Background to the Crisis
In 1979, the European Monetary System (EMS) was founded. Twelve countries now take part in the system: The Netherlands, Germany, Belgium, Luxembourg, Greece, Denmark, France, Ireland, Italy, Spain (since 1989), UK (since 1990) and Portugal (since 1992). The EMS was not designed to act as a fixed rate system, but it was hoped that exchange rate stability could be achieved through the convergence of economic performance. Within the EMS existed the Exchange Rate Mechanism, which contained ten currencies (Greece is not a member of the ERM, while the Luxembourg franc is set at par to the Belgian franc). Between May 1979 and January 1987, there were twelve realignments of ERM-parities, largely due to cost - and price - differentials.
The need for realignments reflected the persistence of inflation differentials across ERM-countries, but no realignments took place between 1987 and 1992. After 1987 the inflation differentials narrowed, though substantial differences remained between Italy, UK and Spain on one hand and Germany on the other. The last realignment in 1987 led to revisions of the EMS arrangements in the Basle-Nyborg agreement. This agreement was based on two assumptions:
(i) Germany would continue to have low inflation and interest rates.
(ii)The Bundesbank would defend exchange rates, provided that members of the system made the necessary interest rate adjustments to protect their currencies.
The first assumption seemed reasonable, who knew that the Wall would be pushed down in two years? The second assumption overlooked the fact that at some point the Bundesbank could be faced with a situation in which its obligation to intervene would be incompatible with control over German monetary conditions. Another part of the agreement was the extension of credit facilities for longer periods. For the first time it was permitted to draw on credits before a currency reached the limit of its EMS band. These innovations led to much confidence, so that the policymakers discarded the realignment option.
Following the Delors report of 1989, the Maastricht Treaty (1991) laid down four convergence criteria for a country to qualify for participation in European Monetary Union (EMU). In the first transition stage, capital controls were eliminated in 1990 to complete the internal market (see appendix 1 for the convergence criteria and the three stages to EMU). This convergence between EMS-countries (on top of the stability of the EMS), strengthened the confidence in the system, and increased the commitment to exchange rate stability. For a time, the no realignment - no controls strategy seemed to work, even in the face of persistent inflation differentials.
The Upheavals in 1992 and 1993
Everything seemed to be going well in the EMS. It not only survived after 1979, but grew and prospered. During the process of ratifying the Maastricht Treaty, strong public doubts about the EMU began to be voiced. In the Danish referendum in June 1992, the result was a small majority against the Treaty. Meanwhile the expectations regarding the French referendum were very uncertain. Indeed, the EMU became a very contentious issue in the French referendum. At the same time, the Bundesbank raised its key interest rate, the discount rate, because of inflationary tendencies resulting from a large fiscal deficit, caused by the unification. This happened at a time when other EMS countries needed lower interest rates to get them out of recession.
Tension within the ERM began to build up from mid-July 1992, concentrating initially on the lira, then on sterling and then on a variety of other currencies. The pressure on the Finnish Markka was so strong that it abandoned its peg with the ECU. Italy raised its interest rates, but still the lira weakened repeatedly. The Bundesbank did not cut its interest rates enough and speculation continued. The pressures on both lira and sterling became so large, that both were suspended from the ERM (and depreciated rapidly afterwards). The foreign exchange markets remained disturbed for the rest of the year, with a renewed outbreak of speculative pressures leading to the abandonment of Sweden's peg to the ECU, devaluation of both the Portuguese escudo and the Spanish peseta in November and the abandonment of Norway's ECU-peg in December. In January 1993, Ireland witnessed pressure (mostly because of the loss of competitiveness vis-à-vis the UK, that devalued in September 1992) and finally devalued by 10%. Germany reduced interest rates in February, March and April and the pressure was eased a bit for the currencies that had not yet been realigned. Particulary in the countries with relative moderate increases in costs and prices, though, the calls for further reductions in interest rates intensified.
The second Danish referendum on Maastricht produced a small majority in favour of the treaty. The new government in France made clear its commitment to the 'franc fort' policy, that is keeping the franc at its existing parity. The government also wanted lower interest rates to relieve the recession, and it appeared willing to challenge the German economic authorities publicly. On 18th June the French money market intervention rate was pushed below the German rates, which received some scepticism in the markets. Speculative pressures within the ERM and against the French franc in particular, developed strongly during the course of July. The Banque de France was forced to raise its interest rate to prevent the franc from falling through its ERM lower band. The Bundesbank did not lower its discount rate, and massive sales of the French franc, Belgian franc, Danish krone, Spanish peseta and Portuguese escudo occurred. Massive intervention was necessary to keep these currencies just above their ERM floor. On 2nd August, EC monetary officials and finance ministers finally agreed that the ERM bands should be widened from 2.25% to 15% (except for the Dutch-German one), though no central parities should be changed. With the wider bands, the system would be less vulnerable to speculation.
The Factors Behind the Crisis
Competitiveness Problems
According to this view, certain countries experienced persistent inflation and rising labour costs. The inflation differentials against Germany were declining, but remained positive. This undermined competitiveness of the traded goods sector. The markets identified the countries with these competitiveness problems and attacked their currencies once devaluation was overdue. The countries whose exchange rates have been under pressure fall into three categories:
(1) Italy, which shows clear signs of deteriorating competitiveness. Italy's labour costs rose progressively through the 1980's and was the first country under attack.
(2) Spain and the UK, which had some competitive difficulties particularly in rising relative labour costs in the period 1987-1991, but evidence is unclear.
(3) Denmark, France and Ireland had no competitive difficulties.
Since only Italy has evident deteriorating competitiveness, the conclusion can be drawn that the divergent movement of prices and labour costs played a limited part in the crisis.
German Unification
German unification generated a real exchange rate (RER) appreciation of the Deutschmark (DM). RER is the exchange rate weighted by the relative price level between other ERM countries and Germany. i.e.
RER = E x (P*/P)
P* : price level in the other ERM countries
P : German price level
So how did this real appreciation occur?
First, the East German wages increased (because of one-to-one conversion) and raised the costs and price of East German output and reduced its quantity. As can be seen in figure 1, the East German wage explosion shifted the German supply curve upwards (from S0 to S1). Second, the large transfers from West to East Germany after the supply shock caused a public sector deficit. The result was a rise in real interest rates in Germany, and the net capital export turned into net capital import. The net capital import caused more world savings being spent on German goods, so that the relative price of German goods had to rise. This expansionary demand shock is reflected in the upward shift of the demand curve (from D0 to D1). (See figure 1).
Figure 1
P/(ExP*)
= S1
1/RER
1 S0
0
D1
D0
q1 q0 German output (q)
The demand shock was merely designed to compensate partly for the supply shock, so the shift in demand was less than the one in supply. As a result, there is a decline in German output, and both shocks resulted in a real appreciation, so that 1 / RER, that is, the relative price of German goods compared to foreign goods (i.e. P / (E x P*) ) has gone up. When you look at the RER (in terms of changes):
Er = En + P* - P
one can see that real appreciation of the DM (Er < 0), without a nominal adjustment (En = 0), has to result in higher prices in Germany (P > 0) or lower prices in the other ERM-countries (P* < 0). As a result of the German aversion to inflation, though, much of the burden needs to be in the form of deflation in other countries. The required decline in inflation was largest in the UK and Italy and they were also the first under attack and the only ones that stepped out of the ERM. The Maastricht Treaty strengthened the Bundesbank's determination to reduce the German inflation rate. The reduction in inflation would be attained by a rise in the nominal interest rate. The other ERM currencies had to follow to be able to attract investors in their currencies. This rise in interest rate was a detrimental thing for the countries that did not have significant inflationary problems, but were experiencing unemployment and were in recession.
Inevitable Policy Shifts
Even if currencies were consistent with the maintenance of ERM parities, the markets could have been anticipating a shift in future policies. They noticed that the policies to defend the existing parities (with the interest rate instrument) gave rise to growing unemployment. The markets therefore expected a relaxation of future policy and the speculators attacked the currency and tested the governments on how much of its potential reserves it was willing to use. The speculators are faced with a 'one way option'; they do not lose by speculating against the currency, even if fears of abandonment of fixed rates proved unjustified.
This explanation can be analysed more systematically by using a one-country version of a model in the tradition of Mundell-Fleming.
Figure 2 : One-country version of Mundell-Fleming model
(1) m-p = ay - bi (money market equilibrium)
(2) y = h (e-p) - kr (goods market equilibrium)
(3) i = r + p
(4) p = cy (naive Phillips-curve)
(5) i = i* + e (uncovered interest parity)
The model reduces to:
(6) e = [ ah (e-p) + (1-kc) (p-m) ]
(7) p = c [ bh (e-p) - k (p-m) ]
(assumed is that = [b + k (a-bc)]-1 is positive, so that the system is a saddle-path stable)
e S1
45
S0
M1
e0=M0 S1
S0
M0 M1 p
This figure represents the long run equilibrium with a money supply Mo and an associated stable convergence path S0So.
When one looks at figure 2 above, it can be seen that the system stays in A with a pegged exchange rate eo= mo if policy is not expected to change. If there is an expectation of a future relaxation of policy (from M0 to M1), the new long run equilibrium will be in D. When markets realise this, they attack the currency, exhausting the government's reserves, forcing them to abandon the peg. The period of floating begins with a jump depreciation from A to B. Following this depreciation, the exchange rate continues to depreciate (in C, the money supply is increased to M1). In this view, the crisis is explained as a consequence of market anticipation of an inevitable shift in monetary policy, provoked by rising unemployment. When governments decided to shift to less restrictive policies, though they were presumably looking at the costs (unemployment) as well as at the benefits (qualifying for monetary union) of the restrictive policies. When the benefits seemed to become less and less (Denmark's initial rejection of the Maastricht Treaty), the governments were more likely to shift their policies, and the markets anticipated these shifts.
Self-fulfilling Speculative Attacks
According to the previous argument, balance of payments crises (this is when speculators acquire a large portion of the Central Bank's foreign reserves as the bank attempts in vain to support its currency) represent an entirely rational market response to persistently conflicting internal (lower unemployment) and external (exchange rate stability) targets. In that view, the exchange rate is only attacked if there already exists a balance of payments problem implying the eventual exhaustion of reserves. The model has a unique equilibrium. According to Obstfeld (1986), though, balance of payments crises may be purely self-fulfilling events rather than the inevitable result of unsustainable macroeconomic policies. In this model, markets do not anticipate the crisis, but provoke it. Without the attack, there would be no balance of payments problem.
For illustration, the single-country model in figure 2 can be used again. The result of an attack depends on the Central Bank's reaction. The Central Bank can be 'wet' (CB cares more about unemployment than inflation) or 'hard-nosed' (CB is inflation-fighter). So, when the CB is 'wet', the money supply will increase from Mo to M1 (see figure 3 below). In case of a speculative attack, the exchange rate would jump to C (overshooting depreciation) and over time, the exchange rate would appreciate towards point B along the stable path S1 S1. This attack is self-fulfilling. The currency is weak because of the lack of credibility of the monetary authority. The depreciation will be just what the speculators wanted. A possible strategy can be the forward sale of the threatened currency. If the speculators sell the weak currency now on a specified date in the future, for the present value of the currency, they will make a profit if the currency depreciates. After the depreciation, they can buy the currency for the new, low price and sell for the high price, as agreed in the forward sale contract.
Figure 3
e S1 45
S0
M1 S2
Mo
M2
S1
S0
S2
p
M2 Mo M1
If the CB is 'hard-nosed', there will be a contraction of the money supply (from M0 to M2, see figure 3) in case of an attack. Now, the currency will immediately appreciate to E (overshooting) and will over time depreciate to D, along the stable path S2 S2. The contraction of the money supply thus results in an appreciation, and if speculators know this is going to happen, they will never attack the currency (if one looks at the forward sale strategy, one can see that the speculators will lose when the currency appreciates after an attack).
The events of 1992 confirm that self-fulfilling speculative attacks can occur in practice. There was/is uncertainty about the exact date of the initiation of stage three and the EMS-countries did not know how much time they would get to recover from the 1992 attacks. If they're not sure about qualifying for EMU anyway, then what is the use in defending the currency by a decline in the money supply (i.e. rise in interest rate), if it hurts the economy for sure but the EMU qualification is not certain. These circumstances created scope for self-fulfilling attacks.
Options for the EMS
A Reconstructed ERM
A narrowing of the bands would be dangerous if it would take place at misaligned rates. A re-narrowing on the basis of existing parities remains a prospect that should be resisted until recovery is well under way and fiscal positions are improving. One easily assumes that any change would be likely to be in the direction that central parities would fall, i.e. that other currencies should devalue against the DM. A variant of this would convert the 15% band into a DM float by reinstating narrow bands first for the non-DM currencies. France would then be the anchor for such a concerted float. This is not so surprising, since France has notably low inflation and quite a good control over its fiscal position. The damage to France (and the other countries involved) with respect to floating down against the DM would be minimal. But the two obstacles here are the Netherlands (stays tied to the DM) and of course the Bundesbank.
If countries succeed in reinstating narrow bands, whether around unaltered or new ones, the management of the ERM has to change, so that the system works in a way which does not invite further crisis. One can distinguish at least two distinct types of reform: capital controls and increased symmetry.
In the early period of the EMS, capital controls played a central role in the successful management of the system. The reimposition of capital controls is again being spoken of. A proposal is the imposition of a 'Tobin tax' on foreign exchange market transactions, designed to reduce speculation. The administratively most feasible option appears to be the incremental reserve requirements on financial institutions taking open positions in the foreign exchange market. However, the political and economic objections to the reimposition of controls are considerable.
If a more symmetric system is to develop, the Bundesbank must be persuaded to set a more 'European' policy. There are two possible routes to achieve this. The first one is that, in the light of the close links of the goods markets, the monetary aggregates in Germany's EMS-partners are likely to prove informative for German inflation. Correcting policy implementation by Germany would lead to more efficient outcomes. The other one is more political, that is for the Bundesbank to accept more often the recommendations of the European Monetary Institute. After all, in stage three the Bundesbank will lose all its power to the European Central Bank (ECB). Without more symmetry, a re-run of the crisis is more likely.
Monetary Union
A monetary union would remove problems like exchange rate stability, monetary asymmetry, speculation and intervention obligations. Currencies cannot fluctuate and a ECB would substitute European monetary policy for a German one, aimed at price stability in Europe and not in Germany alone. To qualify for monetary union, though, the member countries should have participated in the 'normal' bands of the ERM without 'severe tension' for a period of two years. The normal band used to be 2.25%, but the present participants now simply wish to declare the 15% band as normal. The wider band offers no targets for speculation and does not affect the basic criterion of the long term interest rate. However, the necessary framework of the ERM cannot be created overnight. More importantly, Europe is not an optimal currency area in terms of wage and price flexibility, labour mobility and cross-border investments, so it might not be so wise to rush into a monetary union.
A Federation of Central Banks
There are concerns about EMU on the grounds of national sovereignty. Under the gold standard though, it was also possible to stabilise exchange rates within narrow bands. The reason for this was that government's commitments to gold parity were supported by the prestige and the independence of the Central Banks. This is a feature which might be possible to reconstruct. For any restoration of the ERM, the independent status of the Central Banks might be necessary, and it is the case that the Maastricht Treaty requires Central Banks to become independent during the course of stage two.
Floating Exchange Rates
A period of floating rates could be seen as a way of putting the EMS back together with more certainty. It would give Germany time to sort out reunification and overcome its domestic problems in order to resume delivering low inflation and interest rates. Other countries would have the opportunity to loosen tight monetary policies which are pushing up unemployment and aggravating the fiscal situation. There is the worry, though, that floating rates might undo some of the integration that has already been achieved. In Europe, a more co-operative arrangement could be sought to stabilise the behaviour of floating rates, via the co-ordination of monetary policy. This, in fact, is what the European Monetary Institute is designed to do.
Conclusion
Since its initiation in 1979, the EMS has had some notable successes. The absence of realignments between 1987 and 1992 created much confidence in the system, which even increased after the proposed intensified monetary co-ordination ratified in the Maastricht Treaty in 1991. Nevertheless, despite the apparent success, underlying factors were still diverging, and market participants understood what was going on. The factors behind the crisis have been assessed in this paper: competitiveness problems, German unification, inevitable policy shifts and self-fulfilling attacks. The crisis eventually led to a widening of the bands to 15% (except for the Dutch-German one, that stayed at 2.25%). For the future, a re-narrowing of the bands is an option, but in this case, the management of the ERM will have to change, so that the system will work in a way that it will not invite a further crisis.
Appendix 1
The convergence criteria for a country to qualify for participation in EMU are:
The Treaty specified a transition in stages. The first one, beginning with a removal of capital controls in 1990, was meant to reduce inflation and interest rate differentials and the stability of exchange rates. Stage two, beginning on 1st January 1994, was to prepare for monetary union (creation of the European Monetary Institute) and stage three will be marked by the inauguration of monetary union and the establishment of the European Central Bank.
Bibliography
Artis (1993) The prospects for EMU after Maastricht..and Black Wednesday and all of that, Manchester University Press, Manchester
Artis & Lewis (1993) Has the EMS a future?, University of Nottingham, Nottingham
Cobham (1994) European monetary upheavals, Manchester University Press, Manchester
Davis Unsustainable parities: the ERM turmoil, Trinity economic papers, policy paper no.3., October 1994, Dublin
Eichengreen & Wyplosz The unstable EMS, CEPR discussion paper no. 817, May 1993, London
The Economist, 31 July 1993, p.69-70
The Economist, 7 August 1993, p.19-20
Krugman, A model of Balance of Payments crises, Journal of money, credit and banking, 1979, nr.11, p.311-325
Obstfeld, Rational and self-fulfilling Balance of Payments crises, American Economic Review, 1986, nr. 76, p.72-81