The December 1991 Maastricht summit appears to have set the European Community on an irreversible course towards economic and monetary union (EMU). In many respects, the timetable for EMU and the institutional and policy changes that will precede it are now clear, but the economics of the transition and indeed of an established EMU have received less attention and raise many difficult and as yet unresolved issues. A joint CEPR conference with the Banco de Portugal, held in Estoril on 16/18 January, aimed to shed new light on these questions. This conference was organized by Francisco Torres, Economist in the Research Department of the Banco de Portugal, and Professor of Economics and Director of the Centro de Estudos Europeus at the Universidade Católica Portuguesa, and Francesco Giavazzi, Professor of Economics at Università Bocconi, Milano, and Co- Director of CEPR's International Macroeconomics programme. CEPR also acknowledges grants from the Ford Foundation supporting its International Macroeconomics programme and from the SPES programme of the Commission of the European Communities supporting the Centre's research programme on Financial and Monetary Integration in Europe.

The Role of Stage II
The 1989 Delors Committee Report made few recommendations for the transition to EMU. Andrew Crockett (Bank of England) opened the first session with his paper, `The Role of Stage II: The Proposed Treaty Revision', which maintained that the subsequent proposals could be grouped under four broad headings. First, a gradual transition would transfer monetary competence from national authorities to a Community institution, avoiding the need for a `big bang' in Stage III. Second, the UK plan for a `hard ecu' proposed a new `parallel' European currency, which if successful in the market-place would displace national currencies, allowing market forces to determine the pace and direction of EMU. Third, the Bundesbank favours simply reinforcing the procedures for cooperation established in Stage I, with no substantial transfer of competence before a firm decision to move to Stage III. Fourth, a new monetary institution which will emerge as the European Central Bank in Stage III may promote the harmonization of monetary control, but with powers only to persuade during the transition.

Crockett noted that the Maastricht agreement blended elements of the third and fourth approaches. Gradualism was dismissed because attempts to divide monetary sovereignty could lead to confusion and instability; while the hard ecu was rejected largely because of other governments' perceptions that the UK's commitment to EMU was half-hearted. Three main issues remain outstanding: views differ over how prominent a role the new European Monetary Institute (EMI) should play in Stage II; the interpretation of the Maastricht convergence criteria leaves room for discretion; and countries that decide to move to Stage III quickly and those that remain outside must establish a framework for joint discussion of monetary policy.
Vítor Gaspar (Ministério das Finanças and Universidade Nova de Lisboa) argued that Maastricht had put EMU's credibility and irreversibility beyond doubt. The convergence criteria imply a clear commitment to stability that governments have strong incentives to meet: they will probably compete to strengthen their reputation with respect to these criteria.

In a second paper, `The Role of Stage II: The Maastricht Treaty Revisions', Niels Thygesen (Kobenhavn Universitet) noted that the agreement had unambiguously set 1 January 1994 as the starting date and relaxed the conditions to be met by then. All twelve countries can now embark on Stage II, although this is to be `less qualitatively different' from Stage I than initially envisaged. The EMI will oversee the ERM and promote the use of the ecu during the transition and prepare for Stage III. Maastricht's principal institutional innovation was to decree that the head of the EMI will be appointed, not elected by the governors of the national central banks; it cannot be judged in advance whether this will raise its public profile.

Maastricht also resolved uncertainty about the end of Stage II: a majority of EC countries can move to Stage III if they meet the convergence criteria by the end of 1996; otherwise EMU will begin in January 1999 for those countries (whether a majority or not) that satisfy the conditions by mid-1998. Relaxing the requirements for Stage II and setting a binding terminal date will make it harder to move straight from something like the present voluntary and decentralized operations to the highly centralized system designed for Stage III, which will force the EMI's agenda beyond the tasks specified in the Maastricht texts.

Lorenzo Bini-Smaghi (Banca d'Italia) agreed that the EMI has a lot of work to do before it disappears in Stage III. Risks of instability arise from the `Walters effect', which associates high interest rates and strengthening currencies with high rates of inflation, and from increased opportunities for currency substitution. The EMI must also learn how to conduct a common monetary policy and fashion an integrated Community-wide payments system. José Viñals (Committee of Governors of the EC Central Banks, Basle, and CEPR) stressed that a good deal of coordination is already talking place: it is no accident that the ERM has recently coped so smoothly with the shocks of the Gulf war, German unification and the erratic behaviour of the dollar. The convergence criteria may themselves become a cause of currency instability, however, if markets try to guess which currencies will disappear with EMU and which will not. Rudiger Dornbusch (MIT and CEPR) suggested that the point of the Maastricht summit had been to provoke such instability: there is much to be said for threatening countries that are likely to miss the boat with a policy crisis.

Inflation Convergence
Recently many economists have drawn attention to the changing character of the ERM. Governments are now evidently reluctant to seek realignments in circumstances that would earlier have prompted them to do so: the ERM is `more credible' than it used to be. Barry Eichengreen (University of California at Berkeley and CEPR) opened the second session by presenting a paper, `Financial and Currency Integration in Europe: The Recent Statistical Record', by Jeffrey Frankel, Steven Phillips and Menzie Chinn. This used both direct survey evidence of exchange rate forecasts and interest rate differentials to test how the credibility of the ERM's target zones had changed over time, and in particular whether the markets have regarded its currency bands as more credible since the last realignment of its parities in 1987.

Their results indicated that the ERM is indeed more credible than before. Not surprisingly, the guilder's target exchange rate against the Deutschmark was the most credible; while other currencies' credibility was `less than perfect' but had increased recently, and especially since the beginning of 1990. The failure of the simple target zone model to account for the behaviour of ERM member exchange rates could not be attributed to errors in the measurement of expectations, as is often supposed, because their results from survey data were no more supportive than those based on interest rates. The model's failure is more likely due to its implicit and implausible assumption that the credibility of exchange rates is constant.

Paolo Onofri (Università di Bologna) contrasted France's success in adjusting to low inflation with Italy's comparative failure. In France, the ERM discipline was taken seriously, but it was not critical to the achievement of convergence. Italy, on the other hand, was erratic in its attitude to the ERM discipline, even though this was responsible for its progress towards convergence. Rudiger Dornbusch found the authors' conclusions quite reasonable, but he questioned their underlying model, which he thought implied that markets had expected many devaluations of the Deutschmark. Paul Krugman (MIT and CEPR), who had devised the simple target zone model, said that its main weakness was the assumption of rational expectations in the foreign exchange market, for which there is no evidence in favour and a great deal against: it was a pity that this literature had reinvigorated an approach that was largely discredited.

In the next paper, `Contracts, Credibility and Common Knowledge: Their Influence on Inflation Convergence', Marcus Miller (University of Warwick and CEPR) and Alan Sutherland (University of York and CEPR) first noted the surprising persistence of inflation differentials in the ERM, given the perceived lower likelihood of realignments. They developed a model to assess the effects on inflation convergence of contract structures, speed of learning, shocks and perceived rules for realignments. They first found that with overlapping wage contracts and a fully credible exchange rate peg, there is some inflation inertia, but not enough to account for the sluggishness observed in practice. Second, overlapping contracts with a currency peg that is less than fully credible produce sluggish inflation and protracted recessions after the adoption of the peg. Third, introducing informational asymmetry among agents increases the pace of adjustment.

Michael Dooley (IMF) noted that developing countries' experience indicated that markets take a long time to believe an exchange rate peg: a fixed nominal exchange rate may therefore imply a steadily appreciating real rate for a protracted period. Patrick Minford (University of Liverpool and CEPR) noted that if a large overvaluation is undermining the credibility of the peg, a devaluation may increase its credibility. Miller agreed that this was correct, provided this devaluation is seen to be the last. A revaluation of the Deutschmark against the other ERM currencies at the moment of German unification might have met this criterion: German reunification is no everyday occurrence.

In his paper, `Inflation in a Fixed Exchange Rate Regime: The Recent Portuguese Experience', Sérgio Rebelo (Banco de Portugal, Universidade Católica Portuguesa and CEPR) interpreted inflation with a distinction between traded and non-traded goods. Portugal's overall inflation rate fell by less than two percentage points towards the ERM average after the escudo began to shadow the ERM narrow band in 1990. Inflation in the traded- goods sector converged, while inflation in the non-traded-goods sector did not; so an essential part of the story was the persistent rise in the relative price of non-tradables to tradables.
Michael Moore (Queen's University, Belfast) questioned the attempt to account for (nominal) inflation in terms of a (real) change in relative prices and noted that this model excluded external capital accumulation and was therefore unsuitable for capturing the effects of fixing the exchange rate. Finally, the distinction between traded and non-traded goods is endogenous: certain goods will be traded at one set of relative prices, but not at another. Andrew Crockett pointed out that Japan's inflation rate had persistently remained five or more percentage points above that of the US throughout 1950-70, without any loss of competitiveness. Consumer prices shed little light on competitiveness; the prices of traded goods or unit factor costs would do better. The Maastricht convergence criterion was probably far more demanding of Portugal than it needed to be.

Variable Geometry, Fiscal Rules and Politics
One of the most difficult issues in economic policy after EMU is countries' responses to shocks once they can no longer use the exchange rate as a means of adjustment. Many argue that fiscal policy must play a greater role in cushioning the impact of shocks, which seems at odds with the Maastricht convergence criteria. In their paper, `Shocking Aspects of European Monetary Unification', Tamim Bayoumi (Bank of England) and Barry Eichengreen (University of California at Berkeley and CEPR) described and compared the underlying shocks to supply and demand in Europe and the US since 1963. They found that European shocks were significantly more differentiated among countries than US shocks were among regions, which suggests that Europe may find it harder to conduct a successful common monetary policy. For an `inner core' of Germany and its closest neighbours, however, they found that shocks were notably more symmetric than for the Community as a whole.

Giorgio Basevi (Università di Bologna and CEPR) responded that the greater coherence of the shocks affecting the US was not surprising precisely because it is already a currency union. What mattered was how the US would have behaved if it had not been a currency union and how Europe will behave when it is. Lorenzo Bini-Smaghi cited work by the European Commission which suggested that Europe may be a more natural common currency area than the US: its countries are less specialized in production than US regions, so their shocks should be less idiosyncratic.

In their paper, `Tax Smoothing Discretion Versus Balanced Budget Rules in the Presence of Politically Motivated Fiscal Deficits: The Design of Optimal Fiscal Rules for Europe after 1992', Giancarlo Corsetti (Yale University) and Nouriel Roubini (Yale University and CEPR) argued that the fiscal rules adopted at Maastricht will overcome the political biases that might otherwise encourage an over-expansionary policy stance; but they may also prevent governments from running adequate deficits during recessions. Moreover, many countries' inability to meet the rules is already encouraging ministers to interpret them in a `political' way, so strong sanctions against rule-breakers are therefore unlikely to be enforced. This combination of rigidity, implausibility and weak enforcement is likely to prove most unsatisfactory.

Francisco Torres (Banco de Portugal and Universidade Católica Portuguesa) agreed with much of the authors' analysis, but he questioned their conclusions. Portugal's desire to achieve convergence in the period leading up to EMU has been strong, for example, and the balance struck at Maastricht between discipline and flexibility seems about right. Charles Wyplosz (INSEAD, DELTA and CEPR) asked whether it was right to ask for economic, as opposed to political, rules on fiscal convergence; Italy could be asked not to reduce its fiscal deficit, but rather to change its electoral system, which may be the underlying cause of its chronic budget deficit. Guido Tabellini (Università di Brescia and CEPR) suggested that asking Italy to change its budget-making procedures may be a more realistic precondition for its participation in EMU.

Viewed from a public finance perspective, an independent central bank that achieves price stability is by no means an unmitigated blessing: in principle, it may be sensible for governments to raise part of their fiscal revenues through a (small) `inflation tax'. In his paper, `Unanticipated Inflation and Government Finance: The Case for a Common Independent Central Bank', Frederick van der Ploeg (Universiteit van Amsterdam and CEPR) argued that a strongly anti-inflationary European Central Bank entails a trade-off between gains from low inflation and the losses resulting from a sub-optimal mix of tax revenues. The case for an independent central bank is strongest when the stock of outstanding public debt is large, governments care more about cutting inflation than about reducing tax distortions, the underground economy is small (which ensures a broad base for other taxes), and there is little wage indexation.

Francesco Giavazzi (Università Bocconi, Milano, and CEPR) noted that seigniorage had been high in many South European countries in the early 1980s, when some highly indebted governments preferred to finance their debts through inflation rather than taxes. Most of the subsequent reduction in seigniorage was due to financial liberalization, which had forced governments to reduce their banks' reserve requirements, rather than to the fall in inflation.

Geography and Growth
Many have argued that European integration will spur growth, if for example specialization and geographic concentration of economic activities creates beneficial externalities that promote faster growth. In his paper, `Models of Economic Integration and Localized Growth', Giuseppe Bertola (Princeton University and CEPR) argued that such externalities need not imply that concentration is socially optimal. His model distinguished between the aggregate scale economies required to sustain endogenous growth and local scale economies, which may lead to privately optimal but socially excessive concentration. Both geographic and intertemporal distortions must be considered in the planning of a unified European tax structure.

Jo<176>o César das Neves (Universidade Católica Portuguesa) presented data for a variety of pairs of core-and-periphery countries (France and Belgium, the US and Mexico, the European Community and Portugal), which indicated that growth rates in the core and periphery are usually positively correlated and that income levels tend to converge over time. Charles Wyplosz argued that most of the economic integration Europe could expect to achieve had already taken place; the remaining obstacles to trade were few, so Europe already knew that the effects discussed by Bertola do not arise.

Presenting his paper, `Lessons of Massachusetts for EMU', Paul Krugman (MIT and CEPR) examined how Europe might respond to a regional shock similar to the current recession in New England. US regions are more specialized in production than European countries, so the specialization of the latter may increase as integration proceeds. Specialized regions are more vulnerable to shocks, especially those caused by shifts in tastes away from their exports; and capital movements in an integrated area may amplify shocks rather than attenuate them. A disturbing aspect of US growth is that regions diverge over long periods, with no evident tendency to return to any historically `normal' level of relative output or employment. This suggests that EMU and the `1992' project will make US-style regional crises more common and more severe within Europe.

Alessandra Casella (University of California at Berkeley and CEPR) found that Krugman's results contrasted oddly with those presented earlier by Bayoumi and Eichengreen, who had found European shocks to be bigger than those in the US. Krugman had argued that shocks will pose greater difficulties for Europe as labour mobility increases: as regions in recession lose labour more readily, temporary shocks will become permanent. Bayoumi and Eichengreen had argued that it is precisely the lack of labour mobility that will make Europe's shocks harder to accommodate, as regions in recession suffer income reductions that further aggravate inter-regional disparities.

Eichengreen said that he and Bayoumi had agreed with Krugman's finding that US demand shocks were bigger than those in Europe, although their results on the relative size of supply shocks differed. Krugman's claim that US shocks were often permanent was implausible, and it was asserted rather then demonstrated. Employment in California could not grow faster than employment in New England for ever; congestion and high land prices were already inducing firms to move elsewhere. Krugman replied that employment divergence might not last for ever, but it was certainly more than a short-term phenomenon.

Financial Market Integration and Capital Taxation
Many commentators have asserted that currency substitution will jeopardize Europe's financial stability as EMU approaches, but conventional money demand equations do not permit an unambiguous definition of currency substitution. In his paper, `Currency Substitution: From the Policy Questions to the Theory and Back', written jointly with Behzad Diba and Alberto Giovannini, Matthew Canzoneri (Georgetown University) devised a model that explicitly accounts for money demand in terms of market organization and transactions imperfections. If goods and factor markets are integrated but governments continue to run divergent monetary policies, as is currently the case in Europe, the currencies that are relatively costly to carry will disappear, as will the corresponding financial intermediaries; indeed the corresponding firms will disappear as they will no longer be able to sell their output. Canzoneri concluded by recommending that no effort be spared to ensure the convergence of ERM members' monetary policies during Stage II. Possible measures to achieve convergence include a narrowing of the ERM bands to secure the tighter synchronization of national discount rates and supplies of domestic credit.

Lucas Papademos (Bank of Greece) noted that these strong results depended as much on a characterization of substitutability in the goods market as on currency substitution itself. Paul De Grauwe (Katholieke Universiteit Leuven and CEPR) complained that such models do not conform to the facts: they imply that consumers will typically hold foreign as well as domestic currency, while a rise in the cost of holding a currency leads them to migrate smoothly to cheaper ones.
A second troubling aspect of financial market integration is that it may cause investment to flow to countries with the least demanding tax regimes. The extent of this effect will depend on how far the existing tax systems distort investment, whether competition alone will achieve the necessary convergence in tax regimes or intergovernmental cooperation will be required, and how much harmonization of capital taxes is needed. In his paper, `Coordination of Capital Income Taxes in the European Monetary Union: What Needs to be Done?', Peter Birch Srensen (Copenhagen Business School) presented an exhaustive review of Europe's corporate tax systems, which indicated that intra- European tax distortions are now not very large and that most of the gains from Europe's economic and financial integration can be achieved without complete tax harmonization, although certain tax obstacles to cross-border investment should be removed. Tax competition among EC member countries is probably not strong enough to eliminate corporate income tax; but the harmonization of personal taxes on capital income may now be necessary to protect the integrity of the system of personal taxation.

Alberto Giovannini reported that the amount of tax firms pay in many countries is a matter of public relations, so the welfare effects of reform are impossible to gauge. The European Commission hopes to harmonize the tax base and set minimum tax rates, in order to facilitate tax competition, reduce tax rates, and hence reduce the cost of distortions. In view of our current ignorance, this seems an eminently sensible approach. Vítor Constâncio (Banco de Portugal and Instituto Superior de Economica e Gesto, Lisboa) stressed that tax competition need not produce an efficient outcome if competition narrows the tax base by making the collection of taxes on capital income impossible. Axel Weber (Universität Gesamthochschule Siegen and CEPR) agreed that competition may force capital taxes to zero, recalling Germany's experience in 1988-9, when an attempt to impose a withholding tax on interest income had to be abandoned as capital flew abroad.

Conclusions
Opening the final panel discussion, António Borges (Banco de Portugal) observed that EMU is not an end in itself, and the Maastricht convergence criteria are a means of achieving the right sort of EMU. Integration did not begin with `1992', and its results so far have been very encouraging. While further integration may in theory lead to excessive specialization, there is a greater risk that governments' attempts to freeze existing production patterns will prevent the restructuring that is needed for EMU to deliver all its potential benefits. Alberto Giovannini noted that several key issues continued to elude economists' grasp after Maastricht: the factors that drive changes in relative prices; the case for a last realignment; and the circumstances in which a government may devalue but quickly recover its credibility. Otmar Issing (Deutsche Bundesbank) said that the conference, like the Maastricht meeting, had been strong on detailed discussion of the form of EMU but vague on the need for a political union. A smoothly functioning EMU will need an elaborate fiscal transfer system; and there is nothing more political than public finance. Paul Krugman said that economists could not determine whether Europe should proceed with EMU: he found it hard to see economic merits in the convergence criteria but nevertheless favoured EMU because it will promote Europe's political union. Niels Thygesen maintained that good reasons for Stage II include the needs to prepare public opinion and create new institutions. Maastricht had failed to settle two vital questions, however, by not ruling out the possibility of further realignments before 1994 and leaving Europe's post-EMU policy with respect to other countries unclear.

'Adjustment and Growth in the European Monetary Union' edited by Francisco Torres and Francesco Giavazzi
Available from Centre for Economic Policy Research, 90-98 Goswell Road, London, EC1V 7RR
ISBN (hardback) 0 521 44019 X