ESRC/CNRS Workshop
European interdependence and the EMS

In the late 1960s the international monetary system was thought to be too rigid to accommodate differences between countries in rates of inflation and productivity growth. Advocates of free markets called for a flexible exchange rate system that would allow greater national autonomy in formulating monetary policies while ensuring satisfactory performance by individual economies. Today it is generally acknowledged that exchange rates have fluctuated excessively during the decade since the advent of floating exchange rates. Exchange rate movements which appear to be unwarranted by economic 'fundamentals' have caused significant changes in the international competitiveness of a number of countries, and this 'overshooting' has greatly complicated the conduct of domestic stabilization policy. In addition, it is argued, the volatility of exchange rates is an impediment to the smooth functioning of international trade and is held partly responsible for the recent tendencies towards protectionism. It is questionable too whether flexible exchange rates secure monetary autonomy. Some governments have used monetary policy to manipulate the exchange rate in order to achieve domestic objectives. This has led other governments to intervene in foreign exchange markets to prevent spillover effects from destabilizing their own economies. But if the need for some degree of exchange rate fixity is accepted, there remains the question of how more stable exchange rates can be achieved.

This question can be approached in two different ways. The first is to ask whether explicit cooperation between countries in the design of macroeconomic policy would yield gains which would result in exchange rate stability. The experience of policy coordination among the European countries participating in the exchange rate mechanism of the European Monetary System (EMS) could illuminate this issue. This approach highlights problems of modelling the EMS, the behaviour of the EMS itself (i.e. how it has operated in practice, whether it has shown satisfactory performance), the nature of incentives for non-member countries to join the system, and finally the size of possible gains from the EMS. An alternative approach to the problem of exchange rate stability emphasizes achieving the minimum degree of cooperative behaviour among national governments that could bring about an acceptable reduction in short-run exchange rate volatility. The proposals by Williamson and McKinnon for international monetary reform appear to fall into this second category.

The Economic and Social Research Council and the Centre National de la Recherche Scientifique have jointly sponsored a CEPR research programme on 'Economic Interdependence and Macroeconomic Policy in Europe', designed to foster collaborative research by French and British economists. The researchers met in Paris on April 21/22 to discuss the proper strategy for modelling the EMS and evaluating the costs and benefits from membership. They also evaluated recent applied work on gains from policy coordination, and discussed proposals for transatlantic economic cooperation.
In the first session, David Currie, Co-Director of the Centre's International Macroeconomics Programme, discussed the application of game theory to investigate potential gains from macroeconomic policy cooperation within the EMS. In policy coordination models, two contrasting types of equilibrium are usually considered. The first is a 'Nash' equilibrium where each government sets its policy independently but takes full account of the possible responses of other countries. The second is a 'cooperative' equilibrium, where there is explicit cooperation between governments in setting policy.

Currie suggested a third alternative, namely an 'imperfect- information Nash game', in which governments behave as in the Nash equilibrium but, because of information lags, are not fully aware of other governments' policies. Currie noted that exchange rate cooperation within the EMS requires the regular exchange of information between participating governments about other aspects of economic policy as well, thus enhancing their understanding of the spillover effects of their own policy actions on other economies. Currie conceded that it was unrealistic to think of the EMS as a full 'cooperative' equilibrium, because formal cooperation among EMS governments applies only to exchange rate policy. Nevertheless, he argued, the EMS could be considered as a transition from an 'imperfect-information Nash' equilibrium to one characterized by more 'cooperation' among European countries.

In modelling the EMS, Currie emphasized that attention should be paid to the enforcement aspect of the agreement, i.e. whether it will pay EEC governments to sustain the 'rules of the game'. Are there any threats which will be effective against member governments not following the rules? Are such threats equally effective against all members of the system? Finally, Currie raised the question of whether or not the EMS should be modelled as an 'asymmetric' system. Some EEC governments appear to have a 'reputation', in that they are known to honour their economic policy commitments; other countries do not. As a result, governments without a 'reputation' may find it beneficial to enter the system and remain in it, simply to establish their own 'reputation'.

In the workshop's second session, Jacques Melitz (Institut National de la Statistique et des Etudes Economiques (INSEE), Paris, and CEPR) discussed the problem of modelling the EMS. He drew attention to the importance of assumptions concerning the nature of the agreement itself, the environment, the state of information and the differences between 'players' and their relationships. Melitz argued that the EMS should not be analysed as if it were a 'snake' agreement or a 'Bretton Woods' system. It was better characterized as an 'exchange rate union' with 'joint exchange-rate responsibility'. As such it involves greater elements of policy cooperation than did these earlier attempts. The economic environment must be taken into account in assessing the benefits from policy coordination within the EMS, Melitz argued. If prices are not perfectly flexible, it is essential to take into account the transition to 'equilibrium', as well as the possibility of persistent deviations from 'equilibrium', in evaluating the effect of the EMS. Melitz argued that careful attention should be paid to the fact that 'players' in the EMS may not be identical. The countries involved may differ not only in the size of their economy and their policy objectives, but also their access to information.

There is currently a debate as to whether the EMS operates as a 'Deutschmark zone' or an 'ECU zone'. In a Deutschmark zone, the dominant influence of the mark leads one to expect no change in the mark/dollar exchange rate after an EMS parity realignment. In an ECU zone, it is the ECU/dollar rate which one would expect to remain constant during a change of EMS parities. Charles Wyplosz (Institut Europeen d'Administration des Affaires (INSEAD), Fontainebleau, and CEPR) considered the evidence on this question to be ambiguous: movements of the Deutschmark against the French franc and the US dollar suggested that on some occasions the EMS operated as a Deutschmark zone and on others as an ECU zone. Wyplosz found it puzzling that in periods where realignments took place, the interest rate differential between the mark and the franc apparently behaved as if no realignment was expected. Are markets 'irrational', in the sense that they do not fully appreciate how the EMS operates? Other participants pointed out that assets denominated in different currencies might be imperfect substitutes in investors' portfolios. In such circumstances the behaviour of interest differentials can be a misleading indicator of how markets expect the exchange rates to move in the future. It was also suggested that markets might have operated so as to equalize real interest rates, expecting realignments to be governed by deviations from purchasing power parity. Finally, it was argued that markets could have been anticipating a drift within the exchange rate band of the EMS, rather than actual realignments.

Daniel Cohen, (Centre des Etudes Prospectives d'Economie Mathematique Appliquees a la Planification (CEPREMAP), Paris, and CEPR) presented joint work with Philippe Michel on the design of monetary stabilization in open economies. Their work addressed the question of whether governments known to have strong domestic incentives to inflate would adhere to a Deutschmark zone in order to contain their inflation. The model used by Cohen and Michel was a deterministic one, in which the government sought to minimize a (weighted) sum of deviations of output and inflation from their target values. Cohen and Michel had introduced overlapping wage contracts into their analysis; these affected short-run movements in domestic prices. Analysis of Cohen and Michel's model suggested that the government would not seek to reduce inflation by entering into a Deutschmark zone, but would prefer a 'smoother policy', reducing inflation gradually towards its target value. The key to this result appeared to be the dynamics of wage behaviour, which were strongly influenced by the overlapping wage contracts.

The EMS has undoubtedly succeeded in stabilizing nominal exchange rates. Surprisingly enough, however, it appears not to have stimulated trade. Paul de Grauwe (Centrum voor Economische Studien, Louvain) presented empirical evidence on the economic performance of the EMS countries which addressed this paradox. The decline in the trade shares of the EMS countries during 1981-83 has been more pronounced than that of the non-EMS members. De Grauwe argued that this could be attributed to the integration effects of the EEC which, while substantial in the 1960s and 1970s, were only small in the 1980s. He also argued that there was little evidence of greater monetary policy convergence since the establishment of the EMS. A particularly worrying point was that the EMS countries experienced a larger decline in their GDP growth rate in 1973-86 than non-EMS countries. De Grauwe raised the question of whether the EMS imparted a deflationary bias to the European economy because it fell short of full policy coordination among EEC governments. In a fixed exchange rate system, he argued, the absence of explicit policy cooperation may permit inconsistent national policy objectives, as governments seek to run current account surpluses to increase their share in world reserves. This tendency may eventually lead to deflation for everyone. De Grauwe therefore felt that for the EMS to be beneficial, the cooperation objective should be pursued more vigorously. Other participants pointed out, however, that De Grauwe's argument about the deflationary bias of the EMS was in principle incompatible with his observation that little convergence on the inflation front had been achieved. It was suggested instead that the difference in economic performance between the EMS and non-EMS groups might have been the result of a larger 'wage gap' in European countries than in countries such as the United States and Japan.

Real exchange rate variability, it is argued, can damage the growth prospects of a country, not only because it creates increased uncertainty in domestic investment decisions, but also because it may induce traders and exporting firms to seek protection through tariffs and quotas. Marcus Miller (Warwick University and CEPR), discussed UK membership of the EMS in the light of the contrasting behaviour of the real effective exchange rates of sterling and the dollar compared to those of the EMS currencies over the period 1961-86. Since the advent of generalized floating, Anglo-American countries had experienced considerable volatility in their real effective exchange rates. By contrast, fluctuations in the real exchange rates of the Deutschmark and franc had been similar during the period of floating rates and the Bretton-Woods era. EMS membership appeared to hold advantages for the United Kingdom as participation in the system could reduce the present volatility of the real exchange rate of sterling. Furthermore, membership would add credibility to UK inflation targets because of the reputation of the Bundesbank. Miller briefly discussed possible terms for UK entry. First, he argued, the United Kingdom should seek to enter with a value for sterling below its present level, since the pound has not adjusted fully to its recent status as a petrocurrency. Second, the United Kingdom should seek to participate in the EMS with an exchange rate band for sterling which was wider than the normal 2.25%. This was essential: on the one hand the United Kingdom, unlike France and Italy, is free of capital controls, and on the other hand its inflation rate is higher than that of Germany.

Discussion of Miller's paper centred on his arguments concerning the terms of entry. A wide margin of fluctuation for sterling would, it was argued, diminish the credibility of the UK government's commitment to reduce inflation. It was also argued that the size of the inflation differential between the United Kingdom and Germany, unlike that of Italy, was not large enough to justify a wider band; and any initial 'overvaluation' of sterling could in principle be corrected through a realignment after entry.

Francesco Giavazzi (University of Venice and CEPR), speaking on the costs and benefits of the EMS, noted that any system of relatively fixed exchange rates had two undesirable features. First, anticipation of future realignments may lead to substantial pressures on the currencies whose relative values are expected to change: this can only be prevented at the expense of large fluctuations in official reserves. Second, changes in markets' expectations about the future course of exchange rates may necessitate large changes in short-term interest rates. In evaluating exchange rate arrangements like the EMS, therefore, one has to balance the benefits of reduced exchange rate volatility against the costs of fluctuations in official reserves and in short-term interest rates. Giavazzi was concerned that reserve and interest rate volatility could mean that capital controls were necessary for the smooth operation of the EMS. Yet this would impede the full integration of financial markets in the member countries.

Jose Vinals (Stanford University and Banco de Espana) raised the issue of Spanish entry into the EMS. It was less problematic than British entry, he thought: Spain had already had capital controls and, unlike the United Kingdom, was not an oil producer. The benefits for Spain from EMS membership needed investigation, however. Participation in the EMS would reduce the volatility of Spanish competitiveness, and it would be a good opportunity for the government to introduce an anti-inflationary policy package. It would also enable the Spanish government to gain a 'reputation' by sharing the credibility of the Bundesbank's inflation objective. A number of participants were, however, concerned about the consequences for Spanish unemployment, currently 22% of the labour force. If, as argued in earlier sessions, the EMS had been an engine of deflation, would not Spain's entry into the system intensify its current unemployment problem?

Charles Wyplosz discussed the recent devaluation of the French franc. The gap between French and German inflation had been steadily growing since 1983, he noted, and had led to increased pressure on the franc, culminating after the French elections. Wyplosz noted that the April devaluation of 6% against the Deutschmark had not been enough to restore France to its 1979 competitive position relative to Germany.

Mark Salmon (Warwick University and CEPR) discussed some recent evidence on gains from international cooperation established by Andrew Hughes Hallett's simulation study, described in a January lunchtime meeting and reported in Bulletin No. 13. The movement from 'isolationist' to 'cooperative' policies produced a welfare gain equivalent to an increased annual GDP growth rate of 5.8 percentage points for the United States and 3.9 percentage points for Europe. Salmon was concerned, however, that such simulation exercises were unable to resolve some important issues. The gains are calculated by simulating econometric models, which are subject to statistical uncertainties in their estimation. Are the gains from cooperation established by simulating such models 'statistically' significant? How are they influenced by the structure of the model used? In attempting to measure gains to cooperation, should we not take account of the empirical relevance of punishments, threats etc?

Gilles Oudiz (Compagnie Bancaire, Paris, and CEPR) raised similar questions in his discussion of empirical estimation and simulation versus theoretical analysis of simplified models. Oudiz argued that analytical models were unsatisfactory in three respects. First, they typically impose an assumption of 'symmetry': the economic structures of the countries concerned are taken to be essentially identical. Second, analytical models appeared to give undue emphasis to issues such as time inconsistency: there appeared to be little difference between outcomes under time-consistent and time-inconsistent policies. Third, analytical results are particularly sensitive to assumptions relating to the nature of the objective functions of the government and to the parameter values used.

Subsequent discussion focussed on Oudiz's arguments concerning the empirical relevance of the time inconsistency problem. John Driffill (University of Southampton and CEPR) pointed out that time-inconsistent solutions had a marked influence on analytical results in cases where these solutions affected the long-run equilibrium of the system. Further, several participants called attention to difficulties in the empirical estimation of gains from the formation of the EMS itself. How can one best characterize European policy-making? Is it a 'cooperative', a 'Nash', or an 'imperfect-information Nash' equilibrium?

David Coe and Peter Richardson (OECD) discussed the possibility of using OECD empirical models to quantify potential benefits from international policy coordination. These models at present suffer from three main weaknesses: they do not represent the non- OECD world adequately; they contain a 'backward-looking' specification of asset markets; and the large number of equations they contain makes it difficult to identify the source of a given simulation result. The lack of forward-looking behaviour in asset markets in these models was seen as a major weakness, especially for the estimation of potential gains to cooperation. Analysis of a more partial equilibrium nature might be more illuminating than one which involved simulating large OECD models.

David Begg (Bank of England and CEPR) examined John Williamson's proposal for international monetary reform. Williamson's proposal involves the establishment of 10% margins of fluctuation for nominal exchange rates in the short run, with a 'crawling peg' in the long run which took account of changes in 'fundamentals'. Begg felt that Williamson's proposal was unsatisfactory in a number of respects. The number of countries participating in the arrangement was not specified. The plan did not specify who would take the responsibility for maintaining the exchange rate margins in the short run. Williamson's plan effectively relied on the ability of countries to negotiate the appropriate 'equilibrium' exchange rates, but there were serious practical difficulties in calculating such rates. Begg was also concerned that the plan would accommodate differences between countries in inflation and productivity. This would make it more difficult for 'inflation-prone' governments to reduce inflation by acquiring some of the 'reputation' of governments known to have a low propensity to inflate.

John Driffill focussed on McKinnon's plan for world monetary reform. The plan, Driffill noted, reflects an attempt at greater international cooperation, coupled with a desire for simplicity in the policy rules pursued. Driffill argued that the entire plan is based on the view that the present volatility of exchange rates is due to fluctuations in the demand for different currencies. Consequently McKinnon's plan involves the stabilization of exchange rates around an equilibrium path through opposite and symmetrical adjustments in the supply of national currencies, with 'global' money supply growth maintained at a target level. The main 'players' in McKinnon's proposal are the United States, Japan and Germany, with the United States being the 'leader' and Germany and Japan the 'followers'. Warwick McKibbin and CEPR Research Fellow Jeffrey Sachs had recently attempted to evaluate McKinnon's plan. Their results, however, do not appear to indicate substantial gains from cooperation; in some cases cooperation seems to give rise to a worse outcome than Nash equilibrium. Moreover, these results tend to suggest that while McKinnon's monetary rule might be effective in adjusting to monetary shocks, it will respond less effectively to other shocks.

The presentation of these two proposals prompted an interesting discussion. David Currie argued that Williamson's proposal was more acceptable than McKinnon's plan. The former represented in effect an attempt to 'internalize the externalities' of non- cooperative policies in a world of flexible exchange rates. David Vines (University of Glasgow and CEPR) drew attention to the fact that McKinnon's plan involved stabilization of nominal exchange rates while Williamson proposed stabilization of real exchange rates. De Grauwe was concerned that McKinnon's plan might impart a deflationary bias to the system. Cohen argued that the EMS represented a move toward more 'symmetry' in world economic affairs and might encourage Europe to offer more leadership in international monetary matters. It was therefore more acceptable than Williamson's proposal. Other participants thought that Williamson's plan could have the ultimate effect of making governments more responsible for their policies.

Jean-Pierre Danthine (University of Lausanne) contrasted the behaviour of the Swiss economy with that of the rest of Europe, in the final session of the workshop. He drew attention to Switzerland's high rate of growth coupled with an inflation rate close to zero. Workshop participants wondered whether structural factors, rather than economic policies, could explain this difference in economic performance.