eichengreen euro

Journal of Economic Literature ,Vot. XXXI (September 1993), pp. 1321-1357 European Monetary Unification By BARRY EICHENGREEN Vnioersity of California at Berkeley Work on this paper was begun during visits to the International Finance Division of the Board of Governors of the Federal Re.scnc System and the Research Department of the International Monetary Fund and completed during visits to the Bank of France and the Institute for Advanced Study in Berlin. I gratefully acknowledge the support and hospitality of all these institutions while absolviuf], them of responsibility for the lieici expressed here. Research assistance was provided by Ansgar Rurnler and financial support by the Center for German and European Studies of the University of California. For comments on portions of this work I thank Tamin Bayoumi, Lorenzo Bini-Smaghi, Paul De Grauue, Jeffry Frieden, Alexander Italianer, Peter Kenen, Paul Masson, Thomas Mayer, Ronald McKinnon, Jacques Melitz, Richard Portes, Gianni Toniolo, Jiirntrols would be removed, and an EC s\ stem of central banks, modeled loosely on the U.S. Federal Reserve System, would assume control of the monetary policies of the member countries. The size of the EC budget would be increased dramatically, and the Community would coordinate national tax and expenditure programs. SignificantK, the authors of the \\ erner Report were not wedded to a single currency. Though preferring this option to fixed exchange rates between national currencies, tliey suggested that both alternatives were viable and that their benefits were broadly comparable. Elements of the Werner Report were implemented in March of 1972 when EC countries agreed to an arrangement, dubbed "the Snake," limiting bilateral exchange rate movements to 2'/i percent bands. Policy convergence and coordination lagged behind. But when the first OPEC oil shock created different levels of unemployment in different European countries, national governments came under different degrees of pressure to respond in ways that risked inflation. Some currencies were devalued, others revalued. Some countries left the Snake temporarily, others permanentK. Increasingly the arrangement proved incapable of delivering the exchange rate stability that was its central goal (Thygesen 1979). This realization prompted another round of exchange rate stabilization negotiations at the Bremen summit in 1978, leading to the creation of the European Monetary System (EMS) in 1979. The EMS was a modest initiative by the standards of the Werner Report. It sought 1324 Journal of Economic Literature, Vol. XXXI (September 1993) to stabilize exchange rates without at the same time requiring the elimination of international policy divergences either through the application of fiscal and monetary rules or 1)\' empowering the Community to coordinate national policies. Its central element, the Exchange Rate Mechanism (ERM). was designed to accommodate the different policies pursued in different countries. Some participants, including Italy, a charter member of the ERM, and the United Kingdom and Spain, latecomers both, were permitted fluctuation band.s more than twice as wide as the otliers.® Countries were allowed to realign—that is. to change their central rates, essentially the midpoint of tht'ir bands against the other countries—when policy divergences produced balance-of-payments disequilibria. Controls on capital account transactions were retained, giving countries some leeway to run different policies without immediately provoking capital movements in anticipation of realignment. B. The Delors Report The EMS had operated for seven years when in 1986 the members of the Community initialled the Single European Act (SEA). This committed the members of the Community to the creation of an integrated market free of obstacles to the unfettered movement of commodities, capital, and labor by the end of 1992 (Evan Davis, John Kay, and Michael Ridge forthcoming). The SEA was the outgrowth of a June 1985 White Paper which had laid out the goal and the deadline. With the process of European economic integration gathering speed, in 1988 the European Council ap- ' Italy subsequently sliilti'd to the narrow band, which was in place at the time of the September 1992 crisi.s analyzed in Section 8 below. From April 1992. Portugal ha.s also been a ineinlM-r of the ERM with a uiili' fluctuation ni.iriiin pointed a committee, chaired by Jacques Delors, President of the European Commission, to study the feasibility of supplementing the single market with a monetary union.'" The recommendations of the Committee for the Study of Economic and Monetar> Union (1989), published as the Delors Report, provided the framework for intergovernmental negotiations in 1991. Many of its conclusions found their way into the Maastricht Treaty. Nothing is more revealing of continuity and change in discussions of European monetary unification than the similarities and differences between the \\ erner and Delors Reports. The Delors Report, like its predeces.sor, sought to achieve monetary union in less than a decade. Both documents recommended proceeding gradually; like its predecessor, the Delors Report described a transition in three stages. Like the Werner Report, the Delors Report emphasized the need for fiscal harmonization. But there were also differences, reflecting conceptual and political developments in the interim. Where the Werner Report had recommended removing capital controls at the end of the process, the Delors committee endorsed their removal at the beginning. And in a concession to political realities, the Delors report did not propose transferring control of national budgetary policies to the Community. Instead it simply recommended that the Community monitor the "overall economic situation . . . assess the consistency of developments in individual countries with regard to common objectives and formulate guidelines for policy." But to strengthen fiscal and monetary discipline, the Delors Committee proposed rules that would '" The Commission has 17 memlx;rs: one from each of the 12 memlx-r states and a second representative from (Mill of the five large member countries. Eichengreen: European Monetary Unification 1325 firstly, impose efleoti\ e upper limits on budget deficits of iiuli\ idual member countries of the Community . . . [and] secondly, exclude access to direct central l)ank credit and other forms of monetary financing . . . In contrast to this limited degree of fiscal centralization, the Delors Report recommended complete centralization of monetary functions. Where the Werner Report had sketched in broad terms a European system of national central banks joined together in a monetary federation, at the heart of the Delors Report was a new entity, the European Central Bank (ECB), to formulate and execute the Community s single monetary policy. National central banks, though they would not be abolished, would become mere operating arms of the ECB.^^ The authors of the Delors Report (the central bank governors of the 12 Communit\ countries plus 5 independent experts) were quite specific about the new institution's prerogatives and responsibilities. Its priority should be price stabilit>'; only insofar as doing so was compatible with that objective should the ECB support other economic policies of the Community and its member states. It should be responsible for the operation of monetary and exchange rate policies and of the payments system. Finally, while it should not supervise commercial banks, it should help to coordinate the policies of the competent national authorities. Thus, compared to the Werner Report, the Delors Report simultaneously embraced both more and less centralization. There was to be more centralization of monetary control in the hands of a " Committee for the Study of Economic and Monetary L'nion (1989, p. 30). Tliesc issues are discussed further in Sections 2.C and 7. A below. '^Together, the European Central Bank and the national ciiitral banks are referred to as the European System of Central Banks. The national central banks will retain functions that are not allocated to the ECB by the treat\ or do not otherwise interfere with its functions. See .Mberto Giovannini (1992). Community institution to prevent national central banks from executing directives in different ways and so undermining the common monetary policy.'"^ There was to be less central control of national fiscal policies than envisaged in the Werner Report and no e.xtensive transfer to the Community of fiscal functions carried out at the national level." A final contrast between the Werner and Delors reports was the greater attention paid in the latter to mechanism design. The clearest illustration is the Delors Report s insistence on the early introduction of a single currency to insure "the irreversibility of the move to monetary union" (Para. 23). C. The Maastricht Treaty In December 1989, following the appearance of the Delors Report, the governments of the EC member states convened an Intergovernmental Conference to prepare amendments to the Treaty of Rome (the basic law of the European Community). The Conference commenced work in December 1990, one year later producing draft amendments ill the form of a treaty. ^^ Following the Delors Report, the Maastricht Treaty describes a monetary union to be achieved in three stages. But " It has l)een suiiuested by one of the members of the Delors Committee (Thygesen 1989) and the present author (Eichengreen 1992b) that the I'.S. Federal Reser\ e System sufiFered from precisely such di£Bculties in its early years until an appropriate de- Uree of centrali7.ation was established. '' Indeed, a theme of the Delors Report is "subsidiarity." a word meant to capture the notion that public functions should be undertaken at the lowest level of noverninent feasible (see Davis. Kay. and Ridge forthcoming). '^ If the amendments are adopted, the "title" ol the Treat\ of Rome on Economic Polie\ w ill be replaced by a "title" on Economic and Monetary Policy. Other changes will take the form of a series of protocols, including a "Protocol on tlie Statute of the European System of Central Banks and of the European Central Bank." a "ProtcKol on the Statute of the European Monetary Institute, and a "Protocol on the pAtessive Deficits Procedure (Giovanniiii 1992). 1326 Journal of Economic Literature. Vol. XXXI (Septonbcr 1993) where the Delors Report depicted the transitional stages in rather stlu'inatic term.s. the Maastricht Treaty is specific ahout their features. Stage I is to be marked 1)v the removal of capital controls, the reduction of international inflation and interest rate differentials, and tlie increasing stability of intra-European exchange rates. .Member countries must strengthen the independence of their central banks and otherwise bring domestic laws into conformance with the treaty. Thi- inauguration of this stage in Jul\' 1990 was marked by the removal of Europe s most important capital controls. "^ Less progress was made, however, in acliieving convergence of inflation and interest rates and their underlying determinants, and a foreignexchange market crisis in September 1992 led to exchange rate changes not anticipated h\' policy makers and to the reimposition of some capital controls.'' Stage II, starting at the beginning of 1994, Is to l)e characterized by the further convergence of national economic policies and by the creation of a temporiuy entity, the European Monetary Institute (EMI), to coordinate membercountry monetary policies in the final phases of the transition and to plan the move to monetary union. If during Stage II the Council of Ministers, made up of ministers of economics or finance from each national government, decides (by qualified majority, where each country s '*The Delors Report had been rec<'i\ed by the governmeiit.s of the member states at the Madrid Summit in June of 1989, where they decided on the date for initiating Stage 1. The inauguration of Stage I was also marked In a diminishing frequency of excliaiiuc r.itc changes, although tins was an ad hoc policy decision rather than a corollary ol official provisions of the treaty. '^The temporary reimposition iif controls was lawful under the provisions of the SE.\ allowing for their emergency application for no more than six months. These events and their implications for the future of the K.\IU process are discussed in Section 8 below. N'ote is weighted by its size) that a majority of member countries meet the preconditons for monetiir\' union (which are detailed in the Maastricht Treaty and analyzed in Section 7 below), it may rec-ommend that the Council ol Heads of State vote (by qualified majority) on v\liether to inaugurate Stage III, establishing the independent European central bank and transferring to it responsibility for the conduct of monetar\ policy. To prexcnt the indefinite continuation of Stage II, the treat\ requires the EC Heads of State or Government to meet no later than December 31st, 1996 to assess whether a majority of EC member countries .satisfy the conditions for monetar\ union and to decide whether to set a date for the beginning of Stage III. If no date lias been set by the end of 1997, Stage 111 will begin on January 1st, 1999. In the latter case. Stage 111 may proceed with the participation of a minority of EC countries. L'pon the inauguration of Stage III, exchange rates will he irrevocably fixed. The EMI will be succeeded b\ the European Central Bank, which will assume control of the monetary policies of the participating countries. The Council of Ministers will decide when to replace their national currencies with the single European currene\\ It nia>' do so on the first day of Stage III; if it chooses otherwise, the ECB will simply instruct its operating arms, the national central banks, to convert their national currencies into one another at par until these are replaced by the single currency. In its gradual approach, even to the extent of distinguishing three stages, and in its emphasis on policy convergence as a precondition for monetary union, the Maastricht Tieat)' echoes provisions of the \\ erner Report issued more than 20 years ago. The treat>' differs from its predecessor in placing more emphasis on the need for policy convergence and price Eichengreen: European Monetary Unification 132-; stability during the transition to monetary union but less emphasis on central control of fiscal functions once unification is achieved. It pays more regard to the design of institutions and procedures to achieve these goals. * 3. The Rationale for Monetary Unification The attentive reader will have noticed how glibly the last section skipped over the sources of renewed impetus for monetary unification following the passage of the Single European Act in 1986. A cogent case has been made for completing the internal European market in commodities, capital, and labor (see Emerson et al. 1988 for the European Commissions own analysis). Turning 12 segmented European markets into an integrated economy whose constituents can specialize fully in producing goods and services in which they have a comparative advantage and in which factors of production can flow freely to wherever they reap the highest returns is hard to challenge on efficiency grounds. But is there an economic rationale for having a single currency and a European central bank accompanying this process?'® One argument sometimes heard is that separate currencies pose a significant barrier to commodity- and factor-market integration. Travelers cannot help but be impressed by the cost of changing money at airports, where commissions can reach seven percent or more. Currency con- ''' Note that 1 am focusing here on economic justifications. An entirely different line is that the rationale for EMU is political rather than economic. The argument is that monetary unification serves as a stepping stone toward deeper |X)litical integration. Some evidence— for example, German insistence on strengthening the powers of the European Parliament in conjunction with progress on EMU—supports this hypothe.sis. For an introduction to the relevant literature, see David Cameron (1992), Geofirey Garrett (1993). and Wayne Sandholtz (1993). Such political considerations are however beyond the scope of the present paper. version costs of this magnitude would represent a significant barrier to trade across Europe s internal borders and an obstacle to the movement of workers. But airports are not hotbeds of financial competition. Travelers quickly leam that in urban centers, where competition is more intense, they can save significantly on transactions costs. Emerson et al. (1990) estimate that curreney conversion costs average 2.5 percent for travelers, and that they fall to as little as 0.05 percent for transactions in excess of $5 million. Averaging the transactions costs incurred by finns and individuals, they conclude that ciirrenc\ conversion costs come to 0.4 percent of CDP for the EC as a whole. This hardly seems an adequate return on a project riven with uncertainties and risks. Conversion costs, it might be obje'cted, are only the most obvious cost of separate currencies. Also disruptive is the uncertainty associated with the possibility of changes in their relative prices. National currencies necessarily imply exchange rate uncertainty, and exchange rate uncertainty discourages cross-border transactions, according to this argument. Vet the evidence that exchange rate uncertainty or variability discourages international trade is lar from conclusive.'^ In a careflil recent study, Jeffrey' Frankel (1992) considers various determinants of the volume of trade in a cross section of countries, concluding that the effect of exchange rate uncertainty, while present, is quite small. Doubling the standard deviation of the real exchange rate reduces the volume of trade by only 0.7 percent. This is not surprising insofar as the existence of forward markets in foreign exchange permits traders to hedge currencv risk at low cost. '*An influential early study is Peter Htxjper and Steven Kohlhagen (1978). A dated but still-useful survey of the relevant Hterature is IMF (198'ti 1328 Journal of Economic Literature. Vol. XXXI {September 1993) Because the service life of many kinds of plant and equipment exceeds the term to maturity of available forward contracts, exchange rate uncertainty should have a larger effect on cross-border investment than on trade. Robert Morsink and W'illem Molle (1991) report some evidence that exchange rate uncertainty depresses direct foreign investment among EC countries. Yet this argument can cut both ways. Firms with liabilities denominated in several currencies may wish to have assets whose returns are denominated in several cvnrencies as well. To hedge against exchange risk, they may set up plants to produce the same product in different currenc\ areas. David Cushman (1988) reports some evidence of this effect. Other things equal, however, this reduces the efficiency with which investible resources are allocated and, b>' lowering the return on capital, should depress the le\ el of in\'estment. The one study of which I am aware exploring the link between exchange rate variability and the le\el of investment (Kenen 1979) does not find evidence of a statistically significant effect, however. Assume for argument s sake the existence of significant adverse effects of exchange rate variability on trade and investment. Why theMi not minimize this variability simply by stabilizing exchange rates between European currencies instead of abolishing separate currencies altogether? Stabilizing exchange rates was, after all, what the European Monetary S\ stem was all about. Francesco Ciavazzi and Ciovannini (1989) show that its major participants achieved a significant reduction in exchange rate variability over the 1980s. .As EC countries eonduct nearly t\vo-thirds ot their trade with one another, the sucifss of the EMS in stabilizing intra-Etiropean exchange rates goes a long way toward minimizing e.\- change rate variability for the relevant countries. It would seem paradoxical that the European Community, the one part of the world that has succeeded in largely insulating itself from exchange rate \ariabilit\\ is where the call for monetary unification was taken up. A resolution of the paradox is that the Single European Act undermined the viability of the EMS. The EMS was a hybrid of pegged and adjustable exchange rate regimes. Extended periods of exchange rate stability delivered man\' of the benefits of fixed rates, while periodic realignments redressed serious competitiveness problems. But periods of excliantie rate stability punctuated by oixasional realignments were possible onl\ because c apital controls protected central banks' re.serves against speculative attacks motivated by anticipations of realignment. If, for example, France sought to maintain lower interest rates than Cerniain, huge quantities of financial capital did not flow instantaiieousK from Paris to Frankfurt, immediately e.\- hausting the Bank of France s reseivcs. The interest differential had to be large and to be maintained for an extended period before substantial numbers of French arbitragers found it advantageous to incur the cost of circumx entinii French capital controls, thereby forcing a realignment. Thus, in the first deeade of the EMS's operation, monetarv' authorities in the participating countries retained limited policy autonomy. Capital controls took a variety of forms, ranging from taxes on holdings of foreii^ncurrency assets to detailed regulations on the uses to which foreign currency could be put. All of them represented obstacles to completing the internal market. It was hardly feasible to restrict the freedom of Frenchmen to open bank accounts in Cerniany, for examplc\ while c liminating all controls on intra-EC movements of portfolio capital and direct foreign investment, not to mention labor and commodities. Hence controls were a casualtv of Eichengreen: European Monetary Unification 1329 the 1992 program. The SEA mandated their elimination by JuK 1st, 1990, except in Spain and Ireland, which were exempted until December 31st, 1992, and Portugal and Creece, which were exempted until December 31st, 1995. (In addition, it allowed for emergency controls for a period of no more than six months. The Maastricht Treaty, however, rules out their use for any period from the beginning of Stage II on January 1, 1994.) Once the SEA undermined the viability of the "Old EMS," monetary unification followed inevitably.-" Or did it? One can imagine two alternatives. One is floating exchange rates. With floating rates two countries can integrate their economies but retain monetary autonomy. The preceding discussion identifies no clear economic reason whv' factor and commodity markets cannot be integrated while exchange rates continue to float. If the effects of exchange rate variabilitv on cross-border trade and investment ;u-e minimal, then resource allocation will be essentially the same as under fixed rates. Indeed, regional integration initiatives among countries whose exchange rates float, such as Canada, Me.xico, and the United States, the three partners in the prospective North American Free Trade Area, have proceeded with minimal discussion of e.xchange rate stabilization, much less currency unification. (See however Darryl McLeod and John Welch 1991a, 1991b; Bayoumi and Eichengreen forthcoming a: and Ceorge vou Furstenberg and David Teolis 1992.) This may change. Once trade is freed between the U.S., Canada, and \le.\ieo. exchange rate swings whit h adverselv affect competitiveness in the importing country may prompt calls for an exchange rate stabilization agreement to prevent ^Giavazzi and Luigi Spaventa (1990) distinguish the "Old" lie., capital-control ridden) and "Now" (capital-control free) I'.MS exchange dumping." Realizing that dissatisfaction will othei"wise be redirt^cted toward the free trade agreement, policy makers may accede to these demands. Thus, floating is incompatible with integration, this argument implies, not on eflRciency but on political economy grounds (the greater is integration, the larger and more onerous are the distributional consequences of exchange rate changes). The 25 percent depreciati(m of the British pound against ERM currencies like the French franc, the Cerman mark, and the Dutch guilder in the five months following sterling s September 1992 exit from the ERM illustrates this point. This dramatic depreciation of the pound was not offset immediately by changes in domestic-ciurency-denominated labor costs. By February 1993 it had led Hoover Co. to terminate v acuum-cleaner production in France in favor of expanding its operations in Scotland. It had caused Philips Electronics to cease producing cathode tubes in its Dutch plant in favor of Britain. It encouraged S. C. Johnson & Son, a U.S. household-products maker, to shift production from France to plants in Britain. This in turn led EC Commission President Delors to warn the British government that its exchange rate policies were antagonizing other EC countries in a manner incompatible with the privileges of the single market. Angry French officials threatened Britain with exclusion from the single market if it persisted in its abandonment of the ERM. (See for example, Eiic Ipsen 1993.) The linkage between exchange rate stabilization and Europe s Common Agricultural Policv (CAP) is also consistent with the hypothesis. The CAP seeks to maintain minimum domestic-currency prices for particular agricultural commodities in each of the relevant EC countries. (For details see Dimitrios Demekas 1330 Journal of Economic Literature, Vol. XXXI {September 1993) et al. 1988.) A significant change in intra- European exchange rates creates an incentive for producers in the country whose exchange rate has depreciated to export agricultural products to other EC countries, threatening to drive domesticcurrency prices in the importing countries through their CAP floors. Intra- European trade in these commodities is supposed to take place at "green exchange rates" different from market rates precisely in order to insulate domestic markets from these pressures, but adjusting those rates becomes increasingly difficult as exchange rate changes become frequent. And with the removal of border controls as part of the SEA, it will grow harder to enforce the prohibition on transacting at market rates. Thus, greater exchange rate flexibility would pose a fimdamental challenge to the CAP and its distributional objectives.^' The other conceivable alternative to monetary imification if pegged but adjustable exchange rates are precluded by the elimination of capital controls is firmlv fixed exchange rates between existing national currencies. Rather than attempting to create a European currency and a European central bank, why not just fix the relative prices of the national monies of the 12 Community countries once and for all? A sufficiently credible commitment to intra-EC exchange rate stability might fix exchange rates even in the absence of capital controls. Capital would flow in stabilizing rather than destabilizing directions, because speculators would have reason to believe the authorities' stated intention to defend the rate (as analyzed in the recent literature on exchange rate target zones, viz. Paul *' RecentK the EC moved to a system in which farm prices in countries that devalue their currencies are allowed to rise automatically by the ftill amount of the devaluation but ar<' not allowed to fall in strong currency countries. The effect of exchange rate changes is therefore to ratchet up the real cost of the CAP. In any case this arrangement does not offer a solution to the problems posed by floating. Krugman 1991; Robert Flood, Andrew Rose, and Donald Mathieson 1991). The rebuttal is that there exists—and always will exist—a fundamental difference between a single currency and fixed exchange rates between national currencies, namely an escape clause. No one believes that California will devalue against the U.S. dollar when it experiences a recession more severe than the rest of the country, for California has not had its own currency for more than 100 years. Attempting to establish one now would disrupt its relations with the other states and create a host of legal and constitutional problems. This serves as an exit barrieV (as the Delors Committee recognized; see Section 2.B above). In contrast, no matter how earnestly France reiterates its commitment to pegging the franc against the deutsche mark, there remains the possibility that a change in government will lead to a change in policy and a devaluation. In a democracy it is impossible to preclude the possibility than an existing policy instrument like the exchange rate will be utilized. Investing in the sunk costs of a common currency is needed to deter such action.-" Maurice Obstfeld (1992) shows rigorously how the existence of an escape clause can complicate effbrts to stabilize nominal exchange rates even if the authorities pledge to devalue only under exceptional circumstances. Because the contingencies in response to which the escape clause may be invoked are often unobservable, efforts to peg nominal ^ A referee pointed out—correctly—that while the possibility of escaping from a fixed exchange rate agreement is always there, such agreements in the past have lasted for long periods, the pre-1914 gold standard being a prime example. However, as I have argued elsewhere (Eichengreen 1992a), the credibility of exchange rate pegs under the classical gold standard and other such systems derived from a unique set of political conditions (restricted franchise, limited concern with unemployment, etc.) not present in the late 20th centurv. Eichengreen: European Monetary Unification 1331 rates tend to be destabilized by uncertainty about whether or not those contingencies obtain.^ The implication is that pegged exchange rates between national currencies are never perfectly credible; hence they substitute imperfectly for monetary unification. Thus, the Single European Act brought the Community to a crossroads. Its internal logic required the removal of capital controls, undermining the viability of the 1980s-style FMS and leaving a choice between greater exchange rate flexibility and monetary unification. Creater flexibility is fundamentally incompatible with the CAP, which many economists would take as a welcome opportunity to eliminate this inefficient program of agricultural subsidization, although politicians would dismiss them as unrealistic. More fundamentally, wider exchange rate swings would compound the adjustment difficulties associated with completing Europe's internal market. If national industries under pressure from the removal of barriers to intra- European trade find their competitive position eroded further by a sudden exchange rate appreciation, resistance to the implementation of the Single European Act would intensify. The SEA might be repudiated. In this sense and this sense alone, monetary unification is a logical economic corollary of factor- and product-market integration. 4. Possible Adverse Consequences of Abandoning Monetary Autonomy A single currency is unlikely to come without costs. The theory of optimum currenc> areas suggests that the reduction in transactions costs associated with a common currency should be balanced against benefits of retaining monetary independence and exchange rate changes ^ Matthew Canzoneri (1985). ."Vn escape clause can also be destabilizing if the relt'\ant contingencies are not clearly exogenous with respect to policy. as instruments of adjustment. (See Robert Mundell 1961; McKinnon 1963; and Keneu 1969; Yoshihide Ishiyama 1975 provides a survey.) Consider an asymmetric shock (a shift in demand from domestic to foreign products, for example) recjuiring an adjustment in domestic costs (in this case, a reduction) to restore prices and demand to levels consistent with full employment. Altering the exchange rate may be a convenient way of accomplishing this in a decentralized market. This is the "daylight-savings-time" argument for exchange rate changes. Whether it is costly to abandon this instrument depends on two factors: the shocks that participating countries experience and the utility of monetary policy for facilitating adjustment. If monetary policy is incapable of affecting output and employment, then forsaking monetary independence is costless.'^ But if nominal variables display inertia, due to coordination failure, because of the presence of long-term contracts or for other reasons, tlien monetary policy will matter. There is no more contentious debate in economics than whether this is the case. Even the extensive attention that has been lavished on the American data has failed to settle it for the United States.'^ Work on European countries is spottier. And even if V.S. investigators ^ Costlfs.s, that is. aside from the loss of seigniorage in high inflation countries. But growing resistance to inflation in Europe ha.s already brought about a significant reduction in inflation rates and seigniorage revenues (see Table 3 below). On changing attitudes toward inflation, see Susan Collins and C^iavazzi 11993). On the implications of EMU for seigniorage, see Vittorio Grilli (1989). The debate remains unresoht-d because of a fundamental identification problem. Money-output correlations cx>nflate the effects of money on output and output on money, as Christina Roiner and I>a\id Romer (19891 obser\'e. These authors sought to solve this riddle l>\ identifying exogenous monetary impulses on the basis of the proceedings of the V.S. Federal Open Market Committee, and concluded that monetary policy matters for output and employment in the United States. 1332 Journal of Economic Literature, Vol. XXXI {Septcmher 1993) ultimately conclude that monetary policv is efiective, tliere are good reasons to doubt that they would reach the same conclusion for Europe. A point of the large literature on real wage resistance in Europe (e.u., Michael Bruno and Jeffrey Sachs 1985) is that when nominal wages are fully indexed to prices, or real waives are impervious to price-level changes for other reasons, monetary policy will have no output or employment effects. Under these circumstances, European countries sacrifice little by abandoning monetary autonomy. Yet the conclusion of this literature is not that real wages in Europe are invariant with respect to inflation, only that they are less responsive than in the United States. Although OECD (1989) estimates that in North America only 14 to liS percent of a price increase is passed through to nominal wages, wheieas in Europe 25 to 75 percent is passed through, there nonetheless remains some real wage responsiveness to the price level and hence to monetary policy. Even in Gt rmany, where the effect is smallest, the estimated elasticity of real wages with respect to inflation is still 25 percent. Even if monetary policy is effective for redressing niacroeconomic imbalances, it need not follow that sacrificing monetary autonomy is costly. The costs will be negligible when the .same polic\ response is appropriate for all members of the monetary union. If all suffer deflationary disturbances simultaneously, then a unionwide reflationary monetary policy will suffice. The question for those concerned with the costs of monetary unification becomes whether the incidence of disturbances affecting the partners in the monetary union is symmetric or asymmetric. A priori the answer is not clear. Both France and Germany have automotive industries, steel industries, and electronics industries, for example. Because the .same industries operate in many European countries, a sector-specific shock will affect these countries in similar ways. With sh(x;ks to Europe s national economies relatively symmetric in incidence, the costs of abandoning the exchange rate could be low. Analysis of these issues is sparse. Much of it compares sectoral specialization in Europe and the United Stati\s on the groimds that the degree of sectoral specialization in the latter is consistent with monetary union (Krugman 1991; Lorenzo Bini-Smaghi and Silvia Vori 1992). But the degree of sectoral specialization is not a sufficient statistic for the incidence of shocks. Bayoumi and I (forthcoming b) go one step further by comparing the correlations of the GDP growth rates of other EC members with Germany's growth rate over the last 30 years, along with the correlations of the growth rates of other U.S. regions with that for the Mid-East. The latter is 0.68, the former a somewhat smaller 0.58. Though the U.S. Bgure is higher, the differential is not large; it does not suggest that asymmetric shocks will be an insurmountable problem for Europe. Daniel Gohen and Wyplosz (1989) transform real GDP data for France and Germany into sums and differences, interpreting movements in ihe sums as symmetric disturbances, movements in the differences as asymmetric disturbances. They find that s\ mmetric shocks are much larger than asymmetric shocks. Axel Weber (1990) applies their approach to other EG countries, reaching similar conclusions. Output movements are not the same as disturbances, of course; the former conflate information on both shocks and responses to them. Bayoumi and I (forthcoming b, c) use the technique of Olivier Blanchard and Danny Qiiah (1989) to recover disturbances and responses from time series of output and prices. This inEichengreen: European Monetary Unification 1333 volves transforming the residuals from reti^iossions of growth and inflation rates on lagged values of themselves, subject to the assumption that permanent disturbances aflPect both output and price levt-ls in the long run but temporary disturbances have no long-run output eflfect. Using this procedure, the correlations of other countrit-s' permanent disturbances with Germany s averages onK 0.33, compared to 0.46 in the United States, while the correlations of the other EC countries' temporary disturbances with Germany s averages onl\ 0.18, compared to 0.37 in the I'nited States. In contrast to man\ analyses of the raw data, then, this procedure suugests that usytnmetric disturbances may be more pervasive in Europe than the United States. Such empirical work is based iie'cessaril\ on historical correlations. Yet the structure generating those correlations may change with the completion of the single market and monetary union. Country-specific demand shocks caused hy national monetary policies will necessarily be eliminated by EMU. Following Blanchard and Quah in interpreting temporary disturbances as demand shocks and permanent disturbances as supply shocks, Bayoumi and Eichengreen suggest that temporary disturbances attributable to demand-management policy are likely to become more symmetric following EMU, while permanent, or supply, disturbances will have less tendency to change. An additional complication is that European countries ma\' become more specialized in production. The combination of government subsidies and import barriers that have supported the existence of a domestic automotive industry in every large European country, for example, will be eroded b\ the SEA. In sectors characterized by stroni^ agglomeration economies there will be an incentive to consolidate production. This, it is argued by authors like Krugman (forthcoming), will magnify coiiiitrs-specific shocks. But completion of the inteniiil market will also encourage intra-industry trade. In sectors characterized hs scale economies and product differentiation, different varieties of the same product ma\' be produced in a growing number of European countries. Hence, completion of the internal market, by encouragin5i intra-industry trade, may lead to an even greater duplication oi industries across EC countries (Emerson et al. 1990; Daniel Gros and Tlniiestii 1992; Bini Smaghi and Vori 1992). 1^ Chatelier s principle suggests that t'liminating one margin for adjustment (the exchange rate) should encourage adjustment on others. Within the U.S.. who.se regions cannot respond with exchange rate changes, adjustment to region- specific shocks occurs mainly through migration and, to a lesser extent, real wage flexibility (Blanchard and I^wrence Katz 1992). There does not appear to be scope in Europe for comparable adjustment on either margin. It is conceivable that real wages in Europe will become more flexible with monetary union (.see e.g., Henrik Horn and Torsten Persson 1988); ItaK s abolition of the scala mobile in the summer of 1992 as part of its effort to qualify for participation in EMU may be indicative of this tendency, although most observers would be skeptical that real wages will quickly come to exhibit the flexibility characteristic of Ainciican labor markets. Blanchard and Pierre Alain Muet (1993), for example, find that the growing credibility of France s commitment to pegging the franc to the deutsche mark has been accompanied l)y little increase in real wage flexibility. (A dissenting view is Emerson et al. 1990.) Nor is there rea.son to be optimistic that European labor mobility will rise to American levels. Not only is migration 1334 Journat of Economic Literature, Vol. XXXI (September 1993) between European countries significantly lower than between U.S. regions, but migration within European countries is lower as well. Migratory flows between French departements and German lander are only a third to a half those between U.S. states. Eichengreen (1993) uses panel data to estimate the relationship between interregional migration and regional differentials in unemployment and wages for Great Britain, Italy, and the U.S.. confirming that migration is less responsive to such differentials in these European countries than in the United States. The Single European Act will surely enhance both the incentive and capacity to migrate (see e.g., C^uiseppe Bertola 1989). The question is to what e.xtent. Linguistic and cultural differences will remain. It is far-fetched to assume that European labor mobility will rise to American levels in the foreseeable future (Blanchard and Muvt 1993; Davis, Kay, and Ridge forthcoming). This, together with evidence of limited wage flexibility and asymmetric shocks, implies that the loss of monetary autonomy that comes with monetary union will not be without costs. 5. Fiscal Implicatiowi of Monetary Union Having abandoned monetary autonomy, national governments confronted by asymmetric shocks may still have other policy instruments at their command. Faced with a domestic recession but unable to adjust the money supply, a national government should be able to increase public spending or reduce taxes. The standard targe ts-and-instruments framework suggests that the loss of monetary autonomy asscx^iated with EMU will place a premium on fiscal flexibility (.see for example Kenen 1969). Yet there are reasons to worry whether the fiscal independence needed to compensate for the loss of monetar\' autonomy will be available to national policy makers. As discussed above, factor mobility (especially capital mobility but also, to a limited extent, labor mobilit\) will ri.se with economic and monetary union. In turn this may limit the fiscal options that can be pursued unilaterally. This belief has prompted calls for fiscal federalism and fiscal-policy coordination at the level of the monetary imion. This section therefore asks three questions. First, will governments retain fiscal autonomy in an integrated Europe? Second, will it be necessary to supplement decentralized fiscal initiatives with fiscal federalism (also referred to as fiscal coinsurance) at the Gommunity level? And third, will there be a need for greater intt-mational coordination of fiscal policies? A. Fiscal Autonomy Economic and monetary integration could tighten the constraints on fiscal policy. Research suK^tsts that factor- and product-market integration will be more important than monetary union in this respect, and that statutor\ restraints on fiscal policies imposed under the Maastricht Treaty w ill be more important than market forces. Theoretical analyses such as Reuven Glick and Michael Hutchinson (1993) identify channels through which high capital mobility in conjunction with fixed exchange rates or monetary union tighten the government budget constraint. Public spending in tin- current period must etjual the sum of taxes, debt issue, and money creation (seigniorage revenues). Because solvency requires that a government service its debts, the present value of spending (inclusive of service on the initial debt stock) by a solvent government cannot exceed the present value of taxes plus the present value of seigniorage (VVillem Buiter 198.5). By Eichengreen: European Monetary Unification 1335 CHOS Greece Portii^l Spain Italy Source: Emerson S SElGNIORAi;h 198S (1) 2.75 2.23 1.36 1.13 et al. (1990). Note: "1993 sccniuno assumes and reduction in of inflation rates T.\BLE 1 ; REVENUE EFFEtrrs OF MONETARY UNION UNDER ALTERNATIVE S< KNARIOS (AS A PEBCKNTAGE OF GDP) "1993 Scenario" (2) 1.84 1.62 1.20 0.72 EMU Scenario" (3) 0.71 0.71 0.86 0.51 Single Market Effect (•4) = (1H2) 0.91 0.61 0.16 0.41 convergence of reserve ratios at 2%, elimination of interest payments the use of cash due to technological at 2% per annum. change. "EMU scenario" assumes, in addition. EMU Effect 5 = (2)-<3) 1.13 0.91 0.34 0.21 on reserves. convergence limiting the availability of seigniorage revenues, forsaking monetary independence for stable prices restricts the range of feasible fiscal policies. This effect is likely to be small. Table 1 shows, for the four Community members whose seigniorage revenues exceeded one percent of national income in 1988, their actual level in that year along with the European Commission's estimates of how far these revenues would fall with the completion of the single market and monetary union. (Gros and Thygesen 1992 review the other available estimates.) From the perspective of a government wishing to run a budget deficit temporarily for businesscycle- related reasons, the implications for fiscal flexibility are trivial. The integration of goods and factor markets will do more to tighten the budget constraint. Government borrowing today is limited by the taxes that can be levied tomorrow (taxes needed for, among other purposes, servicing the accumulated debt). If capital and labor are freely mobile within the economic union, borrowing today which implies higher taxes tomorrow may induce mobile factors of production to flee to lower-tax jurisdictions, eroding the tax base. Investors will understand that a government's ability to borrow today is limited by its ability to tax tomorrow, and that its ability to tax tomorrow is limited by factor mobility. Hence they will refuse to lend to governments threatening to exceed their borrowing capacity. The more integrated are factor markets, the sooner this will occur. ^^ Several authors have examined U.S. state and municipal bond markets for evidence of the operation of these mechanisms (Eichengreen 1990; Morris Goldstein and GeofiFrey Woglom 1992). In the most recent such study, Bayoumi, Goldstein, and Woglom (1993) find that the required rate of return on state obligations rises sharply with the debt-to-stateproduct ratio, and that state governments are effectively rationed out of the market when the debt-to-gross-state-product ratio exceeds nine percent.^^ Were this ^ For those who place a premium on fiscal flexibility, especially after monetary autonomy is eliminated, the implications are disturbing. In contrast, for tho.se who perceive a bias in the direction of excessive local government sjjending, this may be a healthy form of Tiebout competition. " They attribute the failure of previous studies to identify this effect to simultaneity bias: that higher interest rates discourage debt issue at the same time additional debt issue produces higher interest rates. 1336 Journal of Economic Literature, Vol. XXXI {September 1993) threshold to apply to Europe, all of its national governments would have already exhausted their ability to borrow! Clearly this is absurd. The propensity for mobile factors of production to flee U.S. states whose taxes exceed those necessary to service this debt ratio is partly a function of the debt levels and tax rates that prevail in neighboring jurisdictions. Because all European nations .service debts and levy taxes higher than those of any U.S. state, the threshold at which individual European nations are rationed out of the market will not be ecjually low. In addition, the fact that certain factors of production, notably labor, will remain less mobile in Europe wall allow more variation in tax rates and hence greater fiscal autonomy. Finally, higher debt ratios in Europe have been sustained by the fact that the issuing governments, unlike U.S. states, have been able to require their central and commerc iai banks to hold substantial amounts of their debts.-^ Thus, fiscal autonomy will not be eliminated by economic and monetary union. Indeed, the main reason that fiscal constraints could bind more tightly is statutory restraints on the use of fiscal instruments that will be imposed by the Maastricht Treaty. The treaty limits the budget deficits that countries may run during Stayt- II if they seek to gain admission to the monetary union. The stated limit is three percent of GDP, although this threshold is subject to qualifications. During Stage III, the Council will issue recommendations regarding membercountry fiscal policies. Countries failing to heed (Council recommendations to re- ^ There is also the fact that Eiiro[X!an nations, in c-ontrast to I'.S. states, will still receive the seigniorage revenues of the European central bank (which will \K allocated to them according to a formula agreed to !>> weighted votinii) But assuming that the ECB pursues a jx)lic\ of maintaining staMi' prices, seigniorage revenues are likely to make oiiK a small coiitrihution to their national budgets. duce their budget deficits may incur specific penalties. The Council may require them to publish additional information before issuing bonds and set urities, force them to make non-interest-bearing deposits with the Community, instruct the European Investment Bank to halt lending to them, and impose unspecified fines. Tiiese are likely to be the really binding constraints on national fiscal policies.-' B. Flical Coinsurance Along with concern about the possible loss of fiscal autonomy, tlu-ic is tlie question of whether the highly visible, politicized nature of public discussions of fiscal policy would allow it to be adjusted as smoothly as the relinquished monetary instrument. Both considerations have fueled a debate over tbe necessity of fiscal federalism, or regional coinsurance via the federal fiscal system. The idea is that a system of fedcnil taxes and transfers could be used to automatically shift resources toward jurisdictions suffering negative region-specific shocks. Its intellectual origins can be traced back at least to James Ingram (1959). In the 1970s a Community study group on the role of pul)lic finance in European integration (European Commission 1977) pointed to the need for a significant increase in the size and change in the structure of the EC budget on essentially these grounds. The importance of fiscal coinsurance in existing monetary unions is a disputed issue (see Jacob Frenkel and Goldstein 1991; Wyplosz 1991). Measuring net fiscal transfers to a jurisdiction is far from straightforward. Estimatin,ii the geographical incidence of taxes first requires an assumption about their economic inci- "''Tlie rationale for these fiscal restraints, as descrilwd in the Delors Report and the Maastricht Treat), is discussed in Section 7 A lielim Eichcuiireen: European Monetary Unification 1.337 dence, a question on which contributors to the literature do not agree. In addition, observed tax payments to and transfers from fedeial governments reflect three distinct fiscal functions. First, if, as in many existinji federations, federal tax rates are uniform or progressive while transft IS and other forms of federal expenditure are dispersed fairly evenly across regions, low-income regions will continualh' receive transfers from the rest of the federation; this is the e-cjualization effect of federal fiscal policy.'" Second, the federal tax liabilities of all regions will go down and their transfer receipts, for unemployment in.surance copaN ments and the like, will go up when all regions enter a recession simultaneously; this is the stabilization effect of federal fiscal policy. Third, net transfers from the federal government to a member state will go up wben it enters a recession not experienced by the rest of the federation; this is the regional coinsurance effect of federal fiscal policy. It is this third effect alone that is said to be a necessary concomitant of monetary union. Studies to date have not all succeeded in distinguishing these functions. The first empirical analysis, that of Xavier Sala-i-Martin and Sachs (1992), used data for U.S. census regions to relate tax and transfer payments to movements in pretax personal income, both measured relative to the national average. (They also included real energy prices and a time trend in their regressions, adjusted for simultaneity due to the dependence of state income on taxes and transfers, and normalized their variables by the relevant national averages to control for the stabilization effect.) The elasticities from these regressions were then used to infer ^' III the jargon of the EC, equalization addresses the problem of "cohesion." Tl>e term "equalization t'Bect" is different (and. 1 hope, clearer) than that which I have used in pri'\ii>us papers. the extent of regional coinsurance. Salai- Martin and Sachs found that federal tiLx liabilities decline b\ roughly 25 cents for ever> dollar by whicb regional income falls short of national income and that inward transfers rise by roughly 10 cents. Thus, insurance operates mainly through the tax side of the federal fiscal system. It is substantial. Von Hagen (1992) criticized these results for failing to distinguish insurance and equalization effects. A large share of the net transfers found by Sala-i-Martin and Sachs, he argued, reflected equalization, not insurance. Differencing the data in an effort to isolate the effect of changes in income on changes in transfers, he estimated smaller effects: the sum of the tax and tran.sfer elasticities came to approximately ten percent. His point was pursued by Bayoumi and Masson (1991), who considered both the U.S. and Canadian fiscal systems. Tbey first regressed each region s per capita income net of taxes and transfers on its per capita personal income inclusive of taxes and transfers. (.Both regressors were again normalized by tbe analogous national average to eliminate the stabilization effect.) Tbis equation measures the relationship between personal income before and after federal fiscal flows, with the slope coefficient capturing the si/e of the offset. For the United States, their coefficient of 0.78 indicates that, on average, federal fiscal flows reduce regional income inequalities by 22 cents on the dollar. Wbile somewhat smaller than Sala-i-Martin and Sachs's estimate, this still suggests a substantial effect. Bayoumi and Masson then estimate the same regression after differencing the variables to remove tbe equalization effect. Regressions on the differenced data produce a coefficient of 0.69, suggesting that the stabilization of short-term fluct\iations (31 cents on the dollar) is even stronger than the overall effect. Thus, 1338 Journal of Economic Literature, Vol. XXXI (September 1993) TABLE 2 1992 EC Bi IK:ET (AS OF JI'I.T 1992) In Mill. ECU Percentage of Total Budget Common Agricultural Policy Structural Funds Research and Development, etc. Aid to Countries Outside the EC Administrative Expenses Reserves Total 35.348.000 18.557,399 3.906.683 3,649,929 4,097.741 1.000,000 66.559,752 53% 28% 6% 5% 6% 2% 100% Source: Bulletin der Europ&ischen Gemeinschaften/Kommission No. 5/1992. they provide no support for the critics of Sala-i-Martin and Sachs' results. But those criticisms may still be valid for other federations whose equalization policies are more developed. In Canada, for example, the constitution is generally interpreted as demanding fiscal equalization of regional income differentials. Bayoumi and Masson's analysis for Canada yields evidence of a substantial equalization effect: nearly 40 cents on each dollar. Their estimate of the insurance effect, while slightly smaller than for the U.S., is nonetheless substantial. If the case for fiscal coinsurance is granted, then does the EC have the capacity to undertake it? This is really two questions: does it possess the budgetary resources, and does it possess the political wherewithal? The Community's budget is little more than one percent of EC CNP. The largest share is devoted to the Common Agricultural Policy, which leaves it unavailable for other purposes (Table 2). Much of the remainder is allocated to the Structural Funds, which are targeted at low-income regions within the Community and hence provide more equalization than insurance. James Gordon (1991) estimates that a $1 fall in a member state's per capita income increases its Structural Fund receipts by at most one U.S. cent. At Maastricht a coalition of four low-income countries led by Spain insisted and received assurances that these funds would be increased. But proposals to increase significantly the size of the Community budget, a prerequisite for such a step, ran into resistance subsequently. And, given Gordon's estimates of the relevant elasticities, even a doubling of the Structural Funds would fail to provide regional coinsurance on the U.S. or Canadian scale. An alternative to increasing the Community's budget is to restructure it so as to enhance its capacity to carry out the insurance function. Alexander Italianer and Jean Pisani-Feny (1992) propose a system of transfers from the Community to member states as a function of national GDP and relative unemployment rates. As a country's unemployment rate rises relative to the Community average, so would its transfer receipts. The program would provide regional coinsurance on a scale comparable to that which exists in Canada (where about 20 percent of a decline in a region's relative income is offset). Assuming that transfers are capped once the unemployment differentials reach two percentage points, this system would require adding Eichengreen: European Monetary Unification 1339 to the Coinmiunity s budget no more than 0.25 percent of EC (iDP pven the historical relationship between national unemployment rates. Thus, regional coinsuraticr could be pro\ided without a revolution in European fiscal relations so long as a specific program is dedicated to thf task. Given the budgetary resources to institute a program of regional coinsurance, the EC ma\' still lack the political wherewithal. C:anada, where attempts to continually redistribute income among provinces by direct transfers through the Federal Covernment have poisoned the political chmate, gives grounds for proceeding cautiously. (Indeed, the same point can be made about various federal states in Europe itself) Proponents of regional coinsurance would respond that these difficulties reflect complaints alwut equit\. which arise in turn from programs of equalization (ongoing income redistribution), not coinsurance. In the long run, every regit)n should have occasion to benefit from the latter, providing no grounds for complaint. Every region would benefit, of course, only if the structure of the insurance program addressed moral hazard problems. A program of fiscal coinsurance that provided transfers in response to a rise in unemployment might encourage participating governments to pursue policies that increas(>d unemployment risk. In some federations, this problem is ameliorated In requiring copayments or repayments. In the U.S., for example, though states may borrow from the Federal Unemployment Trust Fund, they must pay interest on those borrowings. These que.stions are superfluous if adequate coinsurance is provided by member states' existing federal fiscal systems. Italianer and Pisani-Ferry estimate that France and Cermany provide more regional coinsurance internally than does the United States. Whereas unemployment insurance is funded primarily at the state level in the U.S., in France and Cermany unemployment insurance s\stems are nationwide.^^ Public contributions (to the retirement system, for example) are highly output elastic and account tor a larger share of GDP than in the United States. And in Germany interregional grants are large and elastic. For all these reasons, fiscal federalism within European nations may go some way toward substituting for explicit coinsurance at the EC level. But if output movements are highly correlated across Ge-rman regions, fiscal federalism within Germany will substitute poorly for regional CH)insurance at the Community level. As yet. there exist no studies of the extent to which national fiscal systems can discharge this responsibility. C. Fiscal Coordination The Maastricht Treaty instructs member states to Vegard their economic policies as a matter of common concern and [to] . . . coordinate them within the Council." As described above, it empowers the Council to formulate guidelines for the economic policies of member states and adopt recommendations for the countries concerned. Importantly, however, no enforcement powers are attached except for the aforementioned "Excessive Deficits Procedure," which is intended to address some quite specific moral hazard problems, not to enforce fiscal policy coordination generally (see Section 7 for further discussion). The lack of enforcement powers may reflect the belief among the treat> s framers that actual situation in the U.S. is .somewhat more complicated. Although each state administers its own unemplo\ment insurance trust fund, it also pays a fraction oi the payroll taxes levied to finauie the program into the Federal Unemployment Insurance Trust Fund. Eichengreen (1992c) provides a further discussion of the implications for Europe. 1340 Journal of Economic Literature, Vol. XXXI (September 1993) economic and monetary union does not strengthen the c;ise for fiscal policy coordination all that significantly. Disputing this position require.s that one articulate an explicit rationale for fiscal coordination. One is to limit tax competition (see Emerson et al. 1988). Assume that the level of taxes is optimalK set initially. Then further economic and monetary integration may tempt local (in Europe, national) governments to lower tax rates to lure enterprises from other regions, augmenting their tax base at the expense of neighboring jurisdictions. If this game of tax competition is played noncooperatively, location decisions will he unaffected in equilibrium, but the level of public services that can be financed b\ the available tax revenues will be depressed. EflBciency will be enhanced if jurisdictions coordinate their tax policies. Factor- and product-market integration certainly increases the scope for such competition. But because, as explained above, mobility in an integrated Europe is likely to remain lower than in existing economic and monetary unions like the United States, so is the scope for tax competition.'- Moreover, those who lielieve that politics biases taxation in Europe toward excessive levels would welcome a healthy dose of tax competition. A second common rationale for fiscal coordination derives from the international macroeconomic spillovers of national fiscal policies (Emerson et al. 1990). Because one country's fiscal policies affect output and employment in others, fiscal policies should be coordinated to internalize these intemational exter- Ndtf also that OMIN insofar as a common currency is a necossan L'oncomtnitant \ Some provision for accountability is admittedly made l)y the Maastriclit Treaty; although the proceedings of the Ciost-rning Board will be confidential, the ECB must publish a quarterly report and submit an annual report to the European Parliament. Members of the Executive Board may be called to testif\ before committees of the Parliament. It is not clear that these measures suffice. Much has been written in other contexts of the Community's "democracy deficit": the notion that EC decision makers are insufiRciently accountable to the European public they represent. Tlie same danger arises in connection with the ECB. As Cooper (1992b, p. 16) puts it, the Europeans have created an instrument [the ECB as constituted unilcr the Maastricht Treaty] that would greatly widen thi' already large democratic gap. The Maastricht agreement would create a powerful body of Platonic guardians to look after monetary affairs, effectively accountable to no one, yet with strong; influence on the course of economic afi'airs. Existing central banks, including those of Germany, the United States, Switzerland, and the Netherlands, while independent of the govemment in power are not independent of politics. German Eichengreen: European Monetary Unification 1343 Chancellors who come into serious conflict with the Bundesbank can, with a sympathetic parliamentary majority, simply change the relevant central bank statute. Accounts of Bundesbank history (e.g., Ellen Kennedy 1991) portray Germany's central bank as not unresponsive to governmental pressure. Statutory changes are more difficult to engineer in congressional systems like that of the United States, where the President, the Senate, and the House of Repre. sentatives must agree. But even there, the White House appears to have significant influence, via moral suasion, over the conduct of monetary policy. The Fed, it is argued, is reluctant to oppose the chief executive elected by the nation (Sherman Maisel 1980). When a conflict arise.s between the Fed and the Congress, Reserve System officials are called up to Capitol Hill to testify beneath the hot lights of the relevant Congressional committee. That these procedures matter is supported by evidence (e.g., Kevin Crier 1984) that the growth rate of the money supply is aflFected by shifts in the membership of the Senate Banking Committee. In periods of particularly intense dispute, bills to limit the Fed s statutory independence are submitted for deliberation and are sometimes the subject of serious debate. This threat, it is thought, balances the central bank s statutory independence against the need to hold it accountable for its actions. These mechanisms wUl work less powerfully in the case of the ECB. While its officials may be required to testify before the European Parliament, that institution has little power compared to, the U.S. Congress or Europe s national parliaments and hence may have little capacity to hold the ECB accountable. The ECB's independence could be modified, for example, not by the European Parliament but only by amending an international treaty, subject to veto by any of 12 signatories (Kenen 1992; Ciovannini 1992). In countries with a reputation for price stability, monetary policy is predicated not just on the central bank's statutory independence but on public support for its policies. The Bundesbank's commitment to price stability rests on the Cerman public's deep-seated aversion to inflation, for example. By embedding in the Maastricht Treat>' measures to insulate the ECB from political pressures at least as strong as those enjoyed by the Bundesbank, the drafters of the treaty responded to the worry that public support for price stability does not run as deeply in other EC countries and to the fear that Cermanv would veto any proposal for a European central bank responsive to the consequent political pressures. Without complementing those measures designed to insure independence with others to guarantee accountability, they may have created a situation where the ECB vdll find it difficult to maintain political support. C. Responsibility for Prudential Supervision Under e.xisting institutional arnin^ements, the domain of the authorities responsible for monetary policy and bank regulation is usually the same: monetary policy is set and banking systems are supervised by national officials. This may be viewed as an implicit contract: national central banks are empowered to extract seigniorage from national monetary systems in return for assuming the costs of running the national pa> ments system, bailing out banks, and in most cases bearing supervisory responsibility. In contrast, the Maastricht Treaty vests the European Central Bank with little regulatory responsibility. It is only to "contribute" to the smooth execution of policies by the competent national au1344 Journal of Economic Literature, Vol. XXXI (September 1993) thorities. The treaty erects barriers to the assumption of additional regulatory power hy the ECB. It is permitted to undertake only such tasks of prudential supervision as are conferred on it by the Council, which must itself act unanimously on a proposal from the European Commission and receive the assent of the European Parliament. The rationale for divorcing monetary policy from prudential supervision is that making the central bank responsible for financial stability may undermine its antiinflationary commitment. Imagine a time when inflation is accelerating and bank balance sheets are weak. As monetary authority, the ECB desires high interest rates; but as bank supervisor it wishes interest rates to be kept low to prevent the banks' debtors from defaulting on their obligations and further weakening bank balance sheets. Price stability may lose out. The solution to this problem in various EC countries—Cerniany for example— is to vest a national agency separate from the central bank with responsibilit\' for prudential supervision. The moral hazard problem for monetary policy is thereby attenuated, while the domains of monetary policy and financial regulation remain the same. But while the Maastricht Treaty proposes to centralize monetary policy at the EC level, beyond encouraging the ECB to "coordinate the policies ot the responsible national authorities ' it makes no provision for shifting bank regulation from member states to the Community. Decentralizing bank regulation within a monetary union may be problematic. First there is the danger of competitive deregulation. European banks have traditionally enjoyed a favored position in their home markets. The SEA will intensify competition be-tween them and allow intermediaries to better exploit economies of scale and scope, ultimately driving some banks out of business. *^ National authorities will be pressured to extend regulatory advantages to domestic banks, as their profits still accrue primarily to domestic shareholders who are domestic residents. But many of the < osts of competitive deregulation, in the form of financial instability, will be incurred by the Community as a whole. The risk of competitive deregulation is addressed by the 1988 Basle Accord, negotiated under the auspices of the Bank for International Settlements, which seeks to prevent the competitive reduction of capital ratios by encouraging the adoption of uniform risk-weighted capital requirements. It provides the basis for the EC's Directives on Solvency Ratios and Own Funds, which are similarly intended to address competitive deregulation problems (Ethan Kapstein 1991). Although capital requirements and liquidity ratios will be standardized in principle, enforcement will still take place at the national level. This leaves open the possibility that rules will be applied in diflPerent jurisdictions with varying degrees of stringency. Lax enforcement could therefore reintroduce all the problems of different regulatory standards. A second problem with decentralizing responsibility for bank regulation is that market integration will blur the borders between national systems. The more banks operate in several European countries, the less clear it will l>e which national authority is responsible for oversight. (Inadequate oversight of the failed * Us. data do not .show much evidence of Cionomies of scale in bankinf^. Tins has not deterred l",iiropean observers from empha.sizing the prospect o) further consolidation following coiiipletion of the single niark( t. For discussion, see Pierre-Andre Chiappori et al. (1991). Eichengreen: European Monetary Ihiification 1345 Bank of Oedit and Commercial International, registered in Luxembourg, may he taken to illustrate this point.) The EC's Second Banking Directive specifies that banks should be supervised by their home countries, although host countries are responsible for liquidity standards. The home-country principle applies only to branches of foreign banks, however, not to subsidiaries separately incorporated under the laws of ht).st countries. Mort'()\er, foreign branches will continue to be covered by host-country deposit insurance schemes. This means that, with the growth of foreii;^!! branching, supervision and deposit protection will become increasingly decoupled: those running deposit-insurance schemes (whose coverage varies enormously across EC nations) and charging the banks insurance premia will lack supervisory authority. Greater centralization of regulatory functions would avoid such confusions. This is not an argument for making the ECB responsible for bank regulation. Rather, it points to the need to vest some Community-level institution with this responsibility. The Maastricht Treaty includes no such provision. Placing responsibility for monetary policy and financial oversight in two separate Community-level institutions may be more problematic when one turns from banking to the payments system, the mechanism enabhng banks to make payments to and receive payments from other banks by using, accounts held at the central bank. (For a discussion of the ]>ayinents system issues raised by EMU, see David F"olkerts-Landau and Peter Garber 1992.) Traditionally, banks have been permitted to run up large overdrafts during the day and to settle at the close of business. When a bank with substantial overdrafts is unable to settle, other banks owed those overdnifts. finding themselves short of liquidity, may encounter difficult>' in setthng their accounts iis well. A cascade of defaults can bring the entire system crashing down. One solution is for the national central bank to stand behind all of the national system's payments, incurring a cost if a bank fails but preventing systematic collapse. This is the policy of the U.S. Federal Reserve System, for example. A fee for overdrafts can be charged to prevent moral hazard from resulting. Such arrangements would be difficult for Europe's national central banks to manage as national payments systems are linked together in an EC-wide payments mechanism. Imagine that British banks incur overdrafts to be settled with creditor banks in France at the end of the day. If the British banks fail, then the French financial system may be put at risk. Will the Bank of England or the Bank of France stand behind the overdrafts of the British bank? To which institution should fees for overdrafts be paid? This problem can be solved b\' giving a Community-level institution with liquidity- creating powers, plausibly the ECB, more prominent responsibility for the operation of the payments system. But while stating that the ECB should concern itself with the operation of the payments system, the treaty does not specify what it should do when that responsibility conflicts with the priority attached to price stability. ^^ •^^ In addition, there is the argument that a central bank which might guarantee the (i\erdrafts of commercial banks would be unwilhng to do so unless it also possessed powers of surveillance and regulation. See Eichengreen (1992a). Another solution would be a real-time uross settlement system, under whiih payments are sctllitl in full throughout the day rather than netted against each other and settled at the end of the day. The central bank might still provide temporary liquidity, hut banks would have to provide collateral before their settlement account was overdrawn. Britain and France are contemplating such systems. 1346 Journal of Economic Literature, Vol. XXXI (September 1993) 7. Who Will Participate? A. The Maastrieht Preconditions Assuming monetary union is desired, who should be entitled to join? Insofar as the eflBciency advantages of a common currency are an increasing function of the number of countries adopting it (Nobuhiro Ki\otaki and Randall Wright 1989), it is desirable that all EC countries participate. At the same time, EC policy makers worry that admitting countries whose monetary and fiscal performance is very difiFerent from that of the rest of the Community will destabilize the monetary union and subject the ECB to inflationary pressure. This fear led the framers of the Maastricht Treaty to specify four preconditions for participating in the monetary union (sometimes referred to as the 'convergence criteria"). Each of them can be (juestioned on economic grounds (David Begg et al. 1991; Buiter, Giancarlo Corsetti, and Nouriel Roubini 1993; Eichengreen 1992a). The first precondition is that a country s inflation rate should converge to a level not too far above that of the Community's low inflation countries. .Specifically, the average rate of CPI inflation over the preceding 12 months must not exceed the inflation rates of the three lowest-inflation member states In' more than l'/2 percentage points. The logic for this condition presumably follows from the fact that, under monetary union, member states will have to run very similar inflation rates. This point should not be overdrawn; Stephen Poloz (1990) shows for Canada, as do Eichengreen (1992c) for the U.S. and De Crauwe and VViin Vanhaverbeke (1991) for Cermany, that the inflation rates of states joined together in a monetar)- union can var\ by as much as IVz percent a year though not persistently in one direction. Leaving aside the appropriateness of the specific threshold, the notion that adjustment to a common inflation rate should precede monetary union may rest on the assumption that wages and other costs art- sufficiently inertial to prevent a rapid reduction in inflation at the time of EMU; unless inflation rates converge earlier, producers will face severe competitiveness problems in the earl\ phases of the union. Bini-Smaghi and Paolo Del Ciovane (1992) show that if the ECB adopts a restrictive monetary policy to promote price stability and counter thtinflationary shock, output will fall most sharply in low inflation countries (which may explain why such countries have been particularly concerned to see the adoption of inflation preconditions). Yet the last occasion on which one would expect to observe inertia in nominal variables is when the monetary policy process is undergoing a fundamental "change in regime" (Thomas Sargent 1986). Knowing that the "process" generating monetary policies has been changed by the establishment of a new monetary authorit\ with different rights and responsibilities, individuals will have every reason to revise their e.xpectations and firms and trade unions to renegotiate their contracts. The second precondition specified in the treaty is that nominal exchange rates be stabilized. Qualifying countries must have maintained their exchange rates within the normal EMS fluctuation bands for two > ears prior to entering the monetary union. This condition might be defended in terms of its contribution to the reputations of the participating governments for valuing and defending their exchange rate commitments. It would rule out a last minute realignment, under which countries whose inflation rates had been relatively high and were otherwise suffering competitive difficulties would devalue so as to enter at a more appropriate parity (Begg et al. 1991; Kenneth Eichengreen: European Monetary Unification 1347 Froot and Kenneth Rogoff 19911. It is not clear what entering the monetary union with a severe competitiveness problem would do to the credibility of a government's commitment to participate permanentl\. In the same way that a record of exchange rate stability should strengthen the credibility of a government's commitment to monetary union, entering with a competitiveness problem suffi< ient to provoke domestic opposition mif^ht weaken the credibility of that commitment significantly. The third precondition specified by the Maastricht Treaty is convergence of interest rates. A qualifying country's longterm interest rates over the preceding year must have been no more than two percentage points above those of the three best performing member states in terms of inflation. If exchange rates are credibly fixed, the only reason interest rates will vary significantly is sovereign default risk.'*" This condition can be rationalized, therefore, if risk of default is a threat to monetary union. This brings us to the fourth precondition, having to do with debts and deficits. The treaty establishes "reference values" for fiscal policy to be met by (iualif>'ing countries. Budget deficits should be no larger than three percent of GDP and gross public debts no larger than 60 percent of GDP.^' In contrast to the other ^ Buiter, Corsetti, and Roubini (1993). This statement is consistent with the obsersation that annual inflation rates can difler by 114 percentage points or more in an economic and monetary union, because those re0onal differentials exhibit little persistence and hence do not affect long-term interest rates. See Manfred Neumann and von Hagen (1992). " Both figures were quite close to the corresponding Community averages when the Maastricht Treaty Wits negotiated. Buiter, Corsetti, and Roubini (1993) suggest that the three percent deficit rutio may have been selected to reflect the fact that the average public investment ratio in Europe is three percent, in conjunction with the Cerman policy rule that public borrowing to finance investment is permissible; similarly, a 60 percent debt ratio is that produced in the steady state when governments deficit finance three conditions, this one is subject to qualifications, reflecting a compromise, one suspects, between those who desired rigid fiscal ceilings and others who preferred flexibility for business-cyclerelated conditions. Excessive deficits will be said to exist only if the deficit ratio exceeds three percent and if in addition either it has not declined "substantially and continuously" to "close to" that level or it cannot be regarded as "exceptional and temporary and . . . close to" the three percent threshold. The debt ratio will be said to be excessive only if it exceeds 60 percent and if in addition it is not sufiRciently diminishing and approaching the 60 percent level at a satisfactory pace.'"*^ The rationale for these conditions is that admitting into the monetary union members displaying inadequate fiscal discipline will subject the ECB to pressure to purchase the debts of the lax countries, with infiationary consequences for the union as a whole. Governments which issue debt in excess of their capacity to service it might expose themselves to a "debt run," in which investors suddenly liquidate their holdings of that government's obligations (Alesina, Alessandro Prati, and Guido Tabellini 1990). The price of its bonds will plummet, and the EGB may feel obliged to purchase them to prevent the entire EG bond market from being demoralized (Giovaimini and Spaventa 1990; Emerson et al. 1990). This swap of money for bonds would fuel inflation. Because the inflationary costs of the bailout are borne three percent of CDP a year (assuming continuous compounding and a five percent annual rate of growth of nominal CDP). These authors also suggest a number of analytical shortcomings of these criteria—for example, that they are defined in terms of gross rather than net debt and thus fail to offset government liabilities against assets *^ Kenen (1992) provides a clear discussion of who will determine whether these conditions are met. 1348 Journal of Economic Literature, Vol. XXXI (September 1993) li\ all members of the monetars union, individual U(>\ t^rnments will have an incentive to run cxc cssi\ e deficits and issue excessive debt. Alternativelv, if the ECB is prohibited from supporting the market in the bonds of an over-indebted go\ernment experiencing a run, EMU may herald a new era of pervasive financial market instability. An objection to this rationale is that the market disciplines borrowers, as described above. As the debt to-income ratio rises, so does the required rate of return on public obligations, deterring excessive borrowing. If they fail to heed the rise in interest rates, governments may find themselves rationed out of the market. Still, confidence in the power of market discipline is not universal. (See for example Begg et al. 1991; and Gros and Thygesen 1992.) Yet neither is it clear that a collapse in the prices of a government s obligations will spill over to other issues, requiring intervention by the ECB. If investors are able to distinguish good from bad credit risks, there is no reason win the entire market should become demoralized. In the U.S.. municipalities like New York (."ity in the 1970s have run into difficulties and found their bonds downgraded without adversely affecting the market as a whole (Craham Bishop 1992). Consequences are limited to the jurisdiction running the excessive deficits, providing no motivation for a central bank bailout and no inflationary threat (Buiter and Kenneth Kletzer 1990). Just as there are no federal restrictions on the debts and deficits U.S. states can in(m, there would be no need for statutory restrictions on the deficits of meml)ers of the ^' Many U.S. states have in place statutiiiv or constitutional balanced-l)udsi"t rules and public debt limitations These. IIOWCMI. are self-imiwsed, .ind tlK'\ are generalK a legacy of 19th cenliii\ experience. Sec Eichengreen (forthcomiiii;). B. Prospects for Participation The treaty as drafted at Maastricht entitles the EC Heads of State or Covernment to initiate Stage III before January 1st, 1999 if and only if a majority of member countries satisfy the convergence criteria. (Recall that, if this condition is not met. Stage III will proceed on January 1st, 1999 even if only a minority of member countries qualify.) On the basis of data for 1992 (Table 3), it would appear that Stage III will commence before 1999 only under vei\' favorable circumstances or a liberal interpretation of the fiscal provisions of the treaty. As of 1991, only three of the 12 EC members clearly satisfied both fiscal conditions. One was the U.K., hardk a steadfast proponent of monetary union. Another was Luxembourg, not one of the major European economic powers. The third was France. V\ hile Germany appeared to \ iolate the three percent deficit limit, capital spending—which is exempt from the three percent limit— accounts for a relatively large share of the German federal budget, so little if any fiscal adjustment would have been re- (]uired. For Stage III to commence before 1999, three additional participants are required. .Although Denmark should havelittle trouble meeting the treaty s debt imd deficit limits, it obtained a waiver enabling it to opt out of the monetary union. The Netherlands, with larger deficits and a higher debt ratio, faces a more difficult task: for it to qualify, fiscal retrenchment would have to proceed without producing even a temporary recession. Spain and Portugal could qualify if their budget deficits were reduced to three percent of national product without interrupting growth. Monetary union before 1999 is conceivable, but only if polEichengreen: European Monetary Unification 1349 MAIN France Germany Italy United Kingdom^ Largest 4 countries^ Belgium Denmark'' Greece' Ireland Luxembourg Netherlands Portugal Spain Smallest 8 countries^ AllEC INDICATORS OF CI Consumer Price Inflation 1991 3.1 4.8 6.3 5.9 4.9 3.2 2.4 19.5 3.2 3.1 3.9 11.4 5.9 5.5 .5 1 1992 2.8 5.0 5.5 3.7 4.3 2.4 2.1 16.0 3.3 2.8 3.3 9.2 5.9 5.0 4.5 TABLE 3 ;)N\I:;II<:I:N< i; PKOBI.KMS General Government Deficit/GDP 1991 -2.1 -3.2 -10.2 -2.7 -4.4 -7.(i -2.3 -17.4 -2.8 1.5 -3.9 -6.7 -4.9 -5.5 -4.6 1992 -2.9 -3.2 -10.4 -6.3 -5.4 -6.8 -2.3 -13.8 -1.8 1.0 -4.0 -5.7 -5.1 -5.2 -.5.3 IN THE f •<>MMI-Nm- Gross Government DebtyGDP' 1991 47.1 41.7 103.5 34.4 55.2 134 4 66.7 115.5 98.0 6.2 79.6 65.3 44.7 75.5 59.4 1992 47.7 42.5 108.5 35.9 56.9 133.4 71.3 114.8 100.0 5.8 80.0 69.5 46.0 77.4 61.3 IN 1992 Term Interest Rates 1991 9.2 8.5 13.0 9.9 10.0 9.3 9.6 23.3 9.2 8.2 8.7 18.5 12.6 11.6 10.3 1992 8.7 7.8 13.5 9.0 9.6 8.8 8.9 21.5 10.0 7.8 8.0 16.5 12.6 11.1 9.9 Source: International Monetar>' Fund (1993, p. 21). ' Debt data are from national sources. They relate to the general government but may not be consistent with the definition agreed at Maastricht. - Debt on fiscal \ car basis. ^Average weighted by 1991 GDP shares ••The debt-GDP ratio would be below 60 percent if adjusted in line with the definition agreed at Maastricht. ' Long-term interest rate is twelve-month tifusiuA bill rate. icy in the relevant countries is directed primarily at this goal."" Barring a capital levy or other equally radical measures, the other EC countries— Belgium, Ireland, Italy, and Greece—have little prospect of reducing their debt ratios to 60 percent by 1999.'*^ •''' Monetary unification before 1999 is also conceivable if the Community is enlarged to include some of the EFTA countries and these IICVN- members qualify for early participation. On this possibility, see Bayoumi and Eichengreen (forthct)minK c) and the references cited therein. •" Because a capital levy would raise the interest rates required of government debt subsequently, it would not be efficacious as this date approaches because it would lead to violation of the interest-rate convergence condition. Only if it were determined that their debt ratios were "sufficiently diminishing and approaching the 60 percent level at a satisfactory pace" could the\' hope to be included. How would the EC Heads of State and Government regard a monetary union centered on Germany and France and including also Denmark, Luxembourg, Spain, Portugal, and the U.K.? Such a union would encompass four of the North-Central European countries usually thought to be prime candidates for participation but exclude two others (the Netherlands and Belgium) that have already taken steps to peg their currencies closely to the deutsche mark. It would 1350 Journal of Economic Literature, Vol. XXXI (September 1993) include three countrie.s (Spain, Portugal, and the U.K.) widely thought to be among the poorer candidates for EMU becau.se their economic structures and policif.s differ from those typical of the Community. Bayoumi and Eichengreen (forthcoming b) show, on the basis of cross-country correlations of supply and demand disturbances, that EC countries seem to divide into two groups, a V-ore" with highly correlated disturbances (Cermany, France, the Netherlands, Belgium, and Denmark), and a periphery" characterized by idiosyncratic disturbances (Spain, Portugal, the U.K., Italy, and (Irtece). These results imply that the composition of the monetary union that is likel\ to be delivered by the Maastricht preconditions is far from ideal on optimum currency area grounds. Even this scenario requires that exchange rate stability not be interrupted by a foreign-exchange market crisis, because two years of exchange rate stability form one of the Maastricht preconditions for EMU. As precisely such a crisis intervened in 1992, it is important to comprehend its origins in order to understand whether it could happen again. S. Future Prospects The crisis in the European Monetary System that erupted in September 1992 cast a pall over the prospects for European monetary unification. This section attempts to draw out the implications of that crisis for the transition to monetary union. Signs of crisis surfaced first at the fringes of the European Community. A domestic banking crisis and economic chaos in neighboring Russia forced Finland to devalue in the summer. This applied pressure to Sweden s balance of payments, because the two economies exported many similar products. To defend the krona, the Swedish Riksbank raised interest rates to stratospheric heights.^ The Italian lira and British pound were the first EMS currencies to be tested. The lira had been weak since the lieyinning of the summer; in earl> September it weakened significantly, and the Italian Covernment was forced on Sunday, September 13th, to devalue by seven percent. This failed to relieve the pressure on the lira and the pound, however. On September 17th, both countries suspended their membership in the ERM and allowed their currencies to float. Spain devalued In fi\e percent and reimposed capital controls. Speculation at^ainst the French franc was repelled onK by the extensive intervention of the Cerman Bundesbank. The markets tested the stabihty of still other currencies like the Irish punt and the Portuguese escudo, leading the first country to reinforce and the second to reimpose capital controls. Thus, by the end of September, much of tlif progress made since 1990 toward the removal of controls and the stabilization of exchange rates was reversed.''^ It is not hard to identify economic imbalances contributing to the trrisis. Although policies diverged less across EMS countries than in the early 1980s, significant differences remained in policy settings and inflation outcomes. Because their effects cumulated, even moderate inflation differentials could erode the * Its efl'ort to defend the krona was successful initially, although a sfictnd trisis later in the autumn proved fatal. Just why instability of non-EMS currencies spilled over to the EMS is not clear. For some time the curri-iiiies of Iwth Finland and Sweden had been "shadowing" the EMJi—in other words, they were pegged to EMS currencies Init did not enjoy the credit lines and other facilities that came with a formal link. At the least, their di£Rculties focuse% would be fa\'orably inclined toward an early start. The final option for the transition is throwing sand in the wheels of international finance. Requiring all institutions taking open positions in foreign exchange to make non-interest-bearing deposits with their central bank, a polic\ used previously b\' countries like Italy and Spain, would slow down adverse speculation (Eichengreen and Wyplosz 1993). The cost would be passed on to currenc> traders, discouraging one-way bets. Siicli measures could not permanently support weak currencies, but they could provide time to organize orderly realignments and therein ensure the survival of the EMS over the remainder of the transition. 9. Conclusion Aboli.shing national monies that in most cases have existed in Europe for a century or more is understandably controversial. Few symbols of national sovereignty are as powerful as coins and banknotes. The EC s Committee of Central Bank Covernors has sought to meet this point l>y considering whether to print a European currency with a common front and 12 distinct back sides featuring 12 national luminaries. Beyond the controversy over s> nibols, there is the concrete question of economic welfare: whether the benefits of monetary unification exceed the costs. I have argued here that tl\e empirical evidence on this point is inconclusive. The sacrifice of national monetary autonomy that conies with establishment of a single currency may involve a serious welfare loss, while direct savings on transactions costs are relatively small. Only if it can be argued that a single currency is a necessary concomitant of the single market, tlie benefits of which are likely to be substantial, can the case for a single European currency be made with confidence. There is no technical reason why a single currency is required to reap the benefits of the single market. In principle, factor- and product-market integration can proceed under floating exchange rates as well as under a common currency, this being the strategy pursued by North American policy makers. The problem with market integration under floating is one of political economy. Some 1354 Journal of Economic Literature, Vol. XXXI {September 1993) national industries will be competed out of business in the c;ourse of creating a truly integrated European market; the notion that integration will admit the chill winds of foreign competition into previously sheltered markets to the point that onl\' the most efficient producers survive is after all one of the central rationales for the single market program in the first place. But if national industries under pressure from the removal of trade barriers find their competitive position eroded further, even temporarily, by ^•apricious" exchange rate swdngs, resistance to the creation of the single market would intensify. It is thus for reasons of political economy, not economic efficiency, that monetary unification is a necessary corollary of factor- and productmarket integration. While market integration could in principle be accompanied by either floating rates or monetary union, the one alternative that is not viable is "fixed" exchange rates between distinct national currencies. Exchange rates exist only to be changed. No matter how earnestly a govemment asserts its commitment to fix its exchange rate, there remains the possibility that a change in govemment will lead to a change in policy and a devaluation. As Portes (1993, p. 2) puts it, " Permanently fixed exchange rates' is an oxymoron." Speculators cognizant of this fact face a one-way bet. In the absence of capital controls, who.se elimination is a necessary concomitant of the single market, they possess almost an infinitely elastic supply of resources with which to extract information by testing the government s resolve. These conclusions suggest that the extended transitional period foreseen by the Delors Committee is problematic. The September 1992 crisis in the European Monetary System exemplifies this fact. It suggests the need to accelerate the schedule set out in the Maastricht Treaty, and for additional measures, like deposit requirements, to regulate the operation of foreign exchange markets over the remainder of the transition. RKI-ERENCKS ALESINA, ALBERTO AND GRILU, VITTOHIO. "The European Central Bank: Reshaping Monetary PoHtics in Europe," in MATTHEW B. CANZONERI, XJTTORIO Grilli, and Paul R. Masson, eds. 1992, pp. 49-T7. ALESINA, AI.KERTO; PRATI, ALESSANDRO AND TABEI.- UNi, GUIDO. "PubUc Confidence and Debt Management: A Model and a Case Study of Italy," in Public debt mananement: Theory and history. Eds.: Ri DICER DoRNBi'.st:ii AND MARIO DRACHI. Cambridge: Cambridge U. Press, 1990, pp. 94- 118. BAYOUMI, TAMIN AND EICHENGREEN. BARRY. "Monetary and Exchange Rate Arrangements for NAFTA."/. Devel. Econ.. forthcoming a. "Shocking .Aspects of European Monetary Unification," in Adjustment and growth in the European Monetary Union. Eds.: FRANCI.SCO TORRES AND FRANCKSC o GiA\ AZzi. Cambridge: Cambridge U. Press, forthcoming b. . "Is There a Conflict Between K{; Enlargement and European Monetar\ Unification?" Greek Econ. Rev., forthcoming c. BAYOI'MI, TAMIM; GOLDSTEIN, MORRIS AND WOC:I.OM, GEOFFREY. "DO Credit Markets Disciphne Sovereign Borrowers? Evidence from U.S. States." Unpub. ms. Intemational .Monetary Fund, 1993. BAYOUMI, TAMIN AND MASSON, PAUL. "Fiscal Flows in the United States and Canada: Lessons for Monetary I'nion in Europe." Unpub. mss. International Monetary Fund, 1991. BEGG. DAVID ET AL. "European Monetary Union— The Macro Issues," in Monitoring European integration: The making of monetary union. London: Centre for Economic Policy Research, 1991, pp. :^66 BELDEN. SUSAN. "Policy Preferences of FOMC Members as Revealed by Dissenting Xotes." J. Money, Credit. Bankina. Nov. 1989, 21(4). pp. 432-41. BERTOLA, GUISEPPE. "Factor Flexibility. Uncertainty, and Exchange Rate Regimes," in M ABCELI.O D E CECCO AND ALBERTO GIOVANNIM, eds. 1989, pp. 95-119. BINI-SMACHI, LORENZO AND DEI . GIOVANE, PAOLO. "Convergence of Inflation, Prerequisite for EMU." Unpub. mss. Banca d'ltalia, 1992. BINI-SMACIII. LORENZO AND VORI, SILMA. "Rating the EC as an Optimum Currency Area: Is It Worse than the U.S.?" Unpub. mss. Banca ditalia, 1992. BISHOP, GRAHAM. "The EC's Public Debt Disease: Discipline with Credit Spreads and Cure with Price Stability," in Fiscal policy, taxation and the financial system in an increasingly integrated Europe. Eds.: DONALD E. FAIH AND CHRISTIAN DE BoissiEu. Dordrecht, the Netherlands: Kluwer Academic, 1992, pp. 207-34. BLANCHARO, OLIVIKR AND KATZ, LAWRENCE. "ReEichengreen: European Monetary Unification 1355 gjonal Evolutions." Brookinns Pap. Econ. Act.. 1992. 1, pp. 1-61. BLAN< MARU. OLIVIER AND MI ET. PIERRE AUMN "Competitiveness Through Disinflation: An Assessment of the French Macroeconomic Strategy," Econ. Policy. 1993. 16. pp. 11-56. BLANCHARD. OLIVIER AND QUAH. DANNY. "The Dynamic Effects of Aggregate Demand and Supply Disturbances." .\mer. Econ. Rci\. Sept. 1989, 7.9(4), pp. 655-73. BiiiNO, Mil:ii\EL AM) SACHS, JEFFREY D . The ccoiii'iitic\ of worlduHde Stagflation. Cambridge: Harvard U. Press, 1985. BRYANT, RALPH i-:I AL., eds. Empirical macroeconomics for interdependent economies. Washington, D.C: BrookiiiKs Institution. 1988. BLITER. Wii.i.EM. "A Guide to Public Sector Debt and Deficits." Econ. Policy. 1985, 1. pp. 4-69. BITTER. VVILLEM; COHSETTI. GIANCAHLO AND ROUuiNi, NOURIEL. "Excessive Deficits: Sense and N(insense in the Treaty of Maastricht," Econ. Policy. 1993. 16, pp. .57-100. BUITER. Win LM \ND KLETZER. KENNETH. "Reflections on the Fiscal Implications of a Common CUirreiic\." Cfntre for Economic Policy Research Discussion Paper no. 41H, 1990. BULLETIN DEK EUROPAISCHER GEMEINSCHAFTEN/ KoMMissioN, No. 5/1992. CAMERON, DAVID. "The Maastricht Agreement on Economic and Monetary Union: Initiation, Negotiation, Implications." Unpub. mss. Ya]<' U., 1992. C^AN/.oNERi, MATTHEW B. "Monetary Policy Games and the Role of Private Information," Amer. Econ. Ret.. Dec. 1985, 7.5(5), pp. 105&-70. CANZONERI, MATTHEW B.; GWLU. VITRIKIO AND NtAssoN. PAUL R.. eds. F.sinblishinga central bank: hsui'.s in Euro})e and lessons from the US. Cambridge: C;ambridi;e U. Press, 1992. CASELUA. ALESSANDRA AND FEINSTKIN, JONATHAN. "Management of a Common Currency." in MARci- i.LO D E CECXO AND ALBERTO GIOVANNINI, eds. 1989, pp. 131-56. CHIAPPORI, PIERRF-ANPRE ETAL. "The Microeconomics of Monetary Union," in Monitoring European integration: The making of monetary union. London: Centre for Economic Policy Research, 1991, pp. 69-120. COHEN, BENJAMIN J. "Bryond EMU: The Problem of Sustainability." Econ. Polit., Jul> 1993. 5(2). pp. 187-203. COHEN, DANIEL AND WVPLOSZ, CHARLES. "The European Monetar\' Union: .\n Agnostic Evaluation: ( I PR Dlscus.sion paper no. 306. 1989. COLLINS, SI SAN M. AND GIAVAZZI, FRANCESCO. "Attitudes toward Inflation and the Viability of Fixed Exchange Rates: Evidence from the EMS," in A retrospective on the Bretton Woods system. Eds.: Mil HAEL D. BoRDO AND BARRY Ell HENCREEN. Chicago: U. of Chicago Press. 1993. pp. 547-86. COMMITTEE OF GOVERNORS OF TIII: CENTRAL BANKS OF THE MEMBER STATES OF THE EUROPEAN ECONOMIC COMMUNITY. Annual Report 1992. Luxembourg: Office for Official Publications of the European Communities, 1993. COMMITTEE FOR THE STUDY OF ECONOMIC AND MONETARY UNION. Report. Luxembourg: Office for Official Publications of the European Communities. 1989. CONFERENCE OF THE REPRESENTATIVES OF THE GOVERNMENTS OF THE MEMBER STATES. "Amendments to the EEC Treaty—Economic and Monetary Union—As Agreed in the European Council of Maastricht on 10 December 1991." Unpub. draft. 1991. COOPER, RICHARD N. "Will an EC Currency Harm Outsiders?" Orbis. Fall 1992a, pp. 517-29. "Whither Europe?" Yale Rev.. 1992b. 80, pp. 10-17. CusHMAN, DAVID O. "Exchange-Rate Uncertainty and Foreign Direct Investment in the United States," WeltuHrtsch. Arch., 1988, 124(2), pp. 322- 36. DAVIS, EVAN; KAY, JOHN AND RIDCE, MICHAEL. "The Economics of European Integration,"/. Econ. Lit.. forthcoming. DE CECCO, MARCELLO AND GIOVANNINI, ALBERTO, eds. A European central bank? Cambridge: Cambridge U. Press. 1989. DE GBAI'WK. PAUL AND VANHANERBEKI., WIM. "IS Europe an Optimum Currency Area? Evidence from Regional Data." CEPR Discussion Paper No 555, 1991. DEMEKAS, DIMITRIOS G. ET AL. "The Effects of the Common Agricultural Policy of the European Community: A Survey of the Literature,"/. Common Market Stud., Dec. 1988. 27(2), pp. 11*^5. EicHENCREEN, BARRY. "One Money for Europe? Lessons from the U.S. Currency Union," Econ. Policy. Apr. 1990, 70, pp. 117-«7. "Should the Maastricht Treaty Be Saved?" Princeton studies in international finance, no. 74, Princeton U.: International Finance Section. Dept. of Economics, Dec. 1992a. "Designing a Central Bank for Europe: A Cautionary Tale from the Early Years of the Federal Reserve System," in MATTHEW B. CANZONERI, VnTOfuoGiuLLi, ANDPAULR. MAS.soN.eds. 1992b, pp. 13-40. "Is Europe aii Optimum Currency Area? in The European community after 1992: Perspectives from the outside. Eds.: SILVIO BORNER AND HERBERT GRUBEL. London: Macmillan, 1992c, pp. 138-61. "Labor Markets and European Monetary Unification," in Policy issues in the operation of currency unions. Eds.: PAUL MASSON AND MARK TAYLOR. Cambridge U. Press, 1993. "The Political Economv of Fiscal Policy Under EMU," Stato e Mcrcato. forthcoming. EIC:HENCREEN, BARRY AND WYPLOSZ, CHARLI.^ "The Unstable EMS," Brookings Pap. Econ. Act., forthcoming. EMERSON, MICHAEL ETAL. "Tlie Economics of 1992." Commission of the European Communities. European Economy. Mar. 1988, 35, special issue. "One Market, One Money." Commission of the European Communities. European Economy. Oct. 1990, 44. special issue. 1356 Journal of Economic Literature, Vol. XXXI (September 1993) European Commission. Report of the study group on the role of public finance in European integration. Brus.sels: European Commission. 1977. Fi.ANDRKAr. MAHC:. "On the Inflationary Bias of Common Currencies; Evidence from the Latin Union Puzzle." Europ. Econ. Rei\. 1993. 47. pp. 501-06. FLOOD. ROBKHTP.; ROSE, ANDREW K. AND MATHIIsoN. DONALD J. "An Empirical Exploration of Exchange- Kate Target-Zones," Carnegie Rochester Conf Ser. Public Policy. Autumn 1991. .35. pp. 7-66. Fill KEHTS-LANDAU, DAVID AND GARBEH, PBTEB. "The ECB: A Bank or a Monetar> Policy Rule?" in MATTHEW B. CANZONERI, VOTORIO GRILLI. AND PALI. R. MASSON, eds m i . pp. 86-110. FRANKEL. JEFFREY A. "Monetary Regime Choices for a Semi-Open Economy." Unpub. mss. U. ot C a l ifornia at Berkeley. 1992. FRENKEL, JACOB A, ANDCOLUMIIN, MOBRIS. "Monetar\ PolicN in an Emerging European Economic and Monetar\ Union," Int. Monet. Fund Staff Pap.. June 1991. 38(2). pp. 356-73. FROOT, KENNETH AND ROCOFF, KENNETH. "The EMS, the EMU, and the Transition to a Common (.urrency," iVBEft Macroeconomics Annual. 1991. 6, pp. 269-317. VON FL'RSTENBER(.. GK<)R<;K M. AND TEOLIS. DAVID P. "Growth and Income Distribution Benefits of North American Monetary Union for Mexico." Unpub. mss. Indiana U.. 1992. GARRETT. GEOFFREY. "The Politics of Maastricht," Econ. Polit. July 1993, .5(2), pp. 105-24. GiAVAZZi. FRANCESCO AND GIOVANNINI, ALBERTO. Limiting exchange rate flexibility: The European monetary system. Cambridge: MIT Press, 1989. GiAV.\zzi, FRANCESCO AND SPAVENTA, LLU.I. "The New EMS," in The European monetary system in the 1990s. Eds.: PAUL DE GRAUWE AND LUCAS PAI'ADEMOS. London: Longmans, 1990, pp. 65-85. GILDEA. JOHN A. "Explaining FOMC Members Votes," in The political economy of American monetary policy. Ed.: Timsrvs MAYER. Cambridge: Cambridge U. Press. 1990, pp. 211-27. Gio\ ANMN[. ALBERTO. "European Monetar>' Refonn: Progress and Prospects," Brookings Pap. Econ. Act.. 1990a, 2, pp. 217-74. "The Transition to European Monetary- Union," E.ssays in international finance. No. 178, Princeton U.: International Finance Section, Dept, of Economics, Nov. 19901). . "Central Banking in a Monetary Union: Reflections on the Proposed Statute of the European Central Bank." Centre for Economic Policy Research Occasional Paper No. 9. Aug. 1992. Gi(n ANNiNi, ALBERTO AND SPAVKNTA, LUICI. "Fiscal Rules in the European Monetary Union: A No- Entr\' Clause." CEPR Discussion Paper No. 516, 1990: GiJCK, REL'VKN AND ikr i i i i v s oN. MiciiAF.L. "Fiscal Policy in Monetary Unions: Implications for Euroix-." Open Economies Rev.. 1993, 4{l). pp. 39- 65. GOLDSTEIN, MORRIS AND WOGLOM, GEOFFREY. "Market- Based Fiscal Discipline in Monetary Unions: Evidence from the U.S. Municipal Bond Market," in MATTHEW B. CANZONEKI, NIITORIO GHILU, AND PAUL R. MASSON, eds., 1992, pp. 228-70. GORDON, JAMES. "TTie Structural Funds and the 1992 Program in the European Community." Unpub. mss. International Monetary Fund, 1991. GRIER, KEVIN. "The Political Econonn of Monetary Policy." Unpub. Ph.D. dissertation. Washington U., 1984. GRILLI, VrrroRio. "Seigniorage in Europe," in .MARCELLO D E CKCCO AND ALIIERTO GIOVANNINI eds. 1989, pp. 53-79. GROS, DANIEL AND THYGESEN, NIEUS. European nwnetary integration: From the European monetary system to European monetary union. Ixindon: Longman. 1992. \oN HA<.I:N. JURCI:N. "Fiscal Arrangements in a Monetary Union: Evidence from the I'S." in Fiscal policy, taxation and the financial system in an increasingly integrated Europe. Eds.: DONALD E. FAIR AND CHRISTIAN DE BOI.SSIEU. Dordrecht, The Netherlands: Kluwer Academic Pub.. 1992, pp. 337-60. HoLTFREKK II. CARL-LI Dwic. "The Monetary Unification PnK'ess in 19th-century Germany: Relevance and Lessons for Europe Today," in MARCELLO D E CEC CO AND ALBERTO GK>\ ANNINI, eds., 1989, pp. 216-41. Hoopf.n. PETLH AND KOHLHACEN, STEVEN W. "The EfiFect of Exchange Rate Uncertainty on the Prices and Volume of international Trade,"/. Int. Econ.. Nov. 1978, 8(4), pp. 483-511. HORN, HENRIK AND PEBSSON, TORSTEN. "Exchange Rate Policy, Wage Formation and Credibility," Europ. Econ. Rev.. Oct. 1988, ,32(8). pp. 1621- 36. IN<;RAM. JAMES. "State and Regional Payments Meclianisms," Quart. J. Econ.. Nov. 1959, 73. pp. 619- 32. INTERNATIONAL MONETARY FUND. "Exchange Rate Volatility and World Trade." IMF Occasional Paper 28, Washington. DC, International Monetary Fund, 1984. INTERNATIONAL M<.>NETAKY FUND. World economic outlook: Interim as.iessinent. Washington, DC: International Monetary Fund, Jan. 1993. IPSEN, EKIK. "EC Grouses Over British Jobs." International Herald Tribune, 5 Feb. 1993. p. 12. ISHIYAMA, VOSHIHIDE, "The Theory of Optimum Currency Areas: A Survey." Int. Monet. Fund Staff Pap.. July 1975, 22(2), pp. 344-83. ITALIANER, .\I.EXANDERANDPISANI-FERRY, JEAN. "Regional Stal>ilization Properties of Fiscal Arrangements: What U-ssons for the Community?" Unpub, mss. Paris and Brussels: Commission of the European Communities and CEPII. 1992. KAP.STF.IN. ETHAN B. "Supervising International Banks: Origins and Implications of the Basle Accord," Essays in international finance. No. 185, Princeton U.: International Finance Section. Dept. of Economics, 1991. KENEN, PETER B. "The Theory of Optimum Currency Areas: An Eclectic \ icw." in Monetary problems Eichengreen: European Monetary Unification 1357 (if the international econmny. Eds.: ROBERT A Mi'NDKi.L AND Ai.E.\.\NOER K. SwoBODA. U. ofChicago Press. 1969. pp. 41-60. "Exchange Rate Variability: Measnrcnnent and Implications." Unpub. mss. Princeton U.. 1979. EMI' after Maastricht. \\a.shinRton. DC: Group of Thirty, 1992. KKNNEOV. r.i.LEN. The Bundesbank: Germany's central bank in the international monetary system. London: Royal Institute of International Affairs. 1991. KlYOTAM, NoBl'IIIKd AND WKIC.HT. RANDAI 1 . "On Money as a Medium of Exchange."/. Polil. Econ.. Aug. i989. 97(4). pp. 827-54. KRI t;MA.N, PAIL. 'Target Zones and Exchan^i- Rate Dynamics," Quart. ]. Econ.. Aug. 1991, 106(3). pp. 669-82. "Integration. Specialization and Regional Growth: Notes on 1992. EMI' and .Stabilization." in Adjustment and growth in the European Monetary Union. Eds : FRAN< iscu TOHRK^ AND FIIANci. s< n)GiAV\/zi. Cambridge: Gambridge U. Press, forthcoming. MAI.SEI., SHERMAN J. Managing the dollar: .\n imiders t icu' of the Federal Resertc Board. London: W. W. Norton, 1980. McKiNNON, RoNALi) "Optimum Currency Art-as." Amer. Econ. Rec. Sept. 1963, 5.i(4). pp. 717-25 Ml l.Koj), DARRYL AND WELCH, JOHN H. "The Gase for a Peso Target Zone, Business Mexico. Nov. 1991a. pp. 19-22. . "Free Trade and the Peso Unpub. mss. Fordham U. and Federal Reser\<.- Bank of Dallas, 1991b. MKADE. JAMKS "The Belgium-Luxembourg Economic Union, 1921-1939," Essays in international finance. No. 25, Princeton U.: International Finance Section, Dept. of Economics, 1956. MONETARY GoMMrrrii OF THE EIIROPHAN COMMUNITIES. Compendium of community monetary ti'xts. Lu.xembourg: Office of Official Publications of l\\v European Gommunitie.s. 1986. MoRsiNK. RoBKUTAND MoLi I . WiLi.EM. "Direct Investment and Monetary Integration." Europ. Econ.. 1991, .special edition no. 1. pp. 36-55. MiMiKi.L, ROBERT. "A Theory of Optimum Gurrency Areas," Amer. Econ. Rev.. Sept. 1961, .5;(4). pp. 657-65. NEUMANN, MAMBEOANDVON HAGKN, Ji RCEN. "Germany," in Handlfook of comparative economic policies. Vol. 3. Eds.: MiiiiELK FRATIANNI ANO DOMINI* K SALVATOKK. Westport, GT: CJreenwood Press, 1992, pp. 299-334. NUGENT. NI.II.I . The goicrniiwtit and politics of the European community. London: Macmillan, 1989. OBSTKELD. MAI HICK. "Destabilizing Effects of Exchange Ratf Escape Glauses." NBER Working Paper No. .3606, 1992. ()I«:ANI/MU)\ FOHE<:ONOMIC('OOPKHATK)N ANI>DIvKi. opMKvr. Flexibility in the labour market. Paris: OEGD, 1986. . Economies in transition: Slnirtural adjustment in OECD countries. Paris: OEGD. 1989. D i , Iii.i.KN ANij SAJ IIS. JEI I HI 1 D. "Macroeconomic PolicN Goordination among the Industrial Economies," Brookings Pap. Ecori. Act.. 1984, 1. pp. 1-64. VAN DER PLC)I:<:. KHKDK.HHK. "ManofcoMoinic Policy Goordination lDuring the Xiu-ious Phases of Economic and Monetary Union." Gominission of the European Gommunities. Europ. Economy. 1991, special issue, pp. 136-64. POLOZ, STEPHI N S. "Real Exchange Rate Adjustiiu'iit Between Regions in a Gommon Gurrency Area." Unpub. mss. Bank of Gaiiada, 1990. PORTES, RK IIAHD liMS and EMU After the Fall," Thr World Economy, 1993. 16. pp. l-Ki. PI'<:KETT. RICH MID S. "Federal Open Market Gominittee Structure and Decisions."/, A/one/. Econ., July 1984, I4i\}. pp. 97-104. RoMEK. GHHIVITNA D . AND ROMER, DA\ ID H. "Does Monetary Policy Matter? A New Test in the Spirit of Friedman and Schwartz," NBER macroeconomics annual. 1989, /. pp. 121-70. SALA-1-MARTIN, .\A\ IER A.ND SACHS. jErKiiEv. "Kiscal Federalism and Optimum Currency Areas: F,\idence for Europe from the United Slates.' in MATTHEW B. GANZONEKI. VrrroRio GIULU, X.ND PAIII. R. MAS.SON. eds., 1992, pp. 195-220. SANDHOLT/, WAYNE. "Monetary Politics and Maastricht," Int. Organ.. Winter 1993, 47{l). pp. 1-39. SARC.ENT, THOMAS. Rational expectiitions and inflation. NY: Harper & Row, 1986. SINN, GERLINDE \NDSINN, HANS-WI.HNKB. Kaltstart: Volkswirtschaftliche Aspekte der deutschen Vcreinigung. Tubingen: J. G. B. Mohr, 1991. TiivcEsEN. Nil I.s. "International Goordination of Monetary Policies with Special Reference to the European Gommunity." in S'eiv approaches in monetary policy. Eds.: JOHN E. WADSWOHTH AND FRANCOIS LEONARD I>I JL'VIGNY. Sijthoff and Noordhoff: Alpen a;m dcr Rijn, 1979, pp. 205- 21. . "Decentralization and Accountabilit\ within the Clentral Bank: .\ny Lessons from the U.S. Experience for the Potential Organization of a Enropean Gentral Banking Institution?" in The ECl' and European monetary integratUm. lOds: PAIH. DE GK.M WK ANDTHEO PEETKKS. London: Maeinillan, 1989, pp. 91-114. UNI.KHEH. HOR.ST ET AL. The European monetary system: Deveh>]mient and prospects. Oetasional Paper 7.3, Washington, DG: International Monetary Fund, 1990. UNPI^D STATES GONIJRESS. "Hearings on Financial Structure and Regulation Before the Sulxommittee on Financial Institutions of the Senate (Committee of Banking, Housing and Urban .Xffairs." 93rd Gongiess. 1st Session, [97.3. WEBER, .\.\EL. l.MU and .\s\iiiinetries and .Xiijustment Problems in the EMS: Some Empirical r.\idence." Gi-ntre for Economic Policy Research Discussion Paper No. 448. 1990. WS l'l.osz, Gii.\RLi:.s. 'Monetary Union and Fiscal Polic\ Di.scipline. " Gommission of the I'uropean Communities. Europ. Economy. 1991. special issue, pp. 165-84.