EMS
Explaining mysterious stability

The European Monetary System (EMS) has weathered its first seven years remarkably well, with only three major realignments between its inception in March 1979 and the realignment of April 1986. On all three occasions, the value of the French franc relative to the Deutschmark fell by between 3.5 and 5.5% from one market day to the next. Any market operator who had foreseen the exact dates of any of these realignments made a fortune. Indeed, Paul Krugman and others have shown that if the market had perfect knowledge, a speculative attack would have come before the first incident. Moreover, the attack would have been of such force and timing that no realignment would have been possible thereafter: France would simply have been forced to leave the system and let its currency float.

There is therefore a certain mystery about the survival of the EMS. The considerable dynamic stability displayed by the EMS is even more puzzling. In the months preceding each of the first three realignments, France lost significant reserves. All three realignments arrested the fall in French reserves, and more significantly, two of them even brought reserve inflows. For the period as a whole, there is no clear downward trend in French reserves, which have now returned to their pre-EMS level. The tendency for the French interest rate to rise relative to the German during periods of stress on the franc or prior to realignments is easily understood. It is harder to explain the pronounced tendency for the French interest rate to stay well above the German one at all times, even right after the realignments when the market can be certain that no further realignment is imminent. The interest rate differential in favour of France at such times ought to provoke a speculative attack on the Deutschmark.

In Discussion Paper No. 96, Philippe Michel and Research Fellow Jacques Melitz attempt to provide a model that reflects these characteristics of the EMS, and in particular displays the same dynamic stability. Their model concentrates on the two largest EMS members, France and Germany. There is perfect capital mobility in the model, but large-scale speculative movements of capital do not play a major role. The authors argue that the characteristic behaviour of the EMS can be reproduced in a model where the only speculation involves 'leads and lags', that is movements in the desired levels of foreign balances held to finance external trade. Michel and Melitz also argue that such a model can be shown to be dynamically stable.

France and Germany each produce a good consumed by both countries; the price of the French good is assumed to rise faster than that of the German, perhaps because of wage pressure. As a result, after every realignment of the EMS parity the competitive position of France worsens as its inflation outstrips that of Germany and the French current account deteriorates. Eventually the parity must be realigned in order to restore the French real exchange rate and its balance of payments equilibrium. Melitz and Michel assume that the two governments deliberately create uncertainty concerning the precise timing of the realignment, and this is a key feature of their model. Each government is assumed to pursue two targets, one for output and one for reserves, but between realignments the domestic interest rate is the only instrument available to policy-makers.

Transactions balances are the key to the behaviour of the foreign exchange markets in Melitz and Michel's model. The French are assumed to hold Deutschmarks in order to finance their purchases of German goods, while Germans hold francs in order to purchase French goods. Since inflation is higher in France than in Germany, the French current account deteriorates over time. As French imports rise, the French require higher Deutschmark deposits in order to finance their purchases of German goods. How large a stock of German deposits the French wish to hold depends on the exchange rate differential and the risk of a fall in the value of the franc when the (inevitable) realignment occurs. The closer is the expected date of the realignment, the greater the exchange rate risk, and holdings of transactions balances will therefore include a speculative component. But in contrast to speculative capital movements, these speculative shifts in transactions balances will be limited by the need to hold some balances in order to carry on trade.

Melitz and Michel then trace the behaviour of their model after a realignment. They assume that the realignment is sufficient to give France a balance of payments surplus initially, during which time it accumulates reserves. If there were no possibility of a further realignment, the model predicts that initially French interest rates will fall and German rates will rise. As the current account deteriorates and reserves fall, however, French interest rates must rise. The possibility of realignment brings the anticipated depreciation of the franc into play. This will tend to produce a higher French interest rate and a lower German one. There will thus be a tendency for the French interest rate to fall at first after a realignment and then rise.

When the realignment occurs, the model correctly predicts that French reserves will rise and German reserves will fall. French reserves increase because the speculative element in transactions balances has disappeared - there is no longer an imminent devaluation of the franc. Prior to the realignment, the prospect of a devaluation in the franc caused German importers to accept higher transactions costs than they would otherwise in an effort to reduce their franc balances. French importers on the other hand allowed their Deutschmark holdings to rise. After realignment, the Germans are anxious to benefit from the convenience of higher franc transactions balances: Frenchmen, having speculated in a fashion, convert some of their Deutschmarks. In addition, the realignment throws the French balance of payments into (temporary) surplus, adding to reserves.

Melitz and Michel find that two conditions are necessary for the model to display the stability which is observed in the EMS. First, the realignment must provide the high-inflation country with an initial advantage in its terms of trade which allows it to build up its reserves in the period immediately following the realignment. In addition, it is essential that policy-makers in the high-inflation country (France) be more concerned with their reserve targets, rather than their output targets, than are the German policy-makers.

Melitz and Michel note that their model does not explicitly model purely speculative capital flows, but is nevertheless able to reproduce the characteristic behaviour of the EMS. They discuss how such speculative movements might be incorporated into their model. They also suggest explanations for the observed absence of speculation within the EMS. Capital controls were not sufficiently stringent to explain the absence of speculation against the franc. They conjecture that a more important reason for the absence of speculation lies in the wide (4.5%) margins of fluctuation in the parities. The ability of the franc to fall in value within the band permits the French interest rate to exceed the German without inducing a speculative attack against the Deutschmark. At a later date, when pressure on the franc grows and the problem becomes one of avoiding a massive movement out of francs, the uncertainty of the date of the next realignment is crucial to the stability of the system. But it is not enough, the authors conclude: the threat of capital controls is also essential. And if this threat is to be effective, it may sometimes be necessary to impose controls.



The Dynamic Stability of the European Monetary System
J Melitz and P Michel

Discussion Paper No. 96, March 1986 (IM)