In June the Council of Europe established a committee, headed by the President of the European Commission, to study the means of achieving economic and monetary union within Europe. This initiative has caught observers by surprise: it does not seem to arise from any immediate threat to the European Monetary System (EMS), which has been able to withstand the sizeable international financial shocks of the early 1980s. It may instead reflect concern with the rapid evolution and integration of the economies and capital markets of Europe.
Previous research on the EMS has concentrated on analysing the behaviour of the System and evaluating its performance, but has paid less attention to its possible evolution and the role that it might play in monetary unification and in the establishment of a European central bank. A June conference on `Monetary Regimes and Monetary Institutions: Issues and Perspectives in Europe', planned with some prescience, did address this question. The conference was the second in an annual series to be held at Castelgandolfo and organized by the Italian Macroeconomic Policy Group and CEPR, with financial assistance from Euromobiliare, Prometeia and the Centro Europa Ricerche. The conference examined the question of a European monetary authority from the perspective of both economic analysis and economic history.

There are at least two separate arguments for the creation of a European central bank. Some observers view the process of monetary unification as desirable in itself, arguing that Europe constitutes what economic theory terms an `optimum currency area'. They note that the EMS has not brought about such unification spontaneously, and that further progress towards unification requires specific political initiatives directed towards this goal.
The second argument relates directly to the way monetary policy has operated during the EMS years. Differences in inflation rates across member countries have not been eliminated, and this has necessitated periodic exchange rate realignments to avoid divergences in relative prices. The `weak-currency' countries have preserved stability in their domestic financial markets by resorting to capital controls: such controls have been essential for the smooth working of the EMS. The complete liberalization of capital markets that is planned for 1992 would then seriously destabilize domestic financial markets, unless market participants perceive countries' commitment to a European monetary union as a credible one. According to this argument, the only credible commitment to a monetary union is monetary union itself.

The EMS: A Dead End?
The EMS was viewed by its creators in 1978 as a transitional step towards monetary unification. The System displayed from the outset some features of an embyronic central bank: unlike the Bretton Woods System and the Snake, the EMS was characterized by a special `money', the European Currency Unit (ECU) and by an institution to control the issue of this money, the European Monetary Fund (EMF).
The EMS has failed, however, to provide the fresh impetus to the process of European union that its creators desired. One symptom of this failure is the lack of any substantial role played by the European Monetary Fund and the ECU. The former has remained just an account at the Bank of International Settlements, used for the clearing of bilateral credits, while the ECU has failed to assume the functions of a European money and has been used mainly as an accounting unit.
Was the failed promise of the EMS due to a defect in its design? Analysis of the regulations governing the EMS suggest that its institutions were clearly not designed to infringe the sovereignty of individual countries' monetary authorities, as would be needed to achieve monetary unification. The rules governing the use of the ECU and the European Monetary Fund, as well as the rules governing intervention, were loose enough to allow independent manoeuvre by individual countries. For example, compulsory intervention in the foreign exchange market, required by the System when two currencies reach the limits of the bilateral fluctuation bands, does not impose any constraint on monetary policies, since countries can freely sterilize reserve flows.

The Costs and Benefits of Monetary Unification
The rules governing the operation of the EMS have carefully excluded any features that could have infringed national monetary sovereignty: this has effectively prevented any further evolution of the System towards a monetary union. But is such a union desirable?
This question leads directly to the issues associated with the theory of optimum currency areas. This theory, first formulated by Robert Mundell, considers the costs and benefits of common currencies and by implication the relative desirability of fixed versus flexible exchange rates. If a system of fixed but adjustable exchange rates will not survive a full liberalization of European capital markets, the theory of optimum currency areas can provide a guide to help determine whether monetary unification might be superior to a return to flexible exchange rates among European countries.
A common currency lowers transactions costs and so is beneficial because it provides a common medium of exchange among countries. Casual empiricism suggests that the benefits might be large: the opportunity cost of holding cash is large, and yet banknotes are an important means of payment in all industrial countries. Similarly, international traders and producers appear to believe that the benefits of a common currency are very high. The system of Monetary Compensatory Amounts in European agriculture provides a good illustration of producers' aversion to exchange rate changes. One explanation for the existence of these Monetary Compensatory Amounts is the concern of agricultural producers with fluctuations in the price of their output induced by exchange rate movements; these are automatically reflected in output prices by the Common Agricultural Policy, which fixes Europe-wide prices in a common unit of account. Unfortunately, even though there are reasons to believe that the benefits of a common currency are significant, we lack economic models that provide a convincing account of the welfare effects of money in modern economies or a sound analysis of the costs that arise if trading nations use different currencies. Similarly, the econometric evidence on the effects of exchange rate uncertainty on international trade is scant.
Mundell's analysis of optimum currency areas identified the costs of common currencies as being those of unemployment and inflation caused by `countryspecific' shocks, which are offset neither by movements of factors across frontiers nor by exchange rate changes. Mundell's analysis does, however, rely on the assumption of downward wage and price rigidity, which prevents adjustment to demand shocks in the goods and labour markets and gives exchange rate realignments the power to affect relative prices. Three of the papers presented to the conference re-examined the theory of optimum currency areas from new perspectives. These papers focused on three issues: the determinants of factor mobility across countries, the costs of relinquishing monetary independence, and the coordination problems of central banks in a monetary union.
In his paper, `Factor Flexibility, Uncertainty and Exchange Rates', Giuseppe Bertola (MIT) re-examin ed an important element of the theory of optimum currency areas. Mundell had argued that downward rigidity of nominal prices would favour exchange rate flexibility between regions subject to different real shocks, if factors of production were immobile. Bertola proposed a new and potentially far-reaching theory of factor mobility itself. Previous analyses have relied on ad hoc assumptions concerning the existence of adjustment costs in order to explain the movement of factors. Bertola argued that uncertainty is likely to be a major determinant of the international mobility of factors and should be taken into account explicitly in the analysis. He noted that recent advances in the microeconomic theory of costly reallocation under uncertainty have shown that the more uncertain the environment, the more reluctant agents will be to undertake adjustments. He attempted to integrate this theory with macroeconomic analysis in order to provide a framework for assessing the costs and benefits of nominal exchange rate stability in a common currency area.
Bertola argued that option pricing theory provided a useful starting point for the analysis of adjustment under uncertainty. In financial markets the holder of a call option has the right to buy a specified asset at a price specified in advance. Option pricing theory suggests that increased uncertainty about the price of the underlying asset will increase the value of the option. Greater uncertainty increases the probability of both high and low asset values, but since the exercise of the option is voluntary, it is only the increased probability of high prices in the future that affects the value of a call option. Bertola noted that option price theory could also be applied to real resource allocation decisions, if these decisions were costly to reverse at a later date. The decision to migrate was one example of such a decision, in which uncertainty arises because of differences in earnings between countries.
Bertola analysed in detail a more tractable problem, however: an individual firm's choice of its stock of capital over time in the presence of uncertainty and costs of adjustment. Bertola's model suggested that factor adjustments were governed by two `trigger' values: when the marginal profitability of capital exceeded one trigger value, more capital was installed; if it fell below the other value, the firm reduced its capital stock. Within the `corridor' between these two values, the firm did not respond to fluctuations in its environment by changing its capital stock. Bertola showed that this region of inaction is wider if the degree of uncertainty about the future is greater or if adjustment costs are larger. This approach, Bertola concluded, suggested how hysteresis might arise in aggregate data: favourable shocks may induce a reallocation of resources that is not revised by a later unfavourable shock.
In general, the reallocation of factors in response to changes in relative prices is larger, the smaller the uncertainty about future changes in relative prices; there might therefore exist increasing returns to stabilization activity. This insight suggests important new directions for empirical and theoretical research, aimed at quantifying the extent to which stabilizing nominal exchange rates might improve welfare. As Bertola emphasized, the analysis is significantly affected and complicated by the assumptions about nominal price stickiness.
Several European countries rely heavily on inflation tax revenues to finance their expenditure; seigniorage accounted for between 6% and 12% of government revenues in Greece, Italy, Portugal and Spain over the period 1979-86. Even without monetary unification, capital market liberalization and converging rates of inflation will force these governments to restructure their finances, with direct and indirect taxes replacing the inflation tax. In `Monetary Policy, Capital Controls and Seigniorage in an Open Economy', Allan Drazen (University of Pennsylvania) analysed the transition to a new steady state with a reduced inflation tax, from the perspective of public finance theory. According to Drazen, the four high-seigniorage countries had not only high inflation rates but also significantly larger monetary bases relative to GDP. The monetary base in these countries was large not because of currency holdings but because of bank reserves, which in turn reflected reserve requirements and other portfolio restrictions imposed on banks by the government. As a result, Drazen argued, capital market liberalization, not reductions in inflation rates, would reduce the size of seigniorage revenues in these countries. Liberalization would mean that governments would not be able to impose reserve requirements which were more restrictive than those in other countries without destroying the competitiveness of local banks.
Drazen presented a simplified theoretical model in which he analysed how monetary policy (i.e. required reserve and inflation policies) should be used in the period before an anticipated capital market liberalization. Should the government try to obtain as much inflation tax revenue as it can before liberalization, or should it choose low inflation at an earlier date? Drazen's model suggested that the choice was complicated when capital accumulation was taken into account. Large inflation tax revenues can be generated by maintaining high reserve requirements, but this is costly to financial intermediaries and so reduces the returns to saving. The accumulation of productive capital will fall as a result, the tax base will shrink, and the government's fiscal position will ultimately deteriorate. The extent to which asset accumulation falls, Drazen concluded, depends on how much the effects of high reserve requirements are offset by other policy actions.

The costs of relinquishing monetary independence were also analysed by Vittorio Grilli (Yale University) in his paper, on `Seigniorage in Europe'. Grilli examined the empirical evidence on the use of the inflation tax by European countries. He noted that the feasibility of a fixed exchange rate system (or a full monetary union) depends in part on its effect on the government finances of its member countries. Economic theory suggests that the optimal financing of government expenditures may require an inflation tax and so may be incompatible with membership of a fixed exchange rate system. Historically this conflict has been one of the principal causes of exchange rate crises, Grilli observed; in his paper he attempted to establish empirically if revenue needs had affected the rate of inflation in European countries. Grilli presented a simple theoretical model of the optimal choice of seigniorage and income taxation by the government. The theory implies that the tax and inflation rates should be `cointegrated': Grilli tested this hypothesis using time-series data for ten European countries.
The results suggested that the countries fell into two groups. Grilli found no evidence of a consistent seigniorage policy for Belgium, Denmark, the Netherlands, Spain and Britain; but for France, Ireland, Italy, Germany and Greece seigniorage was an important component of financing policies. Grilli also found that exchange rate considerations imposed a serious constraint on seigniorage policies for Germany and Britain.
Finally, Grilli tested for the long-run solvency of the ten countries' budgetary policies by examining the behaviour of the stock of debt. His results cast doubt upon the long-run sustainability of the budgetary policies currently pursued by a number of European countries: such potentially disruptive findings highlight the importance of integrating public finance considerations into any discussion of the benefits and costs of common currencies.
The problems of coordinating monetary policies were the focus of the papers by Alessandra Casella (University of California, Berkeley) and Jonathan Feinstein (Stanford University) and by Carlo Carraro (University of Venice and CEPR). In `Alternative Monetary Systems Between Two Countries', Casella and Feinstein provided a general equilibrium analysis of the behaviour of two central banks under three regimes: flexible exchange rates, a common currency issued independently by both banks, and a common currency managed by a jointly controlled central bank. The authors employed a much simplified model, in which each country's welfare was identified with that of a representative citizen, who consumed home and foreign goods and a public good provided by the home country government. Each government sets the amount of the public good that it provides, taking into account the welfare of its own citizens but not those of the other country, and finances this good either by the creation of money or by lump-sum taxation. This gives rise to an externality: an increase in one government's expenditure reduces the total supply of goods for private consumption in both countries. Its own citizens are compensated by an additional supply of the public good but foreigners are not, so that too many public goods are provided.
The framework outlined by Casella and Feinstein suggests that establishing a common currency will be difficult because each country uses a fixed exchange rate to exploit its partners. The problem is illustrated most clearly when there is a common currency issued independently by each country. Both governments find it optimal to levy no taxes and to finance all expenditures through seigniorage: creation of money by one government benefits its own citizens since more of the public good is provided and yet the world price level does not rise proportionally. The common currency creates a second externality, a common inflation rate, and the model is characterized by unbounded inflation: for any amount of currency issued by the foreign government, the home country will wish to supply more.
Casella and Feinstein also outlined a formal model of a common central bank, managed with a system of `proportional representation' in which each country's influence on the Central Bank is proportional to its economic size. They find that even in the presence of a central monetary authority the two countries may prefer a flexible exchange rate: in order to induce the smaller country to participate in the common currency agreement, the smaller country must have an importance in determining bank policy that is greater than proportional to its economic size. Casella and Feinstein argued that this arrangement may be difficult, and that some consideration of political processes was therefore needed in order to advance the analysis.
Theoretical analyses have suggested that the sustainability of policy coordination depends on the nature of policy-makers' objectives and the length of the horizon over which they plan. Carlo Carraro attempted to infer `The Tastes of European Central Bankers' from time-series data on inflation, output growth, and other relevant macroeconomic variables. Carraro assumed that over the period 1977-86, governments and central bankers in the UK, France, Germany and Italy chose policy instruments in order to achieve output growth, inflation, and trade balance targets. By assuming that the observed data for this period were the outcome of a dynamic game in which policy-makers have chosen strategies to maximize their objective functions, Carraro was able to infer what the characteristics of these objective functions must have been. His analysis suggested that central bankers appear to have very short policy horizons, which theory suggests will make cooperative outcomes difficult to sustain; but he did not find significant differences among the central bankers' policy targets, and this tends to facilitate policy coordination.

The Lessons from History
Many of the papers presented at the conference focused on the theoretical costs and benefits of European monetary unification involving some form of common monetary authority. Others relied on historical evidence to provide a perspective on the way in which unification might actually come about, that is, its institutional feasibility. The German and Italian experiences were described in `Monetary Cooperation and the Central Bank Question in the German Unification Process During the Nineteenth Century', by Carl-Ludwig Holtfrerich (Freie Universität, Berlin), and by Valeria Sannucci (Banca d'Italia) in `The Establishment of a Central Bank: Italy in the Nineteenth Century'. Both papers illustrated how monetary unification could result in a `unitary' central bank. The creation of the Federal Reserve System, whose effects were analysed in `The Founding of the Fed, the Real Bills Doctrine and the Destabilization of the Post-1914 Economy' by Jeffrey Miron (University of Michigan), provided an example of the creation of a `federal' central bank.
The German and Italian experiences with monetary unification display some deceptive similarities. In both cases one state actively promoted political unity and, having achieved it through military means, proceeded to establish its monetary system over the whole territory of the unified country. There were important differences, however: before unification Germany had already become an integrated economic area and a unified currency area (based on a silver standard). Italy, on the other hand, was a patchwork of economically heterogeneous states which, at the time of political unification, traded more with foreign countries than with one another. Unlike the German states they were not united by a network of railways, and the two constituents of the new state, Piedmont and the Kingdom of the Two Sicilies, had currencies based on different standards.
The Italian case shows political and monetary unification preceding economic integration. The German case shows economic integration leading to political unification. Historians have argued that Italian unification was a sudden and largely unexpected event, while German unification was a long and gradual process which occupied the best part of a century. This basic difference can go far towards explaining the great difficulties which the new Italian state experienced in the economic and monetary fields, and in particular its difficulties in building a modern banking system around a publicly controlled central bank. On the other hand, the great success of the German Reich can be attributed to the economic and monetary unification which preceded political unity. The influence of the immediate past over the present and future, in the case of both Germany and Italy, seems to have been overwhelming.
Does Europe in 1988 find itself in the situation of Italy in the 1860s or of Germany? Previous attempts at European monetary unification, like that promoted in the Werner Report of 1970, might be compared to the situation in Italy in 1860, when economic and financial unity was not sufficiently advanced to justify the significant step of monetary union. It may be, however, that the Europe of 1988 has come to resemble Germany in the 1860s: economic integration has undoubtedly increased and total intra-European trade has now stabilized at a very high level.

The three historical papers clearly illustrated that political unification need not precede monetary union, either in its nineteenth-century incarnation as free circulation of coins among states or in its present embodiment as joint floating plus liberalization of capital flows. In the absence of political union, however, the decision to establish a central bank to control monetary policy assumes added political significance, as Niels Thygesen (University of Copenhagen) remarked in his contribution to the conference panel discussion.
How can the interests of member states in a currency union be reconciled in the absence of a single political authority? One possibility is the establishment of a decentralized federal monetary authority. The Federal Reserve Act of 1913 illustrates such a mechanism for reconciling such local interests with a central monetary authority. The potential for conflict among local interests in the formulation of monetary policy can also be diminished by reducing the discretionary element in monetary policy; if policy is governed by a rule, then there is less scope for disagreement. The US decision to adopt the Gold Standard and the Reichsbank's switching to the Gold Standard illustrate this solution to the conflict of regional interests. The inelasticity of commodity standards was, however, taken into account by the Federal Reserve Act and by the Bank Act, which permitted some discretionary money creation.
In his paper, Miron argued that the Federal Reserve had used this discretion to pursue a policy of stabilizing interest rates, designed to prevent asset market panics. Although this policy did smooth nominal interest rates, according to Miron, the US economy after 1914 also displayed increased volatility of output and inflation, largely as a result of attempts by the Federal Reserve to stop speculation in stock or commodity markets by restraining monetary growth. The Federal Reserve thus exercised more discretion over monetary policy than its adherence to the Gold Standard might suggest. Miron argued that this discretion of policies had undesirable effects on US macroeconomic performance, though participants at the conference questioned whether other factors might account for increased output and price volatility after 1914.
The regional diffusion of power over monetary policy which characterized the US system took place within the context of a single government and currency. Neither has yet been achieved in Europe. What then should be the shape of a European monetary authority? As Rainer Masera (Banca d'Italia) argued in the conference panel discussion, it may be more useful to follow the Single European Act and to think in terms of a common monetary authority for Europe rather than a European Central Bank. The distinction between a European Monetary Board (as Masera called it) and a European Central Bank is not merely semantic. Without political union a European Central Bank, even one shaped like the Federal Reserve System or the Bundesbank, may never emerge: its creation would require a once-and-for-all abdication of monetary sovereignty, which is very unrealistic to expect from the EC countries.
But will such a loose arrangement withstand the pressures resulting from intra-EC liberalization of capital movements? If 1992 brings about the integration of European banking freedom of establishment by European banks wherever they want on EC territory another element of traditional monetary sovereignty will have been eroded. A full application of the right of establishment for banks and other financial institutions brings with it a series of problems concerning banking supervision and the management of the lender of last resort function, which will require agreements among central banks. These agreements are essential to establish a truly European payments network, which, as Daniel Cohen (Centre d'Etudes Prospectives d'Economie Mathématique Appliqueés à la Planification, Paris, and CEPR) suggests in his paper, `The Problems of a European Central Bank', will make the recycling of balances between European regions much more straightforward. Eventually, intra-European balances of payments would become just a statistical curiosity.
Cohen analysed the question of a European central bank from the perspective of the evolution of the international monetary system since 1945. The Bretton Woods System, he noted, attempted to solve two distinct problems: the first was the enforcement of balance of payments discipline on individual countries: the system was unworkable without such discipline, since without it every country would be tempted to increase its indebtedness without limit. Second, the architects of Bretton Woods intended the system to manage the interactions between the policies pursued by different countries, which had contributed to the deflationary spiral of the 1930s.

Cohen argued that the recent evolution of the international monetary system was guided by the quest for a solution to the problem of external constraint, and that any confusion between the notions of managing the external constraint and creating a supranational currency would only delay the latter as long as the former seems problematic. Cohen considered that the experience of the international monetary system offered clear lessons for Europe. The European currency, if it eventually comes into existence, must be controlled by a central bank radically different from the one proposed by Keynes. Its interventions must be neutral with regard to intra-regional imbalances.
Cohen noted that a European central bank would be technically easy to implement: the real difficulties lay elsewhere. Germany, according to Cohen, feared that a monetary union would not defend a European currency as vigorously as the Bundesbank defended the Deutschmark. Other European countries feared that the difference between their rates of inflation and that of Germany would cause a steady loss of competitiveness relative to German producers.
These reservations could be overcome by a transitional period of three to five years during which time the legal status of European central banks would be harmonized, giving them the same independence as the Bundesbank vis-à-vis their governments. This would immediately facilitate European monetary cooperation. The transitional period could also be used to harmonize the regulation of commercial banks and to establish unified reserve requirements.
The transition period should reduce the risks that an ill-chosen initial parity could entail for one of the participants. This could be achieved by a greater use of the ECU as a unit of account and means of settlement, as European prices and costs would be easier to compare using a single numéraire. Governments could assist this process by allowing use of ECU cheques and notes for national transactions. The success of the transition period ultimately depended on setting a deadline that would encourage private agents to revise their expectations. To overcome any remaining objections, Cohen proposed that the European monetary system could rest on twin institutions: an issuing body, whose President would be known for his monetary conservatism (say a German), and a financing institution, whose Director could draw on budget resources provided in advance to help finance any disequilibria that might arise.

A European Central Bank? Perspectives on Monetary Unification after Ten Years of the EMS, edited by Marcello de Cecco and Alberto Giovannini, will be published by Cambridge University Press in 1989. It will contain revised versions of the papers presented at the Castelgandolfo conference, along with discussants' remarks and the concluding panel discussion. It will also include an introduction by the editors, on which this Bulletin report in based