THE EUROPEAN MONETARY SYSTEM

Since its conception in 1978, the European Monetary System (EMS) has attracted much attention as a framework for policy coordination. Disillusioned by the prospects for global monetary reform and by the performance of the floating exchange rate system, the EMS founding fathers wanted to restore a system of fixed-but-adjustable exchange rates within as large a part as was possible of the European Community. Such a system would protect the large intra-European trade flows against the sharp shifts in competitiveness which were thought likely to occur in its absence. It would also limit the divergence of national inflation rates and permit lower, less variable inflation.
This vision of the evolution of `a zone of monetary stability' contained a strong element of damage limitation. Yet at the time it was rightly regarded as highly ambitious because it brought back into European management the currencies of some countries, France and Italy in particular, which had remained aloof from earlier efforts. The system has subsequently evolved well beyond its original intentions: the exchange rate mechanism has become tighter, monetary policy coordination is now closer and capital mobility higher than was the case in the early years of the EMS.
In the light of these developments, CEPR organized a joint conference with the Banca d'Italia and the Centro Interuniversitario di Studi Teorici per la Politica Economica (STEP) in Perugia in October. The conference took stock of the substantial existing research on the EMS and assessed the contributions it can make to our understanding of how the EMS works, how some of its effects may be measured, and how it might be improved according to certain well defined criteria. Macroeconomic theory and careful use of econometric analysis can illuminate some of the main implications of the particular blend of exchange rate management and external constraints represented by the EMS. These tools can also facilitate the evaluation of the more operational, short-term problems encountered in managing the system and the national monetary policies that underpin it.

The International Environment

In `The EMS, the Dollar and the Yen', Rudiger Dornbusch (MIT and CEPR) criticized three characteristics of the present EMS. Dornbusch questioned whether inflation rates in Southern Europe should converge towards the near-zero level reached in Germany. He also argued that the EC should consider carefully whether to pursue full liberalization of capital movements without allowing more flexibility in the exchange rate for financial transactions. Dornbusch also observed that the process of adjustment in international imbalances is far from complete and that the US dollar may still have some distance to fall: this may pose a major challenge to the cohesiveness of the EMS currencies. More generally, Dornbusch doubted whether a high degree of nominal exchange rate rigidity was either desirable or feasible for all present and potential members of the system.

His argument against further disinflation in Southern Europe was novel: it rested less on the output losses which would result from such a policy than on the sustainability of public debt during a period of declining inflation, when the scope for financing part of the public sector deficit by money creation was diminishing. The aim of stabilizing nominal exchange rates relative to the strong EMS currencies may allow a `Mediterranean' government to lower real interest rates further and generate more growth (together with the associated budgetary improvement) than it could achieve by assuming a less disinflationary stance. But if some countries, notably Italy and the new entrants in the EC, were to follow the Dornbusch prescription and openly accept a permanently higher inflation rate than Germany and other members aiming at price stability, rules would have to be found for accommodating the required exchange rate changes. Dornbusch suggested a crawling peg with frequent, almost continuous adjustments between the two groups of currencies so as to maintain unchanged competitiveness. Domestic monetary and fiscal policies would then have to ensure that inflationary instability does not develop.
Dornbusch argued that this option represented the best framework for widening exchange rate management in Europe, particularly if it were complemented by a system of dual exchange rates. Real exchange rates, or competitiveness, would then be stabilized by the crawling peg, while the rates at which financial transactions take place would not be constrained by intervention limits.
Dornbusch conceded that a moderate transactions tax on all foreign exchange purchases imposed worldwide may constitute a simpler method of deterring excessive capital mobility than a dual exchange rate system, but it would bring the added inconvenience of a tax on current account transactions. He remarked that the EMS countries today benefit from what is in effect a system of dual rates. Most of the trade of EMS participants is with other members and is carried out at tightly managed exchange rates which do not permit sudden shifts in competitiveness, while a large share of members' financial transactions involve floating currencies, notably the dollar.

Disinflation and External Adjustment

Four conference papers offered a detailed assessment of the EMS as a framework for cooperation among its members; of the implications for key macroeconomic variables, notably inflation and output; and of the determinants of the trade imbalances of the three largest participants.
Jacques Melitz (INSEE and CEPR), in his paper `Monetary Discipline and Cooperation in the EMS: A Synthesis', explored the incentives countries have to participate in an arrangement like the EMS. Most observers have tended to focus on the disinflationary mechanism operating on the weaker participants in the system. The monetary authorities in the countries with weaker currencies gain credibility for their disinflationary stance by pegging their currencies to the less inflationary currencies over long intervals, with the prospect that they cannot fully accommodate their excess inflation by devaluing when the EMS parities are realigned. Private agents will therefore interpret an EMS commitment as a more effective constraint than a national monetary target or some other intermediate objective by which an individual country signals its disinflationary intentions.
These benefits of `tying one's hands', to use the expression of Giavazzi and Pagano (see CEPR Discussion Paper No. 135), seem attainable in the context of the EMS. Melitz noted that they are transitory; if a weak-currency country has pursued a disinflationary stance through its link to a stable DM rate in the EMS for some length of time, and national inflation has been reduced to near-German rates, the effort could presumably have been made by the country acting alone, without the external commitment represented by the EMS.
Melitz anticipated the discussion later in the conference on the desirability of a European central bank to replace the present mechanism where countries with weak currencies draw on the reputation of the Bundesbank. Greater internal autonomy for other national central banks could facilitate the construction of a European institution, or might even be regarded as a partial alternative to it.
Melitz also analysed why Germany too has found the EMS attractive. The impetus to launch the system came after all at least as much from Germany as from France. Why does Germany prefer the EMS to the alternative of individual pursuit of price stability? In the latter case Germany would export stability to those countries that want to peg their currencies through their own efforts to the DM, but without undertaking the commitments to sustain these other currencies through intervention and accepting `excessive' money creation and some imported inflation. Melitz developed an extended version of his model in which competitiveness enters the welfare function of both Germany and France. Both countries desire growth in the sector producing traded goods, but Germany finds this easier to achieve than other EMS members because of the steady real depreciation of the DM which results from keeping nominal exchange rates stable while inflation rates have not yet fully converged.
This argument suggested that EMS membership is attractive to Germany because it gains more in terms of gradually improving competitiveness than it loses in terms of upsets to German monetary policy. Melitz argued that these gains may prove more durable than those of Germany's partners pursuing disinflation through the EMS.
Given the emphasis on the convergence of inflation towards a low and stable rate within the EMS, it is natural that the evidence on the determinants of inflation should be examined carefully. Can part of the remarkable deceleration of inflation observable in all participating countries since 1979 or rather since inflation peaked in 1980 after the second oil price shock be attributed to the EMS? This question was approached in both theoretical and empirical terms in the papers by Giavazzi and Giovannini and by Collins.
In their paper, `The Role of the Exchange Rate Regime in a Disinflation: Empirical Evidence on the European Monetary System', Francesco Giavazzi (University of Bologna and CEPR) and Alberto Giovannini (Columbia University and CEPR) first reviewed the recent literature on the channels through which a commitment to the EMS may help a process of disinflation, and then looked for evidence of a downward shift in inflationary expectations following the introduction of the new, supposedly tighter exchange rate regime of the EMS. They estimated a reduced-form system of equations for the quarterly behaviour of wages, prices and output for three high-inflation EMS countries (Denmark, France and Italy), for Germany, and for the United Kingdom over the 1960s and 1970s, prior to the formation of the EMS. They compared the predictions from this model with actual observations of wages, prices and output over 1979-87. The results corresponded broadly to what one would expect: in the three countries trying to squeeze their inflation through EMS membership, wage and price inflation was lower than predicted by the model based on the pre-EMS experience; in Germany, however, inflation was higher than the pre-EMS model predicted. The evidence further suggested that there was no loss in output in Denmark and Italy (rather a gain), some loss in France between 1980 and 1984, and, surprisingly, a substantial loss in Germany. On the whole the evidence was relatively weak; a downward shift in inflationary expectations did appear to have taken place, but only gradually and less dramatically than in the United Kingdom where the actual inflation rate dropped sharply below the predicted rate from 1981 onwards.
In her paper `Inflation and the EMS', Susan Collins (Harvard University) reached a similar conclusion using different data and methods. Collins used the experience of a group of non-EMS countries as a benchmark and found weak evidence that since 1979 inflation rates have converged more rapidly within the EMS than outside. This seemed to occur during the period before 1978 as well, however, and there is evidence only of a gradual effect within the EMS, from about 1983 onwards. There appear to have been two stages to EMS membership. In the initial stage the behaviour of the authorities was modified, because the policy instruments available to them worked differently in a regime with more stable exchange rates. After several years, however, the disinflationary commitment became more credible to the public.
The paper by Stefano Vona and Lorenzo Bini Smaghi (Banca d'Italia), `Economic Growth and Exchange Rates in the EMS: Their Trade Effects in a Changing External Environment', examined the determinants of the exports and imports of goods in the three main EMS economies. Their results permitted them to address a member of questions of central importance to the sustainability of an EMS in which parity realignments have been relegated to a minor role in the adjustment of external imbalances.

Vona and Bini Smaghi found that changes in trade balances for Germany, France and Italy are strongly influenced by relative demand growth in the three countries and by growth outside the EMS. Price elasticities were significant and sufficient for a devaluation to contribute to a trade balance improvement. But devaluation was, even within the partial equilibrium approach adopted by the authors, far less effective as an instrument for shifting demand between the three countries than measures which affect the relative growth of demand more directly, notably fiscal policy.
The authors also concluded that the boom in the United States from 1983-4 and the appreciation of the dollar and several other non-EMS currencies between 1980 and 1985 improved trade balances significantly for the main EMS economies. These economies display a comparable sensitivity to events outside the EMS, but the beneficial effects of these events up to 1985-6 were more valuable to France and Italy, which were in that period trying to disinflate and maintain their parities within the EMS. As these favourable factors have been reversed by the recent depreciation of the dollar and the con vergence of US and European growth rates, difficulties have reappeared in all EMS countries except Germany, where slow growth has implied a rising trade surplus despite the unfavourable external factors.
If demand were to grow in parallel at home and abroad, Germany and Italy would experience no major trade imbalance. For France, however, the income elasticity of demand for imports seems to be well above the elasticity of demand for French goods by her trading partners in the EMS and elsewhere. If growth rates are similar, France will therefore experience a steady deterioration in its trade balance: this must be corrected through an improvement in competitiveness, via price or non-price changes. Vona and Bini Smaghi's paper suggested that tensions, in the form of growing trade imbalances, have been building up within the EMS and may threaten the system.

Capital Mobility and Monetary Coordination

Other papers presented to the conference explored the performance of the EMS in the short term and the degree of monetary coordination required to underpin the system. Most observers have concluded that the EMS has succeeded in reducing exchange rate volatility among its participants. The reduction is obvious if the comparison is made with either the pre-EMS performance of the participating currencies or with a sample of non-EMS currencies. Nor does stability inside the system appear to have been achieved at the expense of more instability vis-à-vis the currencies outside the EMS. Michael Artis (University of Manchester and CEPR) and Mark Taylor (University of Dundee and CEPR) went beyond these straightforward comparisons in their paper, `Exchange Rates and the EMS: Assessing the Track Record'. They computed the conditional variance of EMS exchange rates, taking into account the higher degree of stability in the underlying macroeconomic determinants and without making assumptions as to the specific characteristics of the distribution of exchange rate changes. The more comprehensive approach used by Artis and Taylor suggested similar conclusions: since 1979 the monthly volatility of the nominal and real exchange rates of the DM against five other EMS currencies fell significantly, while similar measures of volatility for the dollar (and to a lesser extent sterling) have increased.
Artis and Taylor also investigated whether the EMS has merely transferred volatility from the foreign exchange markets to domestic financial markets. One plausible hypothesis is that stabilization of the exchange rate has required more activist monetary policies and more flexible use of interest rate diff erentials to sustain the EMS parities. This was not borne out by the empirical work of Artis and Taylor; the volatility of short-term nominal interest rates in all four of the markets considered (Germany, the Netherlands, France and Italy) had fallen since 1979, though the reduction was statistically significant only in the Netherlands and Italy. This is in contrast to what has been observed in countries with floating exchange rates.
Some concern centres on the capacity of the EMS to limit not merely short-run volatility but also longer-term misalignments, i.e. departures from a structure of intra-EMS exchange rates consistent with macroeconomic equilibrium. Under the EMS, a steady real depreciation of some currencies (DM, guilder and Belgian franc) has persisted along with real appreciation of the remaining currencies, though these misalignments are now growing more slowly.
The authors concluded by noting that despite the progress achieved within the EMS, the system has not yet been successful in rendering the participating currencies close substitutes. The simple uncovered interest rate parity condition is closer to being fulfilled between the DM and the dollar than between the DM and other EMS currencies. The risk premium has not been reduced to zero even for the guilder. That suggests that financial markets still question the credibility of the longer-run commitment of the EMS countries to develop into an area of more permanently fixed exchange rates.
The papers presented by John Driffill and Maurice Obstfeld studied whether a fixed-but-adjustable exchange rate system such as the EMS was compatible with perfect capital mobility and a high degree of forward-looking behaviour in financial markets. This analytical extension is timely, because the removal of capital controls has already proceeded well beyond the stage reached during the sample period used for estimation in most of the empirical work reported at the conference. In addition, further steps towards full liberalization of capital movements and of trade in financial services have been proposed by the European Commission and are currently under debate in the Council of Ministers. Can the EMS reconcile stable exchange rates and perfect capital mobility?
John Driffill (University of Southampton and CEPR) in his paper, `The Sustainability and Stability of the EMS with Perfect Capital Markets', suggested that an EMS without capital controls did not have particularly drastic implications for the macroeconomic policies of member countries. Inflation differentials and other indicators of divergence have fallen sufficiently to be accommodated by realignments, at roughly annual intervals, which would be modest enough not to disrupt foreign exchange markets. Analogies to the experience of the final years of the Bretton Woods system, or the literature on collapsing exchange rate regimes inspired by Latin American experiments in currency stabilization and financial liberalization, are likely to be misleading because the residual tensions and the prospects of large speculative gains are much more modest in the present EMS. The EMS has developed a reputation as a relatively robust system. This has been achieved under conditions of less than perfect capital mobility and rests on the notion that EMS governments have gradually precommitted themselves not to allow exchange rates to jump. In order to preserve this reputation, realignments cannot become too infrequent and their timing should continue to be uncertain as well.
Driffill's discussion of the rules of monetary management which could underpin the EMS in the absence of capital controls took as its point of departure a two-country model by Buiter, where the two central banks continuously react to incipient reserve flows by adjusting their domestic credit expansion to accommodate shifts in money demand; hence reserve flows are minimized. A less constraining mechanism can be envisaged for the EMS which has large reciprocal credit facilities for the participating central banks and in which reserves are sizeable. But the stability of the system still requires domestic policy instruments to respond promptly to reserve flows.
In his paper, entitled `Competitiveness, Realignment and Speculation: The Role of Financial Markets', Maurice Obstfeld (University of Pennsylvania) took a more sceptical view of the compatibility of free capital flows and an exchange rate system in which parity changes can rationally be expected from time to time. Obstfeld explored what kinds of equilibria can arise when markets expect a loss of competitiveness of domestically produced goods to lead to realignment. If such behaviour has been observed on the part of the authorities for some time and if realignments have typically restored competitiveness the type of constant real exchange rate maintained through a crawling peg, as suggested by Dornbusch fears of devaluation may force a collapse and an inflationary spiral. If, on the other hand, it is apparent that realignments typically do not fully offset cumulative excess inflation since the previous realignment, then a spiral of inflation and devaluation cannot occur. This appears to have been the case in the EMS since 1983: the realignments of April 1986 and January 1987 between the DM and the French franc were only about half the size required to remove fully the real appreciation of the franc since 1983.
Obstfeld also discussed a second model, in which multiple equilibria are possible if expected realignment is driven by expansionary fiscal policy in the home country. If past behaviour suggests to the private sector that the government will ultimately fully offset the real appreciation resulting from a fiscal expansion, an exchange rate collapse will occur when the expansion is announced but before there is any real appreciation. This line of analysis captured some features of the early EMS experience: even though capital mobility was far from perfect, on occasion capital flight from France followed the announcement of a more expansionary fiscal policy in the 1981-3 period. This perceived risk of engaging in solitary fiscal expansion may have paralysed fiscal policy in some European countries. Obstfeld's analysis provided arguments either for undertaking fiscal action only as part of a coordinated, though possibly differentiated, package in the EMS countries, or for moving to a more tightly managed exchange rate system and ultimately a fixed-rate union in which destabilizing expectations of devaluation and inflation would be effectively curtailed.
Obstfeld's paper raised important issues for the stability of the EMS, which has proceeded to full removal of capital controls before achieving an underlying convergence of inflation rates and the coordination of the macroeconomic policies which are seen to influence convergence. The discussion of Obstfeld's paper suggested that he placed too much emphasis on the likelihood of instability and an inflationary spiral. This seemed to depend more on his assumption of a high degree of foresight on the part of economic agents as to the consequences of government policies and the actions governments will take in the EMS than it did on the assumption of high capital mobility. The uncertainty of the timing and size of EMS realignments and the practice of giving no assurance to `inflationary' countries that they can restore competitiveness fully may well be enough to stabilize the system in practice.
The paper entitled `Interventions, Sterilization and Monetary Policy in EMS Countries, 1979-87', by Cristina Mastropasqua, Stefano Micossi and Roberto Rinaldi (Banca d'Italia), offered detailed insights into the intervention rules and practices in the EMS, as well as case studies of the interaction between the domestic and external components of money creation in four countries (Belgium, France, Germany and Italy). Their empirical work suggested that to date, interventions have not been shared equally between Germany and the other countries studied, particularly since the Bundesbank has not until very recently participated to any major extent in `intra-marginal' interventions.
The authors attempted to measure asymmetries between member countries in the degree to which they have preserved monetary autonomy. One such indicator is the extent of sterilized intervention. Using quarterly data for 1979-86 they found that whereas Belgium, France and Italy have on average sterilized 30-40% of the net change in the foreign assets of their central banks, that proportion is 60-80% for Germany. This rough measure suggested that the Bundesbank has had twice the capacity of its EMS colleagues to sterilize its relatively more modest interventions in the foreign exchange markets in order to keep its preferred domestic monetary aggregate close to target. This difference has been observed despite the existence of some capital controls which should have facilitated sterilization in the three non-German countries. One might expect the asymmetry to widen, ceteris paribus, as one moves towards perfect capital mobility. The authors' findings confirmed an asymmetric interpretation of the EMS, in which Germany pursues a domestic monetary target and sterilizes external flows which are likely to upset its pursuit of this target, while the other EMS countries conduct policy in terms of domestic credit expansion, allowing external flows to influence domestic monetary conditions. Greater symmetry will, the authors concluded, require a degree of consensus on medium-term policy objectives that simply does not exist today.
The extent of this consensus also featured in `Monetary Policy Coordination Within the EMS: Is There a Rule?', by Massimo Russo (IMF) and Giuseppe Tullio (Banca d'Italia), which focused on procedures for decentralized or centralized monetary management in the EMS. Russo and Tullio examined the explicit or implicit adjustment mechanisms under the gold standard and the Bretton Woods system and discussed how the proposals for international monetary reform by McKinnon and by Williamson and Miller could be implemented within a reformed EMS.
The simplest idea, but also the most radical, would be to fix an aggregate monetary objective for the EMS countries combined. Preliminary research reported in the paper suggested that an aggregate EMS money demand function could prove to be at least as stable as the money demand function in Germany. But would it be possible for the EMS countries to agree explicitly on a common inflation objective, such as would be required to set a collective monetary target? How would the national components be assigned in the absence of a fully fledged European central bank?
A second, very interesting possibility is to try to replicate the desirable properties of the first proposal by a decentralized rule in which all participants including Germany set a national objective for domestic credit expansion but agree to refrain from sterilizing the interventions which would be required to sustain fixed exchange rates in such a system. This would remove the asymmetry documented by Mastropasqua, Micossi and Rinaldi. If there were no net interventions in dollars and other non-EMS currencies, aggregate domestic credit expansion in the EMS countries would be equal to total money creation and the participants might then jointly consider whether monetary growth for the area as a whole was acceptable.
The paper also considered two other, partly decentralized procedures inspired by recent proposals to reform the international monetary system: a nominal income target for each EMS participant and a monetary rule for each country, subject to a common inflation target, e.g. stable wholesale or traded goods prices for the EMS on average. These proposals have the virtue of addressing explicitly the issue of the nominal anchor for the system. The second proposal might be suitable as a medium-term objective for a European central bank. There might even be an advantage in declaring them formally in order to confirm the autonomy of such an institution and help to build its reputation. But these proposals faced operational problems in designing the feedback rules needed to link departures from price stability to the changes in national monetary growth rates.

The Future of the EMS

A presentation by Tommaso Padoa-Schioppa (Banca d'Italia) and a concluding panel discussion looked at how the system might evolve. Padoa-Schioppa viewed the achievements of the present EMS as positive and important. Both economic analysis and historical experience led him to predict, however, that the system was likely to prove inadequate as a mechanism for reconciling the contradictions that will arise from trying to move towards the internal market and full-scale financial integration by 1992, while at the same time maintaining a high and increasing degree of nominal exchange rate rigidity and some national monetary autonomy. Of this quartet of policy objectives, national monetary policy was the most likely candidate for abandonment. Padoa-Schioppa re-stated the case for an ultimate European monetary union and a common money, but more constructively, he showed how the EMS could evolve towards that long-run objective in the phase where national currencies still coexist and the decisions of monetary management which have traditionally been national are gradually upgraded to the European level. The role of a parallel currency, the ECU, was critical in building the foundations for European monetary control in the form of a joint monetary base for the participating central banks. He argued for the exercise, at a European level, of the discretionary powers essential to good monetary management once a broad consensus on the basic rules of conduct are established. The conference suggested that we are nearer to that stage than most observers recognize. Padoa- Schioppa argued that further steps in this direction need not await progress towards political unification, such as institutionalized budgetary powers for the union.
The European Monetary System, edited by Francesco Giavazzi, Stefano Micossi and Marcus Miller, will be published by Cambridge University Press in September (ISBN 0 521 36271 7). It will contain revised versions of the papers presented at the Perugia conference, along with discussants' remarks and the concluding panel discussion. It will also include an introduction by Professor Niels Thygesen, on which this Bulletin report is based.