interest rates, or more recently “flows of funds” become the main indicators of the state of the credit markets. The banking school developed this notion, if such a loose and amorphous collection of ideas can be described as “developed.”

Each of the three types of monetary analysis was known in the nineteenth century, and for a time each had a dominant influence on the development of central banking theory and practice. But in the end the banking school view became the established view among bankers and central bankers and was challenged by only a few economists. When the Aldrich Commission in the United States received the testimony of leading bankers and experts on central banking in 1912, the members heard very little about the effects of changes in money on domestic economies and a great deal about the “needs of trade,” “self-regulating productive credit,” and the use of the discount rate to stop an outflow of gold.

They learned, too, about the role of the Bank of England in smoothing the money market to eliminate seasonal fluctuations and its practice, well established by the 1840s, of offsetting the effects of Treasury operations on interest rates. These policies focused attention on short-term market interest rates and were the forerunner of so-called defensive open market operations. Among the by-products of the focus on short-term changes was the increased frequency of discount rate changes and, considerably more important, the belief or opinion that such operations were a main responsibility of central banks.

The promising analysis started by Henry Thornton recognized that money was neutral in the long run but not in the short run. Thornton’s work opened the way to a careful analysis of the differences between central banks and intermediaries, between money and credit, between real and nominal rates of interest, between relative and absolute price changes, and between permanent and transitory changes. He recognized the errors in the real bills doctrine. Later Irving Fisher revived and added to the understanding of these issues, but, like Thornton’s, his work did not influence central bankers until the Great Inflation of the 1970s.

Walter Bagehot did not have a theoretical framework to match Thornton’s. He understood, however, the importance of a lender of last resort. And he emphasized the importance of precommitment by the central bank and of following precommitment with action.

The Federal Reserve’s approach to policy originated in the Bank of England’s nineteenth-century practices and the partially developed theory or framework that the practices attempted to apply. By the end of that century, discussions of central banking confused credit and money, used money market variables as indicators of monetary policy, and denied or cast doubt on the ability of a central bank to induce short-term changes in output or employment by monetary means. Although stabilization of prices and employment was mentioned as a goal of monetary policy in the literature of the early and late nineteenth century, virtually every discussion of the policy of the period concluded that monetary policy was guided by the state of the reserves, not by output, employment, or economic stability.

1. This refers to the economic framework, not the political and administrative framework. The latter is perhaps uniquely American in its blend of public and private enterprise, of centralism and decentralism. A brief description of some of the domestic political forces shaping the Federal Reserve Act is in Dunne 1963.

2. A similar idea returned in the 1960s, when monetary policy was “assigned” to control the gold flow or balance of payments under fixed exchange rates.

3. Viner’s 1965 discussion of Ricardo’s analysis brings out this point and its importance for the policy discussion of the time.

4. A central issue returns many times in monetary history: What is to be included as money? Bullionists and the currency school chose narrow definitions. The bullionists argued that the stock of gold (or silver) bullion determined the price level and the exchange rate. The currency school emphasized the note issue. They wanted a rule tying the note issue to the Bank of England’s gold reserve. See Schwartz 1987a.

5. Laidler (1992, 4) argues that Thornton was perhaps the only classical economist to recognize that monetary impulses contributed to a business cycle, not just a “credit” cycle. Several earlier writers discussed the transitional real effects of monetary changes on real output. Indeed, analyses of monetary effects are among the oldest propositions in economic theory. See Hegeland 1951.

6. The Bank of England received its charter in 1694 to assist in the financing of war with France. See Dowd 1991 for a brief history of the bank and Clapham 1944 for a detailed history.

7. There are a number of estimates for the earlier years. After 1808, country banks required a license to issue notes, so the number of licensees gives a more accurate estimate. The number rose from approximately 700 in 1809 to a peak of 940 in 1814. Thereafter the number of country banks declined, at first because of losses from deflation and after 1826 because of the growth of joint stock and branch banking. By 1842 the number had fallen to 429. See Wood 1939, 14.

8. Bank of England notes did not become legal tender until 1833. London banks stopped issuing notes in 1793.

9. After the crisis of 1825, there were two changes in the arrangements just described. The responsibilities of the Bank of England were more widely recognized, although not acknowledged officially, and bill brokers performed many of the market functions previously performed by London banks, especially the function of absorbing and holding or supplying bills as the market demanded. In recognition of the changed roles of brokers and banks, by 1830 the Bank of England accepted deposits and made advances to the largest brokers. As the system evolved, the London banks no longer borrowed from the Bank of England; instead, the bill brokers borrowed, often for months. See Scammell 1968, 134–42.

10. Hawtrey (1962, 16) argues that Thornton failed to recognize the time dimension in real rates of return (or mercantile profit). This conclusion is based on an incomplete examination. In a speech on the Bullion Report, Thornton (1965, 336 and elsewhere) computes a net rate of return with dimension dollars per dollar per year in the course of his explanation of why an unanticipated inflation increases the realized profits of the borrower. Speaking of the merchants during an inflation, he wrote (336): “There was an apparent profit over and above the natural and ordinary profit on mercantile transactions. This apparent profit was nominal, as to persons who traded on their own capital, but not nominal as to those who traded with borrowed money, the borrower, therefore, derived every year from his trade, not only the common mercantile profit …but likewise the extra profit which he [Thornton] had spoken of. This extra profit was exactly so much additional advantage, derived from being a trader on borrowed capital and was so much additional temptation to borrow. Accordingly, in countries in which currency was in a rapid course of depreciation, supposing that there were no usury laws, the current rate of interest was often, …proportionably augmented. Thus, for example, at Petersburgh, at this time, the current interest was 20 or 25 percent, which he [Thornton] conceived to be partly compensation for an expected increase of depreciation of the currency” (italics added). Thornton then gave examples of the working of this principle from the experiences of Russia, Sweden, France, and America. In his book, (1965, 254) Thornton wrote: “The temptation to borrow, in time of war, too largely at the bank [of England] arises, as has been observed, from the high rate of mercantile profit…. [C]apital, by which the term bona fide property was intended, cannot be suddenly and materially increased by any emission of paper. That the rate of mercantile profit depends on the quantity of this bona fide capital and not on the amount of the nominal value …[is] easy to point out.”

11. On the following pages, 97–100, Thornton offers as one example the panic of 1793 when “many country banks failed. The stock of Bank of England notes, at the start, were not fewer than usual,” but the number became “insufficient for giving punctuality to the payments.” The effect was “to lessen the rapidity of the circulation of notes on the whole, and thus to increase the number of notes wanted.” The remedy was found in issuing Exchequer bills discountable at the Bank of England. Thornton points out as a “fact worthy of serious attention” that the crisis was started by a demand to convert country notes into gold but was brought to an end by an issue of paper (Exchequer bills) that could be turned into banknotes or gold and that “created an idea of general solvency.” In this passage Thornton anticipated Bagehot’s 1962 work on the lender of last resort.

12. See also Thornton 1965, 236–41, where he traces the consequences of an injection of new money for the borrower and for the community. See especially 239 for a brief statement relating the fall in real wages during inflation to forced saving.

13. The idea appears several times. He discusses (1965, 119) the greater variance of nominal prices than of nominal wages that leads agents to regard a fall in price as temporary. The same misperception of a permanent change is used to explain the real effect of currency depreciation. In this case Thornton also invokes misperception of relative and aggregate changes (340).

14. See Viner 1965, 134, for a comparison of Thornton’s views and those of his contemporaries. Keynes was apparently unaware of the extent to which Thornton anticipated his discussion of the demand for money.

15. The “real bills” notion, that credits advanced for productive purposes could not be a cause of inflation, had been proposed by several writers including James Stewart and had been used unsuccessfully to limit the note issue of the Bank of England before the Restriction. Mints (1945, 1) finds the real bills doctrine in writings during the 1770s. He notes (48) that even the earliest statements of the doctrine relate real bills to “elasticity” of the stocks of money or credit and to effective limitation of the note issue.

16. Note Thornton’s careful distinction between credit (borrowing) and money. This distinction was neglected by most writers until Lauchlin Currie wrote in the early 1930s. Notable also in his monetary policy are the principles that the stock of money should grow as the economy expands and that the bank should make temporary advances when there are internal drains. The complete argument of his book makes clear that by the “general trade” he means real output, not “real bills.”

17. Suspension of convertibility came in 1797 following France’s attempt to land troops in Wales. See Dowd 1991.

18. Data in this and in the next several paragraphs are from Wood 1939, 191, and Viner 1965, 174. Viner uses price indexes developed in Silberling 1923, 232–33. I use the data from Gayer, Rostow, Schwartz 1978 instead.

19. For the developments of the theory of money see especially Viner 1965, chaps. 3 and 4.

20. The peak rate of inflation is 20 percent a year compounded from 1798 to 1800. Prices fell an astonishing 20 percent in 1801.

21. The reduction of bank rate was not the only action taken. The gold standard became the legal (de jure) standard in 1821. Also, the bank’s holdings of government securities expanded and the bank lengthened the maturity of eligible private bills from sixty-five to ninety-five days.

Some indication of the effect on the bank of the changes in activity during these years is given by its income from discounts. Scammell (1968, 145) shows the following:

Year: Bank of England Income from Discounts (in thousands of £)

1795–96: 147

1809–10: 914

1815–16: 646

1820–21: 150

1925–26: 303

The figures are, of course, nominal values and therefore overstate the size of the changes in the bank’s real income.

22. See Viner 1965, chart 1 and table 1, 143–44. For Ricardo’s views on devaluation of the pound see Viner 1965, 204. The Gold Standard Act of 1816 repealed bimetallism in England.

23. The government’s budget surplus and deficits are from Wood 1939, table 6, 74–75. A brief discussion of the Dead Weight debt and open market operations is Wood 1939, 80–83.

24. William Huskisson was a director of the bank and had been president of the Board of Trade. He was an active reformer who opposed mercantilism and favored the reforms advocated by Smith and Ricardo.

25. Bagehot called this turnaround “classical” and liked the example so well he repeated much of the passage (1962, 99). Among the many bank failures of the period was Henry Thornton’s bank, Pole, Thornton, and Company, in which his son remained active after Thornton’s death in 1815. Bagehot’s analysis is, of course, similar to Thornton’s discussion of the panic of 1793. See note 11 above.

26. Very similar views about speculation are repeated in the Great Depression of the 1930s. Wood (1939, 83–84) presents a number of quotations from the parliamentary hearings of 1832 to show that the predominant view at this time was that it made very little difference whether the bank increased the circulation by a purchase of Treasury bills or by discounting commercial paper. This view contrasts with the views held by the bank’s directors earlier and criticized by Thornton and the views held by members of the Federal Reserve Board in the 1920s.

27. Bordo and Schwartz 1984 has a thorough discussion of the operation of the gold standard in Britain and other countries during the nineteenth century.

28. The modern version is known as the monetary theory of the balance of payments.

29. The belief appears to rest on statements like the one by Palmer that a 5 percent rate was an offer to lend as much as the market wished to borrow at that rate. Under the usury law, the bank could limit borrowing only by imposing quantitative restrictions once bank rate was at 5 percent.

30. Outstanding country notes were counted as currency for the first time. Further issues were banned, and outstanding issues had to be retired and replaced by Bank of England notes (Wood 1939, 111).

31. The price increase and gold flow were not entirely monetary in origin. The famous Irish potato blight required increased food imports, at higher prices, draining the gold stock.

32. A thorough analysis of the 1847 panic is Dornbusch and Frenkel 1984.

33. The act of 1844 required the bank to publish a weekly statement showing the principal assets and liabilities of the Issue and Banking Departments separately, so the bank’s situation was known within a few days.

34. Hetzel (1987) finds references to the lender of last resort function as early as 1797.

35. At about the same time, the Bank of France also began to change the discount rate more frequently.

36. The belief that the gold standard maintained price stability may have contributed to the error, but that would not explain why the error persisted. I am inclined to the view that central banks, influenced at first by the real bills doctrine and later by habit, looked mainly at money market responses and ignored longer-term effects of their actions.

37. See Bagehot 1962, 80–81 and appendix D. Bagehot recognized that as lender of last resort, the bank should not sell at a time of panic. Exactly one hundred years later, the Federal Reserve at last recognized similar responsibilities. Faced with the prospect of failures by savings and loan associations, the System authorized the reserve banks to lend to the associations if required to prevent failures. The Federal Reserve came to recognize its role as lender of last resort to the entire financial system and thus to the economy. For the contrast with System thinking in 1929, see below, chapter 5.

38. Bagehot (1962, 22) referred to evidence of the effect of interest rates on short-term capital movements: “If the interest of money be raised, it is proved by experience that money does come to Lombard Street, and the theory shows that it ought to come …as soon as the rate of interest shows that it can be done profitably” (italics in the original). Bagehot did not offer comparable evidence of the effect of price changes on the balance of trade, although he discussed some related matters on 69–78 and suggested there that the effect on prices would be delayed.

39. Note that Bagehot assigns priority to maintaining the exchange rate. See also Thornton 1965, 93–99, for a similar argument and analysis of the crisis of 1793. Charles Rist (1940, 404–6) quotes several Frenchmen (Thiers, Burdeau, Vuitry) who recognized, in the middle of the nineteenth century, the role of a central bank as lender of last resort and holder of the reserve and also recognized some of the policies required to carry out these functions.

40. Rist (1940, 291–97) and Hawtrey (1962, 227–31) discuss Marshall’s testimony but fail to see clearly the distinction he tries to draw between nominal and real rates. In the Treatise, Keynes also admits to finding Marshall’s statement to the commission unclear (Keynes 1930, 1:191–92). Fisher (1930b, 1:43 n. 7) recognizes Marshall’s proposition as a weak association between inflation and nominal interest rates.

41. I use his last complete statement (Fisher 1930b), but the argument is not very different from his 1896 tract or his 1907 book. The same argument is made in Fisher 1920, 56–58.

42. The conclusion rests partly on correlation between interest rates and price changes reported later in his book (Fisher 1930b, 410–11) and partly on his finding that nominal rates are less variable than calculated real rates (413–15). His empirical work does not take account of the constraint on sustained inflation imposed by the gold standard.

43. After praising central bankers’ decisions for their “singular judgment and moderation,” Bagehot calls their responses to questions about why they acted as they did “almost classical by their nonsense.” Bagehot (1962, 86) quotes from testimony of the bank’s directors in 1810: “I cannot see how the amount of bank-notes issued can operate upon the price of bullion, or the state of the exchanges; and therefore I am individually of the opinion that the price of bullion, or the state of the exchanges, can never be a reason for lessening the amount of bank-notes to be issued…. Is the Governor of the Bank of the same opinion which has now been expressed by the Deputy-governor? Mr. Whitmore: I am so much of the same opinion, that I never think it necessary to advert to the price of gold, or the state of the exchanges, on the days on which we make our advances.”

44. See Viner 1965, 150–51, for the quotation in the text, the comparison of Thornton and Ricardo on this point, and the views of the directors of the Bank of England.

45. My interpretation of Tooke is based on the discussion in Wood 1939, 56–58. A succinct statement of Sayers’s views is in Sayers 1957. This point and the notion that the relation between components of “liquidity” is ever changing catch essential points of Sayers’s argument on this issue. Most of these issues returned in the early years of the Federal Reserve.

46. Viner notes (1965, 246) that Mill included potential borrowing power as a part of credit.

47. The substance of banking school views on these matters (and on many others) is not strikingly different from the discussion of banking and financial innovation by many contemporary bankers and financial writers. Each major innovation in financial markets brings forth comments designed to establish that the central bank has lost (or will lose) control, so that monetary policy is “impossible” or ineffective. Recall the discussion of Eurodollars in the 1960s.

48. “For in the event of an inflation developing, the note issue is in modern conditions the latest phenomenon in point of time to exhibit symptoms of the disorder which is at work in the economic system. To attempt to maintain monetary health by regulating the volume of the note issue is like attempting to maintain physical health by ordering a drastic operation or amputation after the affliction has run its full course and mortification is setting in. For, generally speaking, the note-issue will not expand—for reasons other than increase in the volume of employment—until the inflationary influences have had time to raise the money rates of remuneration of the factors of production” (Keynes 1930, 2:273, 2:264).

49. Eccles’s statement is in the congressional hearings before the Banking Act of 1935. See, House Committee on Banking and Currency 1935, 321. Asked what would happen if the Federal Reserve printed currency and paid off debt, Eccles replied, “The currency would increase the reserves of the banking system …but the currency would immediately go into the banks and from the banks into the Federal Reserve banks—and you would have—additional excess reserves.”

50. Some, of course, recognized that among all the arguments, there were none that showed an increase in money would not raise prices. See Wood’s discussion of Thomas Attwood (1939, 52–53). Attwood wrote of a depression: “Let them [the public] be glutted with money. They will then seek prosperity and the prosperity of the country will return.” Attwood elsewhere recognized that an unanticipated fall in prices redistributed wealth and intensified depression because prices did not all fall at a uniform rate. Like Thornton, Attwood noted that wages were slow to fall and used this observation to explain unemployment of labor.