The Role of Seigniorage in the 1970’s Inflation and in the Breakdown of the Bretton Woods International Monetary System Report by: Birgit Doerfler December 2001 One of the policy instruments a government has at its hand when thinking about financing its expenditures is the collection of seigniorage, being the real revenues from issuing money to buy goods and non-money assets. “In the United States, for example, the Treasury issues debt and uses the proceeds to make the government’s purchases of goods and services. If the central bank so chooses, it can monetize the debt by printing money and buying it back from the public.”1 The inflation tax can be seen as an essential part of the government’s optimal taxation program. “Seigniorage is alternative to higher ordinary taxation now or, if interest-bearing debt is issued instead, to higher taxation or more seigniorage in the future. Insofar as money creation by the government causes inflation, thereby affecting the real value of nominal assets, seigniorage is a tax on money held by households or by commercial banks for compulsory reserves.”2 Money creation, as every distortionary form of taxation, is inferior to lump-sum taxation as a revenue device for the government. However, the ease of collecting this specific kind of a tax, the low cost of printing money3 and the fact that the burden of seigniorage falls on those holding the currency, which might to a great extent be nonresidents, speaks in favor of it. In the U.S. from 1948-89 seigniorage accounted for less than two percent of total federal government revenues and for approximately 0.3 percent of GNP (Champ and Freeman (2000)). Obstfeld and Rogoff (1999) display very similar numbers for the importance of seigniorage for the United States for a more recent time period. According to them, from 1990 to 1994 seigniorage revenues helped to finance 2.19 percent of the U.S. government spending and amounted to 0.44 percent of GDP. A country wishing to participate in a fixed exchange rate regime has to take into account that considerable constraints on the usage of seigniorage will be introduced. The monetary authority can no longer freely set the rate of money creation to collect seigniorage. Under a fixed exchange rate regime countries have to choose rates of money creation so that the fixed exchange rate can be sustained, which leads to the same inflation rates across participating countries. A fixed exchange rate regime thus bears the danger of high world inflation because each country will 1 2 Obstfeld and Rogoff (1999), footnote 12, p. 523 Spaventa (1989), p. 557 1 have an incentive to inflate as the seigniorage tax base would now be the world’s money balances and not only the real value of the money stock of the respective country. “Because every national government will wish to inflate to collect seigniorage from the citizens of other countries, a large inflation of the world’s money stock will result.”4 Furthermore an interesting question is how much seigniorage the key country of a fixed exchange rate regime can extract from the other participants and whether this can turn into a severe problem for the survivability of this regime. During the Bretton Woods fixed exchange rate system this might indeed have been the case, as several countries thought that the United States forced the others to finance its deficits by abusing the privilege to issue the primary reserve currency of the system. “The larger European countries vexed under what Charles de Gaulle called that "exorbitant privilege," by which the U.S. financed what de Gaulle's advisor Jacques Rueff called a "deficit without tears." The gain the U.S. acquired from the use of the dollar as an international currency was a form of seigniorage, or money tax. Dollar inflation only aggravated the money tax. When the U.S. inflation rate went up, dollar balances depreciated in real terms. In order to keep the same real value of reserves or the same proportion of reserves to imports, countries would have to increase their dollar holdings just to make up for inflation.”5 Countries deciding to take part in a fixed exchange rate regime thus face a trade-off between a stable exchange rate and the loss of control over seigniorage. Fischer (1982) argues that even: “If it is optimal to have a fixed exchange rate, it (the country) should still use its own money to avoid paying seigniorage to the foreign country.”6 A further disadvantage of a fixed exchange rate system is the exposure of individual countries to monetary and real shocks transmitted from the rest of the world via the balance of payments or other channels of transmission. According to Dibooglu (1999) because of the existing capital controls under Bretton Woods, international trade might have been a major transmission channel of macroeconomic shocks. Under this viewpoint it seems astonishing that fixed exchange rates within the provinces of a country, like the United States, work. The monetary organization of the United States is nothing else than a fixed exchange rate regime between 12 district Federal Reserve banks. Every single 3 According to the United States Treasury printing a paper one dollar bill costs approximately 3.8 cents, a 100 dollar bill 4.7 cents. 4 Champ and Freeman (2000), p. 95 5 Mundell (1998) 6 Fischer (1989), p, 296 2 district Federal Reserve Bank prints its own currency and agrees to swap their currency, at any time and in any amount, for any other district’s currency at a fixed rate. The crucial point in the design of this monetary system is that all the Federal Reserve banks decide together on how much money will be printed and the collected seigniorage is pooled and distributed by the U.S. Treasury. Exactly this monetary policy coordination between the Federal Reserve banks distinguishes the monetary system of the United States from the Bretton Woods system and ensures its sustainability. The Bretton Woods exchange rate system The Bretton Woods monetary system was constructed as an adjustable peg combining the favorable features of a fixed exchange rate gold standard and flexible exchange rates. The break down of this anchored dollar standard in the early 1970s might have been partly triggered by the excessive collection of seigniorage by the leading country of the system, the United States. Though Bretton Woods was initiated, among other things, to prevent inflation by tying down gold’s dollar price to $35 per ounce, inflation itself might have been one of its major enemies. “A commonly held view is that expansionary domestic programs in the U.S. and increases in defense spending due to the Vietnam War brought inflation. In order to maintain fixed exchange rates, other countries had to accommodate U.S. expansionary policies by intervening in the foreign exchange market and increase their money supply by importing U.S. inflation.”7 Under the Bretton Woods system the United States set their monetary policy and the other participating countries were responsible for pegging their exchange rates against the U.S. dollar. “… it was primarily the willingness of countries outside the United States to hold their foreign reserves in the form of American dollars that gave the United States the freedom in monetary policy.” 8 Problems started to arise when the United States shifted their primary policy interest from price stability to full employment in the early 1960s. This policy shift made highly expansionist U.S. fiscal and monetary policy in the late 1960s necessary, in order to finance the rapid buildup of social programs and the Vietnam War. “The US inflation rate, which had averaged less than 2% before 1965, rose to 4.4% in 1968, 5.4% in 1969 and 5.9% in 1970.”9 The accelerated money growth in the United States starting from 1965 and its positive correlation with 7 Dibooglu (1999), p. 74 Obstfeld and Rogoff (1999), p. 568 9 Mundell (1994), p. 6 8 3 the evolution of the budget deficit can be seen in Bordo (1993, Fig. 1.32, p. 77). As a consequence, the European countries became increasingly reluctant to import US inflation by buying dollars to keep their dollar exchange rates fixed. The difference between the participating countries’ willingness to accept inflation seemed to have been one of the major discrepancies in the system. “The Bretton Woods system collapsed between 1968 and 1971 in the face of U.S. monetary expansion that exacerbated worldwide inflation. The United States broke the implicit rule of the dollar standard by not maintaining price stability. The rest of the world did not want to absorb dollars and inflate. They were also reluctant to revalue. … Ultimately, the United States attached greater importance to domestic economic concerns than to its role as the center of the international monetary system.” 10 Furthermore the increasing deregulation of capital markets made the procedure of discrete changes in the parities costly and the whole system vulnerable to speculative attacks on weak currencies.11 All this taken together made the system more likely to collapse and led to its breakdown in 1971 when America rejected the conversion of dollars into gold, thus declaring the dollar inconvertible. As far as the possibility to collect seigniorage from nonresidents during the Bretton Woods system is concerned, there are also voices stressing its limitations. “… US dollar liabilities were held as interest bearing assets, which do not yield seigniorage, except in a limited way: central banks did not have much of a choice between accumulating dollar-denominated bills or choosing other assets, thus leaving the former with a competitive advantage that could translate in lower interest rates. Since the demise of Bretton Woods, seigniorage is admittedly much more limited.”12 Garber (1993) even goes further in stating the following “The dollar reserves held by foreign monetary authorities as the ultimate source of liquidity were primarily bank liabilities and money market instruments earning a competitive rate. Foreign central banks therefore did not lose seigniorage on these funds or pay an inflation tax (as in Mundell 1971, chap. 16). Rather, they gave up seigniorage to the United States only to the extent of the same liquidity premium that the price of all money market instruments and deposits in the United States bore – that is, the prorated cost of zerointerest-bearing reserve deposits held at the Federal Reserve by the U.S. banking system. In an inflationary environment, however, the associated rise in the Fed funds rate raises the cost of liquidity if the U.S. banking system cannot proportionally reduce its reserve 10 Bordo (1993), p. 82/3/4 Obstfeld and Rogoff (1999), p. 567/8 12 Bismut and Jacquet (1999), p. 83 11 4 holdings because the cost of the banking system of reserve holdings rises. In this case, the spread between liquid and illiquid securities rises, and holding liquid dollar securities becomes more costly in terms of yield foregone. Since the dollar liquidity is more costly, U.S. holders and official and private foreign holders of liquid dollar securities will reduce their real holdings by switching to less liquid alternatives.”13 Eichengreen (1993) compares M1 growth for the United States and the other G7 countries form 1959 to 1972 (Figure 14.10, p. 646) and cannot find proof for an excessively expansionary U.S. monetary policy, which would lead to a rejection of the monetary hypothesis of the collapse of Bretton Woods. Even though he recognizes some contribution of U.S. monetary and fiscal policies to the Bretton Woods collapse, he argues that tighter U.S. policies, on the other hand, might have led to a shortage of international reserves for the rest of the world. Whether these policies would have prevented the run on the dollar in 1971 or simply delayed it remains debatable. “Ultimately, then, the collapse of Bretton Woods was attributable as much to the structure of the system as to the specific policies pursued. Structural flaws dictated collapse sooner or later; policies at home and abroad determined only the timing of the event.”14 Genberg and Swoboda (1993), on the other hand suggest, that tighter U.S. monetary policy, as one policy option, might have had averted the breakdown of Bretton Woods. “Given the regime, the only way to avoid its untoward consequence, the excessive world rate of inflation, would have been a more moderate increase in the U.S. rate of monetary expansion. Such moderation would have avoided the dramatic increase in dollar reserves outside the United States. That increase could also have been avoided, even given the high actual rate of U.S. monetary expansion, had the rest of the world accelerated domestic credit creation. But, in that case, world inflation would have been as high as, or even higher than, it actually was.”15 The Functioning of the Bretton Woods System The currencies of the participating countries were treated equally. Each one was required to maintain its par value within a one percent margin by intervening in the currency of every other country. As the United States was the only country to peg its currency in terms of gold, Bretton Woods was an anchored dollar standard. The primary targets of the monetary system were the convertibility of the dollar and a stable price level. The role of the United States was to adjust its policies (fiscal and monetary) to maintain convertibility of its currency, whereas Europe 13 Garber (1993), p. 473/4 Eichengreen (1993), p. 650 15 Genber and Swoboda (1993), p. 304 14 5 controlled the composition of the US deficit to achieve inflation targets. The advantage of this system was that by keeping up the convertibility of its currency, the United States, the reserve currency country was made accountable for its policies. Mundell (1994) gives a compact description of the functioning of the Bretton Woods monetary system. “The outer countries peg (directly or, through other currencies, indirectly) their currencies to the inner country's currency (the U.S. dollar) and thus act as residual purchasers or sellers of dollars, while the inner country (the United States) pegs the currency to the ultimate asset (gold), and thus acts as the residual buyer or seller of gold. This means that the size of the U.S. deficit determines the increase in reserves of the rest of the world, while its composition determines the change in reserves of the United States, given the rate of increase of monetary gold holdings in the world. When U.S. monetary policy is very expansive the outer countries have to buy up large amounts of dollars and this has direct and indirect inflationary consequences for the outer countries; similar, when U.S. monetary policy is restrictive there is a scarcity of dollars and this has deflationary consequences for the rest of the world. The U.S. has to buy up any excess supply of gold on world markets and satisfy any excess demand out of its own reserves; failure to do so would result in the dollar price of gold moving away from the dollar parity. Europe, on its part, has to take up any excess of dollars offered to it or supply any excess of dollars demanded; failure to do so would result in the exchange rate moving away from its dollar parity. The boundary conditions are given by the U.S. stock of gold and Europe's stock of dollars; the United States cannot supply gold, nor Europe dollars, they lack. But there is an asymmetry in these conditions because, as long as gold and dollars can be supplied at the U.S. Treasury, Europe has access to additional dollars in exchange for gold. The total reserves (dollars and gold) of Europe therefore constitute Europe's boundary condition, whereas the gold reserve of the United States represents the U.S. constraint. Control of the system rests on U.S. monetary policy, on the one hand, and Europe's golddollar portfolio on the other. When the United States expands the dollar supply it puts upward pressure on world incomes and prices directly, because of interest rate effects and spending changes in the United States, and indirectly because of increases in European reserves. Similarly, when the United States contracts the dollar supply, it puts downward pressure on world prices. Europe's gold-dollar portfolio is the other control variable. When Europe converts dollars into gold it weakens the U.S. reserve position and stimulates or compels a monetary contraction and when it converts gold into dollars it strengthens the reserve position and permits or compels a monetary expansion. Europe's gold-purchase policy thus influences U.S. monetary policy, while the latter "determines" world prices and incomes. When U.S. monetary policy is forcing inflation on the rest of the world, Europe can stimulate or compel a contraction by gold purchases; and when U.S. monetary policy is deflationary, Europe can entice an expansion by gold sales.”16 On the following pages a review of papers related to the research agenda can be found. 6 Grilli, 1989. Exchange Rate and Seigniorage With the use of the inflation tax as a source of revenue for the government, monetary policy has to be seen as an integral part of the government budget constraint. If a country wants to maintain a fixed exchange rate while at the same time conducting inflationary monetary policy this will sooner or later lead to the collapse of the fixed exchange rate system. Grilli’s data for the U.K. government expenditure and seigniorage to GNP ratio from 1870-1990 display a sharp increase of both during fixed exchange rate periods. Nevertheless the advantages of fixed exchange rates roots in the reduced exchange rate volatility, which improves flows of trade because of the missing uncertainty in the evolution of the exchange rates. Grilli sketches scenarios under which a country might find it profitable to switch to a fixed exchange rate regime. “During periods of relatively moderate level of expenditure and low revenue requirements, the distortion introduced by not using inflation tax may be quite small. During these periods, therefore, it would be optimal to fix the exchange rate in order to avoid losses due to speculative activities or to real exchange rate swings. On the other hand, sudden increases in the level of expenditure, or increases in the cost of using alternative tax instruments may make it preferable to incur these welfare losses in return for seigniorage revenues. Once the value of using seigniorage revenues decreases again, monetary authorities will find it optimal to revert back to a fixed exchange rate arrangement.”17 Empirical evidence of seigniorage revenues for the United States and different seigniorage measures Kirschen, 1974. The American External Seigniorage. Origin, cost to Europe, and possible defences The Bretton Woods fixed exchange rate system allowed the U.S. to collect seigniorage from Europe. “…, the American Government has been able to achieve an entirely new type of seigniorage, levied on foreigners. Thus external seigniorage is the advantage resulting, for a country, from the fact that it may impose on independent countries the use of its currency.” 18 The U.S. collected seigniorage from countries in the Bretton Woods system by requiring that they hold an unlimited amount of dollars. According to Krischen this started in 1960 and he estimates 16 Mundell (1994), p. 3/4 Grilli (1989), p. 585/6 18 Kirschen (1974), p.355/56 17 7 its cost for the six European founder members of the EEC of having reached about 9 thousand million 1972 dollars by the end of 1972. “… one half of this sum results from a loss of interest, European de facto long-term loans to the U.S.A being remunerated as if they were short term, and the other half finds its origin in devaluations of dollars held by Central Banks under compulsion.’19 In 1960 an unwritten agreement between U.S and several European countries was reached according to which the Europeans committed themselves to not fully convert its dollar surpluses into gold. More or less severe pressure from the U.S. made the European countries stay in line. Kirschen’s estimate of the yearly average of external seigniorage from 1960-1972 amounts to 5.4 thousand million dollars. He does not include foreign private sector holdings of dollars into his calculations as those could always have been sold to a Central Bank. U.S. liquid liabilities to western European official institutions (Table 2, p. 364) increased from 5.5 in 1959 to 7.5 million thousand dollars in 1961 and between 1969/70 a huge increase of 7.2 thousand million dollars can be observed. The liquid liabilities to banks averaged at yearly 11.79 thousand million dollars. Kirschen calculates the dollars submitted to seigniorage by taking the currency assets of the Central Banks of the Six multiplying it by 0.95, as the dollar holdings amounted to 95% of their holdings in convertible currencies, and subtracting from it the amount of dollars necessary for exchange market interventions. He estimated the amount of dollars used for interventions as 5% of annual imports of goods, excluding services. Dollars submitted to seigniorage more than doubled from 1970 to 1971 and peaked in 1972 with 20 thousand million (current) dollars (Table 4, p. 367). Dollars created by European Central Banks themselves (eurodollars) should not be included in calculation of American external seigniorage. For the calculation of losses of interest he uses the difference between the marginal interest rate in the American economy (rates of return on direct investments in U.S. domestic manufacturing) and the actual interest rate of three-month treasury bills.20 The evolution of the cost of seigniorage at 1972 prices shows a big increase from 1969-1972. This raises the question of whether the U.S. tried to push the line too far in the attempt to finance an increasing amount of government expenditures for the Vietnam War and whether this led to the collapse of Bretton Woods in 1971. Altogether the cost of seigniorage for the six European founders of the EEC between 1960 and 1972 added up to 6.3 thousand million 1972 dollars, which constitutes a mean yearly cost of 485 million 1972 dollars (Table 7, p. 370). 19 Kirschen (1989), p.356 He does not follow the line of other authors to use the difference between short-term and long-term interest rates, as the dollars outside the U.S can be simply seen as credit without limit or repayment. He thinks that this is valid only for the private sector, whereas the dollar assets of monetary authorities are much less liquid and interest rates should thus be higher. 20 8 Another source of costs, are the losses in capital due to devaluations. According to Kirschen’s assumption that the long-term interest rate already included a premium against the 1971 devaluation, he only calculated the capital loss accruing to the ten percent fall of the dollar in 1973. This would amount to 2.7 thousand million dollars of 1973 capital loss. So the sum of both cost sources amounts to 9 thousand million of 1972 dollars, which at 1972 constant prices would have been approximately of the size of the post World War II aid to the six European countries. Under the light of Kirschen’s findings it seems only natural to ask whether the United States by collecting seigniorage from Europe financed its expansionary fiscal policy of the late 1960s and early 1970s. One may reason that this was a permissible way for the United States to get a restitution of the funds provided to Europe after World War II. But at the same time it is questionable whether these actions of the United States jeopardized the Bretton Woods monetary system and subsequently led to its collapse. Barro, 1982. Measuring the Fed’s Revenue from Money Creation Barro highlights where the transfers of the Federal Reserve system to the U.S. Treasury show up in the national account. “The Federal Reserve is treated as a member of the corporate sector. The Fed’s payment to the Treasury (labeled as “interest on Federal Reserve notes”) are regarded as taxes on corporate profits (at nearly a 100% rate in this case). Therefore, the category, “corporate profits tax accruals”, includes as a substantial component the earnings of the Fed.”21 For the 1946 to 1981 period Fed’s payment as a portion of corporate profit taxes show an increasing pattern with the highest value of 22% achieved in 1981. According to Barro measuring seigniorage as the product of interest rate and the amount of high-powered money is the appropriate measure, if the finite maturity of the Fed’s bond portfolio, administrative expenditures net of fees, gold, float, Treasury deposits, and surplus are neglected. In this case, the payments of net revenues to the Treasury would maintain equality between the monetary base and the Fed’s portfolio of securities and loans and the earnings out of the Fed’s portfolio would really be captured by this seigniorage measure. Differences occur because of the sluggishness of the yield on the Fed’s holdings in responding to interest rate changes, because a small part of the 21 Barro (1982), p. 327 9 Fed’s revenues actually goes into surplus and because there are changes over time in gold, Treasury deposits and float.22 The comparison of the 1946 to 1981 reports for the Federal Reserves net revenues, the product of interest rates (from 1953 to 1981 yield on U.S. government securities with 3-year maturities, from 1946 to 1952 yield for 3- to 5-year maturities) and the annual average of high-powered money and the change in money base show substantial deviations from one another. The product of interest rate and annual average high-powered money always exceeds the Fed’s net revenues. The reasons for this gap are an excess of the monetary base over the Fed’s loan portfolio or a sharp increase in interest rates. The difference between the change in the monetary base and the product of the interest rate and the annual average of high-powered money would only disappear, if the nominal interest rate were equal to the growth of the monetary base. In a period of increased real interest rate, the revenues of the Fed will increase without any increase in the rate of prices or the monetary base. As a percentage of federal government’s total tax and non-tax receipts the inflation tax (product of interest rate and average annual high-powered money) is always bigger than net Fed’s revenues and the change in monetary base. For 1981 the inflation tax makes up for 3.8% of total federal receipts, the Fed’s revenues for 2.3% and the change in the monetary base for 1.3%. These are the highest values for the first two measures in the 1946 to 1981 period. The comparison between those numbers shows how essential the choice of the measure of seigniorage is in order to detect its importance as a government revenue source. In relation to GNP the inflation tax takes again the lead over the Fed’s net revenues, with the highest values reached in 1981, 0.8% for the inflation tax and 0.5% for the Fed’s net revenues. Until the early 1960s the different measures of seigniorage in relation to tax receipts but also to GNP experienced a moderate increase, but since then they more than doubled (Table 3). This would suggest that the role of seigniorage as a source of financing government expenditures increased from the late 1960s to the beginning of the 1980s. Jefferson, 1998. Seigniorage Payments for the Use of the Dollar: 19771995 Jefferson extends the paper of Barro (1982) to a longer time period, which allows him to check whether the importance of seigniorage in government financing continued to persist throughout later periods. Furthermore, he takes up the idea that a considerable part of U.S. seigniorage comes 22 Barro (1982), p. 329 10 from the rest of the world, especially from emerging and transitional economies unable to provide their residents with stable national monies.23 His calculations for the overlapping period with the Barro paper, 1977 to 1981, show slightly diverging numbers. The ratios of seigniorage revenue to tax receipts and GNP show the peak of this revenue source in the early and mid 1980s for the net Federal revenues and the opportunity cost measure of seigniorage (inflation tax). The monetary seigniorage measure (change in base money) reached its peak in the early 1990s due to a substantial increase of U.S. currency held abroad. According to Jefferson, the net foreign flows of US currency abroad constitute a measure for the payments that the rest of the world had to make to the US for the use of the dollar. The opportunity cost measure for these payments is the product of the average three-month Treasury bill rate and the foreign US dollar holdings. The ratios of seigniorage (opportunity cost measure) to total tax receipts show that until 1992 domestic seigniorage always outweighed foreign seigniorage revenues but starting from 1993 this relation changed, for example in 1995 domestic seigniorage accounts for 6.7% of total federal taxes, whereas foreign seigniorage for 8%. These findings support the view that the United States actually collects a considerable part of its seigniorage revenues from foreigners, even without a fixed exchange rate system being implemented. But in this case, the proceeds from external seigniorage simply represent payments for the voluntary use of the U.S. dollar in foreign countries. The following reviewed paper by Fischer (1982) gives reasons of why a government might allow its citizens the use of a foreign currency. Fischer, 1982. Seigniorage and the Case for a National Money. In case a country experiences a period of high inflation there exists an incentive for the private agents to substitute their national money for foreign currencies in order to avoid paying inflation tax to its government and to get hold of a reliable store of value and medium of exchange. Whether a country allows this, so called dollarization, depends on the importance of seigniorage as a revenue source for the government. When deciding on whether to impose currency controls or not, and thereby deciding on whether or not to allow the use of foreign money, the government has to consider the costs and benefits of using foreign money. Switching to the use of foreign money makes the government loose one of its revenue sources, seigniorage. Seigniorage is often seen as “a ready source of finance in an emergency, as evidenced by wartime suspensions of the 23 Jefferson (1998), p. 225 11 gold standard.”24 By giving up the domestic currency a government cannot access this easy and fast remedy to short-term government financing requirements. Furthermore, in order to be able to import the foreign money, a country will on average have to run a balance of payments surplus. If a country permits dollarization, it has to be aware that by giving up its monetary policy all the economic policy goals can only be achieved through fiscal policy. Fischer distinguishes three different possible scenarios, between which the government chooses by comparing the costs and benefits of a potential regime switch. A floating exchange rate, which gives the government full control over its monetary policy. A fixed exchange rate with domestic money, where the level of seigniorage the country can collect is determined by the foreign country’s inflation rate. A fixed exchange rate with foreign money, which makes the country pay seigniorage to the foreign country. In case a country decides in favor of a fixed exchange rate, Fischer argues, that the use of domestic money is still superior to the use of foreign money, because it allows for the collection of seigniorage. Nevertheless the benefits in the form of reduced transaction costs in international trade and the discipline imposed on a government by the use of foreign money might outweigh the costs of lost seigniorage opportunity. So there actually might be situations in which it is favorable for a country to give up its domestic currency. Fischer outlines two situations leading to high levels of seigniorage, passive and active seigniorage. In the passive case seigniorage is high because of a rapidly growing economy, here the government has to supply a large amount of money to meet the demand. This is not necessarily accompanied by high inflation. In the active case, seigniorage is excessively used as a revenue device. Fischer mentions several examples for active seigniorage all of them being LDC’s like Argentina, Uruguay, Chile and Brazil. Fischer measures seigniorage as the change in high-powered money (Reserve Money, line 14 in the IMF’s International Financial Statistics, data from 1960s and 1970s) to nominal GNP seigniorag e C C C y P C y p GNP Y CY y PY Where y denotes the real income elasticity of currency demand (assumed either to be one or one-half) and p the price elasticity (assumed to be one). Fischer also calculates the cost of 24 Fischer (1982), p. 297 12 moving from a fixed exchange rate regime to the use of foreign money for different groups of countries under the assumption that bank reserves are held in interest bearing form so that seigniorage would only be collected from the use of currency. The resulting costs in the form of seigniorage paid to the U.S. range from .74 to 1.77 percent of GNP. Looking at the one-time cost of acquiring the initial stock of foreign high-powered money (measured by the currency to GNP ratio), Fischer finds it to be prohibitive averaging at around 8 percent of GNP. Although Fischer proposes actions to reduce the cost of the use of foreign money, he concludes that the use of domestic money is still the cheapest way to organize a countries monetary system. Dollarization might only be a reasonable step to be taken if the government has to be kept under control. Click, 1998 Seigniorage in a Cross-Section of Countries Since, except from lump-sum taxation, all kinds of taxation are distortionary, the government when deciding on whether to use seigniorage or other forms of taxation as a means of revenue collection should choose the less costly method. The optimum tax model predicts higher seigniorage when government spending is higher. In a cross-country study on ninety countries for 1971-1990 Click cannot find a dependence of seigniorage on the ratio of government spending to output. But in case the government needs an easy and fast instrument to collect revenues, seigniorage is always at hand. “If the government cannot repay domestic debt out of conventional taxes, however, it can simply print money. This serves to raise expectations of inflation and further constraint the government’s ability to issue debt. It also creates an incentive for the government to resort to seigniorage not only to finance transitory government spending, but to create unanticipated inflation which devalues the existing stock of debt. However, erosion of the debt may be negligible over long periods of time, as unanticipated inflation is likely to average zero when private agents form expectations rationally.”25 Especially if it is not easy or convenient to change conventional taxation in need to finance transitory government expenditures, the government can always use seigniorage, its residual revenue source. Click’s results seem to underline the assumption that seigniorage is primarily a source for financing transitory spending requirements. The 1971-1990 average annual rate of seigniorage for the U.S. amounts to approximately 2% of government spending and 0.4% of GDP, which lies far below of average amounts of all the considered countries. With both values at the lower end of the range, the U.S. does not use 25 Click (1998), p. 164 13 seigniorage excessively as a means of financing government expenditures. Of course an almost 20 year average takes away all possible variation in reliance on seigniorage during different periods in time. Measures of Seigniorage It is important to distinguish between seigniorage revenues and proceeds from inflation tax. Seigniorage is simply the sum of the change in real money holdings and inflation tax proceeds, which is the total capital loss that inflation inflicts upon holders of real money balances. M t M t 1 M t M t 1 M t 1 M t 1 Pt Pt 1 Pt 1 Pt Pt where M t 1 M t 1 Pt Pt 1 M t 1 Pt 1 Pt Pt Pt 1 “In a growing economy, seigniorage revenue typically exceeds inflation tax revenue, as the government can print money to accommodate a rising demand for real balances without generating inflation.”26 Mundell (1971, chap.15) also separates this growth seigniorage, which is the growth in money balances necessary to support real output growth and avoid deflation, from inflation tax, which are revenues from money issued in excess of real output growth.27 Flipping through empirical work about seigniorage it becomes obvious that different measures of seigniorage have been used. The subsequently reviewed papers should give an overview on which measure to use under which research agenda. Drazen, 1985 A general measure of inflation tax revenues Multiple measures for seigniorage have been proposed in the literature, all of them can be derived as special cases of Drazen’s general measure introduced in this paper. Expanding the money supply opens up two sources of revenue collection for the government. On the one hand there is the revenue accruing from those people already holding real balances and on the other hand there are the profits earned on those, who wish to hold real balances in the future. Drazen measures revenues from seigniorage as the product of the rate of growth of money balances and the per capita level of real balances. This itself is composed of the revenue from 26 27 Obstfeld and Rogoff (1999), footnote 13, p. 523/4. From Bordo (1993), footnote 58, p. 57 14 keeping real money balances constant in the face of growth and the second component being the tax arising from people not changing their money balances in the face of inflation. M M M x m m PL M PL L x denotes total real balances, m real per capita balances, is the rate of monetary expansion, is the rate of inflation and L denotes population. According to Drazen a question is, whether the outstanding real per capita money balances themselves affect the real revenue of the government. The government budget constraint in per capita terms is given by mb kg b is the real value of interest bearing government debt, k g are real assets of the government and is the net government indebtedness, which is the sum of discounted costs of running the monetary authority c and indebtedness above this amount ̂ c . measures the part of previous money issuance immediately spent on consumption. If assets and debt bear the same real interest rate r the government budget constraint can be rewritten in net asset terms, a k g b , ˆ c . To get the net steady state revenue the increase in nominal asset which will give m a holdings necessary to keep the per-capita real value of assets constant in the face of inflation and population growth must be subtracted form the interest on net assets. Thus we get ia ( n)a (r n)a as the profit form the non-interest-bearing outstanding money. The current flow revenue (tax on real balances) together with the revenue from assets purchased form the issued money gives you the total revenues from money creation, m (r n)a . According to Drazen the use of rm as a measure of seigniorage is only appropriate if there are no costs involved in running the monetary authority and if all of the seigniorage revenues are used to buy assets. All the measures of seigniorage seen in empirical work, m , m and im can be derived as special cases from this general measure of seigniorage. When m is used, the focus lies on the tax on existing consumers, m also concentrates on current issuance but takes growth of the economy into account. The measure im is the right one to use, if all the increase in money supply is accompanied by an equivalent retirement of interest bearing debt. If the government is constrained to use the inflation tax for net asset accumulation, c=0 and the focus lies on currently alive individuals the measure becomes m r (m ) im r . “Therefore, if alternative inflation paths are constrained to keep constant (and if r is unchanged), the change in inflation tax revenues when changes is measured by changes in 15 im .”28 In this case the change in revenues from seigniorage is equal to the opportunity cost of holding money. If the government uses the revenues from money creation to buy a certain fraction of the total capital stock of the economy, which would mean that government indebtedness remains constant, then total revenue from money creation is given by m rk . If we concentrate again on the individuals currently alive excluding the supply of money balances to new born consumers, we should use m as the appropriate measure, as rk constitutes simply a transfer from the government to the consumers. Drazen’s paper provides a compact overview of which measures of seigniorge to use under different research objectives and shows that all of them can be derived from his general measure. HONOHAN, 1996 Does it matter how seigniorage is measured? Honohan discusses three alternative measures for seigniorage revenues, which should be chosen according to the focus of research. Rate of growth of money multiplied by the real money base, which measures the actual amount of tax being collected. This cash-flow measure shows the real purchases the government finances by printing money. Rate of inflation multiplied by the real money base, which captures the declining purchasing power of money balances. It thereby represents the capital levy aspect of seigniorage to the amount that it differs from the interest measure by the unanticipated inflation. Nominal rate of interest multiplied by the real money base reflecting consumer’s opportunity cost of holding money instead of other liquid assets and covering the tax aspect of seigniorage. The monetary base is simply the sum of currency and reserves. The inflation and the interest measure of seigniorage differ only if there is surprise inflation. For the steady state analysis there is no difference between those two measures as the real interest rate would be zero, but for empirical work the appropriate measure should be chosen. According to Honohan even the investigation of the taxation and the deficit financing role of the inflation tax call for the use of a different measure. 28 Drazen (1985), p. 329 16 “… the financing and the tax aspects of currency issue should always be sharply distinguished. By issuing new currency, the government is helping to finance its deficit. By arranging that the yield on currency falls far short of the market yield on other liquid assets, the government is implicitly imposing a tax on existing holders.”29 In the case of an open economy with a fixed exchange rate, where the government uses the newly printed money to purchase goods in a foreign country, the private sector does not pay a tax. The spending shows up only in the balance of payments. All that happened was a capital transaction exchanging foreign assets for domestic goods. Using the rate of money growth would disclose this action wrongly as a tax on the private domestic sector, whereas the other two measure do not lead to this conclusion. If the government takes this newly printed money, lets say dollars, to buy goods in a country participating in the fixed exchange rate regime, lets say Germany, this will increase the money supply in Germany. Assume that because the currencies are seen as perfect substitute, the U.S. government takes the issued money directly to German producers, which will accept it as a means of payment. The foreign producers take it back to their central bank to exchange it to DM. This will increase the dollar reserves of the German Central Bank along with the money supply in Germany leading to higher inflation in Germany as well. If the German Central Bank cannot convert those dollars back into DM (or gold, as it might have been the case in the Bretton Woods monetary system) this would force Germany to involuntarily hold reserves in dollars. As far as those reserves earn a lower than competitive rate of interest, the U.S. can collect seigniorage from Germany in the form of lost interest. 29 Honohan (1996), p. 294 17 References Barro, Robert J. 1982. Measuring the Fed’s Revenue from Money Creation, Economics Letters 10: 327-332. Bismut, C. and Jacquet P. 1999. The Euro and the Dollar: An Agnostic View. Tokyo Club Papers, 12: 75-97. www.tcf.or.jp/tcp/tcp12/6IFRIMaster1.PDF Bordo, M. 1993. The Bretton Woods International Monetary System: A Historical Overview. In: A Retrospective on the Bretton Woods System: Lessons for International Monetary Reform. ed. Bordo, M.D. and Eichengreen B., p. 3-108. Chicago: The University of Chicago Press. Click, R. W. 1998. Seigniorage in a Cross-Section of Countries, Journal of Money, Credit, and Banking 30(2): 154-171. Dibooglu, S. 1999. Inflation under the Bretton Woods System: the Spillover Effects of U.S. Expansionary Policies, Atlantic Economic Journal, 27(1): 74-85. Drazen, A. 1985. A General Measure of Inflation Tax Revenues, Economics Letters 17: 327330. Eichengreen, B. 1993. Epilogue: Three Perspectives on the Bretton Woods System. In: A Retrospective on the Bretton Woods System: Lessons for International Monetary Reform. ed. Bordo, M.D. and Eichengreen B., p. 621-657. Chicago: The University of Chicago Press. Fischer, S. 1982. Seigniorage and the Case for a National Money, Journal of Political Economy 90(2): 295-313. Garber, P. M. 1993. The Collapse of the Bretton Woods Fixed Exchange Rate System. In: A Retrospective on the Bretton Woods System: Lessons for International Monetary Reform. ed. Bordo, M.D. and Eichengreen B., p. 461-494. Chicago: The University of Chicago Press. Genberg, H. and Swoboda, A. K. 1993. The Provision of Liquidity in the Bretton Woods System. In: A Retrospective on the Bretton Woods System: Lessons for International Monetary Reform. ed. Bordo, M.D. and Eichengreen B., p. 269-315. Chicago: The University of Chicago Press. Grilli, V. 1989. Exchange Rates and Seigniorage, European Economic Review 33: 580-587. Honohan, P. 1996. Does it matter how seigniorage is measured? Applied Financial Economics 6: 293-300. Jefferson, P.N. 1998. Seigniorage Payments for Use of the Dollar: 1977-1995, Economics Letters 58: 225-230. 18 Kirschen, E. S.1974. The American External Seigniorage. Origin, Cost to Europe and Possible Defences, European Economic Review 5: 355-378. Mundell, R. A. 1971. Monetary Theory. Pacific Palisades, Calif.: Goodyear. Mundell, R. A. 1994. The European Monetary System 50 Years after Bretton Woods: A Comparison Between Two Systems. Paper presented at Project Europe 1985-95, the tenth edition of the “Incontri di Rocca Salimbeni” meetings, in Siena, 25 November 1994. www.columbia.edu/~ram15/ABrettwds.htm Mundell, R.A. 1998. Commentary: The Case for the EURO II, The Wall Street Journal, March 25. phoenix.liunet.edu/~uroy/eco41/mundell/mundell6b.htm Obstfeld, M. and Rogoff, K.1999. Foundations of International Macroeconomics. 4th edition. Cambridge, Massachusetts: MIT Press Spaventa, L. 1989. Seigniorage: Old and New Policy Issues, European Economic Review 33: 557-563. 19