The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
M arch/April 1992 Vol. 74, No. 2 3 M on etary Policy in the Great Depression: W h at the Fed Did, and W h y 29 Seigniorage in the United States: H o w Much Does the U.S G overnm ent M ake fro m M on ey Production? 41 T h e FOMC in 1991: An Elusive R eco very 62 H ow T h e 1992 Legislation W ill A ffe c t European Financial Services THE FEDERAL A RESERVE I1VNR of A r ST.IjOHIS ERRATA In the first article o f this issue, "M one tary P olicy in the G reat D epression: W hat the Fed D id and W hy," the bodies o f figure 3 (U nem p loym ent Rate) and figure 4 (Bank Suspensions) should be sw itched. In addition, the "gold cur rency" line on figure 9 should read "gold m inus currency." W e regret these errors. 1 Federal R eserve Bank o f St. Louis R e v ie w March/April 1992 In T h is Is s u e . . . During the recent recession, there has been considerable discussion of the appropriate policy response by the Federal Reserve. Monetary policy often receives close scrutiny, and recently Congress has considered legislation that would reorganize the Fed’s policymaking structure. In the first article in this Review, "Monetary Policy During the Great Depression: What the Fed Did, and W hy,” David C. Wheelock examines the extent to which the Federal Reserve System's organization affected policy during the Great Depression. Some authors contend that the Fed’s organization caused it to be more receptive to private interests—or to the interests of policymakers wishing to extend their bureaucratic domain—than to the public’s interests. Others argue that leadership changes at the onset of the depression put authority into the hands of inexperienced officials who failed to understand the appropriate policies to counteract the depression. Wheelock finds, however, that organiza tion affected policy little during this episode. Rather, the Fed’s policies can be attributed largely to continued pursuit of a procyclical policy rule and to the gold standard regime, which proved deflationary. * * * In the second article in this issue, “ Seigniorage in the United States: How Much Does the U.S. Government Make from Money Production?” Manfred J.M. Neumann considers one of the oldest and most interesting issues in monetary economics, "seigniorage”—the revenue associated with the creation of money. The author extends a traditional measure of seigniorage with a new measure, "extended monetary seigniorage,” that he has developed. Neumann’s new measure shows the distribution of seigniorage between the central bank and the Treasury. Neumann calculates extended m o n etary seigniorage for the U nited States for the period 1951-90. He estimates that the Treasury's share of seigniorage, which he calls fiscal seigniorage, has amounted to between 1 percent and 2.8 percent of federal spending. He also examines the relationship between inflation and seigniorage and estimates that seigniorage in creases with inflation until the inflation rate reaches about 7 percent, then declines with further increases in the inflation rate. * * * In the third article in this issue, "The FOMC in 1991: An Elusive Recovery,” James B. Bullard presents an overview of recent actions taken by the Federal Open Market Committee, the arm of the Federal Reserve System with the primary responsibility for monetary policy. Since 1991 was a year that began with declines in aggregate economic activity and ended with some slight gains, this article provides a case study of policymaking during the recovery phase o f the business cycle. MARCH/APRIL 1992 2 In the context o f a chronology of FOMC decisionmaking, the author focuses on two key problems faced by the Committee. One is that, be cause of lags in data collection and the difficulty of forecasting, it is hard for the FOMC to assess the strength o f the economy at a point in time. The other is that the magnitude and even the direction o f policy thrust can be a matter of interpretation. The author shows how the FOMC grappled with these two problems through the year and ended up supporting relatively steady policies during the spring, when recov ery seemed likely, and relatively easy policies in the fall, when recovery seemed elusive. * * * In the final article in this issue, "How the 1992 Legislation Will Affect European Financial Services,” K. Alec Chrystal and Cletus C. Coughlin identify and examine the impact of the "1992” regulatory changes that pertain directly to banking and other financial services. The authors view the 1992 reforms as a small step toward the liberalization of the financial services sector. The reforms will prove to be beneficial, but the extent of the gains are unlikely to be large. The reason, they say, is that virtually all of the potential efficiency gains in the financial services sector can be (or have already been) achieved by abolishing exchange controls and allowing foreign firms to enter domestic markets. The key innovation of the 1992 legislation is the split between home country authorization and host country conduct of business rules. This dichotomy will create problems, especially regulatory complications. Whereas wholesale markets already are highly integrated, 12 quite different retail markets will continue to exist in the near future. The authors stress, however, that the advent later in this decade of a single currency for the European Community will cause pressures to revise the regulatory structure so that the conduct of business rules become homogeneous. FEDERAL RESERVE BANK OF ST. LOUIS 3 David C. W heelock David C. Wheelock, assistant professor of economics at the University of Texas-Austin, is a visiting scholar at the Federal Reserve Bank of St. Louis. David H. Kelly provided research assistance. Monetary Policy in the Great Depression: What the Fed Did, and W hy S IXTV YEARS AGO the United States— indeed, most o f the world—was in the midst of the Great Depression. Today, interest in the Depression’s causes and the failure o f govern ment policies to prevent it continues, peaking whenever the stock market crashes or the econ omy enters a recession. In the 1930s, dissatisfac tion with the failure of monetary policy to pre vent the Depression, or to revive the economy, led to sweeping changes in the structure of the Federal Reserve System. One of the most impor tant changes was the creation of the Federal Open Market Committee (FOMC) to direct open market policy. Recently Congress has again con sidered possible changes in the Federal Reserve System.1 This article takes a new look at Federal Reserve policy in the Great Depression. Historical analy sis of Fed performance could provide insights into the effects o f System organization on policy making. The article begins with a macroeconomic overview of the Depression. It then considers both contemporary and modern views of the 1“ The Monetary Policy Reform Act of 1991” (S. 1611) would have abolished the FOMC and thereby ended the voting on open market policy by Federal Reserve Bank presidents. Although hearings on the bill were held, it was not brought to a vote before Congress adjourned at the end of 1991. The Banking Act of 1935 established the role of monetary policy in causing the Depression and the possibility that different policies might have made it less severe. Much of the debate centers on whether mone tary conditions w ere "easy” or “tight” during the Depression—that is, whether money and credit were plentiful and inexpensive, or scarce and expensive. During the 1930s, many Fed officials argued that money was abundant and “cheap,” even "sloppy,” because market interest rates w ere low and few banks borrowed from the dis count window. Modern researchers who agree generally believe neither that monetary forces were responsible for the Depression nor that different policies could have alleviated it. Others contend that monetary conditions were tight, noting that the supply of money and price level fell substantially. They argue that a more aggres sive response would have limited the Depression. Among those who conclude that contraction ary monetary policy worsened the Depression, there has been considerable debate about why present form of the FOMC, whose members include the Board of Governors of the Federal Reserve System and the 12 Reserve Bank presidents. Five of the presidents vote on policy on a rotating basis. MARCH/APRIL 1992 4 Federal Reserve officials failed to respond ap propriately. Most explanations fall into two cate gories. One holds that Fed officials, though wellintentioned, failed to understand that more ag gressive action was needed. Some researchers, like Friedman and Schwartz (1963), argue that the Fed’s behavior during the Depression con trasted sharply with its behavior during the 1920s. They contend that the death of Benjamin Strong in 1928 led to a redistribution of authori ty within the System that caused a distinct de terioration in Fed performance. Strong, who was Governor of the Federal Reserve Bank of New York from the System's founding in 1914 until his death, dominated Federal Reserve policy making in the years before the Depression.2 These researchers argue that authority was dis persed after his death among the other Reserve Banks, whose officials were less knowledgeable and failed to recognize the need for aggressive policies. Other researchers, like Wicker (1966), Brunner and Meltzer (1968), and Temin (1989), contend that Strong’s death caused no change in Fed performance. They argue that Strong had not developed a countercyclical policy and that he would have failed to recognize the need for vigorous action during the Depression. In their view, Fed errors were not due to organizational flaws or changes, but simply to continued use of flawed policies. A second category of explanations holds that the Fed’s contractionary policy was deliberate. Epstein and Ferguson (1984) and Anderson, Shughart and Tollison (1988) contend that Fed officials understood that monetary conditions were tight. Epstein and Ferguson assert that the Fed believed a contraction was necessary and inevitable. When it did act, they argue, it was to promote the interests of commercial banks, rather than economic recover}'. Anderson, Shughart and Tollison emphasize even more the 2Until changed by the Banking Act of 1935, the chief ex ecutive officers of the Reserve Banks held the title “ gover nor.” Today these officers are titled “ president,” while members of the Board of Governors, which replaced the Federal Reserve Board in 1935, now hold the title “ governor.” 3The appendix provides a list of sources for the data used in this article. The GNP and unemployment series used here are standard, but Romer (1986a, 1986b) presents new estimates of GNP and unemployment for the 1920s. Both new estimates exhibit less variability than those tradi tionally used; Romer’s estimate of the unemployment rate in 1929 is 4.6 percent, compared with 3.2 percent plotted here. FEDERAL RESERVE BANK OF ST. LOUIS Fed's interest in aiding its member banks. They argue that monetary policy was designed to cause the failure of nonmember banks, which would enhance the long-run profits of member banks and enlarge the System's regulatory domain. AN OVERVIEW OF THE GREAT DEPRESSION Analysts generally agree that the economic collapse of the 1930s was extremely severe, if not the most severe in American history. To provide a sense of the Depression, Figures 1-3 plot GNP, the price level and the unemployment rate from 1919 to 1939. As the figures show, af ter eight years of nearly continuous expansion, nominal (current dollar) GNP fell 46 percent from 1929 to 1933. Real (constant dollar) GNP fell 33 percent and the price level declined 25 percent. The unemployment rate went from un der 4 percent in 1929 to 25 percent in 1933.3 Real GNP did not recover to its 1929 level until 1937. The unemployment rate did not fall below 10 percent until W orld W ar II.4 Few segments of the economy were unscathed. Personal and firm bankruptcies rose to unpre cedented highs. In 1932 and 1933, aggregate corporate profits in the United States were negative. Some 9,000 banks, with $6.8 billion of deposits, failed between 1930 and 1933 (see figure 4). Since some suspended banks eventually reopened and deposits were recovered, these figures overstate the extent of the banking dis tress.5 Nevertheless, bank failures were numer ous and their effects severe, even compared with the 1920s, when failures were high by modern standards. Much of the debate about the causes of the Great Depression has focused on bank failures. 4Darby (1976) argues that the unemployment rate series considerably overstates the true rate after 1933 because it takes persons employed on government relief projects as unemployed. Kesselman and Savin (1978) offer an oppos ing view. Regardless of which argument is accepted, un employment during the 1930s was exceptionally severe, particularly since there were relatively few multi-income households. 5There was no deposit insurance in these years. The Bank ing Act of 1933 created federal deposit insurance. During the 19th and early 20th centuries a number of states ex perimented with insurance plans for their state-chartered banks, but none was still in existence by 1930. See Calomiris (1989) for a survey of the state systems. 5 Figure 1 Nominal and Real Gross National Product B illio n s o f do llars Figure 2 Implicit Price Index 1929 = 100 MARCH/APRIL 1992 6 Figure 3 Unemployment Rate Number of banks 4000 3500 3000 2500 2000 1500 1000 500 1920 1925 W ere they merely a result of falling national in come and money demand? Or were they an im portant cause of the Depression? Most contempo raries viewed bank failures as unfortunate for those who lost deposits, but irrelevant in macroeconomic significance. Keynesian explanations of the Depression agreed, including little role for bank failures. Monetarists like Friedman and Schwartz (1963), on the other hand, contend that banking panics caused the money supply to fall which, in turn, caused much of the decline in economic activity. Bernanke (1983) notes that bank failures also disrupted credit markets, which he argues caused an increase in the cost of credit intermediation that significantly reduced national output. In these explanations, the Federal Reserve bears much of the blame for the Depression because it failed to prevent 6See Belongia and Garfinkel (forthcoming). FEDERAL RESERVE BANK OF ST. LOUIS 1930 1935 the banking panics and money supply con traction. THE ROLE OF MONETARY POLICY: ALTERNATIVE VIEWS Today there is considerable debate about the causes of business cycles and whether government policies can alleviate them.6 Just as there is no consensus now, contemporary observers had many different views about the causes of the Great Depression and the appropriate response of government. A few economists, like Irving Fisher (1932), applied the Quantity Theory of Money, which holds that changes in the money supply cause changes in the price level and can affect the level of economic activity for short periods. These economists argued that the Fed should prevent deflation by increasing the money supply. At the 7 Figure 4 Bank Suspensions Percent 30,________________________________ 25 20 15 10 5 0 --------- L I----U L ---- I----U ----U ----U ----U ----L U I— ----U U — ----U U— U U— U— — — 1920 1925 other extreme, proponents of "liquidationist” the ories of the cycle argued that excessively easy monetary policy in the 1920s had contributed to the Depression, and that "artificial” easing in response to it was a mistake. Liquidationists thought that overproduction and excessive bor rowing cause resource misallocation, and that depressions are the inescapable and necessary means of correction: In the course of a boom many bad business commitments are undertaken. Debts are in curred which it is impossible to repay. Stocks are produced and accumulated which it is im possible to sell at a profit. Loans are made which it is impossible to recover. Both in the sphere of finance and in the sphere of produc tion, when the boom breaks, these bad commit ments are revealed. Now in order that 1930 1935 revival may commence again, it is essential that these positions should be liquidated. . . ,7 One implication o f the liquidationist theory is that increasing the money supply during a recession is likely to be counterproductive. Dur ing a minor recession in 1927, for example, the Fed had made substantial open market pur chases and reduced its discount rate. Adolph Miller, a member of the Federal Reserve Board, who agreed with the liquidationist view, testi fied in 1931 that: It [the 1927 action] was the greatest and boldest operation ever undertaken by the Federal Reserve System, and, in my judgment, resulted in one of the most costly errors committed by it or any banking system in the last 75 years. I am inclined to think that a different policy at that time would have left us with a different 7Lionel Robbins, The Great Depression (1935) [quoted by Chandler (1971), p. 118], MARCH/APRIL 1992 8 condition at this time. . . . That was a time of business recession. Business could not use and was not asking for increased money at that time.8 In Miller’s view, because economic activity was low, the reserves created by the Fed’s actions fueled stock market speculation, which led in evitably to the crash and subsequent depression. During the Depression, proponents of the liquidationist view argued against increasing the money supply since doing so might reignite speculation without promoting an increase in real output. Indeed, many argued that the Fed eral Reserve had interfered with recovery and prolonged the Depression by pursuing a policy o f monetary ease. Hayek (1932), for example, wrote: It is a fact that the present crisis is marked by the first attempt on a large scale to revive the economy. . . by a systematic policy of lowering the interest rate accompanied by all other possi ble measures for preventing the normal process of liquidation, and that as a result the depres sion has assumed more devastating forms and lasted longer than ever before (p. 130). Several key Fed officials shared Hayek's views. For example, the minutes o f the June 23, 1930, meeting of the Open Market Committee report the views of George Norris, Governor of the Federal Reserve Bank o f Philadelphia: He indicated that in his view the current busi ness and price recession was to be ascribed largely to overproduction and excess productive capacity in a number of lines of business rather than to financial causes, and it was his belief that easier money and a better bond market would not help the situation but on the con trary might lead to further increases in produc tive capacity and further overproduction.9 While the liquidationist theory of the business cycle was commonly believed in the early 1930s, 8U.S. Senate (1931), p. 134. 9Quoted by Chandler (1971), pp. 136-37. De Long (1990) details the liquidationist cycle theory, and Chandler (1971), pp. 116-23, has a general discussion of prevailing busi ness cycle theories and their prescriptions for monetary policy. 10See Temin (1976) for a survey of Keynesian explanations of the Great Depression. FEDERAL RESERVE BANK OF ST. LOUIS it died out quickly with the Keynesian revolu tion, which dominated macroeconomics for the next 30 years. Keynesian explanations of the Depression differed sharply from those o f the liquidationists. Keynesians tended to dismiss monetary forces as a cause of the Depression or a useful remedy. Instead they argued that declines in business investment or household consumption had reduced aggregate demand, which had caused the decline in economic ac tivity.1 Both views, however, agreed that mone 0 tary ease prevailed during the Depression. Friedman and Schwartz renewed the debate about the role of monetary policy by forcefully restating the Quantity Theory explanation of the Depression: The contraction is. . . a tragic testimonial to the importance of monetary forces. . . . Different and feasible actions by the monetary authorities could have prevented the decline in the stock of money. . . [This] would have reduced the con traction’s severity and almost as certainly its du ration (pp. 300-01). Friedman and Schwartz argue that an increase in the money stock would have offset, if not prevented, banking panics, and would have led to increased lending to consumers and business that would have revived the economy. Many disagree with the Friedman and Schwartz explanation, although some recent Keynesian ex planations concede that restrictive monetary policy did play a role in the Depression.1 Other 1 studies, such as Field (1984), Hamilton (1987), and Temin (1989), conclude that contractionary monetary policy in 1928 and 1929 contributed to the Depression. Bordo (1989) and Wicker (1989) provide detailed surveys of the monetaristKeynesian debate about the causes of the Great Depression, and interested readers are referred to them. Since most recent contributions to this literature emphasize the effects o f monetary policy, a new look at the policies of the Federal Reserve during the Great Depression is war ranted. "M o s t criticize the Fed’s discount rate increases and failure to replace reserve losses suffered by banks in the panic following Great Britain’s departure from the gold standard in late 1931. See Temin (1976), p. 170, and Kindleberger (1986), pp. 164-67. 9 Figure 5 Interest Rates 1920 1925 WERE MONETARY CONDITIONS EASY OR TIGHT? A fundamental disagreement within the Feder al Reserve System and among outside observers, even today, is whether monetary policy during the Depression was easy or tight. Most Fed offi cials felt that money and credit were plentiful. Short-term market interest rates fell sharply af ter the stock market crash of 1929 and remained at extremely low levels throughout the 1930s (see figure 5). To most observers, the decline in short-term rates implied monetary ease. Long term interest rates declined less sharply, however, and yields on risky bonds, such as 12The short-term rate series through 1933 is the average daily yield in June of each year on three- to six-month Treasury notes and certificates, and the yield on Treasury bills thereafter. The long-term series is the average daily 1930 1935 Baa-rated bonds, rose during the first three years of the Depression (see figure 5).1 Never 2 theless, the exceptionally low yields on short term securities has suggested to many observers an abundance of liquidity. Other variables also have been interpreted as indications of easy monetary conditions. Rela tively few banks came to the Fed’s discount window to borrow reserves, for example, and many banks built up substantial excess reserves as the Depression progressed (see figure 6).1 To 3 most observers, it appeared that there was little demand for credit and, since most policymakers saw their mission as one of accommodating yield in June of each year on U.S. government bonds. 13Data on excess reserves before 1929 are not available, but they were not likely very large. MARCH/APRIL 1992 10 Figure 6 Borrowed and Excess Reserves of Federal Reserve Member Banks Millions of dollars 1100 ing credit demand, few believed that more vigorous expansionary actions were necessary.1 4 Low interest rates and an apparent lack of de mand for reserves have led many researchers to conclude that tight money did not cause the Depression. Temin (1976), for example, writes: There is no evidence of any effective deflation ary pressure from the banking system between the stock-market crash in October 1929 and the British abandonment of the gold standard in 14The Federal Reserve System’s founders intended that it operate according to the Real Bills Doctrine. Fed credit would be extended primarily through the discount window as member banks borrowed to finance short-term agricul tural or business loans. A decline in economic activity would reduce discount window borrowing, causing Federal Reserve credit to decline. By 1924, System policy had evolved away from a strict Real Bills interpretation, but it FEDERAL RESERVE BANK OF ST. LOUIS September 1931. . . . There was no rise in short term interest rates in this two-year period. . . . The relevant record for the purpose of identify ing a monetary restriction is the record of short-term interest rates (p. 169). Other indicators of monetary conditions, however, suggest the opposite conclusion. Defla tion implied that the value o f the dollar rose 25 percent from 1929 to 1933, which Schwartz probably continued to have considerable influence on many Fed officials. See West (1977) or Wicker (1966) for discussion of the influence of the Real Bills Doctrine on policy over time. 11 Figure 7 Money Supply Millions of dollars (1981) argues reflected exceptionally tight money. Another indicator, the money stock, fell by one-third from 1929 to 1933 (see figure 7).1 5 Friedman and Schwartz contend that: It seems paradoxical to describe as ‘monetary ease’ a policy which permitted the stock of money to decline. . . by a percentage exceeded only four times in the preceding fifty-four years and then only during extremely severe business-cycle contractions (p. 375). justed for changes in the price level, rose shar ply during the Depression (see figure 8).1 While 6 the nominal yield on short-term government securities fell to an exceptionally low level, deflation implied that their real yield rose above 10 percent in 1930 and 1931. Thus, in contrast to the apparent signal given by nominal interest rates, member bank borrowing and excess reserves, the falling money stock and deflation suggest that monetary conditions were far from easy.1 7 And finally, numerous studies point out that the real interest rate, that is, the interest rate ad Many economists now conclude that the Fed eral Reserve should have responded more 15M1 is the sum of coin and currency held by the public and demand deposits. M2 also includes time deposits at com mercial banks. 17Yet another indicator is the real money supply, i.e., the growth rate of the nominal supply of money less the ex pected rate of inflation. Since the price level fell faster than the nominal supply of money (M1 or M2) during the first two years of the Depression, Temin (1976) argues that monetary conditions were not tight. The increase in real money balances was relatively slow, however, which Hamilton (1987) argues was contractionary. 16See Meltzer (1976) and Hamilton (1987), for example. The real interest rate plotted in figure 8 is calculated as the prevailing yield on short-term government securities in June of each year, less the rate of inflation in the subse quent year. Since actual, rather than anticipated, inflation is used to calculate the real rate, it is considered an ex post, rather than ex ante, rate. MARCH/APRIL 1992 12 Figure 8 Ex Post Real Interest Rate Percent vigorously tc the Depression. There is little agreement, however, about why the Fed did not. The next sections examine alternative ex planations for Federal Reserve behavior during the Depression. THE IMPACT OF STRONG’S DEATH: ALTERNATIVE VIEWS Irving Fisher testified before Congress in 1935 that the Depression was severe because “Gover nor Strong had died and his policies died with him. . . . I have always believed, if he had lived, w e would have had a different situation.”1 Ac 8 cording to Fisher, Benjamin Strong had disco vered how to use monetary policy to maintain 18U.S. House of Representatives (1935), p. 534. FEDERAL RESERVE BANK OF ST. LOUIS price level stability, “and for seven years he maintained a fairly stable price level in this country, and only a few of us knew what he was doing. His colleagues did not understand it.”1 In Fisher's view, Strong adjusted the quan 9 tity of money to maintain a stable price level; had he lived, Fisher says, he would have prevented the deflation of the 1930s by not al lowing the quantity o f money to decline. Friedman and Schwartz agree with Fisher that Strong’s death caused monetary policy to change significantly. They argue that Strong’s aggressive open market purchases and discount rate reductions in 1924 and 1927 had quickly al leviated recessions, but that his death produced a sharply different policy during the Depression: 19lbid, pp. 517-20. 13 If Strong had still been alive and head of the New York Bank in the fall of 1930, he would very likely have recognized the oncoming li quidity crisis for what it was, would have been prepared by experience and conviction to take strenuous and appropriate measures to head it off, and would have had the standing to carry the System with him (pp. 412-13). Friedman and Schwartz make a persuasive case. Strong was an experienced financial lead er. He had served as an officer of Bankers Trust Company, and during the Panic of 1907 as head of a committee reporting to J. P. Mor gan that determined which financial institutions could be rescued.2 He was the first governor of 0 the Federal Reserve Bank of New York and emerged as leader of the Federal Reserve Sys tem both because of his personality and stature in the financial community and because o f the relative importance of New York member banks in the international financial market.2 He chaired 1 a committee of Federal Reserve Bank governors that coordinated System open market operations and represented the System in dealings with foreign central banks and Congress.2 It is clear 2 that, with his death, the Fed lost an experienced and forceful leader. Some researchers argue, however, that Strong’s death had little effect on policy. Temin (1989), for example, writes that "The death of Strong was a minor event in the history of the Great Depression" (p. 35). And Brunner and Meltzer (1968) argue that, "While there is some evidence that the death of Benjamin Strong con tributed to a shift in the balance of power w i thin the Federal Reserve. . . we find that a special explanation o f monetary policy after 1929 is unnecessary. . .” (p. 341). The disagree ment between these authors and those such as Fisher, Friedman and Schwartz rests on their views of whether Strong’s policies would have prevented the monetary collapse and Depression. 20Chandler (1958), pp. 27-28. 21The Federal Reserve Act gave the individual Reserve Banks authority to initiate discount rate changes and open market operations. The Federal Reserve Board could ap prove or disapprove these actions, but its role was primari ly supervisory, with no clear authority to determine policy. Because of this, and perhaps because it lacked forceful leaders, the Board did not dominate policy making until af ter the System was restructured by the Banking Act of 1935. See Wheelock (forthcoming) for details of this reor ganization. W H AT WAS STRONG'S POLICY? Much of Strong’s testimony before Congres sional committees, as well as other speeches and writings, suggests that he had developed a poli cy of money supply control to limit fluctuations in the price level. For example, in an unpub lished article dated April 1923, he wrote: “If, as is now universally admitted, prices are in fluenced to advance or to decline by increases or decreases in the total of 'money’. . . then the task of the System is to maintain a reasonably stable volume of money and credit. . . ”2 And, 3 in a speech to the American Farm Bureau in December 1922, he said that monetary policy: . . .should insure that there is sufficient money and credit available to conduct the business of the nation and to finance not only the seasonal increases in demand but the annual or normal increase in volume. . . . I believe that it should be the policy of the Federal Reserve System, by the employment of the various means at its command, to maintain the volume of credit and currency in this country at such a level so that, to the extent that the volume has any influence upon prices, it cannot possibly become the me ans for either promoting speculative advances in prices, or of a depression of prices.2 4 These statements suggest that Strong would not have permitted the money supply collapse or deflation that occurred after 1929. Other aspects of Strong's testimony, speeches and writings give different or ambiguous im pressions of his views, however, making it difficult to infer what policies he would have advocated during the Depression. In testimony before the House Banking Committee in 1926, Strong described the relationship between Fed policy and the quantity of bank deposits, dis cussing in detail the multiplier relationship be tween bank reserves and deposits.2 But he also 3 they made it difficult to price new debt issues and a grow ing understanding of the impact of open market operations on economic activity, led the Banks to form a “ Governors Committee” to coordinate open market operations. This committee was replaced in 1923 by the Open Market In vestment Committee, which Strong headed until his death. 23“ Prices and Price Control,” in Burgess (1930), pp. 229-30. 24Quoted by Chandler (1958), p. 200. 25U.S. House of Representatives (1926), pp. 334-35. 22lnitially, each Reserve Bank determined its own open mar ket operations. But Treasury Department complaints that MARCH/APRIL 1992 14 testified that, "when it comes to a decline of price level, the origin of which can not be at tributed to a credit policy, this effort that you make by a credit policy to arrest a fall of prices may do more harm than good. . . ,”2 It is also 6 difficult to interpret his writing that "the task of the System is to maintain a reasonably stable volume of money and credit, with d u e al low an ces f o r sea so n a l flu ctu a tio n s in d em an d, f o r n orm al annual g row th in the country's d ev elo p m ent. . . an d with su ch allow an ce a s m ay b e im p o s e d b y th o se g reat cycles o f p ro sp erity an d d ep ressio n . . . .”2 What sort o f allowance for 7 fluctuations does he mean? This statement could be read as advocating an increase or a d e c rea se in money in response to a decline in economic activity. The latter is suggested by the following statement: "there should be no such excessive or artificial supplies of money and credit as will simply permit the marking up of prices when there is no increase in business or production to warrant an increase in the volume of money and credit.”2 This sounds like the warnings by 8 some officials during the Depression that mone tary expansion would be inflationary or cause speculation b e c a u s e economic activity was low. Strong also seems to have concluded that the deflation from mid-1920 to 1921 had positive effects: The deflation which took place in the United States following the collapse of prices resulted in extricating the reserve system—the whole monetary system of the country—from a posi tion of permanent entanglement. . . and I think that was one of the fortunate results of the policy. . . . One of the results of this liquida tion. . . has been to put this country on as sound or a sounder monetary basis than any other country in the world, without the in troduction of a lot of money or credit into cir culation, based solely upon the Government debt to the bank of issue. I mean to explain that there have been offsetting advantages to that deflation. . . ,2 9 26lbid, p. 577. 27“ Prices and Price Control,” April 1923, in Burgess (1930), p. 230 (italics added). 28From a speech to the American Farm Bureau in 1922 [quoted by Chandler (1958), p. 200]. 29U.S. House of Representatives (1926), p. 309. 3°Federal Reserve credit is supplied by Fed purchases of securities and discount window lending (member bank bor rowing). It consists also of some miscellaneous compo- FEDERAL RESERVE BANK OF ST. LOUIS This quotation suggests that Strong might have found similar offsetting advantages to the defla tion that followed the stock market crash in 1929 and might have been reluctant to expand the money supply through purchases of govern ment securities. These quotations illustrate the ambiguity of many of Strong's statements and the difficulty of inferring what policies he would have pur sued in the 1930s. To determine whether mone tary policy was changed by Strong's death, it is probably more instructive to examine the poli cies he actually implemented. Tw o aspects of Strong’s policies have received attention from scholars studying Federal Reserve behavior. First, beginning in the early 1920s, the System offset or "sterilized” gold flows and other changes in reserve funds by al tering the volume o f Fed credit outstanding.3 0 This policy limited fluctuations in bank reserves and, thus, in the money supply and price level. According to Friedman and Schwartz (1963), pp. 394-99, however, the Fed permitted gold out flows during the Depression to reduce bank reserves and the money supply and more than offset gold inflows. What had been an essential ly neutral policy, therefore, became a contrac tionary policy after Strong's death. Miron (1986) argues that a similar change oc curred in the Fed’s accommodation of seasonal credit and currency demands. From the Sys tem's inception, Federal Reserve credit was sup plied to prevent seasonal demands from draining bank reserves and increasing interest rates. According to Miron, the Fed was less ac commodative after 1928, which contributed to the frequency o f financial crises during the Depression.3 1 Beyond the offsetting o f gold and currency flows, a second aspect of Strong’s policies has received considerable attention. In 1924 and 1927, the Fed made large open market pur chases and discount rate reductions that were followed by increases in bank reserves and the nents, such as float. In this era, the Federal Reserve pur chased both U.S. government securities and bankers ac ceptances (at fixed acceptance buying rates), and discount window lending consisted of both rediscount of eligible paper and advances to member banks at the discount rate. 3 Miron does not test this claim except to show that Federal 1 Reserve credit was somewhat less seasonal after 1928 than before. 15 money supply. Friedman and Schwartz (1963) argue that the Fed's purpose was to combat recessions and that its failure to respond as ag gressively during the Depression reflected a dis tinct change in System behavior. Other researchers, however, such as Wicker (1966) and Brunner and Meltzer (1968), find no incon sistency in Fed behavior, arguing that the com paratively weak response to the Depression was in fact predictable from the policy strategy de veloped by Strong. The Sterilization P olicy Before entering W orld W ar I, the United States absorbed large gold inflows that added directly to bank reserves and caused a signifi cant money supply increase.3 Although inflows 2 ceased after America entered the war, bank reserves and the money supply continued to in crease rapidly as Federal Reserve credit was ex tended to help finance the war. After the war, gold outflows reduced the reserves of the Reserve Banks, leading them to raise their dis count rates and thereby restrict credit to mem ber banks.3 The resulting decline in Fed credit 3 coincided with a sharp decline in the money supply and deflation.3 4 Following the violent inflation-deflation cycle of 1917-21, the Fed began to intervene to pre vent gold flows from affecting bank reserves.3 5 In testimony before the House Committee on Banking and Currency, Strong gave a clear ex planation of this policy, presenting charts show ing the relationship between gold flows, Fed credit, bank reserves and the price level.3 He 6 explained: In the old days there was a direct relation be tween the country's stock of gold, bank deposits and the price level because bank deposits 32The accompanying shaded insert discusses the sources and uses of reserve funds and explains the mechanics of the Fed’s sterilization policy. 33The Reserve Banks were required to maintain gold reserves of 40 percent against their note issues and 35 percent against deposits. A discount rate increase was in tended to reduce discount loans and, thus, the Fed’s note and deposit liabilities, as well as encourage gold inflows as investors sought higher yields in the United States. 34ln January 1916, the All Commodities Price Index stood at 112.8 (1913=100). In April 1917 (when the United States entered the war), it was at 172.9. At its peak in May 1920, the Index was at 246.7. It then fell to a low of 138.3 in January 1922. 35Before the war, the Fed lacked the resources to offset gold inflows, so sterilization was impossible. By the end of 1921, the Reserve Banks had sufficient reserves to reduce were. . . based upon the stock of gold and bore a constant relationship to the gold stock, and the volume of bank deposits and the general price level were similarly related. But in recent years the relationship between gold and bank deposits is no longer as close or direct as it was, because the Federal Beserve System has given elasticity to the country's bank reserves. Reserve Bank credit has become the equivalent of gold in its power to serve as the basis of bank credit. . . . Hence. . . the present basis for bank credit consists of gold plus Federal Reserve credit. Federal Reserve bank credit is an elastic buffer between the country’s gold supply and bank credit.3 7 Strong credited the Federal Reserve System for preventing inflation in 1921 and 1922: As the flow of gold imports was pouring into the United States in 1921 and 1922, many economists abroad, and in this country as well, expected that this inward flow of gold would result in a huge credit expansion and a serious price inflation. That no such expansion or infla tion has taken place is due to the fact that the amount of Federal Reserve credit in use was diminished as the gold imports continued. Thus, in the broad picture of financial events in this country since 1920, the presence of the Reserve System may be said to have prevented rather than fostered inflation.3 8 Figure 9 illustrates the policy of offsetting gold and currency flows during the 1920s.3 9 The shaded insert on pages 18-19 describes the mechanics of this policy. Since gold is a source of banking system reserves, gold inflows, unless offset, add to the stock of reserves. A gold in flow thus has the same effect on reserves as a their discount rates, and individually they began to pur chase government securities. By 1923, there seems to have been a conscious effort to offset gold flows [Fried man and Schwartz (1963), pp. 279-87]. 36These charts are reproduced by Hetzel (1985), p. 7, who examines Strong’s unwillingness to support legislation that would require the Fed to adopt a price level stabilization rule. 37U.S. House of Representatives (1926), p. 470. 38lbid, p. 471. 39ln practice, the Fed also offset changes in other sources and uses of reserve funds, but gold and currency flows were the most substantial; the others can be ignored for il lustrative purposes. MARCH/APRIL 1992 16 Figure 9 Gold and Currency Sterilization Millions of dollars Federal Reserve purchase of securities. Currency held by the public is a use of reserves: in creases in public currency holdings reflect reserve withdrawals from banks. Thus, if not offset, an increase in currency would correspond to a decrease in bank reserves. The difference between gold and currency is plotted in figure 9. It is clear that net increases (decreases) in this difference were largely offset by declines (in creases) in Fed credit outstanding, so that total bank reserves changed relatively little. It is also clear that Benjamin Strong’s death did not interrupt the offsetting of gold and cur 40The multiplier plotted in figure 10 equals (1 +k)/(r + k), where k is the ratio of currency held by the public to de mand deposits and r is the ratio of bank reserves to deposits. The multiplier is defined as the money supply (here M2) divided by the monetary base, or “ high-powered money,” which is the sum of bank reserves and currency held by banks and the public. FEDERAL RESERVE BANK OF ST. LOUIS rency flows, at least until the fourth quarter of 1931. The money supply contraction and defla tion during the first two years o f the Depres sion were not caused by a decline in bank reserves. Instead, as figure 10 illustrates, the money supply fell because the money multiplier declined.4 This was particularly true beginning 0 in the fourth quarter of 1930, when banking panics caused marked increases in the currency-deposit and reserve-deposit ratios.4 1 The relative stability o f bank reserves ended abruptly in September 1931. On September 21, Great Britain left the gold standard. Speculation 4 Friedman and Schwartz (1963), pp. 340-42, conclude that 1 the money supply decline between August 1929 and Oc tober 1930 was caused by a decline in the monetary base This decline was due to a decrease in currency, not bank reserves. Thereafter, the base rose, but less than neces sary to offset the sharp decline in the multiplier. 17 Figure 10 Money Supply and Base Multiplier Multiplier M illions o f dollars January 1929 to February 1933 50.000 7 _____ 48.000 Multiplier 46.000 44.000 42.000 40.000 38.000 36.000 34.000 32.000 I 1929 I 1930 I 1931 I 1932 1933 I Monthly Data that the United States would soon follow led to a large withdrawal of foreign deposits from American banks and a consequent gold out flow.4 In the six weeks ending October 28, 2 1931, the gold stock declined $727 million (15 percent).4 The Fed raised its discount and ac 3 ceptance buying rates, hoping that an increase in domestic interest rates would halt the gold outflow by raising the relative yield of U.S. financial assets. This action was hailed as demonstrating the Fed’s resolve to maintain gold convertibility of the dollar, and the gold outflow ceased. 42lf the United States had left the gold standard, it is likely that the dollar would have depreciated against gold and other currencies that remained linked to gold. This would have meant an immediate loss of wealth in terms of gold for anyone holding dollar-denominated assets. Banks continued to lose reserves, however, as depositors panicked and converted deposits into currency. Member banks were able to partially offset the reserve outflows by borrowing and by selling acceptances to the Reserve Banks, al beit at the recently increased discount and ac ceptance buying rates. But the Fed made only trivial purchases o f government securities, and, in all, Federal Reserve member banks suffered a $540 million (22 percent) loss of reserves be tween September 16, 1931, and February 24, 1932.4 4 44Non-borrowed reserves declined $1112 million (52 percent), while discount loans (borrowed reserves) increased $572 million, 43Board of Governors of the Federal Reserve System (1943), p. 386. MARCH/APRIL 1992 18 The Federal Reserve Balance Sheet and Reserve Sterilization A simplified version o f the Federal Reserve System’s balance sheet on December 31, 1929, is shown below. The principal assets of the Federal Reserve were its gold and cash reserves and Fed credit outstanding. The lat ter consisted of member bank borrowing (bills discounted),1 bankers acceptances held by the Reserve Banks (bills bought),2 U.S. government securities held by the Reserve Banks and a miscellaneous component, made up primarily by float. The principal liabilities of the System were Federal Reserve notes outstanding, deposits of member banks, and deposits of the U.S. Treasury and others, such as foreign central banks. Most System transactions involve member commercial banks and directly affect member bank reserves. If the Fed makes an open market purchase o f government securities from a member bank, for example, it pays for the securities by crediting the member bank's deposit with the Federal Reserve. Since a deposit at the Fed is the principal form in which banks hold their legal reserves,3 an open market purchase adds directly to bank reserves.4 Many Federal Reserve transactions are in itiated by commercial banks. When the Unit ed States was on the gold standard, the Fed held substantial gold reserves, and transac tions in gold w ere common. For example, suppose gold coin was deposited by a cus tomer of a member bank. The bank could send the coin to its Federal Reserve Bank and receive an increase in its reserve deposit of that amount. The Fed’s gold reserves and member bank deposits would increase by the same amount. Suppose instead that a member bank was experiencing large cash withdraw als and needed extra currency. It could re quest currency, in the form of Federal Reserve notes, from its Reserve Bank and pay for the currency with a reduction in its reserve deposit. Hence, as Federal Reserve notes outstanding increased, bank reserves would decline by the same amount. The Fed could offset, or "sterilize,” the im pact of one transaction on bank reserves FEDERAL RESERVE BANK OF ST. LOUIS Federal Reserve System Balance Sheet December 31,1929 (billions of dollars) Assets Gold and cash reserves Federal Reserve credit Bills discounted Bills bought Government securities Other Other assets Total assets Liabilities and Capital Federal Reserve notes Deposits Member bank Other Other liabilities Capital accounts Total liabilities and capital $3.01 1.58 $0.63 0.39 0.51 0.05 0.87 $5.46 $1.91 2.41 $2.36 0.05 0.69 0.45 $5.46 with a second transaction having the opposite impact on reserves. For example, if the Fed sold government securities in the amount of a gold inflow, there would be no net change in aggregate member bank reserves. The open market sale would reduce reserves just as the gold inflow added to them, leaving no net reserve change. Similarly, a bank could borrow reserves from its Reserve Bank to pay for Federal Reserve notes needed to satis fy withdrawal demands, and thus avoid drawing down its reserve deposit. In this case, Federal Reserve credit (bills discounted) would increase by the amount of the increase in Federal Reserve notes outstanding, and bank reserves would not change. Note that, in this case, the Fed did not initiate the o ff setting transaction. Indeed, much of the sterilization of gold and currency flows dur ing the 1920s and early 1930s was at the in itiative of member banks, although it was definitely the Fed’s intent that sterilization occur. Federal Reserve sterilization of gold and currency flows from January 1924 to Febru ary 1933 is illustrated in figure 9. Note that increases (decreases) in Federal Reserve 19 credit accompanied declines (increases) in the net of gold and currency outstanding, and thus bank reserves changed comparatively lit tle. In 1930, for example, member bank reserves rose from $2395 million in Decem ber 1929 to $2415 million in December 1930, an increase of just $20 million. Over the same 'Loans to member banks consisted of discounts and ad vances. Many commercial loans, which often were called “ bills,” were made on a discount basis; hence, when they were endorsed by a bank and sent to a Reserve Bank in exchange for reserve balances, they were “ re-discounted” by the Reserve Bank at the prevailing discount rate. Alternatively, bills were used as collateral for direct advances to member banks, hence the term “ bills discounted.” 2The terminology is confusing because “ bills” in this case refer to bankers acceptances, not to the promisso ry notes that member banks used as collateral for dis count loans. months, there was an increase of $259 mil lion in the monetary gold stock and a $120 million decline in currency in circulation. The gold inflow and decline in currency would have added $379 million to bank reserves, but Fed credit declined by $370 million to offset their impact almost entirely.5 gal reserves. At other times, vault cash has also counted. “ Even if the Fed were to purchase securities from some one other than a member bank, bank reserves would still increase once the check issued by the Fed to pay for the securities was deposited in a member bank. Open market security sales reduce bank reserves since, ultimately, a member bank reserve deposit is reduced to pay for the securities sold by the Fed. 5The gold inflow, decline in currency and decline in Fed eral Reserve credit do not sum exactly to the change in bank reserves because of the effect of other, small transactions affecting reserves. 3From 1917 to 1960, such deposits were the only form in which member banks were permitted to hold their le- The Fed’s failure to fully offset the gold and currency outflows suffered by banks permitted the money supply contraction to accelerate. Fed officials claimed that the Reserve Banks’ lack of reserves precluded government security pur chases to offset the reserve losses suffered by banks.4 The Reserve Banks w ere required to 5 maintain gold reserves equal to 40 percent of their note issues and reserves o f either gold or “eligible paper” against the remaining 60 per cent.4 Since gold outflows had reduced the Sys 6 tem's reserve holdings, and since the System lacked other eligible paper, Fed officials asserted they could not increase Fed credit by purchas ing government securities, which w ere not eligi ble collateral. Friedman and Schwartz (1963), pp. 399-406, dispute the Fed’s justification for not buying government securities. They argue that the Sys tem had sufficient gold reserves and, in any event, that the Federal Reserve Board had the power to suspend the reserve requirements 45See the Board of Governors of the Federal Reserve Sys tem. Annual Report (1932), pp. 16-19. 46Eligible paper consisted of either bankers acceptances or commercial notes acquired by direct purchase or pledged by member banks as collateral for discount loans. See Board of Governors of the Federal Reserve System (1943), pp. 324-29, and Friedman and Schwartz (1963), p. 400. 47Why the Fed undertook these purchases is unclear, espe cially if fear of undermining the gold standard explains temporarily. Epstein and Ferguson (1984), pp. 964-65, contend, however, that Fed officials did feel constrained by a lack of gold. Wicker (1966), pp. 169-70, suggests that Fed officials feared that open market purchases would weaken confidence in the Fed’s determination to maintain gold convertibility and thereby renew the gold outflow. In any case, the Glass-Steagall Act o f 1932 re moved the constraint by permitting government securities to serve as collateral for Federal Reserve note issues. In March 1932, the System began what was then the largest open-market purchase program in its history.4 Between 7 February 24 and July 27, 1932, the Fed bought $1.1 billion of government securities. Member bank reserves increased only $194 million in these months, however, because of renewed gold and currency outflows and a reduction in member bank borrowing. Moreover, the supply of money continued to fall because o f a sharp decline in the money multiplier (see figure 10).4 8 why purchases were not made immediately following Bri tain’s departure from gold. Friedman and Schwartz (1963), pp. 384-89, argue that the Fed succumbed to pressure from Congress, while Epstein and Ferguson (1984) con clude that pressure from both Congress and commercial banks was important. ^D urin g these months, both the reserve-deposit and currency-deposit ratios rose. M ARCH/APRIL 1992 20 The Fed ended its purchase program in July 1932, largely because officials believed it had done little good.4 Bank reserves continued to 9 increase, however, as gold inflows were not o ff set by a corresponding reduction in Fed credit outstanding. Although the money supply ceased to fall, it also failed to rise significantly. In early 1933, large gold and currency outflows caused a renewed money supply decline.5 On this oc 0 casion, the crisis was stopped by Franklin D. Roosevelt’s decision to declare a Bank Holiday and suspend gold shipments. In essence, the Fed’s failure to insulate the banking system from gold outflows and panic currency with drawals had caused the president to act to pre vent further reserve losses. While failure to sterilize gold and currency outflows in 1931 and 1933 was inconsistent with previous actions, it did not represent a fundamental change in regime. Fed officials ap parently believed strongly in the gold standard, and there seems to have been no discussion of following Great Britain o ff gold. Benjamin Strong had been a committed advocate of the gold standard, and it seems doubtful that he would have proposed actions that might have weakened it.5 As an institution, the Federal 1 Reserve System was willing to forego short-run stability to preserve the gold standard, which it saw as its fundamental mission.5 2 Reserve sterilization constituted one aspect of System policy begun under Strong, and the Fed deviated little from the policy after his death, at least until the fo u rth q u arter o f 1931. In fact, from the stock market crash in October 1929 to Britain’s departure from gold on September 21, 1931, the Fed did little but offset gold and cur rency flows. It certainly did not make large open market purchases, despite the deepening depression. On the surface, this lack o f vigor appears at odds with the relatively large open market purchases the Fed made during the minor recessions of 1924 and 1927. 49Banks’ excess reserves increased substantially during the months of the open market purchases, which many saw as idle balances that were unneeded and potentially inflation ary. See Friedman and Schwartz (1963), pp. 385-89. As discussed below, Epstein and Ferguson (1984) suggest that pressure from commercial banks contributed to the Fed’s decision to end the program. 50The money supply fell both because of a decline in reserves and a decline in the money multiplier induced by panic deposit withdrawals. 51Strong testified before the House Banking Committee in 1928 that, “ When you are speaking of efforts simply to FEDERAL http://fraser.stlouisfed.org/ RESERVE BANK OF ST. LOUIS Federal Reserve Bank of St. Louis Strong’s Countercyclical Policy The Fed’s actions in 1924 mark its first use of open market operations to achieve general poli cy objectives. In that year, the Fed purchased $450 million of government securities and cut its discount rate (in three stages) from 4.5 per cent to 3 percent. In testimony before the House Banking Committee in 1926, Benjamin Strong listed several reasons for these actions, including the following: 1) To accelerate the process of debt repay ment to the Federal Reserve Banks by the member banks, so as to relieve this weaken ing pressure for loan liquidation. 2) To give the Federal Reserve Banks an asset which would not be automatically liquidated as the result of gold imports so that later, if inflation developed from excessive gold im ports, it might at least be checked in part by selling these securities, thus forcing member banks again into debt to the Reserve Banks and making the Reserve Bank discount rate effective. 3) To facilitate a change in the interest rela tion between the New York and London mar kets. . . by establishing a somewhat lower level of interest rates in this country at a time when prices w ere falling generally and when the danger of a disorganizing price ad vance in commodities was at a minimum and remote. 4) By directing foreign borrowings to this market to create the credits which would be necessary to facilitate the export of com modities. . . . 5) To render what assistance was possible by our market policy toward the recovery of sterling and the resumption of gold payment by Great Britain. 6) To check the pressure on the banking situ ation in the west and northwest and the resulting failures and disasters.5 3 stabilize commerce, industry, agriculture, employment and so on, without regard to the penalties of violation of the gold standard, you are talking about human judgment and the management of prices which I do not believe in at all.” [Quoted by Burgess (1930), pp. 331.] See also Temin (1989), p. 35. 52See Temin (1989), pp. 28-29 and 78-87, and Wheelock (1991). 53U.S. House of Representatives (1926), p. 336. 21 The Fed undertook a second large purchase program in 1927, purchasing $300 million of government securities and reducing the dis count rate again. Strong left no written justifica tion for these operations. Friedman and Schwartz (1963) argue that they were made in response to a recession, and that the 1924 pur chases had also been intended to bring about a domestic recovery. Wicker (1966), pp. 77-94 and 106-16, challenges this interpretation, arguing that the actions were motivated by international considerations. According to Wicker, the pur chases in 1924 were intended to encourage the flow of gold to Britain by reducing U.S. interest rates relative to those in London, with the goal of assisting Britain’s return to the gold standard. The 1927 purchases were intended to help Bri tain through a payments crisis, again by direct ing capital toward London; these purchases followed closely a meeting between Strong and European central bank heads. Chandler (1958), p. 199, argues that both domestic and international goals were important in 1924 and 1927, and Wheelock (1991), ch. 2, finds empirical support for this view. Wheelock also shows that, relative to the decline in eco nomic activity, the Fed made substantially few er open market purchases in 1930 and 1931 than it did during 1924 and 1927. This might reflect a significant change in System behavior between the 1920s and early 1930s. But an analysis of the Fed's policy methods suggests that its anem ic response in 1930-31 might also be explained as the consistent use of a single strategy. During the early 1920s, the Fed developed a strategy of using open market operations and discount rate changes to affect the level of member bank discount window borrowing. Fed officials observed that, when the System pur chased government securities, member bank borrowing tended to decline by nearly the same amount and, similarly, that open market sales led to comparable increases in member bank borrowing. But, while the Fed’s operations had little impact on the total volume of Fed credit outstanding, they appeared to have a significant 54lbid, p. 468. 55Presumably, discount rate changes alone could achieve the same impact on interest rates, but the Fed preferred to precede discount rate changes with open market opera tions. Strong testified that “ the foundation for rate changes can be more safely and better laid by these preliminary operations in the open market than would be impact on money markets. According to Chan dler (1958): Federal Reserve officials soon discovered. . . much to their amazement at first, that open market purchases and sales brought about marked changes in money market conditions even though total earning assets of the Reserve Banks remained unchanged. When the Federal Reserve sold securities and extracted money from bank reserves, more banks were forced to borrow from the Reserve Banks, and those al ready borrowing were forced more deeply into debt. Since banks had to pay interest on their borrowings and did not like to remain continu ously in debt, they tended to lend less liberally, which raised interest rates in the market pp. 238-39). Strong testified that “the effect of open market operations is to increase or decrease the extent to which the member banks must of their own initiative call on the Reserve Bank for credit. . . .”5 Security purchases led to less mem 4 ber bank borrowing and lower interest rates, while sales increased borrowing and rates.5 5 Strong believed that monetary policy could stimulate economic activity by easing money market conditions: . . .[W]hen we have very cheap money, corpora tions and individuals borrow money in order to extend their businesses. That results in plant construction; plant construction employs more labor, brings in to use more materials. . . . It may cause some elevation of wages. It creates more spending power; and with that start it will permeate through into the trades and the general price level.5 6 Chandler (1958) and Friedman and Schwartz (1963) conclude that under Strong’s leadership the Federal Reserve System attempted to stimu late economic activity during recessions by promoting monetary ease (cheap money). This explains why Strong listed “to accelerate the process of debt repayment. . . by the member banks” as a reason for the open market pur- possible otherwise, and the effect is less dramatic and less alarming. . . than if we just make advances and reductions in our discount rate.” [U.S. House of Represen tatives (1926), p. 333]. 56U.S. House of Representatives (1926), pp. 578-79. MARCH/APRIL 1992 22 Table 1 Fed Policy During Three Recessions (dollar amounts in millions)________ Date AIP GS 1929 Jul Oct 1930 Jan Apr Jul Oct 1931 Jan Apr Jul Oct 124 118 106 104 93 88 83 88 82 73 $147 154 485 530 583 602 647 600 674 733 1923 Apr Jul Oct 1924 Jan Apr Jul Oct 1925 Jan 106 104 99 100 95 84 95 105 229 97 91 118 274 467 585 464 1926 Oct 1927 Jan Apr Jul Oct 1928 Jan 111 107 108 106 102 107 306 310 341 381 506 512 DR DL DL(NYC) $1096 885 501 231 226 196 253 155 169 614 $319 74 39 17 0 6 5 0 0 74 4.5 4.5 4.5 4.5 4.5 3.5 3.0 3.0 658 834 873 574 489 315 240 275 123 143 121 85 45 13 28 32 4.0 4.0 4.0 4.0 3.5 3.5 663 481 447 454 424 465 84 76 78 59 75 94 5.0% 6.0 4.5 3.5 2.5 2.5 2.0 2.0 1.5 3.5 Variable definitions: AIP: Index of Industrial Production (seasonally adjusted); GS: Federal Reserve government secu rity holdings; DR: discount rate of the Federal Reserve Bank of New York; DL: discount loans (member bank borrowing) of all Federal Reserve member banks; DL(NYC): discount loans of reporting banks in New York City. chases in 1924. Strong used the level of mem ber bank borrowing to determine the specific quantity of security purchases necessary to bring about monetary ease: Should we go into a business recession while the member banks were continuing to borrow directly 500 or 600 million dollars. . . we should consider taking steps to relieve some of the pressure which this borrowing induces by pur chasing Government securities and thus ena bling member banks to reduce their indebtedness. . . . As a guide to the timing and 57Presentation to the Governors’ Conference, March 1926 [quoted by Chandler (1958), pp. 239-40]. 58Wicker (1969) agrees that, to the extent that the Fed responded to domestic conditions, it used member bank borrowing as a guide. See also Meltzer (1976). FED ERAL RESERVE BANK OF ST. LOUIS extent of any purchases which might appear desirable, one of our best guides would be the amount of borrowing by member banks in prin cipal centers. . . . Our experience has shown that when New York City banks are borrowing in the neighborhood of 100 mil lion dollars or more, there is then some real pressure for reducing loans, and money rates tend to be markedly higher than the discount rate. On the other hand, when borrowings of these banks are negligible, as in 1924, the money situation tends to be less elastic and if gold imports take place, there is liable to be some credit inflation, with money rates drop ping below our discount rate. When member banks are owing us about 50 million dollars or less the situations appears to be comfortable, with no marked pressure for liquidation. . . ,5 7 Table 1 compares Federal Reserve actions dur ing the 1924, 1927 and 1930-31 downturns. The Fed’s index of industrial production indicates the severity of each recession. Following the stock market crash in October 1929, the New York Fed purchased $160 million of government securities and, by the end o f December, the Sys tem had purchased an additional $150 million. But, from January 1930 to October 1931, the Fed made only modest purchases, particularly in comparison with those made in 1924 and 1927, when the declines in economic activity were less. The relatively small purchases in 1930 and 1931 appear consistent, however, with the use of member bank borrowing as a policy guide. This, according to Brunner and Meltzer (1968), explains the Fed’s failure to respond aggressive ly to the Depression.3 Member bank borrowing 8 fell substantially following the stock market crash in October 1929 and averaged just $241 million from January 1930 to August 1931. Bor rowing by reporting member banks in New York City averaged just $8 million over the same months. Thus, by Strong's guidelines, money was exceptionally easy and substantial open market operations w ere unwarranted. The Fed’s use o f member bank borrowing as a guide to monetary conditions could explain 23 why it permitted the money supply to decline sharply during the Depression. During a reces sion, loan demand declines and banks have few er profitable investment opportunities. Consequent ly, the demand for borrowed reserves declines. If this decline in demand is not offset, total reserves and the money supply fall. In a minor recession, as in 1924 and 1927, member bank borrowing falls little. The Fed’s guidelines would have suggested that monetary conditions were relatively tight and, in response, it would have made large open market purchases. In a severe economic downturn, as in 1930-31, however, member bank borrowing may fall substantially. But, by Strong’s rule the Fed would have made few open market purchases. Thus, ironically, this strategy could result in a greater contrac tion in the supply of money, the more severe a decline in economic activity.5 9 If, as Brunner and Meltzer (1968) argue, Sys tem officials followed Strong’s prescription to use the level of bank borrowing to guide policy during the Depression, then it seems that the Fed made no fundamental change in policy after Strong’s death. Although Friedman and Schwartz (1963), pp. 362-419, believe that Strong would have re sponded aggressively to the Depression, they agree that a majority of Fed officials interpreted the low level of member bank borrowing in 1930 and 1931 as signaling monetary ease. They contend, however, that officials o f the New York Fed understood the flaws in using member bank borrowing as a policy guide and would have pursued appropriately expansionary poli cies if they had the authority.6 0 that the Policy Conference was established to wrest power from the New York Bank. And they show, pp. 367-80, that New York officials proposed more expansionary actions, particular ly in 1930, than were accepted by the rest of the System. Wicker (1966) finds, however, that Harrison ceased to advocate open market pur chases once New York banks were no longer borrowing reserves. Thus, while the Federal Reserve would likely have pursued somewhat more expansionary policies had New York offi cials held more authority, the modest open mar ket purchases of 1930 and 1931 were apparently consistent with the guidelines out lined by Strong. In sum, during the Depression, the Federal Reserve continued to sterilize gold and currency flows and made limited open market purchases and discount rate reductions in response to the economic decline. Notable deviations from these policies occurred, such as the incomplete sterili zation of gold outflows during the crises of 1931 and 1933. But it seems likely that mone tary policy would have been somewhat more responsive to the Depression, particularly in 1930, had officials of the Federal Reserve Bank of New York been able to dominate policymak ing in the way Strong had before his death.6 1 The general thrust of policy, however, appears consistent with that of Benjamin Strong. INTEREST GROUP PRESSURE EXPLANATIONS OF FED BEHAVIOR In March 1930 the Open Market Investment Committee, which consisted of five Reserve Bank governors, was replaced by the Open Mar ket Policy Conference, in which representatives o f all 12 Banks participated. The Investment Committee had been led by Benjamin Strong, and then by George Harrison, Strong’s successor as governor of the Federal Reserve Bank of New York. Friedman and Schwartz, p. 414, contend Until recently, most studies of Fed behavior have concluded that policymakers failed to per ceive a need to take expansionary actions, despite deflation, rising unemployment and widespread bank failures. Some researchers now argue, however, that Fed officials were quite aware that their policies were contribut ing to the contraction. These researchers con clude that policymakers responded to interest group pressure and their own desire for in- 59lndeed, except for a brief decline in M1 in 1927, the abso lute quantity of money did not fall in 1924 or 1927, although its rate of increase declined. The Fed's strategy and the consequences of using bank borrowing as a policy guide are examined in greater detail in Wheelock (1991), ch. 3. policies, particularly in 1927, had contributed to stock mar ket speculation and the crash and depression that fol lowed. See Wheelock (1991), ch. 4, for analysis of disagreements among System officials during the Depression. 60See also Schwartz (1981), pp. 41-42. 6 It is by no means clear that Strong could have retained 1 this degree of influence, as many officials believed that his MARCH/APRIL 1992 24 fluence, rather than the public interest. Epstein and Ferguson (1984), for example, contend that a combination o f ideology and conflicting in terests explain the System’s policy. And Ander son, Shughart and Tollison (1988) argue that "the restrictive monetary policy o f the Fed in the 1929-33 period was not based on myopia but instead on rational, self-interested behavior" (p. 4). Epstein and Ferguson (1984) focus their study on the Federal Reserve’s $1.1 billion open mar ket purchase program of 1932. They ask why the Fed waited so long to begin an expansionary program, what had changed to cause the Fed to begin the program when it did, and what led to the decision to end the program. To the first question, Epstein and Ferguson (1984) conclude that the liquidationist business cycle theory was dominant among Fed officials. Liquidationists believed that depressions were “vital to the long-run health o f a capitalist econ omy. Accordingly, the task of central banking was to stand back and allow nature’s therapy to take its course.” (p. 963). This certainly was the opinion of some key officials, such as George Norris, who argued at the September 25, 1930, meeting of the Open Market Policy Conference: We believe that the correction must come about through reduced production, reduced in ventories, the gradual reduction of consumer credit, the liquidation of security loans, and the accumulation of savings through the exercise of thrift. These are slow and simple remedies, but just as there is no ‘royal road to knowledge,’ we believe there is no short cut or panacea for the rectification of existing conditions. . . . We have been putting out credit in a period of depression, when it was not wanted and could not be used, and will have to withdraw credit when it is wanted and can be used.®2 Norris clearly believed that monetary policy had been too stimulative and was interfering with the natural process of liquidation and recovery. Strong and other officials apparently held similar views during the recession of 1920-21. According to Wicker (1966): 62Quoted in Chandler (1971), p. 137. 63See also Friedman and Schwartz (1963), pp. 249-54, Wick er (1966), pp. 57-66, and West (1977), pp. 173-204. “ Quoted by Chandler (1958), p. 239-40. FEDERAL RESERVE BANK OF ST. LOUIS In the view of System officials the money sup ply in 1920 was redundant (excessive) and should decline to restore the ‘proper’ relation ship between prices, credit, and volume of production. The term most frequently used to describe this process was ‘liquidation,’ the necessity for which was not disputed by either the Board or by any other Federal Reserve offi cial including Benjamin Strong. . .(p. 49). Most researchers argue that Strong’s views changed significantly after the 1920-21 episode, however. Chandler (1958) writes: Like most other Federal Beserve officials, [in 1920-1921 Strong] believed that some deflation of bank credit was essential and that some price reduction was inevitable and desirable. Within three years, Strong himself had rejected many of these ideas. A much smaller business recession in 1924 led him to advocate large and aggressive open-market purchases of govern ment securities and reductions of discount rates to combat deflation at home as well as to en courage foreign lending (p. 181).8 3 In rejecting the importance of Strong’s death, Epstein and Ferguson (1984) implicitly deny that Strong sought to prevent loan liquidation during recessions by pursuing monetary ease or that he subscribed to the countercyclical policy guide lines he presented to the Governors Conference in 1926: "Should we go into a business reces sion. . . we should consider taking steps to relieve. . . the pressure. . . by purchasing Government securities. . . .”6 4 Epstein and Ferguson emphasize two addition al reasons for the timing and extent of Fed ac tions during the Depression. First, in contrast to Friedman and Schwartz, they conclude that a lack o f gold reserves did keep the Fed from making open market purchases in the fourth quarter of 1931. They argue further that, while the Glass-Steagall Act of 1932 lessened the problem for the System as a whole, some o f the Reserve Banks were reluctant to continue the purchase program in 1932 because they lacked sufficient gold reserves.6 5 Second, Epstein and Ferguson argue that Fed concern with member bank profits contributed 65Each Reserve Bank was required to maintain its own reserves. Pooling was not permitted, although the Banks could lend to one another. 25 to the timing and extent of open market pur chases in 1932. During the first two years of the Depression, leading bankers generally ar gued for loan liquidation and lower wages. But the sharp increase in interest rates in the fourth quarter o f 1931 reduced the value of bond portfolios and threatened the solvency of many banks. Bankers then began to press the Fed to support bond prices. Epstein and Fergu son argue that "a major goal of the [purchases of 1932] was to revive railroad bond values. . . and bond prices in general." (p. 967). Just as constituent pressure contributed to the decision to make open market purchases, it also seems to have caused the program’s end. Dur ing the Depression banks generally had shifted their bond portfolios toward short-term maturi ties. And, while the need to support bond prices was paramount in early 1932, as the year progressed short-term interest rates fell sharply and bank earnings declined. The decline in earnings was especially acute in Boston and Chicago because banks in those cities had un usually large holdings of short-term securities. Epstein and Ferguson conclude: “That the gover nors o f the Boston Fed and, especially, the Chicago Fed should be early critics of the refla tion program is therefore no mystery.” (p. 972). Declining interest rates and questions about the willingness o f the United States to maintain the dollar's gold convertibility led to deposit withdrawals by foreigners, causing commercial banks to raise further doubts about the pur chase program: “The continued loss of gold and deposits put many New York banks in an in creasingly uncomfortable position. . . . Many complained that the reflation program had 'demoralized money and exchange markets’.”6 6 Thus, pressure from member banks experienc ing falling earnings and deposit outflows and the desire of some Reserve Banks to protect their gold reserves caused the System to aban don its program of open market purchases.6 7 Epstein and Ferguson (1984) were the first to explain Federal Reserve behavior during the Depression as a response to pressure from com mercial bankers. Anderson, Shughart and Tollison (1988) push this view to the extreme, arguing that the principal aim of Fed policy dur ing the Depression was to enhance the long-run profitability of member banks by eliminating nonmember competitors. This, in turn, benefited the Fed by increasing the proportion o f the banking system under its regulatory control: The fall in the money supply presided over by the monetary authority between 1929 and 1933 eliminated a large number of state-chartered and small, federally-chartered institutions from the commercial banking industry. The profits of those banks that survived. . . rose significantly as a result. Coincidentally, the monetary con traction expanded the proportion of the com mercial banking system within the Fed’s bureaucratic domain. Thus, rather than representing the leading example of bureaucratic ineptitude, the Great Contraction may instead be the leading example of rational regulatory policy operating for the benefit of the regula tors and the regulated.6 8 Nonmember banks made up 75 percent of the banks that suspended operations between 1930 and 1933. Failures were highest among small in stitutions located in rural areas. Policymakers typically argued that such failures were caused by bad management or transportation improve ments that made many banks redundant. George Harrison, Governor of the New York Reserve Bank, for example, testified before the Senate Banking Committee in 1931 that: . . .with the automobile and improved roads, the smaller banks. . . with nominal capital, out in the small rural communities, no longer had any reason really to exist. Their depositors wel comed the opportunity to get into their automo biles and go to the large centers where they could put their money.6 9 66Epstein and Ferguson (1984), p. 975. 68Anderson, Shughart and Tollison (1988), pp. 8-9. 67ln a comment on Epstein and Ferguson, Coelho and Santoni (1991) present econometric evidence suggesting that banks did not suffer reduced profits as a result of the Fed’s 1932 purchases, and they question whether pres sure from commercial banks caused the Fed to end its program. Indeed, they even doubt that expansionary policy was ended since the monetary base continued to rise. Ep stein and Ferguson (1991) present additional qualitative evidence showing that banks thought that low interest rates had reduced their earnings. 69U.S. Senate (1931), p. 44. MARCH/APRIL 1992 26 From the Federal Reserve's inception, Fed offi cials argued that it was important that all banks join the System. Benjamin Strong argued in 1915 that “no reform of our banking methods in this country will be complete and satisfactory to the country until it includes all banks. . . in one comprehensive system."7 Policymakers were 0 likely less concerned with the failure of non member banks than they were with the health of member banks.7 But it remains to be shown 1 that Federal Reserve policies were deliberately intended to cause the failure o f thousands of nonmember institutions. Anderson, Shughart and Tollison (1988) argue that “the Great Depression. . . was a by-product o f economically rational behavior on the part of Federal Reserve member banks seeking rents through the elimination of their nonmember rivals.” (p. 9) Member banks did not capture the Fed directly, they argue, but rather exerted pressure through members of the House and Senate Banking Committees. To test this hypoth esis, the authors regress deposits in failed nonmember banks in each state on dummy varia bles indicating whether a state was represented on the House or Senate Banking Committees.7 2 They find that nonmember bank losses were higher if a state had a representative on the House Banking Committee. This, they argue, supports their view that the Fed deliberately caused nonmember bank failures to be highest in states having a congressman on the Banking Committee, thereby enhancing the long-run profits of the member banks that remained. The Fed’s payoff came in 1933 when it was freed from having to return a portion of its revenues to the Treasury.7 3 This explanation of Federal Reserve policy provokes a number of questions. Left unclear, for example, is why member banks had more influence over Congress than nonmember banks. Nor is it explained how the Fed was able to af fect the fortunes of nonmember banks in partic 70Quoted by Chandler (1958), p. 80. 71Friedman and Schwartz, pp. 358-59, also make this point. 72They include deposits in failed member banks, total bank deposits, and other control variables as additional regressors. 73Anderson, Shughart and Tollison (1988), pp. 16-17. 74Anderson, Shughart and Tollison (1990) reply by pointing out that member bank share prices rose dramatically rela tive to those of nonmember banks during 1930. Even the FEDERAL RESERVE BANK OF ST. LOUIS ular states with the tools at its disposal. The Fed cannot control the destination o f reserve flows generated by open market operations, and the discount window was not open to nonmember banks. Perhaps Fed officials could have selec tively restricted credit to member banks that lent to nonmember banks in particular states, but it is doubtful that such a circuitous route would have had a large impact on losses. Huberman (1990) casts further doubt on the Anderson, Shughart and Tollison (1988) view. He notes that, although 75 percent of banks that suspended during the Depression were nonmember banks, the ratio of nonmember to member bank suspensions from 1930 to 1933 was lower than it had been during the 1920s. Nonmember banks that suspended, moreover, reopened twice as often as member banks. Membership in the Federal Reserve System grew at a comparatively low rate during the Depression. Santoni and Van Cott (1990) also report evi dence contrary to the Anderson, Shughart and Tollison hypothesis. They calculate a share price index for large New York City member banks and show that, relative to both the wholesale commodity price index and the Standard and Poor’s index, bank share prices declined sub stantially from 1929 to 1934. They show also that the index of bank stock prices was not af fected by changes in the money supply, suggest ing that monetary policy did not enhance the fortunes of banks in their sample.7 4 CONCLUSION The Federal Reserve’s failure to respond vigorously to the Great Depression probably cannot be attributed to a single cause. Each of the explanations discussed in this article clarifies certain points about Fed behavior during the Depression. A number of contemporary observ ers, both within and outside the System, attrib- share prices of “ small” member banks rose relative to those of nonmember institutions. Neither study tests whether the profits of surviving member banks were en hanced in the long run, although the former note that in 1936 the average national bank profit rate was higher than in 1925. Unfortunately, it is impossible to determine whether the increase in bank profits was caused by the demise of nonmember banks or by New Deal reforms, such as deposit interest rate ceilings, deposit insurance, and increased chartering requirements, which reduced competition and enhanced bank charter values. 27 uted some of the blame to what they viewed as excessively easy monetary policy during reces sions in 1924 and 1927. They argued that the Fed’s actions had promoted stock market specu lation and led inevitably to the crash and Depression. The best policy during the Depres sion, according to these observers, was to pro mote loan liquidation and wage rate reductions, to allow recovery on a “sound basis.” While those officials subscribing to the liquidationist view did not win approval o f open market sales, they were able to prevent significant open mar ket purchases until 1932. It is likely that the Fed would not have made large purchases even then without pressure from major bankers and Congress. The most important explanations of Federal Reserve behavior during the Depression, however, appear to be the dedication of policymakers to preserving the gold standard and their attachment to policy guides that gave erroneous information about monetary condi tions. Benjamin Strong’s death robbed the Sys tem of an intelligent leader at a crucial time and undoubtedly imparted a contractionary bias to monetary policy during the Depression. It seems clear, however, that Strong’s death did not cause a fundamental change in regime. Strong believed in the gold standard, and he would not likely have done anything to jeopardize gold convertibility of the dollar. There was also little deviation from either the gold sterilization or the countercyclical policy rules that Strong had developed during the 1920s—at least until the fourth quarter of 1931, when maintenance of the gold standard became the overriding goal of policy. Thus, while leadership changes and in terest group pressure probably had some effect, monetary policy during the Depression was not fundamentally different from that of previous years. Federal Reserve errors seem largely at tributable to the continued use of flawed policies. REFERENCES Anderson, Gary M., William F. Shughart II, and Robert D. Tollison. “A Public Choice Theory of the Great Contrac tion,” Public Choice (October 1988), pp. 3-23. _______ . “A Public Choice Theory of the Great Contraction: Further Evidence,” Public Choice (December 1990), pp. 277-83. Belongia, Michael T., and Michelle R. Garfinkel, eds. What Do We Know About Business Cycles? (Kluwer Academic Publishers, forthcoming). Bernanke, Ben S. “ Nonmonetary Effects of the Financial Crises in the Propagation of the Great Depression,” Ameri can Economic Review (June 1983), pp. 257-76. Board of Governors of the Federal Reserve System. Annual Report (various issues). _______ . Banking and Monetary Statistics, 1914-1941 (1943). _______ . Federal Reserve Bulletin (various issues). Bordo, Michael D. “ The Contribution of A Monetary History of the United States, 1867-1960 to Monetary History,” in Michael D. Bordo, ed., Money, History, and International Finance: Essays in Honor of Anna J. Schwartz (The Univer sity of Chicago Press, 1989), pp. 15-75. Brunner, Karl, and Allan H. Meltzer. “ What Did We Learn from the Monetary Experience of the United States in the Great Depression?” Canadian Journal of Economics (May 1968), pp. 334-48. Burgess, W. Randolph, ed. Interpretations of Federal Reserve Policy in the Speeches and Writings of Benjamin Strong (Harper and Brothers, 1930). Calomiris, Charles W. “ Deposit Insurance: Lessons from the Record,” Federal Reserve Bank of Chicago Economic Per spectives (May/June 1989), pp. 10-30. Chandler, Lester V. Benjamin Strong, Central Banker (Brook ings Institution, 1958). _______ . American Monetary Policy 1928-1941 (Harper and Row, 1971). Coelho, Philip R. P., and G. J. Santoni. “ Regulatory Capture and the Monetary Contraction of 1932: A Comment on Ep stein and Ferguson,” Journal of Economic History (March 1991), pp. 182-89. Darby, Michael R. “ Three-and-a-Half Million U.S. Employees Have Been Mislaid: Or, an Explanation of Unemployment, 1934-1941,” Journal of Political Economy (February 1976), pp. 1-16. De Long, J. Bradford. “ 'Liquidation' Cycles: Old-Fashioned Real Business Cycle Theory and the Great Depression,” National Bureau of Economic Research Working Paper No. 3546 (December 1990). Department of Commerce. Historical Statistics of the United States (1960). Epstein, Gerald, and Thomas Ferguson. “ Monetary Policy, Loan Liquidation, and Industrial Conflict: The Federal Reserve and the Open Market Operations of 1932,” Journal of Economic History (December 1984), pp. 957-83. _______ . “Answers to Stock Questions: Fed Targets, Stock Prices, and the Gold Standard in the Great Depression,” Journal of Economic History (March 1991), pp. 190-200. Field, Alexander J. “A New Interpretation of the Onset of the Great Depression,” Journal of Economic History (June 1984), pp. 489-98. Fisher, Irving. Booms and Depressions (Adelphi, 1932). Friedman, Milton, and Anna J. Schwartz. A Monetary History of the United States, 1867-1960 (Princeton University Press, 1963). Hamilton, James D. “ Monetary Factors in the Great Depres sion,” Journal of Monetary Economics (March 1987), pp. 145-69. Hayek, Friedrich A. von. “ The Fate of the Gold Standard,” in Roy McCloughry, ed., Money, Capital and Fluctuations, Ear ly Essays of Friedrich A. von Hayek (Routledge and Kegan Paul, 1984), pp. 118-35. Hetzel, Robert L. “ The Rules Versus Discretion Debate Over Monetary Policy in the 1920s,” Federal Reserve Bank of Richmond Economic Review (November/December 1985), pp. 3-14. Huberman, Douglas A. “An Alternative to ‘A Public Choice Theory of the Great Contraction,” ' Public Choice (Decem ber 1990), pp. 257-68. Kesselman, Jonathan R., and N. E. Savin. “ Three-and-a-Half Million Workers Never Were Lost,” Economic Inquiry (April 1978), pp. 205-25. Kindleberger, Charles P. The World in Depression, 1929-1939, 2nd ed. (University of California Press, 1986). Lebergott, Stanley. Men Without Work: The Economics of Unemployment (Prentice-Hall, 1964). MARCH/APRIL 1992 28 Meltzer, Allan H. “ Monetary and Other Explanations of the Start of the Great Depression,” Journal of Monetary Eco nomics (November 1976), pp. 455-71. U.S. House of Representatives. Stabilization. Hearings Before the Committee on Banking and Currency. 69th Congress, 1st Sess. (GPO, 1926). Miron, Jeffrey A. “ Financial Panics, the Seasonality of the Nominal Interest Rate, and the Founding of the Fed,” American Economic Review (March 1986), pp. 125-40. ______ _ Banking Act of 1935. Committee on Banking and Currency. 74th Congress, 1st Sess. (GPO, 1935). Romer, Christina. “ Spurious Volatility in Historical Unemploy ment Data,” Journal of Political Economy (February 1986a), pp. 1-37. _______ . “ New Estimates of Prewar Gross National Product and Unemployment,” Journal of Economic History (June 1986b), pp. 341-52. Santoni, G. J., and T. Norman Van Cott. “ The Ruthless Fed: A Critique of the AST Hypothesis,” Public Choice (Decem ber 1990), pp. 269-75. U.S. Senate. Operation of the National and Federal Reserve Banking Systems, Committee on Banking and Currency, 71st Congress, 3rd Sess. (GPO, 1931). West, Robert Craig. Banking Reform and the Federal Reserve 1863-1923 (Cornell University Press, 1977). Wheelock, David C. The Strategy and Consistency of Federal Reserve Monetary Policy, 1924-1933 (Cambridge University Press, 1991). Schwartz, Anna J. “ Understanding 1929-1933,” in Karl Brunner, ed., The Great Depression Revisited (KluwerNijhoff, 1981), pp. 5-48. _______ . “ Banking Act of 1935,” in John Eatwell, Murray Milgate, and Peter Newman, eds., The New Palgrave Dic tionary of Money and Finance (Macmillan Press, forth coming). Temin, Peter. Did Monetary Forces Cause the Great Depres sion? (W. W. Norton, 1976). Wicker, Elmus R. Federal Reserve Monetary Policy, 1917-1933 (Random House, 1966). _______ . Lessons from the Great Depression (The MIT Press, 1989). _______ . “ Brunner and Meltzer on Federal Reserve Mone tary Policy During the Great Depression,” Canadian Journal of Economics (May 1969), pp. 318-21. U.S. Bureau of Labor Statistics. Wholesale Prices 1890 to 1925 (GPO, 1926). _______ . “ What Caused the Great Depression 1929-1937? A Survey of Recent Literature,” unpublished manuscript (Indiana University, August 1989). Appendix Data Sources All Commodities Price Index: Bureau of Labor Statistics (1926), pp. 24-25. Gold stock: Board of Governors of the Federal Reserve System (1943), pp. 369-77. Bank reserves: Board of Governors of the Federal Reserve System (1943), pp. 369-77, (member banks) and Friedman and Schwartz (1963), table A-2 (all banks). Gross national product: Department of Com merce (1960), H istorical Statistics o f the United States, series FI (current dollar) and F3 (cons tant dollar). Bank suspensions and deposits in suspended banks: Board of Governors of the Federal Reserve System (1943), p. 283. Currency stock: Board of Governors of the Federal Reserve System (1943), pp. 369-77. Discount rate (Federal Reserve Bank of New York): Board of Governors of the Federal Reserve System (1943), pp. 440-41. Federal Reserve credit and its components: Board of Governors of the Federal Reserve System (1943), pp. 369-77, and ibid, pp. 136-44 (discount loans of reporting New York City member banks). Federal Reserve System balance sheet: Board of Governors of the Federal Reserve System (1943), p. 331. FEDERAL RESERVE BANK OF ST. LOUIS implicit price index: Department of Commerce (I960), H istorical Statistics o f the United States, series F5. Index of Industrial Production: Board of Governors of the Federal Reserve System (1937), A nnual R eport, pp. 175-77. Interest rates: 1) Baa-rated: Board of Governors of the Federal Reserve System (1943), pp. 468-70; 2) long-term (daily average yield in June of each year on U.S. government bonds): ibid, pp. 468-70; 3) short-term (daily average yield in June o f each year on U.S. government three- to six-month notes and certificates (1919 to 1933), and on Treasury bills (1934 to 1939): ibid, p. 460. Money supply: Friedman and Schwartz (1963), table A-l, col. 7 (M l) and col. 8 (M2). Unemployment rate: Lebergott (1964), p. 27. 29 M anfred J.M. Neumann Manfred J.M. Neumann, a professor of economics at the University of Bonn, Germany, was a visiting scholar at the Federal Reserve Bank of St. Louis. Courtenay C. Stone made many helpful comments on an earlier draft. Lora Holman and Richard I. Jako provided research assistance. Seigniorage in the United States: H ow Much Does the U.S. Government Make from Money Production? M ONEY IS CERTAINLY one of the greatest inventions of mankind. As Brunner and Meltzer (1971) have noted, its vast social productivity arises from the enormous reduction in transac tions and information costs that it provides by serving as a standardized medium of exchange.1 Of course, these benefits, like those o f any other good or service, are not provided at zero cost. The revenue received from producing and maintaining a nation’s money stock covers its production costs and, perhaps, some profit as well for its producers. In monetary economics, the revenue from money creation is called "seigniorage.” Unfor tunately, this term has been subject to a variety of interpretations in the literature. After review ing several traditional definitions, this article de velops a new seigniorage measure, extended monetary seigniorage, and shows how it is dis tributed between the Federal Reserve, member banks and the U.S. Treasury during the 1951-90 period. Then, it examines the relationship be tween inflation and seigniorage during this peri od and shows that this relationship is analogous to the well-known “Laffer curve” that relates tax rates and tax revenues: seigniorage in creases as inflation rises until the inflation rate reaches about 7 percent; thereafter, inflation and seigniorage are inversely related. Indeed, for each percentage point rise in inflation above 7 percent, the U.S. Treasury’s share o f seig niorage fell, on average, by $1.4 billion (meas ured at 1982/84 consumer prices). REVENUE FROM MONEY CREA TION: SOME ANALYTICAL CONCEPTS The term "seigniorage” dates back to the early Middle Ages, when it was common for sover eigns o f many countries to finance some of their expenditures from the profits they earned 1See Brunner and Meltzer (1971). MARCH/APRIL 1992 30 from the coinage of money. In the money litera ture, seigniorage has often been used inter changeably for either the total revenue or the profit derived from money production and maintenance. Of course, revenues and profits are identical only if costs are zero. Although theoretical analysis can be simplified by assum ing that costs are zero, this assumption cannot be maintained in empirical applications. Since this article focuses on the empirical issues as sociated with seigniorage, the total revenue, cost and profits associated with money production must be carefully distinguished and the relevant notion of seigniorage must be clearly defined. In the analysis that follows, seigniorage is de fined as the revenue associated with money production and maintenance, rather than the resulting profit. Also, the focus is on the reve nue accruing to the government and, therefore, on the creation of monetary base rather than the creation of deposits by private depository institutions. Monetary theorists have used two main con cepts of seigniorage in analyzing its relationship to inflation. These concepts are termed “oppor tunity cost seigniorage” and "monetary seig niorage.” Opportunity Cost Seigniorage As its name indicates, opportunity cost seig niorage defines seigniorage as the total “oppor tunity costs” of money holders. It asks the question, What additional real income would in dividuals have earned if they had held interestearning assets instead of non-interest-earning money? The real interest earnings foregone by holding money are called its opportunity cost. Real opportunity cost seigniorage (sQ is: ) (1) s0 = rB/P, where B denotes total base money holdings, r is the representative nominal rate of return on as sets other than base money and P is the con sumer price level. This concept of seigniorage has been used as an elegant tool of theoretical analysis.2 Its ana lytical attraction is that it derives the value of 2See Bailey (1956), Johnson (1969), Auernheimer (1974) and Barro (1982). 3For a different view, see Gros (1989), p. 2. He interprets equation 1 to represent “ the interest savings the govern ment obtains by being able to issue zero interest rate securities in the form of currency.” This interpretation, FEDERAL http://fraser.stlouisfed.org/ RESERVE BANK OF ST. LOUIS Federal Reserve Bank of St. Louis seigniorage from the individuals’ valuation of the services of money. It does this by identify ing seigniorage with the interest income that in dividuals voluntarily forego by holding some of their wealth as money instead of as earning as sets. This concept, however, presents some problems when it is used for empirical studies of seigniorage. To make the concept of opportunity cost seig niorage operational for empirical analysis, some actual nominal rate o f return must be chosen as the measure of the representative rate of return (r) in equation 1. Estimates of seigniorage will differ widely depending on which rate of return — for example, the federal funds rate, the average yield on government bonds or the rate o f return on stocks of, say, the computer industry — is used. Thus, the problem is to de termine a weighted average of observable asset returns that meaningfully approximates the true opportunity cost o f money holders. There is also a conceptual problem with using this definition of seigniorage: opportunity cost seigniorage does not equal the monetary author ity’s actual revenue from money creation.3 Be cause the structure of the monetary authority’s portfolio differs markedly from the asset struc ture preferred by private investors, opportunity cost seigniorage does not provide a measure of the gains to the monetary authority from money creation and maintenance. Monetary Seigniorage The concept of monetary seigniorage permits a more straightforward and unambiguous em pirical measurement. Monetary seigniorage (s*,) is defined as the net change in base money out standing (AB), deflated by the consumer price level (P): (2) s* = AB/P. Monetary seigniorage measures the transaction value of non-monetary assets that money holders trade in to the monetary authority to obtain the desired increase in their base money balances (AB). Because the data necessary to calculate this measure are easily available, the concept of however, is valid only if the nominal rate of return (r) equals the effective yield on government debt and operat ing costs are zero. 31 monetary seigniorage has been widely used and measured by monetary economists.4 Extended M onetary Seigniorage Unfortunately, the traditional concept of monetary seigniorage does not provide a com plete account of the government's revenue from base money provision. It abstracts from the ac tual process of base money creation and, there fore, neglects the fact that the total flow of revenue in addition depends on the asset struc ture of the central bank. The total flow of seigniorage to the govern ment consists o f two components. The first is the real value of the non-monetary assets that the central bank receives from the public in ex change for an increase in the monetary base. This is measured by the traditional concept of monetary seigniorage as defined above. The se cond component is the interest earnings the central bank receives on its stocks of non government debt. Since domestic private and foreign debtors have to service the debt held by the central bank, there is a flow of seigniorage to govern ment even if the public does not desire to in crease its cash balances. It is important to note, however, that only the interest earnings on non-government debt qualify. The Treasury’s payment of interest on its debt held by the cen tral bank is an inside transaction between gov ernment institutions that does not affect the resource transfer from the private money holders to government. Finally note that the central bank occasionally realizes capital gains (losses) by subsequently selling assets in the open market at higher (lower) prices than it had purchased them. To take these additional components of the revenue from base money production and main tenance into account, let the interest rates on the monetary authority’s holdings of private domestic debt (D) and official foreign debt (F) be denoted by d and f, respectively, and unrealized capital gains by GR Then, the extended mone . tary seigniorage, sM is: , (3) sM = s* + (dD + fF + G„)/P. 4See Friedman (1971), Calvo (1978), Fischer (1982), Dornbusch (1988), Grilli (1989) and Klein and Neumann (1990). 5“ Foreign deposits” include the demand balances of for eign central banks, the Bank of International Settlements, Extended monetary seigniorage encompasses the traditional measure o f monetary seigniorage. The new concept provides the seigniorage meas ure best suited for this study for two reasons. First, it directly measures the total real net flow of assets that the Federal Reserve System and U.S. Treasury receive from their monopoly over base money production; second, it can readily be computed from available data. EXTENDED MONETARY SEIG NIORAGE IN THE UNITED STATES: A DETAILED FRAME W ORK FOR ANALYSIS An analysis of extended monetary seigniorage in the United States can begin at either of two points: the "sources” side shows us how the gains were achieved, while the "uses” side tells us who received the gains. Sources o f U.S. Extended Monetary Seigniorage From the sources side, the extended monetary seigniorage is shown by equation 3. In more de tail, it can be written as: (4) sM = (AB + dD + fF + G„)/P, where B = C + R„ + RF . It is important to note that, in this analysis, the monetary base (B) is defined more broadly than is usually the case. Here, official foreign deposits at the Federal Reserve System (RF are added to ) the usual monetary base components of curren cy in circulation (C) and reserves of depository institutions (RB This expanded definition is ap ).5 propriate because the Federal Reserve obtains seigniorage from producing foreign deposits in precisely the same way it does from producing deposits for domestic depository institutions. Uses o f U.S. Extended Monetary Seigniorage To develop the uses side of extended mone tary seigniorage, two financial accounts are uti lized: (1) the combined Federal Reserve-Treasury "monetary” balance sheet and (2) the income statement of the Federal Reserve System. foreign governments, and international organizations, like IMF and World Bank; it excludes the Treasury’s Exchange Stabilization Fund. MARCH/APRIL 1992 32 The U.S. monetary authorities’ combined bal ance sheet can be written in first-difference form to show the changes that have occurred over some specific time period as follows: (5) AA fr + AD + AF + ACc + AOA = AB + ARg + AK. The left-hand side of equation 5 describes the ch an g es in the Federal Reserve’s assets that sup ply funds: outright purchases o f U.S. Treasury and federal agency obligations (AAF ), loans to R depository institutions via the discount window and government securities bought under repur chase agreements (AD), the acquisition of gold, special drawing rights, and foreign exchange (AF), issuance of coin by the Treasury (ACc) and other Federal Reserve net assets (AOA).6 The right-hand side of equation 5 describes the ch an g es in the factors that absorb these funds: the monetary base (AB), deposits o f the Treasury (ARg) and the Federal Reserve System's capital accounts (AK).7 Again, note that the mon etary base definition used in this analysis in cludes foreign deposits (RF held at the Federal ) Reserve. The Fed’s income statement is summarized in equation 6. The left-hand side describes the Fed’s current income and expenses that give rise to its net revenue; the right-hand side of equation 6 shows how the Fed’s net revenue is distributed. (6) dD + fF + aAp„ + GR + G^ - OCF R = + ^FR + Y GN . As noted earlier, d and f represent the inter est rates that the Federal Reserve receives on its loans to the domestic private sector and its in ternational assets, respectively; similarly, “a” 6ln contrast to their practice of valuing the domestic assets at the original purchase price, the Federal Reserve Banks mark their foreign exchange holdings to the market. As a consequence, reported changes in the stock of the Feder al Reserve System’s international assets include net pur chases at the actual transaction values (AF) and valuation gains (or losses) on the previous stock of these assets if foreign currency prices have changed. Because these valuation changes do not directly increase or absorb reserves, they are not included in equation 5. Incorporat ing them explicitly would simply introduce the same value on both sides of equation 5 and, hence, they would be “ netted out” of the analysis. 7RGincludes Treasury cash holdings. 8The Federal Reserve’s international assets include the na tion’s gold stock on which it receives no interest earnings. FEDERAL http://fraser.stlouisfed.org/ RESERVE BANK OF ST. LOUIS Federal Reserve Bank of St. Louis denotes the average interest return it receives on its portfolio of government securities bought outright.8 The next two terms are the "realized” profits (Gr) that the Fed receives from sales of its bonds and foreign assets at prices above those that it paid for them, and the "unrealized” profits (G,;) that result from the Fed’s practice o f marking the prices of its foreign exchange holdings to their market value. This accounting practice was introduced in 1978; before foreign exchange holdings w ere valued at historical rates.9 The term (OCF ) measures the current operat R ing costs or expenses of the Reserve Banks and the Federal Reserve Board minus the fees and reimbursements that the System collects for the services it sells to the banking industry, the Treasury and other government agencies. These service fees and reimbursements are "netted out” to remove receipts and expenses that are presumably unrelated to the Federal Reserve System’s monetary authority role. The right-hand side of equation 6 shows how the Federal Reserve’s net revenue (Y) is distri buted. The Fed pays its member banks statutory dividends (YB on their paid-in capital and uses ) an amount YF , which is equal to .5AK, to raise R the Reserve Banks’ surplus capital to the level of its member banks’ paid-in capital. The remain der of the System’s net income is transferred to the U.S. Treasury under the heading o f "In terest on Federal Reserve notes” (YG ).1 N 0 Subtracting equation 6 from 5 and using the identity that the current issuance of coin (ACc) equals the operating cost of the U.S. Mint (OCM ) plus the profit to the Treasury on the issuance of coin (YG ) yields: C 9Most European central banks do not mark their foreign ex change holdings to market values; instead, they evaluate their holdings at the lowest market price that occurred since they were acquired. As a result, their income state ments never show unrealized profits from their foreign ex change holdings; however, they show unrealized losses whenever the prices of foreign currencies fall below their acquisition values. 10As an historical aside, prior to 1933, the income transfer to the U.S. Treasury was effected as a franchise tax based on a provision of Section 7 of the Federal Reserve Act. This provision was repealed in 1933 to permit Reserve Banks to restore their surplus accounts, after they had been cut to one-half by the enforced subscription to the Federal Deposit Insurance Corporation, founded in 1933. 33 (7) AB + dD + fF + GR = (OCF + OCM + y fr+ y b R ) "*■ ^G,N "*■ Y g,c — — AK( ) + A(Afr + d + f + OA - K) - GU ( where: B = C + RB + RF and , ^G ~ ACc — OCM ,C , Y fr = .5AK. Dividing equation 7 through by the consumer price level (P) and using the definition of ex tended monetary seigniorage shown in equation 4 yields: (8) sM = sc + s„ + sG + s, + sL , where: sc = (OCF + OCM R )/P sB = Yb /P S = (YGc + YG\ + AA i!{ — sA fr — ARg)/P, G j s, = A(D + F + OA - ,5K)/P and sL = ( —Gu )/P. THE USES OF EXTENDED MONE TARY SEIGNIORAGE Equation 8 shows how the extended monetary seigniorage in each period is used: (1) sc is the cost o f providing the public’s desired real base money balances, including the costs associated with monetary policy and the Federal Reserve’s contribution to bank supervision, (2) member banks receive sB the statutory dividends, (3) the , government receives sGfor spending purposes, (4) the Federal Reserve uses s, to increase its portfolio of assets other than government debt and (5) the Fed uses sLto make up for booklosses resulting from adverse changes in asset prices. It is useful to consider in detail the seignior age distributed to the U.S. government, which may be termed “fiscal seigniorage.”1 Fiscal seig 1 niorage can be written in two different ways. 11See Klein and Neumann (1990) 12ln the theoretical literature, it is usually taken for granted that the net revenue received from the Federal Reserve (y g .n - a A f r ) cannot be negative since the government receives back as part of the transfer (YGN the interest ) paid on its debt to the central bank (a A F ). This conclu R The first way, using the government’s budget constraint, is (9a) sG = (G - T + aA0 - AA0 )/P, where (G - T) is the government’s primary budget deficit or surplus and aA0is the govern ment’s interest expenditure on its debt held out side the System (A0). Equation 9a shows that fiscal seigniorage is the portion of the govern ment’s deficit that is not financed by borrowing from the public (AA0 This means that fiscal ). seigniorage contributes to the finance of the primary budget deficit and of the interest ex penditures on debt held by the public (outside the Federal Reserve System). The second way o f writing fiscal seigniorage, as shown in equation 8 above, is (9b) sG = [Ygc + (Ygn - aAF ) + A(AF - RG R R )]/P. Equation 9b breaks down fiscal seigniorage into three source components: the net revenue from issuing coin (YG ), the net revenue received C from the Federal Reserve (YGN- aAF ), and the R net borrowing from Reserve Banks (AAfr- ARg).1 2 The treatment of net borrowing as a source o f fiscal seigniorage is not an obvious one, since the Fed does not lend directly to the Treasury; instead it purchases Treasury securities in the open market. From a purely technical point of view, the Treasury receives the borrowed funds from the public on the date of security issue, not from the Fed at the later date when the public resells the securities to the Federal Reserve. The above treatment of borrowing can be justified by the following considerations: First, from the economic point of view, what counts is not the first but the final placement of the Treasury securities. Thus, if the security dealers do not hold but resell the Treasury securities to Reserve Banks after a short duration, it is, in fact, the Fed that supplies the borrowed funds to the Treasury. At the same time, these trans actions permit the security dealers to buy an other load of new debt from the Treasury. Second, the bulk of the Federal Reserve’s pur- sion, however, holds only if the costs of the monetary authority are assumed away. In the United States, for ex ample, the Treasury’s interest payments to the Fed typical ly exceed the Treasury’s income received as “ interest on Federal Reserve notes.” MARCH/APRIL 1992 34 Figure 1 Extended Monetary Seigniorage and Operating Cost (in 1982/84 Consumer Prices) Billions of dollars chases of Treasury securities is at the short term end of the maturity spectrum. During the 1980s, for example, 83 percent of the securities purchased had maturities o f less than one year. For this large portion of newly acquired debt, the time difference between the public's buying and reselling plays no significant role in the em pirical analysis below based on annual observa tions. Third, some portion of new Treasury debt issued with maturities exceeding one year is also purchased by the Federal Reserve during the year of its issue. Finally, any approximation error with respect to the annual time unit ceases to play a role when annual average data for de cades are examined below. EMPIRICAL ANALYSIS Figure 1 shows the magnitudes of the extend ed monetary seigniorage (sM and the monetary ) authorities’ operating costs (sc) as measured in 1982/84 dollars, from 1951 to 1990. During the FEDERAL RESERVE BANK OF ST. LOUIS Billions of dollars past four decades, the annual real value of ex tended monetary seigniorage has generally risen, while ranging over that period from -$ 6 billion in 1954 to $31 billion in 1986. As figure 1 indi cates, a comparatively small portion of this amount—only about 7 percent on average—was used to cover the costs of producing the mone tary base by the monetary authorities. Conse quently, the government’s production of base money has resulted in sizable and rising net profits, which are equal to the difference be tween the two curves in figure 1. This result is shown in somewhat different form in table 1, which provides annual average data for each of the past four decades. During the 1980s, the annual real net profit from base money production and maintenance averaged more than $14 billion, while, during the 1950s, it averaged $1 billion per year. The steepest in crease in extended monetary seigniorage oc curred in the 1960s, when it jumped to almost $10 billion annually, on average, up sharply 35 Table 1 The Uses of Extended Monetary Seigniorage1 1951-60 Extended monetary seigniorage sM - Operating cost, sc = Net profit - Dividend payment to member banks, sB - Book-loss on foreign exchange, sL 1961-70 1971-80 1981-90 $1,555 $9,847 $14,373 $14,948 454 695 1,039 1,101 9,152 13,334 14,202 67 100 101 96 n.a. n.a. 233 7462 -3 7 9 - Investment in loans and foreign assets, s, -1 ,3 9 7 -1 ,1 5 9 1,855 3,283 = Fiscal seigniorage, sG $2,431 $10,211 $11,355 $11,202 $1,498 $9,683 $14,414 $13,945 1.0% 2.8% 2.1% 1.6% Memo: Traditional monetary seigniorage, Fiscal seigniorage as percent of real federal on-budget spending 'Annual averages in millions of dollars, 1982/84 consumer prices. 2Net of revenue from priced services ($501 million). 3Starting in 1978. from its 1950’s level. About 80 percent of this rise resulted from drastic increases in average reserve requirements during the 1960s. As table 1 shows, the average annual total operating costs of the monetary authorities in creased by about 50 percent during the 1970s from its earlier levels. This rise primarily reflects the Federal Reserve’s efforts to in troduce a variety of services in the 1970s as sociated with the payments mechanism. The monetary seigniorage used for covering operat ing costs fell in the 1980s when the Fed began to charge explicitly for these services.1 3 Dividend payments to member banks on their paid-in capital (sB which run about $.1 billion ), per year, and the System's accounting losses on its holdings o f foreign exchange (sL represent ) 13For example, if the cost of priced services was added to the operating costs for the 1980s, it would raise the figure shown by almost 75 percent. 14During the 1970s, the Fed’s realized (as opposed to ac counting) losses on foreign exchange amounted to about $148 million per year in real terms. These losses resulted from foreign exchange intervention attempts to stem the fairly negligible uses o f the total monetary seig niorage. As table 1 indicates, these accounting adjustments began in 1978, when the Fed start ed valuing its foreign exchange holdings at cur rent market prices.1 4 During the 1980s, the Federal Reserve System accou nted an annual valuation gain on its for eign exchange holdings averaging $380 million. This gain reflects the appreciation of the Deutschmark and yen against the dollar from 1985 to 1987 and again in 1989/90. Occasionally, the Fed also realized profits on foreign ex change holdings; they averaged $151 million per vpar rhirinff thp 1980s As in all countries, the bulk of the extended monetary seigniorage went to the government, The average annual flow o f fiscal seigniorage sharp decline in the value of the dollar during the 1970s. They appear, of course, on the sources side of the seig niorage equation and, hence, reduce the total monetary seigniorage collected; see equation 4. MARCH/APRIL 1992 36 Table 2 The Components of Fiscal Seigniorage1 1951-60 1961-70 1971-80 1981-90 $2,295 $10,525 $12,117 $10,462 -5 2 9 -54 1 -1 ,2 3 3 455 591 428 1,092 587 74 -2 0 1 -6 2 1 -3 0 2 $2,431 $10,211 $11,355 $11,202 Interest received on FRnotes $1,671 $5,214 $10,489 $15,954 Interest paid to Federal Reserve $2,200 $ 5,755 $11,722 $15,499 U.S. government debt bought outright by Fed + Interest received on FRnotes net of interest paid to Fed + Coin issued net of outlays of U.S. Mint -C h a n g e in Treasury’s deposits with Fed = Fiscal seigniorage Memo: 'Annual averages in millions of dollars, 1982/84 consumer prices. rose from $2.4 billion during the 1950s to $11.2 billion in the 1980s. The dominating source component of fiscal seigniorage is the outright acquisition of government securities by the Fed. Just how important it is can be seen in table 2; for all practical purposes, it matches the total. This observation underscores the fact that the seigniorage flow to government must not be identified with the Fed’s payment to the Treasu ry of "interest on Federal Reserve notes.” In ser vicing the debt held by the Fed, the Treasury makes interest payments of roughly the same order of magnitude as the Fed pays to the Treasury (see the bottom lines in table 2).1 In 5 deed, the Fed’s portfolio of U.S. government securities can be interpreted as an interest-free loan to the Treasury. While during the 1970s and the 1980s fiscal seigniorage amounted to 80 percent of mone tary seigniorage collected, it even exceeded the 15Barro (1982) was not aware of this, when he identified the interest on Federal Reserve notes as the revenue from money creation. But note that Barro would be correct if the Fed, like the Deutsche Bundesbank, would mainly hold earning assets other than government debt. FEDERAL http://fraser.stlouisfed.org/ RESERVE BANK OF ST. LOUIS Federal Reserve Bank of St. Louis total flow during the 1950s and the 1960s. How was it possible that the government consumed more seigniorage than was collected? The an swer to that question is asset substitution in favor of U.S. government debt: for a given base money stock, the Fed can reduce its loans to the banking sector or sell foreign assets and use the proceeds for buying outright government debt. For example, if the Fed replaces foreign assets worth $100 million with Treasury bills of the same amount, it foregoes foreign interest earnings of, say, $5 million. As a result, the cur rent flow of fiscal seigniorage is increased by $95 million (see equation 9b).1 To be sure, this 6 is a one-time effect. During subsequent periods, the flow of fiscal seigniorage will be smaller than otherwise, because of the lost stream of foreign interest earnings. As table 1 indicates, the observed differences between the annual flows of monetary and fis cal seigniorage are largely due to asset substitu tion. During the 1950s and the 1960s, the Fed 16While the discussed transaction reduces the Fed’s interest earnings on foreign assets, it raises the interest earnings on the portfolio of Treasury securities. But, as noted above, this does not affect the net flow of fiscal seig niorage. 37 raised the annual flow of fiscal seigniorage above the flow of extended monetary seigniorage by replacing non-government debt worth more than $1 billion each year, on average, with U.S. government securities. The reverse policy was chosen thereafter so that the annual flow of fiscal seigniorage fell behind monetary seignior age by an average of almost $4 billion during the 1980s. While the continuous flow of fiscal seigniorage helps to finance the federal budget, it is a fairly small source of funds. On average, it contrib uted about 2 percent to the finance of federal expenditures over the past 40 years. Seigniorage and Inflation In the monetary economics literature, seig niorage is often discussed and analyzed in terms of an “inflation tax,” a term that was coined by Milton Friedman (1953). This association reflects the fact that, other things the same, a nation’s monetary authorities can increase monetary seigniorage by increasing the supply of base money relative to its demand. Because the re sulting rising price level reduces the real value of the public's base money holdings, the public will demand more nominal base money balances to make up for the price-level-induced decline in its real cash balances. As a result, the price rise produces an increase in monetary seigniorage. Extended monetary seigniorage, however, will not rise in some fixed proportion to inflation; the demand for real cash balances is inversely related to the rate of inflation. Hence, the in crease in seigniorage associated with higher and higher inflation becomes smaller and smaller; eventually, some inflation rate is reached at which monetary seigniorage is maximized. Thereafter, higher inflation will reduce the level o f seigniorage as the inflation-induced effect dominates the price-level effect on the public’s demand for real cash balances. flation rate has passed a certain threshold. Thus, the predicted relation resembles the shape of the Laffer curve in public finance where the revenue from the income tax first rises with the effective tax rate but begins to decline once the disincentive effect of too high a tax rate be comes dominant. Monetary theorists have applied profit-maxi mizing conditions for a monetary authority to generate the seigniorage-maximizing rate of in flation.1 This concept, however, is unlikely to 7 yield much insight in the actual behavior of monetary authorities. While, in history, mone tary authorities of several countries have re peatedly produced larger inflations in the attempt to accommodate fiscal problems, central banks, in general, are not profit-oriented organi zations, and ascribing profit-maximizing motives to them is misleading.1 8 Thus, instead of looking for some theoretically justified story about inflation and the motives of the Federal Reserve System, we are better o ff by simply looking at the "stylized facts” about the relationship between inflation and the ex tended monetary seigniorage in the United States. Figure 2 provides one way of assessing this rela tionship. The points in the diagram show the rate of inflation and the associated monetary seigniorage in each year from 1951 to 1990. As expected, the data reveal an initial positive rela tionship between inflation and extended mone tary seigniorage. As also expected, however, this positive association slowly disappears, then be comes negative for sufficiently high rates of inflation. Thus, the empirical relationship resem bles the shape of a Laffer curve. The curve drawn in figure 2 shows the re sults of estimating extended monetary seignior age (sM as a quadratic function of the rate of ) inflation ( tt): (10) sM = aQ + ajTi - a 2n2 + e, In sum, monetary theory predicts that seig niorage rises with inflation but falls once the in where e is a white-noise residual. The estimated parameters imply that monetary seigniorage be- 17Consider the traditional concept of monetary seigniorage, as defined above by equation 2. It can be rewritten as: (a) Si, = (AB/B)(B/p). um in which y and r are constant. This implies: n = AB/B. Maximizing equation a subject to equation b yields the seigniorage-maximizing rate of inflation: Next, assume a standard money demand function: (b) B/p = y exp[ - A.(r + n)], (c) n°p' = 1/A. 18For a different view, see Toma (1982). where (y) is real income, (r) is the real rate of interest and n the inflation rate. Finally, assume a steady-state equilibri- MARCH/APRIL 1992 38 Figure 2 Extended Monetary Seigniorage and Inflation (1982/84 Consumer Prices) B illions of dollars Billions of dollars P ercent gins to decline once inflation exceeds a rate of 7.9 percent.1 9 From the point of view of the U.S. govern ment, fiscal seigniorage is more interesting than monetary seigniorage because the former meas ures what the government actually receives for budget finance. As figure 3 shows, fiscal seig niorage is quite similarly related to the level of inflation, except that it reaches a maximum at an inflation rate of 7.2 percent.2 0 These estimates suggest that high inflation, at least more than 7 percent or 8 percent per year, has been less profitable for the U.S. 19Estimating equation 10 with a dummy for 1986 yields: a0 = -9 7 9 (-0 .5 5 ), a, = 4,325 (5.94), and a, = 269 (4.73), numbers in parantheses are t-values. R 2 =.62, DW = 1.52. 20The respective curve is computed from estimating equa tion 10 with fiscal seigniorage as the dependent variable. Denoting the estimated parameters by b yields: b0 = 1,233 (0.56), b, = 3,123 (3.51), b2 = 217 (3.07), R 2 = .28, DW = 1.88. FEDERAL http://fraser.stlouisfed.org/ RESERVE BANK OF ST. LOUIS Federal Reserve Bank of St. Louis government when monetary or fiscal seig niorage alone are considered. This is demon strated in table 3, where average annual seigniorage flows are compared over different inflation ranges. During the high inflation years, when inflation exceeded 9 percent annually, fis cal seigniorage averaged 7.7 billion per year. This is about 5 percent lower than it averaged during the low inflation years and even 45 per cent lower than during the medium inflation years.2 1 CONCLUSION The government's monopoly in issuing base money yields profits that facilitate its fiscal 21While it may be tempting, given the evidence presented in table 3, to conclude that the U.S. government should prefer inflation in the 4.6 percent to 9 percent annual range, it would be a mistake to do so. First, there are other social (and governmental) gains from lower inflation that are not examined here. Second, and perhaps more relevant, the U.S. rate of inflation has been below 4.5 per cent for 25 of the 40 years between 1951 and 1990. 39 Figure 3 Estimated Seigniorage and Inflation (in 1982/84 Consumer Prices) Billions of dollars Billions of dollars Percent Table 3 Seigniorage and Inflation1 Inflation range - 0 . 3 to 4.5% 4.6 to 9.0% 9.1 to 13.6% Monetary seigniorage, sM $8,056 $15,030 $11,106 Fiscal seigniorage, sG $8,096 $11,117 $ 7,685 Number of years Average inflation rate 25 10 5 2.3% 6.2% 11.1% 'Annual averages in millions of dollars, 1982/84 consumer prices. MARCH/APRIL 1992 40 finance. This paper developed a new measure of monetary seigniorage and presented a frame work for analyzing and measuring the total seigniorage flow from base money production and its allocation to various uses, including fis cal finance, in the United States. In addition, the paper analyzed the relation ship between monetary and fiscal seigniorage and inflation in the United States from 1951 to 1990. While it is well-known that, within certain limits, governments are able to increase their seigniorage flows through higher inflation, the limits to such actions are not as well-known. The evidence presented here suggests that the U.S. government's fiscal seigniorage declines when the rate of inflation exceeds 7 percent. In deed, in those years when inflation exceeded 9 percent, the U.S. government’s fiscal seig niorage fell short of the levels achieved when U.S. inflation rates were less than 4.5 percent. REFERENCES Auernheimer, Leonardo. “ The Honest Government’s Guide to the Revenue from the Creation of Money,” Journal of Po litical Economy (May/June 1974), pp. 598-606. Friedman, Milton. “ Discussion of the Inflationary Gap,” Essays in Positive Economics (University of Chicago Press, 1953). Bailey, Martin J. “ The Welfare Cost of Inflationary Finance,” Journal of Political Economy (April 1956), pp. 93-110. ________“ Government Revenue from Inflation,” Journal of Political Economy (July/August 1971), pp. 846-56. Barro, Robert J. “ Measuring the Fed’s Revenue from Money Creation,” Economic Letters Vol. 10 (1982), pp. 327-32. Grilli, Vittorio. “ Seigniorage in Europe,” in Marcello de Cecco and Alberto Giovannini, eds., A European Central Bank? Perspectives on Monetary Unification after Ten Years of the EMS (Cambridge University Press, 1989). Brunner, Karl, and Allan H. Meltzer. “ The Uses of Money: Money in the Theory of an Exchange Economy,” American Economic Review (December 1971), pp. 784-805. Calvo, Guillermo A. “ Optimal Seigniorage from Money Crea tion: An Analysis in Terms of the Optimum Balance of Pay ments Deficit Problem,” Journal of Monetary Economics (August 1978), pp. 503-17. Dornbusch, Rudiger. “ The European Monetary System, the Dollar and the Yen,” in Francesco Giavazzi, Stefano Micossi and Marcus Miller, eds., The European Monetary System, (Cambridge University Press, 1988). Fischer, Stanley. “ Seigniorage and the Case for a National Money,” Journal of Political Economy (April 1982), pp. 295-313. FEDERAL RESERVE BANK OF ST. LOUIS Gros, Daniel. “ Seigniorage in the EC: The Implications of the EMS and Financial Market Integration,” IMF Working Paper (Washington, D.C., January 23, 1989). Johnson, Harry G. “A Note on Seigniorage and the Social Saving from Substituting Credit for Commodity Money,” in Robert A. Mundell and Alexander K. Swoboda, eds., Mone tary Problems of the International Economy (University of Chicago Press, 1969). Klein, Martin, and Manfred J.M. Neumann. “ Seigniorage: What is It and Who Gets It?” Weltwirtschaftliches Archiv (Heft 2/1990), pp. 205-21. Toma, Mark. “ Inflationary Bias of the Federal Reserve Sys tem: A Bureaucratic Perspective,” Journal of Monetary Eco nomics (September 1982), pp. 163-90. 41 James B. Bullard James B. Bullard is an economist at the Federal Reserve Bank of St. Louis. David H. Kelly provided research assistance. The FOMC in 1991: An Elusive Recovery \ J- JLS 1991 BEGAN, the U.S. economy was in the second quarter of a downturn in aggregate economic activity. Real output, as measured by the gross national product (real GNP), had fallen in the fourth quarter of 1990 at a 2.5 percent annual rate; the first quarter of 1991 would turn out to be even worse, with output falling at a 2.8 percent annual rate. As the year wore on, the pace of real output growth turned posi tive, although it seemed to stall somewhat toward the fourth quarter. The recession and the subsequent slow recovery put pressure on the primary policymaking group of the Federal Reserve, the Federal Open Market Committee (FOMC), to take action to spur greater output growth in the short term.1 This paper provides a chronologically based assessment of the Com mittee's policymaking in this environment. As such, it provides a case study in the making of monetary policy during the recovery phase of the business cycle. Generally speaking, the FOMC has well-defined goals but faces two daunting uncertainties when making decisions. One is that the immediate past, current and future path o f real output is not easily surmised by considering current data. This inhibits the Committee’s ability to assess the state of the economy in a timely fashion and, thus, to make short-run policy decisions. Secondly, the Committee has a difficult time assessing its own policy stance at a point in time, primarily because alternative measures of policy actions sometimes send conflicting signals. In order to put FOMC decision making into context, a framework for discussion is outlined. The framework is intended to provide a basis for thinking about how monetary policy is made and why, in a broad brush sense, certain con cerns reign paramount in Committee delibera tions. The framework is meant to be qualitative and suggestive, so as not to belabor the techni cal details of theoretical arguments. To understand the FOMC’s decision making in 1991, a general framework or reference point is useful in order to put into focus the arguments presented for various policy actions. For the most part, FOMC members and the Board staff, the primary participants in these meetings, pre sent broad points of view and avoid technicali ties. Disagreement, when it occurs, is often a matter o f the interpretation of recent economic The next section provides the framework for understanding FOMC decision making. The chro nology is presented in the subsequent section. T h e fin al section p rovid es sum m ary com m ents. A FRAMEWORK FOR ANALYZING FOMC POLICY ACTIONS 1See the shaded insert on the following page for a discus sion of the structure of the FOMC in 1991. MARCH/APRIL 1992 42 The Organization of the FOMC The Federal Reserve System consists of 12 regional Federal Reserve Banks located in major cities across the country, with the ad ministrative offices o f the Federal Reserve Board o f Governors in Washington, D.C. The Federal Reserve Board consists of seven mem bers, and each o f these members has voting rights on the Federal Open Market Commit tee (FOMC). The president o f the New York Federal Reserve Bank also is a permanent voting member of the FOMC. The remaining 11 Reserve Bank presidents attend meetings and present views, but only four of the 11 have voting privileges at any one meeting. The voting rights are held for terms of one calendar year and rotate among these presi dents annually. The Committee typically meets eight times per year, as it did in 1991, and sometimes consults by telephone between scheduled meetings. At the end of each meeting, the Committee agrees on a directive to issue to the Federal Reserve Bank of New York; the developments, but sometimes concerns the amount of weight to attach to certain broadly theoretical arguments. Before beginning an anal ysis of Committee deliberations, it is therefore helpful to consider, in a nontechnical way, the ideas that underlie Committee debate. A frame directive is implemented by the Manager of Domestic Operations. The directive contains instructions for the conduct of open market operations until the next regularly scheduled meeting. A summary of each FOMC meeting is released to the press within a few business days following the next regularly scheduled meeting and is subsequently published in the F ed era l R eserv e Bulletin. The summary, known as the "Record of Policy Actions of the Federal Open Market Committee,” is pre pared by the Board staff and approved by the Board. It typically contains: (1) a synopsis of recent economic data, (2) a review of re cent open market operations and money mar ket conditions, (3) a Board staff projection of likely near-term economic developments, (4) a summary of Committee deliberations, (5) the policy directive along with a Record of votes and any dissenting comments, and (6) a sum mary o f other policy matters discussed. The Federal Open Market Committee seeks monetary and financial conditions that will foster price stability, promote a resumption of sustainable growth in output, and contrib ute to an improved pattern of international transactions.2 w o r k o f this sort w as presen ted in Bullard (1990), and this section briefly describes that approach. F O M C Monetary P olicy Objectives The Committee states its goals for monetary policy repeatedly in documents released to the public throughout the year. In particular, at the conclusion o f each meeting, the Committee is sues a directive which contains, with other in formation, a statement o f the following type: 2Federal Reserve press release, March 29, 1991, p. 22. The Federal Open Market Committee releases its record of policy actions (in the remainder of this article, simply ‘‘the Record” ) for a particular meeting shortly after the next regularly scheduled meeting. The press releases referred to in the remainder of this article are dated May 17, 1991; July 5, 1991; August 23, 1991; October 4, 1991; November 8, 1991; December 20, 1991; and February 7, 1992. All press releases will be referred to by month. FEDERAL http://fraser.stlouisfed.org/ RESERVE BANK OF ST. LOUIS Federal Reserve Bank of St. Louis This statement of objectives has three parts. The first goal, to foster price stability, is based on the idea that FOMC policy, over long periods of time, can influence the inflation rate. The se cond objective, to promote sustainable growth, is associated with the idea o f countercyclical monetary policy, in particular that the FOMC can influence real output over short time hori zons, say, less than a year.3 The third part of the statement of objectives, an improved pattern of international transactions, is more oblique, 3The statement of objectives quoted above is noteworthy for the inclusion of the words “ a resumption of.” Available information suggested that real output was declining at the time this statement was released, and hence the more standard phrase, “ . . . promote sustainable growth in out put,” was modified. Later in the year, when real output again appeared to be growing, albeit rather slowly by historical standards, the term “ resumption” was dropped. 43 and beginning about midyear, this phrase was dropped from the Committee’s statement. Therefore, for the last five meetings o f the year, the statement of objectives was: The Federal Open Market Committee seeks monetary and financial conditions that will foster price stability and promote sustainable growth in output.4 In this article, focus will be placed on this last statement o f objectives, as the price stability and real output goals are consistently mentioned throughout the year, and the international goal is not. Controlling Inflation A widely held view among economists is that inflation, over long time periods, is closely relat ed to the rate of money growth within a coun try. In fact, the idea is that the rate o f money growth eventually translates directly to the rate of inflation. The theory, which dates to Hume (1742), is broadly supported by international cross-section evidence, which shows that coun tries choosing high rates of money growth over a decade or more tend to have the highest infla tion rates. The United States, for instance, has experienced an average annual rate o f inflation o f 5.4 percent in the 1980s, which was associat ed with an average annual money growth (M2) rate o f 7.5 percent. Iceland, on the other hand, experienced 32.1 percent average annual infla tion over the same period, associated with a money growth rate o f 38.2 percent. Similarly, Mexico had 50.1 percent average annual infla tion associated with money growth of 45.9 per cent. A look at other countries in which data are available reveals similar patterns. Of course, despite the fact that such views are widely held, the theory is incomplete. It is not clear, for instance, what constitutes a suffi ciently long time period. In addition, as the ex amples given above indicate, the relationship is far from exact, even over a decade or more. Fi nally, the theory by itself gives no indication of what to expect from, say, a short but intense burst of money growth. Despite these caveats, the theory and evidence are sufficiently strong to suggest that, in the long run, inflation is a policy-induced phenomena, and thus that con trol of inflation is an important aspect of central bank policymaking. “•February press release, p. 16. 5March press release, p. 15. Sustaining Real Output Growth The notion that monetary policy actions can significantly affect the growth o f real output over short time horizons, such as a quarter or a year, is deeply seated among macroeconomists. It is also controversial and largely unresolved. Nevertheless, the Committee has generally adopted the view that monetary policy actions do have material effects on real output growth within the following few quarters. It is not difficult, for instance, to find statements in the Record that attest to members' views in this regard. For instance, at the February meeting, there was talk that “inadequate monetary stimu lus . . . could fail to cushion possible further deterioration in the economy.”5 Similarly, in March, “if the economy was indeed near its recession trough, additional easing would not be necessary,” in May, "the System’s earlier easing actions . . . had provided a good deal of insur ance against cumulative further weakening in business activity;” and in July, "policy was posi tioned to foster a sustainable economic expan sion.”6 Statements of a similar sort can be found throughout the year. Since the Committee operates in an environ ment in which the short-term effects of mone tary policy on real output are taken for granted, in this paper these effects are simply assumed to exist and to be substantive, with due notice to the ongoing debate on that topic in academic circles. Generally speaking, it will be assumed that an "easing” of monetary policy is associated with a temporary gain in output (relative to what would have occurred without the easing) a few quarters hence, and that a "tightening” of policy has the reverse effect. The R ole o f Forecasts in ShortRun Policy Actions It is important to note that these postulated real output effects occur only with a lag, which many economists suppose is at least one quarter and may be as long as a year or more. The no tion of lagged policy effects is an important theme in Committee debates, as it forces mem bers to form opinions about the path of real output several quarters into the future. Such forecasting is notoriously difficult. The inability to forecast accurately tends to produce uncer6Respectively, the May press release, p. 12, the July press release, p. 12, and the August press release, p. 14. MARCH/APRIL 1992 44 tainty among policymakers when choosing ap propriate short-run policy actions. If the members of the FOMC were concerned solely with long-run policy, as would be suggest ed by the available theory and evidence on in flation, they would presumably have much less concern for current forecasts o f real activity over the next few quarters. But the Committee is not concerned solely with inflation, as their statement of objectives attests; therefore, the Committee members and Board staff have an acute concern for the daily goings-on of the U.S. economy. In fact, the Record consists pri marily of a recitation of recent economic devel opments as captured in various measurements produced by the Federal Reserve or the U.S. government, often with an associated inference about what seems to be in store for real activi ty. The idea that policy actions taken today will affect real output in the not-too-distant future drives the concern for up-to-the-minute informa tion about the status of economic activity. Shortrun forecasting is a necessary ingredient of any strategy based on the notion o f significant short-run monetary policy effects on real output. Measuring the Policy Stance The Committee also has some difficulty in as sessing the policy stance at a point in time be cause various measures of the thrust of policy can give conflicting signals. The Record and other Federal Reserve documents describe poli cy in terms of whether it is "tight” or “easy.” These vague terms, which have a long history o f use within the Federal Reserve, cannot be as sociated directly with System actions. This has created the situation in which two observers, and indeed Committee members themselves, can easily disagree about the thrust of monetary policy at a point in time. To see how this might be so, consider how most monetary policy ac tions are implemented. Commercial banks must maintain reserves against certain types of deposits. Reserve posi tions are maintained on a two-week basis, so that a particular depository institution might find itself either over- or under-supplied with reserves at a point in time. These reserves are traded among banks on a daily basis in the fed eral funds market, and the interest rate in this 7For a recent discussion, see Garfinkel and Thornton (1991). FEDERAL http://fraser.stlouisfed.org/RESERVE BANK OF ST. LOUIS Federal Reserve Bank of St. Louis market is the federal funds rate. The Federal Reserve can supply or drain reserves from the system through intervention in this market. Such op en m arket o p eration s are carried out by the Federal Reserve Bank of New York based on the directives from the FOMC. Given a conventional downward-sloping de mand for reserves, the Federal Reserve can in crease (decrease) the federal funds rate by decreasing (increasing) the supply of reserves. Total reserve supply is subject to close control by the System. A standard analysis relates the sum of total reserves and currency, the m o n e tary base, and measures of the money supply, such as M2, by a proportional factor called the m on ey m ultiplier.7 Thus, increases in total reserves, increases in the money stock and decreases in the federal funds rate are simply different aspects of the same mechanism in this simple framework and are associated with the term "ease.” “Tight” policy would involve move ments of these variables in opposite directions. Since, generally speaking, lower long-term in terest rates are associated with higher rates of investment and greater consumer purchasing, which in turn are associated with higher levels of real output, easy policy is often thought to be stimulative in the short run. Tight policy is viewed as having the reverse effect. Of course, the federal funds rate is only a short-term in terest rate, and there might be some question whether movements in a short rate will actually be reflected in the spectrum o f interest rates. Theories on this topic abound, and, as will be apparent in the following chronology, members of the Committee sometimes disagree about the net interest rate effects of a lower federal funds rate. The essential problem in practice is that total reserves, the federal funds rate and the money supply can, and sometimes do, give conflicting signals about whether monetary policy is actual ly easy or tight. Casual consideration of the above analysis suggests why this might be so. In particular, the demand for reserves is probably shifting over time in response to the level of economic activity, implying that the federal funds rate could rise or fall even when reserves are constant. Additional reserves, therefore, need not signal a lower federal funds rate. M2 growth might also be affected by vagaries in real activity. 45 These and other concerns produce the funda mental problems in assessing the FOMC’s policy stance at a point in time. The Federal Reserve also occasionally lends reserves to commercial banks directly through the discount w indow . The rate on these loans, the discount rate, is administered by the Board of Governors, and the FOMC plays no official role in changing it. The volume of loans made by the Federal Reserve at the discount rate is relatively small, so that the direct impact of a change is viewed as relatively unimportant. In the past, however, the discount rate has been set somewhat below the prevailing federal funds rate, so that, in 1991, when the federal funds rate declined through most of the year, a lower discount rate was sometimes taken by market participants as a signal of a lower feder al funds rate at some point in the future. In fact, in 1991, the Committee often allowed a lower discount rate to sh o w through to the fed eral funds rate, meaning that the funds rate was allowed to fall when the discount rate was lowered. Therefore, discount rate changes play an important role in the following chronology, even though the discount rate is not, strictly speaking, under the jurisdiction of the FOMC. Based on the simple framework outlined above, the three indicators of policy that will be considered in this paper are the federal funds rate, the M2 monetary aggregate and total reserves.8 The behavior of these indicator varia bles in 1991 is summarized in figures 1-3; refer ence will be made to these graphs throughout the chronology. Figures 4-7 provide a synopsis of the recent behavior of several other key variables—namely, real output, total nonfarm payroll employment, industrial production and consumer confidence. The framework that will be used to summa rize FOMC decision making is now complete. The Committee states its major objectives fre quently: to control inflation and maintain sus tained growth in real output. International evidence suggests that low inflation rates can be achieved by maintaining low rates of money growth. The real output effects of monetary policy are less certain, but summaries of Com 8While there are many other possible indicators of mone tary policy, in this article these three are the only ones considered. 9A summary of FOMC actions in 1991 is contained in table 1. mittee deliberations indicate that members be lieve temporary easing can mitigate downturns in economic activity. Pursuit of this objective re quires an assessment of the current and future path of real output, but knowing whether the incoming data signal a change in direction for the economy is complicated by lags in data releases and errors in economic forecasts. To summarize FOMC policy actions, some measure of the monetary policy stance is required. Since various measures sometimes suggest differing interpretations of the thrust of monetary policy, several indicators will be employed. A CHRONOLOGY OF FOMC DECISION MAKING IN 1991 On the whole, the chronology in this section indicates that the FOMC became increasingly pessimistic about the prospects for a sustained recovery as 1991 progressed, and this led to particularly aggressive easing actions late in the year.9 In the first meetings of 1991, when a substantial amount o f information suggested a decline in real output, members were neverthe less hopeful that easing implemented since the December 1990 meeting would be enough to lay the groundwork for a moderate recovery begin ning in the spring and summer. In fact, several directives in the first half o f the year called for steady policy with no bias toward ease, although some easing actions actually w ere implemented according to the Record.1 Beginning in August, 0 in an atmosphere o f increasing concern about the strength of the recovery, the Committee turned to asymmetric language toward ease in the directive. The trend toward easing actions peaked in the November directive, with the C om m ittee vo tin g to su pport im m ediate ease with additional bias toward ease during the intermeeting period. As emphasized in the chronology, however, merely outlining the content of Committee direc tives does not provide a complete summary of monetary policy during this period. At times, for instance, the thrust of policy is open to in terpretation. In addition, policy changes are sometimes implemented via other methods, such as intermeeting conference calls, or in concert with discount rate changes. 10The Committee sometimes uses so-called asymmetric lan guage in the directive, which is one way of making policy changes contingent on intermeeting developments. This phenomena is also sometimes described as “ bias” in the directive. MARCH/APRIL 1992 46 Table 1 Important Dates in the Chronology of 1991 FOMC Actions The following summary is based solely on statements in the Record regarding policy actions. See the text for a discussion of measures of the thrust of policy. Early January. An easing action is implemented. February 1. The Board of Governors approves a reduction in the discount rate to 6 percent from 6.5 percent. The FOMC allows the entire amount of the cut to show through to the federal funds rate. February 5-6. The FOMC meets. The target range for M2 growth is kept at 2.5 to 6.5 percent. The directive calls for an unchanged policy with some bias toward ease depending on intermeet ing developments. Early March. An easing action is implemented. March 26. The FOMC meets. The directive calls for an unchanged policy without bias. End of April. The Board of Governors cuts the discount rate from 6 percent to 5.5 percent. The FOMC allows part of the 50 basis-point decline to show through to the federal funds rate. May 14. The FOMC meets. The directive calls for an unchanged policy without bias. July 2-3. The FOMC meets. The target range for M2 growth is kept at 2.5 to 6.5 percent. The directive calls for an unchanged policy without bias. Early August. An easing action is implemented. August 20. The FOMC meets. The directive calls for an unchanged policy with some bias toward ease depending on intermeeting developments. Mid-September. The Board of Governors lowers the discount rate from 5.5 percent to 5 percent. The FOMC allows part of the decline to show through to the federal funds rate. October 1. The FOMC meets. The directive calls for an unchanged policy with some bias toward ease depending on intermeeting developments. End of October. An easing action is implemented. November 5. The FOMC meets. The directive calls for immediate ease with bias toward addition al ease depending on intermeeting developments. November 6. The Board of Governors lowers the discount rate to 4.5 percent. An easing action is implemented in concert with the discount rate cut. Early December. An easing action is implemented. December 17. The FOMC meets. The directive calls for an unchanged policy with bias toward ease depending on intermeeting developments. December 20. The Board of Governors lowers the discount rate by a full percentage point to 3.5 percent. The FOMC allows partial show-through to the federal funds rate. http://fraser.stlouisfed.org/ RESERVE BANK OF ST. LOUIS FEDERAL Federal Reserve Bank of St. Louis 47 Figure 1 Weekly Federal Funds Rate Percent Jul 90 Aug O ct Nov Dec Feb 91 M ar May Jul Aug O ct Nov Dec Feb 92 Vertical lines represent FOMC m eeting dates Figure 2 13 Week Growth Rates of M2 Annualized Percent Change Vertical lines represent FOMC m eeting dates MARCH/APRIL 1992 48 Figure 3 Intermeeting Growth of Total Reserves1 A n n u alized P e rc en t C h ange A u gust 1990 to February 1992 40 -2 0 A u g 90 O ct N ov Dec Feb 91 M ar M ay Jul Aug O ct 1D a ta a re Board series, ad ju sted fo r res e rv e req u irem en ts. V e rtic a l lin es re p rese n t F O M C m eetin g d ates Figure 4 Private Forecasters’ View of Real Output1 An nualized P e rc en t Change 'A v a ila b le data and Blue Chip forecasts FEDERAL RESERVE BANK OF ST. LOUIS N ov D ec Feb 92 49 Figure 5 Total Nonfarm Payroll Employment M illio n s o f P e rs o n s Figure 6 Industrial Production A n n u a liz e d P e rc e n t C h a n g e J a n u a ry 198 9 to J a n u a ry 1992 MARCH/APRIL 1992 50 Figure 7 Consum er Confidence In d e x L e v e l 1 9 8 5 = 100 Meeting o f February 5-6, 1991 In keeping with standard practice and Con gressional requirements, the FOMC took up a review of long-range policy at the February meeting. As in the past, most of the discussion focused on target ranges for monetary and debt aggregates, primarily the M2 monetary aggregate. In July 1990, the Committee had tentatively set the 1991 target range for M2 growth at 2.5 to 6.5 percent, measured from the fourth quarter o f 1990 to the fourth quarter of 1991. As can be seen from figure 2, the recent 13-week growth rates for M2, at the time of this meet ing, had mostly been below the target band.1 1 In recent years, the FOMC has pursued a strategy of gradually reducing the target ranges for M2 growth, usually in increments of 0.5 percent per year, with the idea of eventually at taining a range consistent with price stability.1 2 At this meeting, however, "most of the mem bers indicated a preference for affirming the ranges that had been established on a tentative basis in July.”1 The essential reason for the in 3 terruption in the usual sequence was the weak economy, in particular, that “lowering [the tar get range]. . . could lead to concerns about the System’s objective of fostering an upturn in bus iness activity.”1 On the other hand, increasing 4 the ranges, perhaps in an effort to show reces sion-fighting resolve, was also viewed with sus picion, since it “could raise questions about the System’s commitment to its anti-inflationary goals.”1 After further debate on these points, 5 the Committee agreed to a directive calling for maintenance of the tentative M2 target ranges, that is, with the lower end at 2.5 percent and the upper end at 6.5 percent.1 6 In terms of short-run policy, that is, policy for the immediately upcoming intermeeting period, " O f course, it is not necessarily of concern that the 13-week M2 growth rates sometimes fall outside the target band; it is the growth rate over the entire year that is of im portance. 13March press release, p. 14. 12March press release, p. 14. 16March press release, p. 17. FEDERAL RESERVE BANK OF ST. LOUIS 14March press release, p. 15. 15March press release, p. 15. 51 the developments since the previous FOMC meeting were an important consideration. The December 1990 directive had called for some in itial easing along with additional easing should conditions warrant. The Board o f Governors, which exercises authority over discount rate changes separately from the FOMC, voted to lower the discount rate to 6.5 percent immedi ately following the December meeting, and, ac cording to the Record, some of this decline was allowed to show through to the federal funds rate.1 Further easing followed in January, and, 7 when the Board of Governors approved a fur ther discount rate cut on February 1 to 6 per cent, the entire amount of the cut was allowed to show through to the federal funds rate.1 As 8 figure 1 shows, the federal funds rate, although relatively volatile, had dropped approximately 100 basis points during the intermeeting period. By this measure, dramatic ease had taken place during the intermeeting period. The intermeeting growth rate of total reserves, graphed in figure 3, also seemed to indicate substantial ease in the period since the December meeting. The annualized growth rate for this period was in excess of 30 percent. M2 growth had begun to pick up somewhat by February, with the 13-week growth rate moving slightly higher than 2.5 percent at the time o f the meet ing. According to the Record, "the continuing weakness in M2. . . appeared to reflect in part heightened concerns about the financial condi tion of many depository institutions in the wake o f the closing of privately insured banks and credit unions in Rhode Island and the failure of the Bank o f New England.”1 9 At the time of this meeting, it appeared that real output had declined in the fourth quarter of 1990 and was on a path of further decline in the current quarter. In particular, total nonfarm payroll employment fell in January, on the heels o f a December decline. Industrial production declined sharply in the fourth quarter of 1990, as had capacity utilization. Consumer spending, “partly reflecting [a] lackluster. . . holiday sea son,” was weak in the fourth quarter.2 These 0 factors w ere somewhat offset by a relatively strong nonagricultural export performance.2 1 The Board staff's forecast for real output pre pared for the February meeting suggested that "some further decline in economic activity” was likely in the near term.2 The staff forecast as 2 sumed that the war in the Persian Gulf, which was just getting under way, would be short lived, perhaps lasting a few months, and that further disruption of oil supplies would be avoided for the foreseeable future. The economy was expected to begin growing again "subse quently,” aided by export growth, falling oil prices and interest rates, and improved con sumer confidence.2 3 Committee members saw the economic situa tion at the time o f the February meeting as marked by “heightened. . . uncertainties,” due in part to the outbreak of war in the Persian Gulf.2 4 In general, members saw a “relatively mild recession followed by a moderate upturn in eco nomic activity. . . as a reasonable expectation.”2 5 "Risks," however, “w ere clearly on the down side,” and even a “relatively long recession could not be ruled out.”2 In assessing the out 6 look, members w ere particularly concerned about business and consumer confidence. In dices of sentiment were already at low levels and were poised, in the eyes o f the Committee, to go lower, owing not only to the unfolding conflict in the Middle East, but also to “financial excesses of the past decade.”2 Nonetheless, not 7 all of the news was gloomy, as the Committee noted that a lower spectrum o f interest rates, lower oil prices and a depreciating dollar would probably contribute to a rebound in aggregate activity.2 On the inflation front, “several mem 8 bers stressed that the slowing in monetary growth over a period of years was likely to be reflected increasingly in lower inflation.”2 9 According to the Record, "the considerable easing of monetary policy. . . [in] recent months” encouraged members to endorse unanimously 17March press release, p. 4. 24March press release, p. 7. 18March press release, p. 4. 25March press release, p. 7. 19March press release, p. 6. 26March press release, p. 8. 20March press release, p. 2. 27March press release, p. 9. 2 March press release, p. 3. 1 28March press release, p. 10. 22March press release, p. 6. 29March press release, p. 13. 23March press release, p. 7. MARCH/APRIL 1992 52 an unchanged policy for the intermeeting period ahead.3 In particular, "sufficient time had not 0 yet elapsed for the effects of the lower [interest] rates to be felt in the economy or indeed to any measurable extent in the growth of the monetary aggregates.”3 While many members mentioned 1 sluggish M2 growth as an area of concern, most seemed to agree with a staff analysis that sug gested faster rates of growth by the end of March, given a steady policy course. Some members, however, reiterated a call for a "rela tively high priority [on] achieving satisfactory rates of growth in reserves and money.”3 2 The degree of bias in the directive, if any, was a slightly more contentious issue. One view held out for a tilt toward ease in the weeks ahead, owing primarily to “the downside risks to the economy and the potential for inadequate mone tary growth.”3 Some members were especially 3 concerned that there would be "a high premium on avoiding any tendency for the weakness in the economy to cumulate because [of|. . . the se vere consequences of a potentially deep and prolonged recession.”3 An alternative view held 4 by some members was that, while easing might be necessary in the future, there were "con siderable risks of overreacting” and that “condi tions for a recovery in economic activity already appeared to be in place.”3 The former view, 5 however, carried the day, and the directive con tained bias toward ease, giving "special weight to potential developments that might require some easing during the intermeeting period.”3 6 Meeting o f March 26, 1991 During the intermeeting period, the bias toward ease in the directive was acted upon. According to the Record, "in early March, in response to information suggesting that econom ic activity had continued to decline through February, pressures on reserve positions were eased slightly.”3 The indications that economic 7 activity was weakening further included a sharp decline in total nonfarm payroll employment and a fall in industrial output.3 Accordingly, 8 the federal funds rate, depicted in figure 1, fell to a level just over 6 percent by the time of the March meeting. Money growth, as measured by 13-week M2 growth rates, continued to pick up during this period and seemed to indicate ease relative to previous rates, as outlined in figure 2. According to the Record, members cit ed "the strengthening in M2 growth in February and March [as] a welcome development follow ing an extended period of limited expansion.”3 9 Total reserves, shown in figure 3, were more puzzling during this period, as they actually declined, indicating a net drainage of reserves from the system instead of an injection. The reserves measure, therefore, seemed to indicate a relatively tight intermeeting policy. The Board staff expected a resumption of real output growth within a few months o f the March meeting.4 Positive factors cited included 0 the end of the war in the Persian Gulf (which presumably would brighten consumer and busi ness attitudes), lower nominal interest rates and oil prices, and expected improvement in ex ports.4 The staff felt that "reduced availability 1 of credit” and "a moderately restrictive fiscal policy” were factors restraining near-term growth.4 2 The Committee’s assessment of the outlook was essentially in agreement with that of the Board staff. Members were especially en couraged by the improvement in consumer con fidence at the end of the war and felt that "an upturn in economic activity was widely expect ed.”4 Many members emphasized, however, 3 that little hard evidence of growth in real out put had accumulated thus far and that, in fact, "there was some risk that the recession could deepen considerably further.”4 Many members 4 did not share the staff’s optimism about a sig nificant contribution to the recovery coming from export growth, as such effects were likely to be “curtailed by slower growth abroad.”4 Ac5 30March press release, p. 18. 39May press release, p. 12. 3 March press release, p. 18. 1 40May press release, p. 6. 32March press release, p. 19. 41May press release, p. 6. 33March press release, p. 20. 42May press release, pp. 6-7. 34March press release, p. 20. 43May press release, p. 7. 35March press release, p. 20. 44May press release, p. 7. 36March press release, p. 21. 45May press release, p. 10. 37May press release, p. 4. 38May press release, pp. 1-2. FEDERAL http://fraser.stlouisfed.org/RESERVE BANK OF ST. LOUIS Federal Reserve Bank of St. Louis 53 cording to the Record, real output growth had slowed in Germany and Japan, and “some weak ening in activity apparently had occurred in several other major industrial countries.”4 6 Most members supported the notion, with regard to short-run policy action, that sufficient easing had already taken place to foster a recov ery.4 In fact, some members commented that 7 "the most likely direction of the next policy move was not clear at this point and. . . caution was needed before any action was taken.”4 In par 8 ticular, further easing was a “possibility” due to “prevailing uncertainties,” but, “if the economy was indeed near its recession trough, additional easing would not be necessary.”4 Firming was 9 viewed as "premature,” although there might be a "potential need to tighten reserve conditions promptly if emerging economic and financial conditions. . . threatened progress toward price stability.”5 Given these considerations, all of the 0 members agreed to support a directive calling for an unchanged policy in the weeks ahead.5 1 While the members "expressed a range of views” relating to the possible degree of bias in the directive, the directive issued was symmet ric.5 As noted in the Record, the symmetry 2 represented a deviation from the policies adopted since July 1990, as virtually all of the directives since that time had been biased toward ease.5 3 The policy shift was consistent with the "assess ment that the risks to the economy. . . were now more evenly balanced.”5 In one view, the 4 recession had bottomed out, and therefore little could be achieved through further easing. In fact, "policy adjustments should be made only in the event of particularly conclusive evidence. . . that the recession might be deeper. . . than an ticipated.”5 An alternative view was that down 5 side risks still predominated and that “the Committee should react relatively promptly” should real output appear to decline further.5 6 One member suggested that recoveries tend to be stronger than expected, and therefore the “[greatest] risks were in the direction of too much ease and o f persisting or increasing inflation.”5 7 In this view, the bias in the directive should be toward a tighter policy, especially considering the lags in monetary policy effects. Considering all o f these views, however, the Committee elected to issue the symmetric directive. The Committee also discussed the interaction between discount rate changes and open mar ket operations as a technical matter of operat ing procedure.5 The policy in recent years has 8 been to keep the discount rate somewhat below the federal funds rate. Discount rate changes, which again are governed directly by the Board, “usually had been allowed to pass through auto matically to the federal funds rate” in the recent past, although there were some exceptions.5 9 Therefore, actions implemented by the Board alone might influence open market operations without explicit FOMC approval. Comments by members indicated the practice of show through should be continued, in general, “but that con sultation among members of the Committee would be particularly appropriate in [some] cir cumstances.”6 In particular, the members men 0 tioned cases in which a partial show through was more appropriate or particularly large policy actions were being considered.6 1 Meeting o f M ay 14, 1991 Immediately following the March FOMC meet ing, a steady open market policy was main tained.6 As figure 1 shows, however, the 2 federal funds rate began declining immediately after the March meeting; the Record notes, “the rate was under downward pressure at times from market expectations of some further eas ing.”6 At the end of April, two weeks before 3 the May meeting, an easing action was imple mented when the Board voted to reduce the discount rate to 5.5 percent and a portion of the drop was allowed to show through to the federal funds rate. Federal funds traded at about 46May press release, p. 3. 55May press release, p. 13. 47May press release, p. 11. 56May press release, p. 13. «M ay press release, p. 12. 57May press release, p. 14. 49May press release, p. 12. 58May press release, p. 14. 50May press release, p. 12. 59May press release, p. 14. 5 May press release, p. 11. 1 60May press release, p. 14. 52May press release, p. 13. 6 May press release, p. 14. 1 53May press release, p. 13. 62July press release, p. 4. 54May press release, p. 13. 63July press release, p. 4. MARCH/APRIL 1992 54 5.75 percent as the FOMC convened in May. Quarterly money growth rates had begun to slow at the time of this meeting, as shown in figure 2, but remained squarely within the target band. Reserve growth had resumed, eliminating some of the puzzle of the sharp decline recorded in the previous intermeeting period. The easing action was taken in response to "indications of [continuing] weakness in the economy.”6 The Record describes incoming 4 data as "mixed,” perhaps broadly suggestive that real output growth might be flat or slightly positive after declining in the fourth quarter of 1990 and the first quarter of 1991.6 While total 5 nonfarm payroll employment fell again in April, the rate of decline was less than in previous months. Industrial output was flat in April. The available data on foreign economies suggested that they grew at a relatively slow pace in the first quarter. Calling an upswing in economic activity “immi nent,” the Board staff at this meeting forecast a recovery fully under way in the summer months of 1991 and continuing through 1992.6 The 6 growth in real output was expected to be slower than that experienced during other postwar recoveries.6 Restraint in the recovery was sug 7 gested, according to the staff, by "the absence of further significant impetus from net exports” and "moderately restrictive” fiscal policy, at all levels of government.6 The staffs changing view 8 on the contribution of net exports, relative to its forecast from the previous meeting, was con sistent with the evidence that major foreign in dustrial economies continued to slow in the first months of 1991. The Committee "generally viewed a business recovery in the months ahead as a reasonable expectation.”6 While most members felt that 9 signals were "mixed,” many felt that "a variety o f developments appeared to have laid the groundwork for a recovery.”7 An important 0 factor would be the evolution of consumer and business sentiment.7 Many members seemed to 1 concur with the staff forecast that the strength of the recovery was questionable, as "current conditions did not point to major sources of stimulus.”7 Some members, however, did discuss 2 inventory investment and housing construction in such a role. As for inflation, "the members continued to express confidence that the ongo ing effects of earlier monetary policy actions and reduced monetary growth over an extended period. . . would tend with some lag to exert a favorable restraining effect on prices.”7 3 The Committee unanimously supported a directive with the thrust of policy unchanged from that of the previous meeting, and virtually all members supported instructions avoiding bias.7 At this point, in members’ eyes, “a steady 4 monetary policy appeared to. . . [reflect] an ap propriate balancing o f the risks of an overly stimulative policy that would threaten progress against inflation versus the risks of a deepening recession or an overly delayed recovery.”7 5 While some members felt that the costs o f a further fall in real output were much more se vere than an unexpectedly strong burst of growth, most agreed that the easing actions that had already been taken, given the presumed lags in effects on real output, w ere enough to insure against a further downturn.7 In particu 6 lar, “the System’s commitment to the goal of reducing inflation argued for a cautious ap proach to any further easing at a time when the economy might be close to its recession trough.”7 7 The growth rate of the Committee’s primary monetary aggregate, M2, was a point of discus sion at the May meeting. The Board staff pre pared a report suggesting that M2 growth would improve in the summer following a flat performance in April. Members showed some 64July press release, p. 4. 73July press release, p. 11. 65July press release, p. 1. 74July press release, p. 12. 66July press release, p. 6. 75July press release, p. 12. 67July press release, p. 6. 76July press release, p. 12. 68July press release, p. 6. 77July press release, p. 12. 69July press release, p. 7. 70July press release, p. 7. 7 July press release, p. 7. 1 72July press release, p. 8. FEDERAL RESERVE BANK OF ST. LOUIS 55 concern that, in particular, "subnormal mone tary growth might be an indication that mone tary policy was still too tight.”7 For this reason, 8 according to the Record, “a number o f members underscored the desirability of achieving mone tary growth within the Committee’s ranges for the year.”7 9 Meeting o f July 2-3, 1991 The midyear meeting of the FOMC included a review of long-term objectives as required by law. The target range for M2 was the focus of discussion. The growth of this monetary ag gregate was slowing at the time of this meeting, to the point where the 13-week growth rate had dropped to just over 2 percent, as illustrat ed in figure 2. For the year, however, M2 growth was in the middle of the target range, thanks to faster growth rates earlier.8 Neverthe 0 less, members felt that the “growth of this ag gregate thus far in 1991 had fallen short of what might have been expected on the basis of historical relationships with nominal income and interest rates.”8 Furthermore, “the reasons for 1 the shortfalls were not fully understood.”8 The 2 view of the Committee seemed to be that there was simply a good deal of uncertainty sur rounding the behavior of M2, but that “the four-percentage point range provided adequate leeway for any adjustments that might be need ed.”8 As in February, the Committee decided to 3 leave the target range unchanged. With regard to short-run policy, operations had focused on maintaining the existing policy stance since the last meeting. The federal funds rate seemed to bear this out, as the weekly average rate shown in figure 1 remained steady for the most part at about 5.75 percent, except for a 50 basis-point spike on the week of the July meeting. According to the Record, there was some upward pressure on interest rates in the intermeeting period due in part to "expecta tions that no further easing of monetary policy was likely in the near term.”8 While figure 2 4 shows that M2 growth continued to slow, meas ured on a 13-week basis, intermeeting reserve growth as captured in figure 3 appeared satis factory. The Board staff forecast “considerable growth” through the end of 1991.8 Again at this meet 5 ing, the staff felt that this phase of the recovery would be slow relative to past experience. The restraint was attributed, in part, to weakness in nonresidential construction, which would be “depressed by high vacancy rates,” and "fairly restrictive” fiscal policy, again at all levels of government.8 6 Members of the Committee "generally agreed that a recovery very likely was under way.” While “puzzling aspects” were noted, it was also pointed out that “ sources of strength in an eco nomic expansion often have been difficult to an ticipate near a cycle trough.”8 The Committee’s 7 policy of moderate money growth over the last several years was expected to pay o ff in the form of lower inflation in the upcoming quart ers.8 Many members agreed with the staff 8 regarding the restrictive effects of fiscal policies at all levels of government relative to past recovery phases. In particular, “despite burgeon ing [federal] borrowing requirements in the near term, cutbacks in defense spending and other efforts to curb expenditures” had the ear marks of a restrictive fiscal approach.8 9 The Committee unanimously supported a directive that called for an unchanged policy in the weeks until the next meeting. According to the Record, “an unchanged policy course [was viewed as offering] the greatest promise of reconciling the Committee’s goals of sustaining the nascent business recovery while also foster ing further progress against inflation.”9 I m m i 0 nent policy change was viewed as "unlikely," despite "obvious. . . uncertainty.”9 Recent slug 1 gishness in M2 growth, which can be seen in figure 2 as the declining 13-week growth rates in May and June, was a concern o f some mem bers, who commented that perhaps "monetary policy had not been eased sufficiently in recent 78July press release, p. 13. 85August press release, p. 5. 79July press release, p. 13. 86August press release, p. 6. 80August press release, p. 13. 87August press release, p. 6. 8 August press release, p. 13. 1 88August press release, p. 7. 82August press release, p. 13. 89August press release, p. 12. 83August press release, p. 14. 90August press release, p. 18 84August press release, p. 4. 9 August press release, p. 18. 1 MARCH/APRIL 1992 56 months.”9 It was pointed out, however, that 2 other measures, “especially. . . reserves,” seemed to show growth that was relatively strong. Most members felt that "the behavior of M2 . . . did not call for any policy adjustments at this point.”9 In any event, the staff projected faster 3 M2 growth in the near future, under a presump tion of an unchanged policy stance. Meeting o f August 20, 1991 growth rate of real output for the second half of the year, however, was now forecast to be somewhat lower than previously suggested. The staff outlook emphasized "a cyclical swing from substantial liquidation to modest accumulation in business inventories” as a stimulus for recov ery. As in previous forecasts, restrictive fiscal policy was thought to be retarding real output growth rates from their more usual cyclical pattern. As the Committee convened in August, the “recovery was proving to be sluggish.”9 While 4 operations during the intermeeting period ini tially had been directed toward maintaining ex isting policy, an easing move was implemented in early August.9 One reason for the unsche 5 duled action was weakness in M2 growth; as can be seen in figure 2, the 13-week growth rates were approaching zero at the time of the easing action and had turned negative by the time of the meeting. By any of the measures of the policy stance considered here, however, an easier policy was not obvious. Federal funds, which had been trading at about 5.75 percent, moved only slightly lower by the time of the meeting according to the weekly averages graphed in figure 1. M2 growth continued to falter as the Committee convened. Intermeeting total reserve growth was flat, perhaps suggest ing a relatively tight policy. FOMC members saw an economy that was “uneven," although they appeared to agree with the staff in principle that real output growth would be positive over the next several quart ers. In particular, a “ sustained expansion . . . was still viewed as a reasonable expectation, [but] many members now believed that the risks were tilted toward the downside.”1 0 The coup 0 attempt in the Soviet Union, the outcome of which was unknown at the time of the meeting, added in the view of the Committee additional uncertainty to the outlook. Closer to home, weakness in M2 growth was cited as “a matter of increasing concern to the extent that it im plied. . . a faltering economic expansion.”1 1 0 Again at this meeting, according to the Record, the FOMC seemed to concur with the staff as sessment that fiscal policy effects would proba bly be "somewhat negative” for the immediate future.1 2 0 The Record again describes the information on the economy at this juncture as "mixed,” but generally suggestive that sluggish growth in real output would continue in the near term.9 In 6 dustrial production increased in July, in part be cause of a rise in automobile production. July total nonfarm payroll employment increased slightly, as did retail sales, but business fixed in vestment declined in the second quarter and was expected to remain weak.9 Interest rates 7 generally fell in the intermeeting period; one reason cited, in the case of short-term Treasury securities, was the attempted coup in the Soviet Union.9 8 The shift toward pessimism in the Commit tee's outlook is reflected in private sector fore casts from August 1991 and December 1991 for quarterly growth in real output, as illustrated in figure 4. As of August, projections for the fourth quarter of 1991 and the first quarter of 1992 were relatively robust, although perhaps somewhat low relative to previous recoveries. By December, however, the projected growth rates for these quarters had dropped substan tially, far below that which might have been ex pected based on historical experience. The Board staff forecast a “moderate expan sion over the next several quarters.”9 The 9 92August press release, p. 19. The Committee voted to issue a directive maintaining current policy for the immediate fu ture but with an asymmetry toward ease. Ac cording to the Record, "an immediate easing 98October press release, p. 4. 93August press release, p. 19. "O cto b e r press release, p. 5. 94October press release, p. 4. ' “ October press release, p. 6. 95October press release, p. 4. 101October press release, p. 7. 96October press release, p. 1. ' 02October press release, p. 11 97October press release, p. 2. FEDERAL http://fraser.stlouisfed.org/ RESERVE BANK OF ST. LOUIS Federal Reserve Bank of St. Louis 57 move would be premature because the most re cent economic information, although mixed, still suggested a moderate rate of economic expan sion.”1 3 Advocates of asymmetry argued that 0 "risks. . . w ere largely on the side of a weakerthan-projected economy” and that the Commit tee should “react promptly” if any cumulative decline should become apparent.1 4 Some mem 0 bers disagreed, however, because they ’’were concerned about responding to what might prove to be short-lived fluctuations in the eco nomic data and anecdotal information.”1 5 Even 0 these members, however, could accept some asymmetry toward ease in the directive.1 6 0 The Record indicates that members paid "con siderable attention” in their discussion to sag ging M2 growth rates.1 7 While members found 0 explanations "difficult to disentangle,” some saw the slow growth to be of "little import” if it merely reflected shifts into alternative invest ment instruments that are not counted in the broad monetary aggregates.1 8 On the other 0 hand, the “behavior. . . might be indicative of. . . a monetary policy stance that was too tight.” The Board staff analysis continued to forecast some pick-up in the growth o f M2 in the near term.1 9 0 Some members suggested that, in current cir cumstances, the Board should emphasize move ments in M2 more explicitly when "guiding possible intermeeting adjustments in policy.” The majority, however, did not support this idea, for at least three reasons. One was that broad monetary aggregates like M2 were viewed by some as "unreliable indicators” of real output growth paths. Another was that nar rower measures, such as M l and total reserves, "might be more indicative o f the underlying thrust of monetary policy." Finally, some mem bers felt that including stronger reference to monetary aggregates in the directive might “mis construe the views o f many members,” who might not be willing to support a policy response to "aberrant fluctuations” in money growth.1 0 1 Meeting o f O ctober 1, 1991 Again, the bias in the August directive was acted upon during the intermeeting period. The Board voted to lower the discount rate to 5 per cent in mid-September and part of the 50 basispoint decline was allowed to show through to the federal funds rate. Accordingly, federal funds traded at about 5.25 percent by the time the Committee met in October, as shown in figure 1. Monetary growth, as measured by the 13-week M2 growth rate displayed in figure 2, was actually negative at the time of the meet ing, but apparently the fall in this growth rate had stalled somewhat relative to the decelera tion apparent in the graph since the spring of the year. Figure 3 indicates that intermeeting growth in total reserves had resumed, almost reaching the rates observed at the May and July meetings; by this measure, policy appeared to have been eased somewhat since August. According to the Record, the information on the economy reviewed at the October meeting suggested a continuing recovery, but one that was “uneven across sectors.”1 1 Total nonfarm 1 payroll employment had been essentially flat since March. Industrial production had increased in August. Consumer spending was rising, but retail sales fell in August. Overseas, the growth rates o f the Japanese and German economies fell in the second quarter, although real output growth appeared to have strengthened in other large economies.1 2 1 The Board staff projected continued recovery, tempered by downside risks and somewhat slow relative to previous cyclical upturns because of "persisting weaknesses in some sectors of the economy.”1 3 In this forecast, consumer spend 1 ing would be a significant positive factor, with "a swing from inventory liquidation" providing an "additional boost.”1 4 Other sources of stimu 1 lus included business equipment spending and housing construction.1 5 Dampening factors 1 were still seen on the fiscal policy side and also in commercial construction, where high vacancy rates were viewed as a deterrent to building.1 6 1 103October press release, p. 12. " “October press release, pp. 14-15. 104October press release, p. 12. "'N o ve m b e r press release, p. 1. 105October press release, p. 12. " 2November press release, pp. 1-3. 106October press release, p. 13. " 3November press release, p. 5. 107October press release, p. 13. " 4November press release, p. 6. 108October press release, p. 13. "^Novem ber press release, p. 6. 109October press release, p. 14. " 6November press release, p. 6. MARCH/APRIL 1992 58 Committee members seemed to agree with the staff prognosis, viewing the fledgling recovery as somewhat threatened.1 7 According to the 1 Record, "members commented that the anecdo tal reports on economic conditions and on busi ness and consumer sentiment continued to have a generally negative tone that did not appear to be fully consistent with the available economic statistics.”1 8 Reports on business attitudes in 1 particular seemed to suggest that key participants in the economy thought momentum in economic activity was stalled.1 9 Members were concerned 1 about risks to the recovery arising from "finan cial strains in the economy” as well as slow money growth. On the whole, however, the Committee appeared to feel that "the prospects remained favorable for a sustained expansion [in economic activity] at a moderate pace over the next several quarters.”1 0 2 In the discussion about operating instructions for the upcoming few weeks, all of the members of the FOMC supported language leaving the policy stance initially unchanged. According to the Record, “the present policy stance provided an appropriate balance between the risks of a faltering economic expansion and the risks of little or no progress toward price stability.”1 1 2 Some previous easing had not yet filtered through to effects on real output growth.1 2 2 Several members asserted, however, that the Committee should remain "particularly alert to indications of renewed weakening in business activity,” in part because they felt a second downturn might be less responsive to monetary stimulus.1 3 Other members emphasized the ad 2 verse consequences o f easing too much, focus ing on the prospect of higher long-term interest rates due to increased inflationary expectations which might then retard growth.1 4 On balance, 2 however, a steady course proved to be the con sensus. The slow growth of M2 continued to be a concern. While some members emphasized spe cial factors that might be depressing otherwise robust growth, such as the resolution of the cri sis in the thrift industry, others felt that the broad monetary aggregates "needed to be moni tored with special care.”1 5 As at previous meet 2 ings, the Board staff continued to predict some pick-up in the growth of M2, even in the ab sence of further easing action. As for contingencies in the directive, most of the members "indicated a preference for a directive that was biased at least marginally toward easing.”1 6 The downside risks cited 2 earlier provided the primary justification in the majority view. A minority preferred a symmet ric directive, citing likely cumulative effects from previous easing actions as a sufficient safeguard against further declines in real out put.1 7 Nevertheless, these members indicated a 2 willingness to accept an asymmetric directive.1 8 2 In the discussion, some members in the majori ty emphasized that “there should be no strong presumption that any easing would be under taken during the intermeeting period ahead.”1 9 2 Meeting o f N o v e m b e r 5, 1991 Because the recovery appeared to be weaken ing during the intermeeting period, an easing action, consistent with the bias toward ease con tained in the October directive, was implement ed at the end of October.1 0 According to the 3 Record, a key concern in taking this action was evidence on “flagging consumer and business confidence.”1 1 The federal funds rate fell after 3 the easing action to just above 5 percent by the time of the November meeting. The 13-week growth rates of M2 accelerated substantially, even before the easing action, and turned posi tive during the intermeeting period. Growth for the year remained near the lower end of the Committee’s range. Intermeeting total reserve growth was substantial, hitting the second highest level of the year, as outlined in figure 3. All measures of the policy stance therefore seemed to indicate at least some ease. The Board staff, concerned about "recent reports on business and consumer confidence 117November press release, p. 6. 125November press release, p. 12. 118November press release, p. 7. 126November press release, p. 12. 119November press release, p. 7. 127November press release, p. 12. i2°November press release, p. 6. 128November press release, p. 12. 1 1November press release, p. 11. 2 129November press release, p. 12. 122November press release, p. 11. 130December press release, p. 4. 123November press release, p. 11. 1 1December press release, p. 4. 3 124November press release, p. 11. FEDERAL http://fraser.stlouisfed.org/RESERVE BANK OF ST. LOUIS Federal Reserve Bank of St. Louis 59 combined with other information,” continued to forecast an expanding economy in the quarters ahead but made an “appreciable markdown” in the expected rates of growth in real output relative to past forecasts.1 2 The staff saw the 3 downside risks to the forecast as “predomi nant.”1 3 In particular, real output was expected 3 to grow quite slowly over the winter months, with the more robust growth normally associated with cyclical upswings a possibility in the spring of 1992 or later.1 4 Except for the decline in the 3 measures o f sentiment contributing to less con sumption than previously predicted, the staff foresaw the same sources of strength and the same retarding factors that w ere given appreci able weight in previous forecasts. FOMC members were concerned about the re cent developments in the measures of confidence, but "generally concluded that the available eco nomic data appeared consistent with continuing, albeit sluggish, expansion in overall economic activity.”1 5 Some commented that the measures 3 of business and consumer sentiment "had to be viewed with caution because they had tended in the past to be coincident rather than leading in dicators o f economic activity.”1 6 In terms of 3 downside risk, several members indicated con cern for "the vulnerability of the expansion stemming from the troubled condition of many financial institutions,” while others felt risks were symmetric or even on the upside.1 7 Some 3 members noted that any potential downturn was expected to be confined to the fourth quarter of 1991 or the first quarter of 1992 and that "a resumption of growth next year. . . [was] a reasonable expectation.”1 8 Given the 3 lags associated with short-run policy actions, these members believed that stimulus already in the economy should be given a chance to take effect, and any actions taken to stimulate real activity within the quarter might be viewed as somewhat late.1 9 3 At the end of the meeting, a majority of the voting members supported a proposal to ease immediately and to bias the directive toward further ease should conditions warrant. While recognition that "monetary policy had been eased considerably over the course of recent months” was forthcoming from most members, many felt that "further modest easing . . . [might] provide some added insurance” against a decline in real output.1 0 The majority felt that 4 additional easing would help bolster consumer confidence and lead to further declines in key long-term rates.1 1 The Record indicates that 4 there was "considerable” discussion of a proposal to make "a somewhat stronger move,” mainly because "small moves would lack the visibili ty. . . needed.”1 2 4 A minority of members argued against sub stantial easing.1 3 The notion that confidence 4 could be appreciably affected by monetary poli cy actions was questioned.1 4 Long-term interest 4 rates, it was argued, might well increase on a substantial easing move, as fears of rekindled inflation took hold among investors. Several members also reiterated that several easing steps recently taken should be allowed to work through the economy before further action was taken.1 5 4 Meeting o f D ecem b er 17, 1991 On November 6, the Board of Governors ap proved a 0.5 percentage-point reduction in the discount rate, and a "slight easing” was carried out in concert with this move.1 6 The bias in the 4 November directive was acted upon in the intermeeting period, as "an additional slight easing” was implemented in early December.1 7 The se 4 cond move was made, according to the Record, "as economic indicators continued to point to a faltering recovery and growth of the broad monetary aggregates remained sluggish.”1 8 The 4 132December press release, P- 6. 133December press release, P- 6. 143December press release, p. 12. 134December press release, P- 6. 145December press release, p. 12. The timing of the easing action was also discussed “ at some length,” as the Treasury auction beginning the day of the FOMC meeting would normally not be an appropriate time to intervene in the market for reserves. In particular, an immediate action could hurt the participants in the auction. Nevertheless, the Committee did not wish to delay action, and the direc tive contained no delay. 135December press release, P- 7. 136December press release, P- 8. 137December press release, P- 7. 138December press release, P- 7. 139December press release, P- 7. 144December press release, p. 12. ' ““ December press release, P- 11. ,46February press release, p. 4. 141December press release, P- 11. 147February press release, p. 4. 142December press release, P- 11. 148February press release, p. 4. MARCH/APRIL 1992 60 federal funds rate fell 0.5 percentage points be tween the November and the December meet ings, indicating substantial ease. The 13-week M2 growth rates continued to accelerate during the intermeeting period, as illustrated in figure 2, also indicating a relatively easy policy. Similar interpretations are possible for intermeeting to tal reserve growth, which, while down somewhat from the previous meeting, was still robust. Growth in real output appeared at this meet ing to be slow and perhaps waning.1 9 Depressed 4 levels of business and consumer confidence, a fall in November industrial production and weakness in consumption expenditures led the Board staff to suggest, according to the Record, that “a pause in the recovery. . . might extend into early 1992.”1 0 Faster growth was expected to return 5 at that time in part because o f “the cumulative effects of declines in interest rates in recent months.”1 1 The staff felt that key sources of 5 growth would be provided by “increases in residential construction, somewhat larger con sumption expenditures and some pick-up in busi ness equipment spending.’’1 2 Restrictive fiscal 5 policy was still viewed as a key element retarding growth relative to what expectations might war rant based on historical relationships in past recoveries.1 3 5 The members seemed to agree with the staff that past policy actions would eventually lead to increased growth and that “the economy might well remain quite sluggish over the months im mediately ahead.”1 4 Focus was placed on the 5 "evident pause in the business recovery and its interaction with very gloomy business and con sumer sentiment.”1 5 Factors that had been pre 5 viously identified as dampening growth "had in fact proved to be stronger and more persistent than anticipated.”1 6 The measures of sentiment 5 combined with some anecdotal reports on busi ness confidence received “considerable empha sis” in the Committee’s deliberations, although the reasons behind the dismal attitudes were "difficult to ascertain.’’1 7 Growth in the mone 5 tary aggregates was viewed as a positive sign by some members.1 8 5 In the discussion o f short-term policy for the period immediately ahead, the Committee sup ported a directive that left unchanged the policy stance for the time being, but which contained an “especially strong presumption” that an eas ing action would be necessary, "unless improve ment in the economy became evident fairly promptly or there was significant evidence of a pick-up in M2 growth.”1 9 Some members again 5 argued for "a more substantial policy move at some point.”1 0 They hoped that "a larger and 6 more visible policy action. . . would have greater effectiveness in part because it would be more likely to bolster confidence.”1 1 6 Other members argued for more deliberate policymaking.1 2 According to the Record, they 6 "expressed reservations about the urgency to ease in the near term” and suggested that “monetary policy could do little” to alter the fac tors that w ere restraining the economy at this point.1 3 In this minority view, the fact that the 6 extent of recent easing actions was substantial and the effects on real output were yet to be realized was given a good deal of weight. Any easing action needed to be coordinated with the Board’s approach to the discount rate. On December 20, the Board voted to move the discount rate down by a full percentage point. The FOMC then considered, in a telephone con ference, reactions to the move and decided to allow part of the cut to show through to the federal fu nds rate.1 4 A s sh ow n in figure 1, the 6 funds rate fell below 4 percent on a weekly average basis by the end of the year. SUMMARY In 1991, the FOMC operated in an environ ment in which growth in real output was resuming. This paper has therefore provided a case study of the making of monetary policy during the recovery phase of the business cycle. 149February press release, p. 1. 157February press release, p. 7. 150February press release, p. 6. 'ssfebruary press release, p. 8. 15'February press release, p. 6. 159February press release, p. 11. 152February press release, p. 6. 160February press release, p. 12. 153February press release, p. 6. 1 1February press release, p. 12. 6 154February press release, p. 7. 162February press release, p. 12. 155February press release, p. 7. 163February press release, p. 12. 156February press release, p. 7. 164February press release, pp. 16-17. FEDERAL http://fraser.stlouisfed.org/ RESERVE BANK OF ST. LOUIS Federal Reserve Bank of St. Louis 61 Relative to past cyclical upswings in economic activity, growth in 1991 was slow, and the recovery itself seemed at times elusive. The Committee states its objectives on a regu lar basis, and members support policy actions primarily based on their assessment of the out look for inflation and real output. Since growth in economic activity was sluggish in 1991 and since inflation was low by recent standards, the Committee’s objective of sustaining real output growth played a predominant role. Repeatedly, members wrestled with arguments about the lagged effects of monetary policy actions, noting that if the economy had bottomed out, easing to mitigate real output declines would be unneces sary. Still, at times, incoming data seemed to suggest a renewed decline in economic activity, and the Committee took actions throughout the year in the hope of avoiding this possibility. Measuring the thrust of monetary policy at a point in time was a continual topic of discussion at FOMC meetings in 1991. While the federal funds rate played a dominant role in this capac ity, the Committee devoted a considerable amount of time to analyses of M2 growth, which seemed to falter at times during the year. In general, conflicting signals of the thrust o f mon etary policy played a significant role in Commit tee deliberations. During the first half of 1991, the FOMC dis played considerable optimism that a recovery would begin and gain momentum as the year progressed. Three o f the first four directives of the year called for an unchanged policy without bias, although, as indicated in the chronology and table 1, some easing was implemented dur ing this period. Beginning about August, how ever, the Committee's confidence in the recovery began to wane. The four directives issued in the second half of the year all contained bias toward ease, as Committee members expressed deep concern about declines in industrial production and consumer confidence. By the end of the year, the FOMC had approved a number of easing actions designed to provide insurance against further declines in real output. REFERENCES Blue Chip Economic Indicators. 1991 Issues. Bullard, James B. “ The FOMC in 1990: Onset of Recession,” this Review (May/June 1991), pp. 31-52. Federal Reserve. Press Releases, Record of Policy Actions of the Federal Open Market Committee, dated March 29, 1991; May 17, 1991; July 5, 1991; August 23, 1991; October 4, 1991; November 8, 1991; December 20, 1991; and February 7, 1992. Garfinkel, Michelle R., and Daniel L. Thornton. “ The Multipli er Approach to the Money Supply Process: A Precaution ary Note,” this Review (July/August 1991), pp. 47-64. Hume, David. Essays Moral, Political, and Literary (London: Oxford University Press, 1742). MARCH/APRIL 1992 62 K. Alec Chrystal and Cletus C. Coughlin K. Alec Chrystal is the National Westminster Bank Professor of Personal Finance at City University, London. Cletus C. Coughlin is a research officer at the Federal Reserve Bank of St. Louis. Kevin White provided research assistance. H ow The 1993 Legislation W ill Affect European Financial Services T h e EUROPEAN ECONOMIC Community (EC) was created by the Treaty of Rome of 1957. Its intention was to create an integrated "Common Market” within which goods, services, labor and capital would move freely. In its early years, the implementation o f the Treaty of Rome focused on eliminating tariff barriers on trade in goods between the member countries. Barriers affecting capital movements and trade in serv ices were neglected, while those affecting labor mobility, such as lack of recognition of profes sional qualifications across member countries, were greatly reduced but not eliminated. that pertain directly to banking and other finan cial services.2 A major initiative to eliminate all remaining barriers to intra-EC trade began in 1985. This is referred to as the "single market program” or "1992,” its target date for completion (in reality, the end of 1992).1 The legislation underlying the single market program affects virtually every product area. This paper examines one key por tion of the legislation: the regulatory changes The commitment to eliminate the remaining EC trade barriers was formalized in the Single European Act (SEA), which was signed in 1985 and came into force on July 1, 1987. (See the shaded insert on pages 64-65 for additional high lights on EC history and a description of institu tions and legislative instruments.) The SEA defines both the goal—"an area without internal •For a recent overview of 1992, see Boucher (1991). 2Grilli (1989b) summarizes the numerous restrictions affecting international trade and investment transactions in the financial services sector, both in the EC and in other deve loped countries. FEDERAL http://fraser.stlouisfed.org/ RESERVE BANK OF ST. LOUIS Federal Reserve Bank of St. Louis In 1985, this sector accounted for 6.4 percent o f total output and 2.9 percent o f employment.3 Since the sector provides services for other sec tors, the integration of EC financial markets will affect efficiency not only within the financial services sector, but also in sectors using finan cial markets. 1992: GRADUAL RATHER THAN SUDDEN CHANGE 3See Emerson et al. (1988) for additional details on the economic dimensions of the financial services sector, 63 frontiers in which the free movement o f goods, persons, services and capital is ensured”—and the target date—the end of 1992. It also incorpo rates reforms to speed up decision-making within the EC by establishing "qualified majority voting” to decide most issues o f the reform process.4 In 1985, the EC Commission produced a White Paper entitled "Completing the Internal Market.” It listed numerous measures thought to be neces sary for the completion of the program, many of which have not yet been adopted.5 Because o f the large number of required measures, all barriers cannot be eliminated at once.6 A SINGLE MARKET IN FINANCIAL SERVICES: THE CORE REGULATORY CHANGES Before the 1980s, no systematic attempts had been made to reduce trade barriers in financial services. Although services had been addressed when the EC was formed in 1957, the implemen tation of intra-EC free trade in services had been neglected. Moreover, trade in financial services had not been covered by multilateral negotiations under the General Agreement on Tariffs and Trade (GATT). (This may change in the current Uruguay Round of negotiations.) The large number of proposals and the time necessary to consider a given proposal contrib ute to 1992 being a process rather than an event. Each directive must go through a com plex process of discussion, first within the Com mission and then in the Council of Ministers. Member state governments must be informed at each stage because they wish to consult with the domestic parties that will be affected. Parlia ments of member states, as well as the Europe an Parliament, also comment on each proposal. Finally, each agreement has to be ratified and reflected in the legislation of each member state. More important, many countries maintained exchange controls for capital account transac tions long beyond when they liberalized current account transactions.7 Without a free flow of financial capital to balance the flows o f goods between countries, "free” trade is constrained by capital controls. That is, financial services, which include a range o f banking, investment and insurance services, cannot be freely provid ed across borders if access to foreign exchange is restricted. A typical EC directive could take three years from first draft to Council ratification, with another two years or so for full implementation. Only measures close to adoption in early 1992 (or already adopted) will be implemented by the end of 1992; and measures not yet drafted will not be implemented before the mid-1990s. Thus, an important step before removing specific restrictions on cross-border trade in financial services is to remove all exchange con trols. Such a step was provided for by the Coun cil Directive of June 24, 1988—The Capital Liberalization Directive—which removes con trols on all capital flows within the EC and, for 4Key (1989) notes that under qualified majority voting, the number of votes of each member is weighted roughly ac cording to its population. To adopt legislation, 54 votes out of a total of 76 are required. 5According to Hill (1991), as of December 1991, 65 of the 282 measures outlined in the White Paper remained to be adopted. A goal of the EC Commission was to have all measures adopted by year-end 1991 to allow member na tions to convert the directives into national legislation. Problems with the directives are also occurring at the na tional level. For example, Italy has converted only half of the relevant directives into national law. 6Capie and Wood (1990) stress that gradual deregulation of the financial system is unlikely to cause instability. The history of deregulation, they note, reveals that only rapid changes in regulation threaten the stability of the financial system. A ccording to Bannock et al. (1972), exchange controls are government policies that attempt to control the purchases and sales of foreign currencies undertaken by the resi dents of a specific country. For example, the Exchange Control Act of 1947 restricted the purposes for which for eign currencies could be bought by British residents and limited the use and retention of foreign currencies and gold they acquired. MARCH/APRIL 1992 64 An O verview of the European Community The European Community (EC) is a group ing o f 12 member states.1 These are the origi nal six signatories of the Treaty of Rome in 1957—France, Italy, Belgium, Luxembourg, the Netherlands and West Germany—plus six countries that joined later—Denmark (1973), Ireland (1973), the United Kingdom (1973), Greece (1981), Portugal (1986) and Spain (1986). Further expansion of the EC to in clude Austria and Sweden, as well as other countries, is a strong possibility. Key dates and events in the history of the EC, including the recent Maastricht Summit Accords on monetary and political union, are provided in the accompanying table. E C Institutions There are four major EC institutions. The C om m ission is the civil service of the EC. It is divided into 23 functional areas (Directorates General). There are 17 commissioners who are responsible for managing these areas. The Commission proposes new laws and poli cies and is responsible for implementing deci sions made by the Council. The Council is the ultimate decision-making authority. It is a committee whose members represent their own national governments. The Council makes final policy decisions based on Commission proposals. Participants at Council meetings vary according to subject matter. For example, if the topic is finance, then the finance ministers of the 12 member nations attend. If the topic involves the exter nal relations of the EC, then foreign ministers attend. Council meetings involving heads of state occur twice a year. The chairmanship of the Council rotates among member states in alphabetical order for six-month periods. In some areas, such as for most labor and taxation questions, unanimity is required; for most single market issues, however, "qualified majority voting” is used. The E u ropean P arliam ent is a chamber of elected representatives from all member states. It offers opinions on most European 1For more details on the EC, see Rosenberg (1991). FEDERAL http://fraser.stlouisfed.org/ RESERVE BANK OF ST. LOUIS Federal Reserve Bank of St. Louis legislation but it has no formal legislative powers. The E u ropean Court o f Ju stice is a body of 13 judges, including at least one from each member country. The Court rules on applica tions and interpretation of EC laws. Judgments o f the Court are binding on each member state. Legislative Instruments There are four main legislative instruments. To become effective, legislation generally must be "adopted” by the Council. In some circumstances, however, the Commission may make laws itself. Typically, this will involve legislation that is required to implement previous Council decisions. One instrument is regulations, which are le gally binding on all member states whether or not ratified by national parliaments. If a regulation conflicts with national law, the regulation dominates. A second instrument is directives, which are legally binding only as to their ultimate effect; it is up to member states to decide how to implement the rules in their own national legislation. Virtually all of the 1992 program is being implemented by directives. D ecisions are the third instrument. Decisions, which are more narrowly focused than directives, are legally binding on all those to whom they are directed. All deci sions with financial implications are enforcea ble in courts of member states. Finally, recom m en d ation s (or opinions) have no legal support but merely state a view about some desirable condition or policy change. EC legislation normally is subjected to a lengthy process of consultation and discus sion before it is adopted by the Council. The "right of initiative” lies with the Commission. Once the Commission has drafted a proposal, there are consultations with all affected par ties both directly and via the relevant minis tries o f member states. The European Parliament also has the right to be consulted and is given the opportunity to propose amendments. 65 Major Post-War Steps Towards European Integration 1947 Customs Union formed between Belgium, Netherlands and Luxembourg - "Benelux” . 1948 Organization for European Economic Cooperation (OEEC) formed to administer U.S. aid for rebuilding post-war Europe. 1951 France, West Germany, Italy and Benelux form European Coal and Steel Community (ECSC) providing for a “ Common Market” in these products. 1957 Treaties of Rome establish the six-member (Belgium, France, Italy, Luxembourg, Netherlands and West Germany) European Economic Community (EC) and the Europe an Atomic Energy Community (Euratom). 1960 European Free Trade Association (EFTA) formed to promote free trade between nonEC Western European countries - Austria, Britain, Denmark, Finland, Iceland, Norway, Portugal, Sweden and Switzerland. 1962 Common Agriculture Policy (CAP) started. 1963 Britain’s application to join EC vetoed by President de Gaulle. 1965 France boycotts EC in protest at excessive speed of integration moves. 1968 Customs union completed. 1970 “ Werner Report” calls for Economic and Monetary Union within Europe - including a single currency. 1972 European exchange rate “ Snake” arrangement formed, but the United Kingdom leaves the Snake after six weeks. 1973 United Kingdom, Denmark and Ireland join the EC. 1979 European Monetary System (EMS) formed - establishing the Exchange Rate Mechan ism (ERM) and the European Currency Unit (ECU). Britain joins EMS but not ERM. 1979 First direct elections to European Parliament. 1981 Greece joins EC. 1983 Common Fisheries Policy established. 1985 White Paper on completing the internal market published. 1986 Spain and Portugal join EC. 1987 Single European Act comes into force. 1989 Delors Report calls for Economic and Monetary Union - including a single currency. Undertakings for Collective Investment in Transferable Securities took effect. 1990 United Kingdom joins ERM and Capital Liberalization Directive and Second Non-life In surance Directive took effect. 1991 Maastricht Summit Accords on monetary and political union. The third and final stage of Economic and Monetary Union will begin by January 1, 1999. A single European currency will begin by this date (possibly as early as January 1, 1997). An independent European Central Bank will be set up six months before the single currency. 1993 Second Coordinating Bank Directive, Own Funds Directive, Solvency Ratio Directive and Second Life Insurance Directive take effect. Council Directive on Investment Serv ices in the Securities Field and the Capital Adequacy Directive likely take effect. The enforcement o f EC laws is the respon sibility of the Commission. W here breaches of EC laws are suspected, the Commission may issue a formal letter of notice to the govern ments of member states. Where this proce dure proves insufficient, the Commission may refer the issue to the European Court of Justice. MARCH/APRIL 1992 66 the most part, on capital flows between an EC member and a non-member. For most member states, this directive was to apply from July 1, 1990.8 The deadline has been met, though several countries, like the United Kingdom, Germany, the Netherlands and Denmark, had eliminated explicit controls before 1988.9 Various approaches have been used to quanti fy the integration o f international financial mar kets. One way to see the effects of the relaxation of capital controls is to examine interest rates on comparable financial instruments in different countries that are denominated in the same cur rency. The elimination of capital controls should allow capital flows to equalize these interest rates.1 This is exactly what has happened in 0 the EC countries that have already eliminated capital controls. Figure 1 presents evidence for the United Kingdom, which abolished exchange controls as of October 24, 1979, and undertook a series of domestic liberalization measures in the 1980s. The U.K.’s deregulation has caused the Eurosterling-London Interbank Offer Rate (LIBOR) spread to collapse near zero.1 Similar 1 evidence exists for other EC countries that have liberalized.1 2 This evidence suggests that most of the effects of liberalizing capital flows for some, but not all, countries have already been realized, rein forcing the point that 1992 is a series o f changes. There are, however, additional gains possible from the 1992 process. One is that 1992 will make it less costly for financial firms from one member country to be authorized to provide services in other EC countries. New financial services, as well as lower prices for existing services, might also occur. Before discussing these potential gains, we will summarize the major directives that pertain directly to financial services. 8lreland, Spain, Greece and Portugal have until the end of 1992 to comply, with the latter two having the option to de lay compliance until 1995. 9According to Blundell-Wignall and Browne (1991), the in tegration of financial markets internationally began in the mid-1970s with the removal of capital controls in Germany, the United States and Canada. Japan and the United Kingdom relaxed capital controls in the late 1970s, while France, Italy and some other EC countries realized the complete elimination of controls by the middle of 1990. 10This result is analogous to the effect of eliminating trade barriers on goods. When a country eliminates a tariff on a specific good, the difference between the price of the good in the country’s domestic market and that in the interna tional market should narrow. FEDERAL RESERVE BANK OF ST. LOUIS MAJOR DIRECTIVES The major directives of the 1992 program for financial services can be divided into four cate gories: banking, investment services, undertak ings for collective investments and insurance.1 3 B a n k i n g . Efforts at EC coordination did not be gin with the Single European Act for any of the four categories of financial services. Rather, the SEA has accelerated the process of harmonizing regulations. For example, the First Banking Coordination Directive, which was approved by the Council in December 1977, required mem ber states to establish systems for authorizing and supervising credit institutions.1 4 A second example is the Consolidation Super vision Directive of June 1983, which required that credit institutions be supervised on a con solidated basis. Any credit institution owning 25 percent or more of the capital o f another finan cial institution was to be supervised on a con solidated basis by the authorities in the owning institution’s home state. Another provision man dated the exchange o f information between su pervising authorities to obtain an overview of a consolidated company’s affairs. To assist this su pervisory cooperation, the Bank Accounts Direc tive of December 1986 harmonized accounting rules for credit institutions. In the 1992 legislation, the Second Coordinat ing Banking Directive (2BD) is the primary bank ing directive. The 2BD allows any credit institu tion authorized in one member country to es tablish branches and provide banking services anywhere in the EC. While this so-called "com mon passport" allows home-country authoriza tion, the credit institution must conform to all local laws. Thus, the host country's business rules, such as reporting requirements and res- "T h e two interest rates are ones charged by banks to other banks for three-month loans denominated in British pounds. The Eurosterling rate pertains to loans made out side the United Kingdom and the LIBOR applies to loans made inside the United Kingdom. ,2See Blundell-Wignall and Browne (1991) for charts similar to figure 1 for Germany, the Netherlands and France. 13See U.K. Department of Trade and Industry (1991) for a summary of EC Directives relating to 1992. 14We refer to credit institutions rather than “ banks” because these regulations include institutions other than banks. These would include the European equivalent of thrifts. 67 ■■ F igure 1 Difference Between the Three-Month Eurosterling and Libor Rates P e rc e n t M o n t h ly D a ta trictions on permissible products and activities, must be followed. The 2BD also gives the commission some in fluence in authorizing institutions from outside the EC—the so-called "Reciprocity Clause.” The first, but not the final, draft of this clause created much controversy and is partly responsible for the label "fortress Europe” that has inappropri ately been associated with the 1992 program. (See the shaded insert on page 68 for additional discussion of this topic.) The 2BD is supported by the Own Funds Directive and the Solvency Ratio Directive. The former provides common definitions for the components of the capital base; the latter uses these definitions to establish minimum asset ra tios to be met by all credit institutions. All three directives become effective on January 1, 1993. I n v e s t m e n t S e r v i c e s . A related, but more problematic, set of measures deals with invest ment services. This category covers all aspects of the markets in tradeable securities, including investment banking, stock brokerage and the or ganization of the exchanges themselves. The key elements of the 1992 program are formulated in the Council Directive on Investment Services in the Securities Field and the Capital Adequacy Directive, neither of which has been adopted formally. Until recently, observers generally thought both directives would begin operation at the same time as the banking directives because the 2BD gives banks (and other credit institutions) the right to do securities business throughout the EC on a single passport basis. As time pass es, this simultaneity becomes less likely. If an identical single passport is not extended to non- MARCH/APRIL 1992 68 The Second Banking Directive and Fortress Europe One o f the great concerns, often heard out side the EC, is that the 1992 program will lower barriers to internal trade but at a cost o f higher external trade barriers. The 1992 program does not introduce new barriers to trade in goods between Europe and the rest of the world. Nonetheless, a mistaken belief persists that access to the EC market will be harder after 1992. This belief stems partly from the "Reciprocity Clause” in early drafts of the Se cond Banking Directive. This required the Commission to evaluate all applications for new subsidiaries where the parent company was based outside the EC. The Commission would have had the power to delay approval if the other country did not offer "m irror im age” reciprocity. Mirror image reciprocity would have required that EC firms be al lowed to operate in foreign countries, just as they could at home, before access would be offered to nationals o f that country. This would have been very restrictive. For exam ple, because there is no legal separation be tween investment banking and commercial banking in the EC, it would have required abolition of the Glass-Steagall Act in the United States before U.S. banks could gain access to the EC. This requirement was weakened in later drafts of the directive. The final directive simply calls for negotiations with third coun 'See Title III, Article 9, paragraph 4 of the 2BD. In offi cial documents, the 2BD is the “ Second Council Direc tive of 15 December 1989.” 2 See Title III, Article 9, paragraph 6 of the 2BD. FEDERAL http://fraser.stlouisfed.org/ RESERVE BANK OF ST. LOUIS Federal Reserve Bank of St. Louis tries (that is, countries outside the EC) in the event that EC firms are denied "effective market access.” The critical criterion now is that EC firms should not be discriminated against in third markets—they should be ac corded “national” treatment. “W henever it ap pears to the Commission . . . that EC credit institutions in a third country do not receive national treatment offering the same competi tive opportunities as are available to domestic credit institutions and the conditions of effec tive market access are not fulfilled, the Com mission may initiate negotiations in order to remedy the situation.”1 If negotiations about unfair treatment in a non-EC country have been initiated, approval o f EC market access by credit institutions from that country may be delayed by up to three months. After this time, the Council must de cide whether such delays should continue. This procedure will not apply to any firm al ready authorized to trade in an EC country. Finally, this intervention in the approval process must not contravene “the Community’s obligations under any international agree ments, bilateral or multilateral, governing the taking-up and pursuit o f the business of credit institutions.”2 The general structure of the reciprocity clause in the Second Banking Directive is expected to be copied for the other major areas of financial services, in cluding investment services and insurance. 69 bank securities firms at the same time, they will be at a disadvantage. A key problem in formulating regulations in investment services has been that the range of activities covered is much more heterogeneous than in the banking area.1 Arguments have 5 arisen about which activities to include and how much capital should be required for different lines o f business. Initial proposals, for example, incorporated such high capital requirements that some businesses objected strongly. Non bank securities houses argued that the require ments were so onerous, their business would be driven outside their countries. Universal banks, on the other hand, feared they would be at a disadvantage if securities houses had lower re quirements than banks.1 The latest drafts of 6 the directives incorporate a compromise that ap pears acceptable to both camps. Banks will be permitted to treat their securities business separately and calculate capital requirements under the investment services rules rather than the banking rules. Another point of controversy concerns the provision o f compensation schemes for inves tors. A commission recommendation in 1986 suggested the establishment of compensation schemes for depositors (that is, deposit insur ance) in credit institutions. In the wider area of investment services, the position of compensa tion schemes is even less clear. Some countries, like the United Kingdom since the implementa tion of the 1986 Financial Services Act, have compulsory compensation schemes for invest ment business, while many others do not. This position raises potential anomalies in crossborder business. A final sticking point in the Investment Serv ices Directive relates to the monopoly of or ganized stock exchanges over securities trading. Some countries, like France, have argued for the official stock exchange to have a monopoly. 15Another reason for the relatively faster agreement on banking is that bank regulation had already been well worked out globally—through the Bank for International Settlements and formalized in the Basle Agreement. The 1988 Basle Agreement replaced differing national regula tions for measuring capital adequacy by a single, interna tionally accepted standard. The goals were to strengthen the soundness of the international banking system and re move regulatory differences that affected the international competitiveness of banks. See Blanden (1988). 16Generally speaking, EC countries did not have counter parts to U.S. banking regulations that limited their spread geographically or their lines of business activity. As a result, a small number of large banks evolved. For exam Without a monopoly, the present French system could not be used throughout the EC. Others, especially the British, are strongly opposed. U n d e r t a k in g s f o r C o l l e c t i v e I n v e s t m e n t s . In contrast to the banking and investment services directives, the directive governing Undertakings for Collective Investment in Transferable Securi ties (UCITS), which are open-ended mutual funds, has already come into effect. The Council Direc tive on the coordination o f laws relating to UCITS took effect in October 1989. The directive estab lishes minimum requirements for authorization of UCITS and permits their marketing through out the EC. This freedom is subject to the usual proviso that the host state be notified and local marketing rules be obeyed. Minimum require ments are established for adequate risk spread ing, the separation of trustees from managers and the specification of acceptable investments. Before it was implemented, there was some concern that the UCITS Directive would lead to a migration of UCITS managers to countries, like Luxembourg and Ireland, with the most favorable tax treatment. It is too early to deter mine whether this expectation is correct. To counteract this possibility, however, efforts were made to reduce tax differences. For exam ple, the British budget of 1989 reduced taxes on unit trusts. I n s u r a n c e . A final set of directives on financial services deals with insurance. Insurance pro vides examples of 1992 initiatives already in ef fect as well as those many years away. The primary directives are the Second Non-Life In surance Directive and the Second Life Insurance Directive. The Second Non-Life Insurance Directive es tablishes freedom of services for cross-border business within the EC. This freedom, however, applies only for large commercial risks. What is ple, German banking is dominated by a small number of banks engaging in normal commercial banking as well as buying and selling stocks for others, underwriting new stock issues and owning stock on their own behalf. In fact, German banks are represented on the boards of directors of many companies. In the United Kingdom, merchant banks specialized in the securities business, while com mercial banks had the bulk of deposits. Since the deregu lation of British financial markets that began on October 27, 1986, known as the Big Bang, U.K. commercial banks have gone universal in that they have merchant bank sub sidiaries and are expanding into insurance services, espe cially life insurance. Belgium is the only EC country that separates investment and commercial banking. MARCH/APRIL 1992 70 referred to as "mass risk,” which includes most things insured by people other than their lives— theft and fire damage to personal property— remains subject to numerous restrictions. A new, more liberal regime applies to all marine, aviation and shipment risks, and other fire, property and financial risks for situations in which the policy holder is a large commercial company. Here, the insurer has an obligation to notify the authorities (in the insured company’s country), but may write the business directly. For all other businesses, the authorities in each country may continue to control the terms of authorization, premiums, policy conditions and reserve assets. This Directive took effect in July 1990 and, hence, the large commercial risk market has ef fectively achieved the single market position al ready. Unlike banking, this directive did not create a common passport. Thus, branching in other countries is not freely permitted, and es tablishment still requires authorization in each member state. Two draft "Framework Direc tives” for life and non-life insurance appeared in 1991 and 1990, respectively. These would estab lish the single passport for insurance; the fact that the first drafts o f these directives did not emerge earlier, however, suggests that they will not be in operation until 1995 at the earliest. Only modest progress has been made on life insurance so far. The Second Life Insurance Directive was adopted in November 1990 for implementation on May 21, 1993. It only goes a small way, however, toward creating a single market in life insurance. A liberal regime is provided for, but only in cases where the con sumer takes the initiative in buying a life insur ance policy from a firm in another member country. In all other cases, the restrictive re gime applies, under which the insurer may be required to obtain special approval (depending upon local law) and the policy terms may be proscribed. Under the most recent draft of legislation in volving life insurance, whose date of implemen tation has yet to be agreed upon, insurance com panies are permitted to advertise, but they may not approach consumers directly. It also is pos sible that “local” asset backing for the policy 17For example, the U.K. Financial Services Act of 1986 re quires any firm selling investment products in the United Kingdom to register with either the Securities and Invest ment Board or a recognized regulatory organization. The FEDERAL RESERVE BANK OF ST. LOUIS may be required. This means that, for example, an Italian firm selling insurance in Germany would have to back its German policies with German securities. This draft o f the legislation also restricts the role of brokers. For three years after implementation, member states will be able to forbid consumers from seeking poli cies from other member states through brokers. Considerable resistance exists in some quart ers to the creation of a genuine single market in life insurance. The basic conflict arises be cause some countries—notably Germany—have had a very conservative attitude to life insur ance, while others—like the United Kingdom— have been very innovative. German insurance companies have typically invested in safe fixedinterest securities, and innovation in the indus try has been strictly controlled. The United Kingdom, in contrast, allows its firms to invest across a range o f assets including property and equities. Thus, the typical British firm’s portfolio is riskier than its German counterpart, but has a much higher average yield, producing signifi cantly lower prices for British products. The Com m on Passport Before discussing the reform process, an im portant distinction must be made between wholesale and retail financial markets. As demonstrated above, the globalization of inter national financial markets in the 1970s and 1980s has already led to highly competitive wholesale capital markets across many EC countries. These markets, in which financial firms deal directly with each other, experienced considera ble competitive pressures in the past 20 years. Faced with the choice of deregulation or the loss of firms to less-regulated environments in other countries, most nations dismantled much of the regulatory structure in wholesale finan cial markets. Retail markets, in which consumers deal with firms to borrow money, purchase insurance and trade stock, are quite different and present the biggest problem for deregulation. These markets retain a myriad of complex regulatory structures and external barriers that are gener ally justified on the grounds that they protect the small consumer.1 Regardless of whether 7 firm must conform to a complex set of rules, subject itself to inspections and pay membership charges, which in clude investor compensation schemes. 71 domestic officials actually believe this or are simply disguising their protection of domestic firms, the abolition o f regulations to increase cross-border trade and competition in retail financial markets is the primary challenge of the 1992 program. Starting with the existing regulatory struc tures in each member country, the central prin ciple guiding deregulation is that regulators in each member state are competent to judge which firms are "fit and proper” to do business in the industry. Once a firm has been authorized by the regulatory authority in its home cou n tryso-called home authorization—it is automatically authorized to do business in any other member country and is said to have a "common passport.” Previously, many countries have allowed firms from other EC countries freedom of establish ment, but this freedom has been subject to a separate process of approval in each country.1 8 The abolition of this requirement, therefore, will make it easier for firms to establish subsidi aries in other member countries. Home authorization, however, is not the end of the story. Firms operating outside their home states still have to obey “host country conduct of business rules.”1 In other words, foreign 9 firms must obey all the local regulations about the nature o f acceptable products and the way in which they may be advertised and sold. For example, France does not allow interest pay ments on checking deposits, while most other EC countries do. The fact that business rules will continue to differ across countries limits the extent to which there will be a genuine single market. The various rules increase the costs of crossborder activity and are sometimes even anti competitive. For example, the business rules in some member states define which products can be sold and their respective prices. Thus, one of 1B For example, Emerson et al. (1988) note that each EC country allows freedom of establishment for foreign banks; however, the conditions under which this may be done vary substantially across countries. High establishment costs make it difficult for a foreign bank to enter and com pete successfully with an existing domestic retail bank. Ad ditional obstacles in certain countries, like Italy and Spain, are restrictions on foreign acquisitions and involvement with domestic banks. 19For an alternative interpretation of the implications of home authorization in the context of the 2BD, see Key (1989). In our view, home authorization aplies to the issue of a license and prudential control, but it does not apply to the main incentives for attempting to enter new markets—the introduction of new products not offered by local firms—is not guaranteed. Regulatory Complications f r o m the Com m on Passport The move to a common passport will compli cate the regulatory process.2 At this point, only 0 hypothetical situations can be offered to suggest the potential difficulties. While firms require authorization only in their home states, the regu latory authorities of other nations have to moni tor the activity of these firms within their do main because they are responsible for consumer protection and adherence to business rules. To illustrate, suppose a German bank estab lishes a subsidiary in the United Kingdom after 1992 on the basis of its German banking license. It takes deposits and makes loans in British pounds sterling. As the German banking authori ties are responsible for prudential supervision, the bank must file the reports required by these authorities. The bank, however, must also register with the Bank of England, fulfill all reporting requirements and conform to all Brit ish banking regulations in the United K ingdom including reserve requirements and banking codes o f practice. It must also pay regulatory fees just as any British bank must do. The lower costs of establishing an office in the United Kingdom may increase the regulato ry burden of both the British and German authorities. Suppose, for example, the German bank gets into difficulties, like a run on deposits, or is involved in a breach of rules, like fraud. Clearly, both British and German authorities will have to get involved to resolve the problem. In deed, a bank with branches (or subsidiaries) across Europe could draw 12 sets of regulators into a dispute over its operations. The number o f regulators would rise even further if the any behavior that falls under conduct of business rules. Home authorization is much different than home control. Even though a bank is given a license to operate abroad by its home authorities, the bank’s subsidiaries will have to obey all the laws attached to banking practice in the foreign countries in which they operate. 20Capie and Wood (1990) make a similar point that the Se cond Banking Directive will make supervision and regula tion much more complicated. They speculate, however, that this complexity may cause a change in regulation from detailed supervision to one in which central banks are primarily lenders of last resort. MARCH/APRIL 1992 72 Table 1 Deposit Insurance in the EC1 Coverage2 Limitations (in U.S. dollars as of July 6, 1990) Country Belgium Denmark France Germany Ireland Italy Luxembourg Netherlands Spain United Kingdom Deposits in foreign currency Deposits in domestic branches of foreign banks Deposits in foreign branches $14,706 39,708 72,033 30% of bank’s liable capital 16,206 659,385 14,706 21,486 14,789 35,730 No — No Yes No Yes Yes No — No Yes — Yes — Yes — Yes — Yes — No No No — Yes — Yes No No — SOURCE: Bartholomew and Vanderhoff (1991). 1Greece and Portugal have no formal systems of deposit insurance. 2The ” indicates no information was available. bank’s activities spread beyond banking into securities or insurance. It is also noteworthy that the British authori ties have no power to withdraw the banking license if the bank transgresses business rules in the United Kingdom. Even though the Bank of England could stop a bank from trading tem porarily, a high degree of communication and cooperation between regulators of the member countries will be required to manage such a problem. Eventually, there might be a formal regulatory agency that operates on a community-wide basis. The preceding example, which pertains to all member countries, is relatively simple in com parison to the regulatory issues that might arise when services are provided across national borders. Suppose the German bank takes de posits and makes loans in sterling with retail customers in the United Kingdom only by mail or telephone from its head office in Frankfurt. In this case, the German bank need not register with the Bank o f England, but has an obligation to conform to British conduct of business rules. This means that the Bank of England must mo nitor this business in some way. While cases like this may be of trivial quantitative significance (especially in retail trade), they also may gener ate the greatest regulatory headaches, in terms FEDERAL RESERVE BANK OF ST. LOUIS of allocating regulatory responsibilities for the monitoring and enforcement o f standards of business practice. Such jurisdictional problems may be greatest where deposit insurance is involved. Table 1 summarizes the deposit protection schemes for commercial banks in the EC. The amount of protection for depositors varies substantially across countries. This may influence where a specific deposit may be made. The high level of protection in Italy could attract large depositors. By the same token, the different levels of pro tection may confuse depositors. A Spanish depo sitor, who made a deposit in a French branch in Spain that fails, for example, may mistakenly be lieve that the French deposit insurance scheme applies. Since deposit insurance is politically sen sitive, controversy is not difficult to envision. The EC Commission has drafted a proposal, not yet published, for the harmonization of deposit insurance, but any changes are unlikely to take effect before the mid-1990s. The almost complete harmonization of regula tory standards is inevitable when transactions within an industry are predominantly of an in ternational nature. By itself, however, 1992 is unlikely to make the transactions in European retail financial markets to be primarily interna tional. Thus, the regulation of retail financial 73 markets in Europe involves a compromise be tween host country control and the creation of a single market. Harmonization of business rules will not be complete and, in some cases, may not be even close. Product Innovation The potential gains from removing barriers to the spread o f new products across borders seem to be positive and potentially quite large. Lower-cost producers of financial services prod ucts would prosper at the expense of less effi cient firms that now survive only because of regulations that limit competition by foreign firms. Consumers would benefit from having a greater variety of products from which to choose and would pay lower prices for them. The basic problem is the resistance by some countries to relaxing domestic regulation o f an industry. Frequently, a country’s business rules inhibit product innovation. For example, current German regulations restrict the introduction of new insurance products into Germany. Even with a common passport, a foreign insurance firm faces a major deterrent to entering the German market. Taken together, German citizens and foreign insurance firms clearly would benefit from free trade in new products, but it is also clear that some German insurance companies would suffer from the influx of competition. This is the area where the least progress has been made in the 1992 program. In view of the time required to reach and implement EC deci sions, as well as the current controversy about these decisions, the potentially large gains from product innovation and lower prices in many financial services will not be realized any time in the near future. POTENTIAL BENEFITS OF THE SINGLE MARKET The preceding discussion raises doubts about how sizable the gains will be from the 1992 legislation in the financial services sector; however, we do not provide an estimate of the 21To reiterate, we are not questioning the gains from the abolition of exchange controls; rather, we are questioning the gains from the common passport in light of the con tinuation of different conduct of business rules. 22ln theory, the abolition of trade barriers for goods traded among a group of countries may or may not yield net benefits. An elementary demonstration of this result can be found in Coughlin (1990). gains themselves.2 These doubts are at odds 1 with the potential gains estimated in the Cecchini Report, the best-known attempt to measure such gains.2 This report found substantial 2 potential gains from the creation of a single market in many industries.2 The gains from the 3 liberalization of the financial services sector, which are presented and examined below, were found to be substantial as well. Financial Services: The Estimated Gains o f Eliminating Trade Barriers The reduction of trade barriers can generate gains via a number o f routes, all of which are driven by increased competitive pressures. For example, the reduction of trade barriers will al low firms with lower production costs to ex pand their production, increasing total output and economic welfare. Other gains can be real ized as larger markets increase the opportuni ties to use certain production technologies that lower per-unit production costs. Finally, in creased competition tends to drive down profit margins, eliminate waste and stimulate the de velopment of new products and less costly methods to produce existing products. Ultimate ly, the competitive pressures will allow con sumers throughout the EC to consume (use) more financial services at lower prices per unit. The competitive pressures resulting from 1992 are expected to narrow the price differences of a financial service across the EC. As part of the Cecchini Report, Price Waterhouse calculated prices across eight EC countries for the 16 financial services—seven banking services, five insurance services and four securities services— listed in table 2. The average o f the four lowest prices for each service was chosen as the likely price after the elimination of trade barriers. The potential price declines for financial serv ices are listed in table 3. Exactly how much of this potential decline will be realized is difficult to estimate, so an expected decline (with a plus/minus 5 percentage-point range) was de fined as one-half of the potential decline. 23The Cecchini Report estimates that the gains from com pleting the internal market range from 4.3 percent to 6.4 percent of gross domestic product in the EC. See Cough lin (1991) for an examination of the approach used in the Cecchini Report as well as other approaches used to esti mate the economic effects of 1992. MARCH/APRIL 1992 74 Table 2 List of Standard Financial Services or Products Surveyed Name of standard service Description of standard service Banking services 1. Consumer credit 2. Credit cards 3. Mortgages 4. Letters of credit 5. Foreign exchange drafts 6. Travellers checks 7. Commercial loans Annual cost of consumer loan of 500 ECU. Excess in terest rate over money market rates. Annual cost assuming 500 ECU debit. Excess interest rate over money market rates. Annual cost of home loan of 25,000 ECU. Excess in terest rate over money market rates. Cost of letter of credit of 50,000 ECU for three months. Cost to a large commercial client to purchase a com mercial draft for 30,000 ECU. Cost for a private consumer to purchase 500 ECU worth of travellers checks. Annual cost (including commissions and charges) to a medium-sized firm of a commercial loan of 250,000 ECU. Insurance services 1. Life insurance Average annual cost of term (life) insurance. 2. Home insurance Annual cost of fire and theft coverage for house valued at 70,000 ECU with 28,000 ECU contents. Annual cost of comprehensive insurance, 1.6 liter car, driver 10 years experience, no-claims bonus. Annual coverage for premises valued at 387,240 ECU and stock at 232,344 ECU. Annual premium for engineering company with 20 em ployees and annual turnover of 1.29 million ECU. 3. Motor insurance 4. Commercial fire and theft 5. Public liability coverage Brokerage services 1. Private equity transactions Commission costs of cash bargain of 1,440 ECU. 2. Private gilt transactions Commission costs of cash bargain of 14,000 ECU. 3. Institutional equity transactions Commission costs of cash bargain of 288,000 ECU. 4. Institutional gilt transactions Commission costs of cash bargain of 7.2 million ECU. SOURCE: Emerson et al. (1988), p. 102. Using the expected price declines for financial services, the gains for the eight EC countries ex amined are estimated to be 21.6 billion ECU, which is 0.7 percent of their gross domestic product.2 The distribution o f these gains across 4 the EC are listed in table 4. One's confidence in these estimates, as acknowledged in Emerson et al. (1988), should not be great. First, the price comparisons themselves can be questioned. Products such as “credit” and “life insurance” More important, even if price differences exist for identical products, it is far from clear that the 1992 legislation will eliminate such differ- 24The ECU, which stands for the European Currency Unit, is composed of the weighted averages of the currencies of the 12 member countries and is the unit of account for the EC. Even though much negotiation remains, the ECU is likely to become the single currency of the EC. For a brief history of the ECU, especially recent developments, see Tyley (1991). One ECU was equal to $1.29 on February 11, 1992. FED ERAL http://fraser.stlouisfed.org/ RESERVE BANK O F ST. LOUIS Federal Reserve Bank of St. Louis have been priced as if the characteristics are the same in each country. For example, no at tempt has been made to adjust for theft and mortality differences across countries, and, hence, it is not clear that homogeneous products are compared. 75 Table 3 Potential and Expected Price Declines for Financial Services Country Belgium France Germany Italy Luxembourg Netherlands Spain United Kingdom Potential price fall 23% 24 25 29 17 9 34 13 Table 4 Estimated Gains Resulting from the Expected Price Reductions for Financial Services Range of expected fall Country 6-16% 7-17 5-15 9-19 3-13 0-9 16-26 2-12 SOURCE: Emerson et al. (1988), p. 104. ences. The reason is that business rules will continue to differ from country to country, thereby impeding trade in financial services and limiting potential gains to levels below those es timated in the table.2 Thus, the value of the 5 single passport is diminished considerably by the inability of firms entering new markets to offer a full line of products and services. Grilli (1989a) has also raised doubts about the estimates in the Cecchini Report on the likely effects of liberalization on wholesale and retail banking throughout the EC. Grilli doubts whether a perfectly competitive market structure is an accurate approximation of retail banking post-1992. Much evidence suggests that banks have market power in their retail markets that will not be eliminated by the 1992 legislation. For example, within the same country, which is already a homogeneous regulatory and institu tional environment, the terms o f a deposit con tract, such as the interest rate paid on a time deposit, frequently vary across banks. In addi tion, the transaction costs of switching between domestic and foreign bank accounts will remain after 1992, and a business relationship with a local bank will remain less complicated than with a foreign bank. Furthermore, Grilli argues, 25Evidence that supports this view was highlighted by Grilli (1989b). For individual financial services, he noted that the price dispersion across countries that had already liberal ized, like Germany, Belgium, Luxembourg, the Nether lands and the United Kingdom, was no less than across the remaining EC members. Total (million ECU) Belgium France Germany Italy Luxembourg Netherlands Spain United Kingdom Total 685 3,683 4,619 3,996 44 347 3,189 5,051 21,614 Percentage of GDP 0.7% 0.5 0.6 0.7 1.2 0.2 1.5 0.8 0.7 SOURCE: Emerson et al. (1988), p. 106. the use of other, more appropriate market struc tures produces smaller estimated gains from 1992 than those based on perfect competition. The bottom line is that the estimates in the Cecchini Report are probably optimistic. Of course, the absence o f better estimates precludes any quantitative statements about the degree of overstatement. Single Currency The preceding discussion, including the esti mates in the Cecchini Report, has presumed that 12 currencies continue to exist within the EC, albeit tied together by the exchange rate target zones of the European Monetary System (EMS). Thus, far from there being a single mar ket in financial services, there will continue to be 12 quite separate markets at the retail level. Within those markets, firms will operate separa ble portfolios and most retail customers will stick almost exclusively to their domestic en vironment.2 6 The creation o f a single currency, which was agreed upon at Maastricht, the Netherlands, in deposit and loan book in each currency. For example, a Dutch bank with a subsidiary in Greece will use drachma deposits rather than guilder deposits to fund drachma loans. 26Separable portfolios means that a bank with subsidiaries in more than one member state will operate a matched MARCH/APRIL 1992 76 December 1991 will induce major changes, ir respective of the regulatory regime.2 Obviously, 7 the foreign exchange market—and with it the costs of currency conversion—among the EC members will be eliminated. Closely related is the fact that the international accounting of many businesses will be simplified by the elimi nation of multiple currencies. On the other hand, many contracts will have to be rewritten. For example, a long-term bond contract that re quires interest and principal payments in a specific currency, say French francs, will have to be modified. Generally, retail customers will continue to do business with familiar institutions in their own countries, while wholesale market arbitrage and potential competition ensure that product prices are brought closely into line throughout the EC. These competitive pressures will lead to changes in the regulatory structure so that the conduct of business rules become more similar and, in some cases, identical; otherwise, firms in some countries will be at a competitive disadvantage relative to firms in other countries.2 It is 8 difficult to predict exactly how business rules will be harmonized for each financial service and, thus, how extensive the potential gains from a “free” single market will actually be. A more homogeneous and unitary monitoring mechanism is likely, although its full implica tions are equally hard to anticipate. Nonetheless, the gains from a single market are more likely to be realized if monetary union is achieved. CONCLUSION The goal of 1992 is to create a single Europe an market, a goal that encompasses the finan cial services sector. Our assessment is that the 1992 reforms are a small step toward the liber alization o f the financial services sector. Clearly, 1992 will contribute to the realization of some gains, especially in countries that have previous ly resisted liberalization. Nonetheless, serious doubts exist about how extensive the changes will be in the near future and, thus, the magni tude of the gains to be realized overall. In reali 27A recent issue of The Economist (“ The Deal is Done,” 1991) characterizes the Maastricht Treaty as important as the Treaty of Rome because it lays the foundation for a much closer union of countries via a single currency, a common foreign and defense policy, common citizenship and a parliament with power. A summary of the Maastricht Treaty as it pertains to monetary union can be found in “ Mapping the Road” (1991), page 5. FEDERAL http://fraser.stlouisfed.org/ RESERVE BANK OF ST. LOUIS Federal Reserve Bank of St. Louis ty, the 1992 legislation will not cause major changes. The reason is that virtually all of the potential efficiency gains in the financial serv ices sector can be (or have been) achieved through the combination of the abolition of ex change controls and the freedom of foreign firms to enter domestic markets. In fact, the former was implemented in July 1990 (in all but Spain, Portugal, Greece and Ireland). The key innovation o f the 1992 legislation is the split between home country authorization and host country conduct o f business rules. This dichotomy will create problems. Whereas wholesale markets already are highly integrated, not just within Europe but at the global level, 12 quite different retail markets will continue to exist in the near future. This segmentation means that many existing regulatory burdens will re main; however, regulatory complications may multiply as numerous domestic and EC authori ties become involved in the supervision of a sin gle firm. Finally, in some markets, like insurance, rigid regulation of domestic markets will delay any implementation o f the current model o f a framework directive until well beyond 1992. The greatest boost to financial market integra tion, once markets are open, will be the use of a single currency. With a single currency, pres sure will mount to revise the regulatory struc ture so that the conduct of business rules are homogeneous. Major changes in the regulatory structure lie ahead. It is these changes that will create a sin gle market and allow for the realization o f sub stantial gains in the next century. REFERENCES Bannock, Graham, Ron Baxter, and R. Rees. A Dictionary of Economics (Penguin Books, 1972). Bartholomew, Philip F., and Vicki A. Vanderhoff. “ Foreign Deposit Insurance Systems: A Comparison,” Consumer Finance Law: Quarterly Report Vol. 45 (1991), pp. 243-48. Blanden, Michael. “ Ironing Out Those Troublesome Bumps,” The Banker (February 1988), pp. 56-59. 28Not surprisingly, the U.S. legal system has had considera ble experience with conflicting laws and regulations across states. The Uniform Commercial Code is an excellent ex ample of states reaching general agreement on numerous laws. See Levine (1976) for additional details. 77 Blundell-Wignall, Adrian, and Frank Browne. “ Increasing Financial Market Integration, Real Exchange Rates and Macroeconomic Adjustment,” OECD Department of Eco nomics and Statistics Working Paper No. 96, 1991. Boucher, Janice L. “ Europe 1992: A Closer Look,” Federal Reserve Bank of Atlanta Economic Review (July/August 1991), pp. 23-38. Capie, Forrest H., and Geoffrey E. Wood. “ Financial Struc ture in a Changing Regulatory Environment: Europe after 1992,” in Game Plans for the '90s (Federal Reserve Bank of Chicago, 1990). Coughlin, Cletus C. “ Estimating the Economic Effects of 1992,” 1991 International Trade and Finance Association proceedings, forthcoming. _______ . “ What Do Economic Models Tell Us About the Ef fects of the U.S.-Canada Free Trade Agreement,” this Review (September/October 1990), pp. 40-58. “ The Deal is Done,” The Economist (December 14, 1991), pp. 51-54. Emerson, Michael, Michel Aujean, Michel Catinat, Philippe Goybet, and Alexis Jacquemin. The Economics of 1992 (Oxford University Press, 1988). Grilli, Vittorio. “ Europe 1992: Issues and Prospects for the Financial Markets,” Economic Policy (October 1989a), pp. 388-411. _______ . “ Financial Markets and 1992,” Brookings Papers on Economic Activity (No. 2, 1989b), pp. 301-24. Hill, Andrew. “ EC Leaders Reminded of Single Market Goal,” Financial Times, December 9, 1991. Key, Sydney J. “ Mutual Recognition: Integration of the Finan cial Sector in the European Community,” Federal Reserve Bulletin (September 1989), pp. 591-609. Levine, Mark Lee. Business and the Law (West Publishing, 1976). “ Mapping the Road to Monetary Union.” Financial Times, De cember 12, 1991. Rosenberg, Jerry M. The New Europe: An A to Z Compendi um on the European Community (Bureau of National Af fairs, 1991). “ Second Council Directive of 15 December 1989,” Official Journal of the European Communities, December 30, 1989. Tyley, Robert T. "The ECU: A New Global Currency,” Inter national Economic Insights (July/August 1991), pp. 38-40. U.K. Department of Trade and Industry. The Single Market: Financial Services, March 1991. MARCH/APRIL 1992 Federal Reserve Bank of St. Louis Post Office Box 442 St. Louis, Missouri 63166 The Review is published six times p er yea r b y the Research and Public Information Department o f the Federal R eserve Rank o f St. Louis. Single-copy subscriptions are available to the public f r e e o f charge. Mail requests f o r subscriptions, back issues, or address changes to: Research and Public Information Department, Federal R eserve Rank o f St. Louis, P.O. Rox 442, St. Louis, Missouri 63166. The views expressed are those o f the individual authors and do not necessarily reflect official positions o f the Federal R eserve Rank o f St. Louis o r the Federal R eserve System. Articles herein may b e reprinted provided the source is credited. Please provide the Rank’s Research and Public Inform ation Department with a copy o f reprinted material.