Chapter

CHAPTER 6 Credibility and Asymmetries in the EMS

Author(s):
International Monetary Fund
Published Date:
September 1990
Share
  • ShareShare
Show Summary Details
Author(s)
Michele Fratianni and Jürgen von Hagen 

I. Introduction

In light of the skepticism with which it was initially received, the European Monetary System (EMS) has surprised many by the sheer fact of its survival. Indeed, the majority of the economics profession—arguing on the strength of conventional open-economy macroeconomics—predicted the inevitable failure of an exchange rate arrangement where high-inflation France and Italy would have to coexist with low-inflation Germany.1 Yet, the EMS is the only lasting formal exchange rate arrangement among the industrialized countries since the breakdown of the Bretton Woods system. Moreover, the EMS is now regarded as a launching pad for a European monetary union, as is clearly stated in the Delors Report (1989).

What are the benefits and achievements of the EMS? Despite a 10-year experience with the system, the literature has yet to reach a consensus on this question. Two different views, although not mutually exclusive, have emerged over the years. The first regards the EMS as a shock absorber and an instrument to reduce the variability of output and inflation. The second sees the EMS primarily as an instrument to alter inflation trends in the region.

The shock-absorber interpretation of the EMS stresses the cooperative character of the system and regards the fixed exchange rate mechanism of the EMS as a substitute for more complete, welfare-improving monetary policy coordination. Complete coordination, though preferable, does not work for much the same reasons that cartels break down: once an optimal cooperative policy has been agreed upon by the group, each individual country finds it best for its own benefit to “cheat” and deviate from the arrangement.

Arrangements for international policy coordination therefore need an observable policy standard that is easy to monitor to prevent cheating, such as a fixed exchange rate (Canzoneri and Gray, 1985; Melitz, 1985; Laskar, 1986; and Fratianni and von Hagen, 1990). According to this interpretation, policy coordination in the EMS improves the system’s response to aggregate shocks from outside the region, such as oil price shocks and swings in U.S. monetary and fiscal policy. That is, exchange rate policies in the EMS serve as a shield against external disturbances.

At the same time, the EMS has no obvious force to alter its members’ inflation trends. So long as realignments are possible, inflation trends can be chosen in accordance with national preferences. The early literature supporting the EMS rested on this view and emphasized two structural characteristics that would make the EMS a new kind of exchange rate arrangement: symmetry, in the sense that all member countries would share equally the burden of adjustment to balance-of-payments disequilibria; and flexibility, meaning that no country would be forced to adopt an inflation trend other than its own, most preferred one. Symmetry and flexibility were embedded in the EMS through the creation of the basket currency, the European currency unit (ECU), an indicator of divergence, a system of financial support, and a provision for realignments (Fratianni and von Hagen, 1989 and 1990).

More recently, a second interpretation of the EMS has emerged and gained popularity: the credibility hypothesis. The key ingredient in this hypothesis comes from Barro and Gordon’s (1983) analysis of central bank credibility, according to which equilibrium inflation rates are determined by the credibility of a central bank’s commitment to price stability. From this perspective, the large inflation differentials in the EMS of the late 1970s reflect widely different degrees of credibility among the central banks in the region. More specifically, the Deutsche Bundesbank appears to have a comparative advantage in credibility over its counterparts in the more inflation-prone countries, like France and Italy. Based on its proven commitment to price stability in the past and its relatively large degree of political independence, the Bundesbank is able to pursue a low-inflation monetary policy. Other central banks, in contrast, are locked into unfavorable, high-inflation equilibria and are unable to convince the public of their commitment to carrying out a domestic disinflation program.

According to the credibility hypothesis, the high-inflation countries have used the fixed exchange rate mechanism in the EMS to transfer their monetary policy autonomy to the Bundesbank and thus to signal their own new commitments to price stability. By borrowing credibility from the Bundesbank, the high-inflation countries can reduce, relative to a purely domestic strategy, the cost of disinflation measured in terms of output and employment losses.

This interpretation of the EMS is similar in spirit to the proposal by Rogoff (1985) and Thompson (1981) to appoint a conservative central banker to reduce inflation expectations and, hence, equilibrium inflation rates. Exchange rate policies are a vehicle to achieve the common goal of domestic price stability.

The credibility interpretation of the EMS yields a plausible answer to the puzzle of why the larger countries in the EMS, France and Italy, have sought membership in a system in which Germany appears to be the center country. Indeed, it implies that the EMS is characterized by a structural asymmetry, namely that monetary policymaking in the system is dominated by the Deutsche Bundesbank. We call this implication the “German Dominance Hypothesis.”

In this paper we review and assess the credibility hypothesis in three stages. In Section II we discuss the theoretical aspects of the hypothesis, based on a simple, two-country model of monetary policymaking. Our main conclusion there is that the credibility hypothesis does not yield a plausible explanation of the EMS. In Section III, we take up the implication that monetary policymaking in the EMS is dominated by the Bundesbank. We argue that, in contrast to common practice, German dominance cannot be inferred from inflation data. We propose a discriminating test of the hypothesis based on two empirical models of monetary policy action in the EMS policy game, instead of the results of such action, that is, inflation rates. In the first empirical model, which is more short-run oriented, we select domestic money market rates as the relevant variable representing policy action. In the second model, which has a longer-run orientation, we consider monetary base growth rates. The tests disconfirm German dominance. The rejection of the latter does not imply that the EMS works symmetrically. In Section IV we evaluate whether the EMS countries behave asymmetrically in the general sense that each contributes more or less than its “fair” share to the EMS policy game. Section V summarizes the main points and conclusions.

II. Credibility in the EMS: Theory

During the second half of the 1980s there has been a growing consensus among European economists and policymakers that EMS membership has been important in the fight against inflation because of the role of Germany in the system. Based on Germany’s strict adherence to the principle of price stability, the Bundesbank has provided the nominal anchor to the EMS. The other members, in particular the larger and more inflation-prone France and Italy, have used this anchor to reduce their own inflation trends.

Theoretically, this view is derived from the analysis of central bank credibility (Barro and Gordon, 1983). The Bundesbank has high credibility as a central bank committed to price stability. The Banque de France and Banca d’Italia, in contrast, have low credibility. By pegging the deutsche mark price of the franc and the lira, France and Italy can adopt the Bundesbank’s policy standard. This leads necessarily to a loss in policy autonomy that is offset by a credibility gain. The end result is a lower disinflation cost than under the alternative of announcing and pursuing an autonomous stabilization program.

A Model of Monetary Policymaking and German Dominance in the EMS

Much of the literature on the credibility hypothesis has dealt with central bank credibility in static, two-country models of a monetary policy game between two central banks and their private sectors (Artus, 1988; Collins, 1988; Giavazzi and Giovannini, 1987, 1988, 1989; and Giavazzi and Pagano, 1988).2 The characteristic of this game is that the central banks choose their optimal policies after nominal wages have been set by the private sector. This leaves them with an opportunity to deviate from previously announced policies and create monetary surprises to gain extra output and employment. Being aware of this opportunity, the private sectors behave in a way to eliminate the central banks’ incentives to renege on their expected policies. In this context a monetary policy strategy is called credible if and only if it is a “sub-game perfect” Nash equilibrium, that is, the central bank has no incentive to deviate from it once nominal wages have been set. Here, we follow this approach to credibility in order to review its theoretical merits.

The model we use is adapted from Canzoneri and Henderson (1988) and Fratianni and von Hagen (1990). It focuses on the interaction of central bank policies and private sector wage-setting in two countries, Germany and Italy. In this model, displayed in Table 1, the private sectors in each country set money wages at the beginning of the period on the basis of the expected money supply, leaving the monetary authorities with the opportunity to generate monetary surprises. These authorities behave in the manner described by Barro and Gordon (1983); namely, they have employment targets, n¯ and n¯*, which differ from the natural employment levels—equal to zero in the logs. By assumption, the Italian central bank has a higher employment target than the German central bank: n¯*>n¯>0. The private sector is aware of the authorities’ preferences and, since no precommitment is assumed possible, sets expectations so as to eliminate the incentive for monetary surprises. With independent monetary policies in both countries, the sub-game perfect Nash equilibrium becomes

Table 1.A Simple Two-Country Model
There are two identical economies described by (1) a short-run production function, (2) a demand-for-labor function, (3) a money market, and (4) an aggregate demand function for goods and services:
where the unstarred country is low-inflation, high-credibility Germany and the starred country is high-inflation, low-credibility Italy. Variables are expressed in natural logarithms with the meaning:
y = output
yn = ln(1/(1 − α))
ω = nominal wage rate
p = price level of domestic output
n = employment level
m = the money stock
yd = demand for output
r = real exchange rate = nominal exchange rate + p* − p.
The consumer price levels are defined as q = p + γr and q* = p* − γr. With the initial conditions q1=q1*=0, they are equivalent to the inflation rates. Their reduced-form solutions are
where ∈ = γ2δ (1 − α).
Similarly, with m- 1 = m*- 1 = 0, m and m* are the money supply growth rates. The authorities aim at target values of employment and the overall price level:

In equilibrium, Italy’s inflation rate is higher than Germany’s because the Banca d’Italia has a higher employment target than the Bundesbank. Substituting equation (7) into (6), we obtain the associated utility levels

In equilibrium, inflation yields no employment benefits. The mere fact that the authorities aim at employment levels different from the natural ones translates into an inefficiency. The authorities would be better off if they could commit to credible noninflationary policies.

Can Italy improve on the utility level implied by equation (8) by fixing the price of the lira to the deutsche mark? Giavazzi and Giovannini (1987) and Giavazzi and Pagano (1988) argue for such a strategy on the grounds that Italy can borrow credibility from the Bundesbank and achieve disinflation at a lower cost than through a purely domestic policy. To prove this point these authors compare Italy’s utility in equation (8) with the utility attainable under fixed exchange rates, taking Germany’s policy in equation (7) as given. Pegging the exchange rate in this model requires that Italy adopt the same money growth rate as Germany, m* = m. In this sense, Germany dominates monetary policymaking in the EMS. Italy’s welfare, under these circumstances, would increase by

But this conclusion rests on the flawed premise that German policy would not change with a fixed exchange rate. However, the fact that Italy adopts the exchange rate target and follows German monetary policy makes the Bundesbank the Stackelberg leader of the game. This alters the Bundesbank’s policy constraint so that the policy described in equation (7) is no longer optimal for Germany. Specifically, the fixed exchange rate lowers the cost of a German monetary surprise, since part of its domestic price level effect can be exported to Italy, and thereby increases the German employment-inflation trade-off (see Canzoneri and Henderson, 1988; and Fratianni and von Hagen, 1990). Consequently, the Bundesbank has a higher incentive for monetary surprises under fixed exchange rates. The private sector recognizes this change and sets me accordingly. The sub-game perfect Stackelberg equilibrium for the EMS is

The Bundesbank’s “brand name” as an inflation fighter loses credibility in the EMS. In turn, this loss affects the utility levels of both countries:

The EMS and the High-Inflation Country

The equilibrium equation (11) shows that the high-inflation country’s incentive to join the EMS is ambiguous, once the endogeneity of Germany’s policy is recognized. Italy’s change in utility from joining the system is

so that Italy benefits from the EMS only if (ε+α)/α<n¯*/n¯

, a condition that links the structural characteristics of the economies to the central bank preferences. For a sufficiently small value of ∈ the condition will hold. But ∈ is small either when γ, the income elasticity of import demand, is small and/or when 1/δ, the real exchange rate elasticity of import demand, is high. In the transition period from the current stage of the EMS to the long-run objective of a complete internal market, “Europe 1992,” and European Monetary Union, the two constraints contradict each other in the sense that, while a high value of 1/δ is consistent with a higher degree of integration, a low value of γ is not.

EMS Discipline

Even if the above condition is satisfied, however, it is not clear that joining the EMS is a feasible strategy for the Banca d’Italia. The crucial question is, what makes markets believe in its commitment to a low-inflation exchange rate target when the commitment to a low-inflation domestic monetary target is reputed to be impossible by assumption? Generally, EMS membership with the adoption of the German money rule is not credible because it is not a sub-game perfect strategy. To see this, consider the Italian central bank’s marginal utility of a monetary surprise, once expectations, hence wages, have been set according to the EMS (see equation (10)):

The marginal utility is positive if n˜*/n˜>1+ε/α

, which is exactly the condition assuring that Italy gains from the EMS. Given EMS expectations, the Banca d’Italia has an incentive to abandon the exchange rate target, making its commitment to the EMS not credible. Hence, the public will not form expectations according to equation (10). In sum, without any further modification, Italy can expect no gains from joining the EMS.

One way to overturn this negative conclusion is to assume that, once the EMS has been put into place, there is an additional, “political” cost of breaking the arrangement (for example, Melitz, 1988). Let P* be the loss in utility to the Italian central bank of reneging on her EMS commitment. The new marginal utility of a monetary surprise now becomes dU*/dm*P*

. To make the EMS credible, the political cost must at least be equal to

that is, it must balance the incentive for a monetary surprise. The larger the weight of the employment target and the larger the difference in the pre-EMS inflation biases, the larger the political cost to Italy of leaving the EMS must be to make EMS membership feasible. Since this political cost is not an intrinsic part of the exchange rate arrangement itself, it is erroneous to call the EMS a “disciplinary device” (Giavazzi and Pagano, 1988). The discipline has to come from some other source. The superiority of EMS membership as a monetary strategy for Italy thus hinges entirely on the assumption that, by creating new political threats to the policymaker, the EMS can alleviate the institutional deficiency behind the inflation bias.

Melitz (1988) explains the nature of the political cost, arguing that devaluations under a fixed exchange rate system cost electoral support. However, even assuming that electoral considerations are of direct relevance for the monetary authorities, it is empirically unclear how important international monetary arrangements are for voting behavior. Two episodes in recent EMS history are of interest.

The first one is the end of the French “Mitterrand experiment” with expansive demand management in a recessionary international environment (1981–83), often considered a prime example of EMS discipline (for example, Giavazzi and Giovannini, 1989). In their account of the events, Sachs and Wyplosz (1986) argue that EMS membership was decisive for convincing the French Government to adopt a more restrictive stance in 1983. The political threat of being forced to leave the EMS induced the Government to adopt policies that were consistent with its EMS commitment.

Yet, the French example is not entirely convincing. As Sachs and Wyplosz explain, the French experiment resulted in unsustainable external imbalances and demanded correction in any case. The relevant political alternative was, therefore, whether or not this correction should take place inside or outside the EMS. A recent review of the case in OECD (1988) argues that, based on the French experience in the 1970s, adjustment under a flexible exchange rate was perceived to be even more harmful than adjustment within the EMS. De Grauwe (1989) corroborates this assessment by showing that, in the 1980s, disinflation in the non-EMS countries was more like shock therapy, compared with the more gradual process that took place in the EMS countries. Furthermore, the EMS offered recourse to financial assistance. Thus, rather than providing a threat, the EMS may actually have been regarded as the lesser evil in the unavoidable policy adjustment.

Similarly, the history of the recent Italian disinflation does not lend much support to the political cost hypothesis of the EMS. Fiscal dominance—that is, a high degree of monetization of budget deficits—prevailed in the 1970s. The Italian Treasury could automatically expand the monetary base through an overdraft facility and rely on the central bank as the residual buyer of government debt at Treasury auctions. In addition, the Banca d’Italia was hampered in its desire to offset these “automatic” sources of monetary base creation by the absence of a secondary market for government debt. In 1976, the Treasury began offering securities to the public and the central bank regained some monetary independence by destroying the monetary base through open market sales. Additional independence was obtained in 1981 when the authorities (including the Treasury) freed the Banca d’Italia from the obligation of being the residual buyer of government debt, a decision the Italian press defined colorfully as the “divorce.” The reason for the “divorce” had little to do with the EMS. As Tabellini (1988, p. 97) puts it:

The Secretary of the Treasury and his advisers, perhaps even more than the Bank of Italy, were convinced that the “divorce” would have positive repercussions on the incentives of the government and the political authorities [to reduce the budget deficit].

Giavazzi and Spaventa (1989), on the other hand, regard the defeat of the 1984 national referendum called for by the opposition and militant labor unions against the government’s limit to wage indexation as the decisive event in lowering inflation expectations. In their own words (pp. 154–55):

EMS membership might have helped by providing a justification for unpopular policies, but the data strongly suggest that the new exchange-rate regime did not automatically produce an improvement of the output-inflation trade-off. Governments had to prove that they were prepared to bear the cost of unpopularity before price-setters became convinced that the commitment to the new monetary regime was lasting.

In sum, the decisions underlying both the “divorce” and the indexation referendum were not inspired by fear of the political cost of abandoning EMS membership.

The EMS and the Bundesbank

Germany’s change in utility from EMS membership is

which represents a welfare loss resulting from the higher equilibrium inflation rate. Credibility cannot be transferred freely among central banks. The model cannot explain German membership in the EMS.

A first way to tackle this problem is based on the potential conflict between the Bundesbank and the German Government. According to German law, the decision to join the EMS was actually made by the Ministry of Finance. One may therefore argue that EMS membership was forced on the Bundesbank, perhaps as a way to weaken its political independence (Vaubel, 1980). But this only redirects the question: Why would the Government accept the higher rate of inflation? Credibility considerations suggest that, in contrast, the Government should keep the Bundesbank independent to minimize the inflation bias of German monetary policy (Rogoff, 1985; and Thompson, 1981).

Giavazzi and Pagano (1988) and Melitz (1988) have presented models that enlarge policymakers’ preferences by a real exchange rate target, so as to resolve the conflict between the Ministry of Finance and the central bank. The assumption is that the authorities have a target for the ratio of traded to nontraded goods in domestic output, which reflects their inclination to favor domestic export industries. This ratio rises with a real depreciation of the home currency. The inclusion of a real exchange rate target makes the German authorities willing to accept more inflation and less monetary discipline.

The principal problem with this approach is that Germany cannot obtain a permanent real devaluation in the EMS. The resulting trade imbalances would make the EMS unsustainable in the long run. As markets realize this, speculative attacks on the EMS parities would follow and lead to the early breakdown of the system (Wyplosz, 1986). Sustainability of the EMS requires that, on average, the deutsche mark be neither over- nor undervalued.

Another explanation for German membership is that the Bundesbank seeks greater financial integration in the EMS. As noted by Giavazzi and Pagano (1985), Germany has historically borne the brunt of speculative capital flows originating from the U.S. dollar market. Under fixed exchange rates, increased substitutability between deutsche mark- and other EMS-currency-denominated assets would produce a more even distribution of speculative capital flows.3 The problem here is that—particularly around realignments—the stability of the EMS depends heavily on capital and exchange controls in the weak-currency countries (Wyplosz, 1986; and Fratianni and von Hagen, 1989), which runs counter to the German interest in financial integration.

Credibility in the EMS: Are “Side-Payments” Feasible?

The essence of our arguments so far is that, to induce German membership in the EMS, the system must somehow compensate the Bundesbank. In other words, the EMS game requires “side-payments” to the German authorities. The French and Italian agreement to greater liberalization of their home capital markets may be an example of such side-payments. But such compensations necessarily lower the high-inflation countries’ incentive to join the system. A necessary condition for side-payments is that Italy’s welfare gain in the EMS exceeds the German loss. To determine under which circumstances this condition holds, add equations (12) and (12a):

The aggregate change in utility is positive if (1+/α)2<0.5[1+(n˜*/n˜)2]. Only under this condition would Italy be willing to join the EMS and sufficiently compensate the Bundesbank for its role in the system. For γ2δ3.5α/(1α),, this condition is more restrictive than the condition of Italy alone gaining from EMS membership (see equation (12)). In light of the discussion above, there is no reason to assume that the above restriction will hold under a high and rising degree of economic integration among the member countries.

III. Credibility in the EMS: Empirical Tests of German Dominance

The proponents of the credibility hypothesis of the EMS have focused primarily on the reduction and convergence of inflation rates in the system during the 1980s. This evidence is presented in Table 2. Average inflation rates in the EMS, measured in terms of consumer price indexes, peaked in 1980 at 11.6 percent, and declined steadily from then on to 2.2 percent in 1987. At the same time, the standard deviation of inflation rates among the members fell from 6.2 percent to 1.9 percent, a first indication of convergence. Since the German rate of inflation was always at the low end of the system, this evidence has been interpreted as showing that EMS inflation rates were pulled toward the German rate. Supporters of the hypothesis have inferred from this that the Bundesbank dominates the system.

Table 2.Consumer Price Inflation: EMS and Non-EMS Countries(percent)
EMS197719781979198019811982198319841985198619671986
Belgium7.14.54.56.67.68.77.76.34.91.31.51.2
Denmark11.110.19.612.311.710.16.96.34.73.64.04.6
France9.49.110.713.813.411.89.67.45.82.53.32.7
Germany, Fed. Rep of3.72.74.15.46.35.33.32.42.2-0.20.21.2
Ireland13.67.613.218.220.417.110.58.65.43.83.12.2
Italy17.012.114.821.217.816.514.710.89.25.94.7
Netherlands6.54.24.26.56.75.92.83.32.20.10.00.8
Average8.85.98.511.611.210.07.96.04.92.62.2
Standard Deviation4.95.34.46.24.64.44.43.22.71.71.9
Non-EMS
Australia12.37.99.110.19.711.110.14.56.78.78.67.2
Austria5.53.63.76.46.85.43.35.73.21.71.41.9
Canada8.09.09.110.212.410.85.84.34.04.14.44.0
Finland12.77.87.511.612.09.38.47.15.92.94.1
Greece30.544.145.558.550.649.186.130.832.023.016.4
Japan8.03.83.68.04.92.61.82.32.00.60.00.7
Switzerland1.31.13.64.06.55.73.02.93.40.71.41.9
United Kingdom15.88.313.418.011.98.64.65.06.13.44.24.9
United Stales6.57.611.313.510.46.23.24.33.62.03.64.0
Average7.77.09.612.439.76.43.94.23.82.33.2
Standard Deviation4.13.74.75.14.14.27.22.32.72.32.1
Average without U. S.9.36.27.310.98.76.74.84.13.62.72.7
Standard Deviation6.09.96.57.76.26.411.03.54.63.63.6
Note: Averages and standard deviations are weighted with the following. 1982 real GNP-based weights (percents): EMS: Belgium 4.55, Denmark 2.97, France 28.84, Germany 34.42, Ireland 0.99, Italy 21.07, Netherlands 7.22, Non-EMS: Australia 3.08, Austria 1.23, Canada 5.61, Finland 0.94, Greece 0.71, Japan 20.03, Switzerland 1.79, United Kingdom 8.98, United States 57.63.Source: IMF, International Financial Statistics.
Note: Averages and standard deviations are weighted with the following. 1982 real GNP-based weights (percents): EMS: Belgium 4.55, Denmark 2.97, France 28.84, Germany 34.42, Ireland 0.99, Italy 21.07, Netherlands 7.22, Non-EMS: Australia 3.08, Austria 1.23, Canada 5.61, Finland 0.94, Greece 0.71, Japan 20.03, Switzerland 1.79, United Kingdom 8.98, United States 57.63.Source: IMF, International Financial Statistics.

On a closer look, however, the data are less convincing. The lower part of Table 2 shows the rates of inflation for a group of non-EMS countries in the OECD during the same period. Their inflation performance was similar to that of the EMS in the 1970s and 1980s (Ungerer et al., 1986; and Fratianni, 1988). Average inflation in this group peaked in 1980, too, and declined from then on to about 3 percent in 1987. As in the EMS, there is notable convergence of inflation rates at the low level: the standard deviation of inflation rates declined from 5.1 percent to 2.1 percent. Disinflation and convergence is thus by no means a special characteristic of the EMS.4

There are at least two alternative interpretations of the data in Table 2 that give no credit to the EMS in Europe’s disinflation. One is that all industrial countries, whether in the EMS or not, have responded during the 1980s to common external shocks and constraints that facilitated disinflation, such as the tightening of U.S. monetary policy in the early years of the decade, and the decline in oil prices. The other is that policymakers inside and outside the EMS, dissatisfied with the stagflationary consequences of active demand management policies pursued in the 1970s, adopted more disinflationary policies in the 1980s (Chouraqui and Price, 1984). Both alternative hypotheses have the same implication for the inflation performance inside (and outside) the EMS: they predict that inflation rates decline and converge at the lower level. They are therefore observationally equivalent to the credibility hypothesis with respect to the data presented in Table 2. That is, it is impossible to infer or test the credibility hypothesis empirically on the basis of inflation rates.

Testable Implications of German Dominance in the EMS

In view of this observational equivalence, we propose to test the hypothesis by focusing on monetary policy actions—rather than inflation rates, which reflect the outcomes of monetary policy actions—and by testing the hypothesis of German dominance in the EMS implied by the credibility hypothesis. Our test is based on the premise that the German dominance hypothesis implies a specific structure of central bank “reaction functions” in the EMS. Let DY be the 7x1 vector of variables summarizing all the central bank actions in the EMS, DX a matrix representing changes in domestic variables believed to influence the behavior of the monetary authorities, and DW the 7×1 vector representing changes in “world” variables also believed to influence the EMS countries’ monetary authorities. The latter can be modeled as a system of linear dynamic equations:

where A(L), B(L), and C(L) are polynomial matrices in the lag operator L ; b is a fixed intercept vector; andet is a vector of residuals with properties E(ei)=0,E(e2i)=σii, i = country 1,…,7, E(etet)=diag(σii)

and E(etet*)=0 for tt*. Furthermore, the leading coefficient of aii(L) is set equal to one. These restrictions force the entire set of interactions among EMS monetary policies to take place within the matrix A(L). The test of German dominance, consequently, focuses to a large extent on the properties of this matrix.5

The Four Components of the German Dominance Hypothesis

German dominance consists of four separate hypotheses. First, it implies that other EMS countries do not react directly to monetary policies occurring outside the EMS. The world at large can influence other EMS countries only through its impact on German monetary policy. Defining country 1 as Germany, this can be stated as6:

H1: World insularity ci = 0, i = 2,…,7.

Rejecting H1 means that monetary policies in other EMS countries are influenced directly by what goes on outside the EMS, over and beyond what is implied by the German rule.

Second, German dominance implies that each EMS country reacts only to Germany and not to other members’ policies:

H2: EMS insularity aij = 0, ij, i, j = 2,…,7.

Rejecting H2 signifies that EMS countries interact independently with one another, which is inconsistent with their joint adherence to the path set by the Bundesbank.

Third, German dominance implies that monetary policy in a member country depends on German policy and we must reject

H3: Independence from Germany ail = 0, i = 2,…,7.

Fourth, to make German dominance meaningful, Germany itself must not be influenced by the monetary policy actions of other members:

H4: German policy independence ali = 0, i = 2,…,7.

Strong and Weak Forms of German Dominance

In the above formulation, German dominance does not allow any short-run deviation of the other members’ policies from the path prescribed by the Bundesbank. We call this the strong form of the hypothesis. A less restrictive, and more realistic view of German dominance, is to allow short-run deviations from the Bundesbank rule, but not long-run ones. This weak form of German dominance can be formally written as:

H1W: g(ci) = 0, i = 2,…,7

H2W: g(αij) = 0, i # j, i, j = 2,…,7

H3W: g(αil) = 0, i = 2,…,7

H4W: g(αli) = 0, i = 2,…,7,

where g(.) defines the sum of all coefficients of the lag polynomial ci(L) or αi(L).

The Choice of Policy Variables

Central banks in the EMS commonly express and assess their short-run actions in terms of the resulting changes in domestic money market interest rates. On a weekly and monthly frequency, money market interest rates embed a high degree of policy exogeneity. Thus, modeling monetary policy actions and interactions on the basis of money market interest rates will produce a picture of the policy game in the EMS in the short run that comes closest to how policymakers perceive it. Beyond this short-run horizon, changes in interest rates incorporate the endogenous response resulting from changes in economic activity and inflation rates. From a longer-run perspective, the growth rate of the monetary base is an appropriate policy variable, since the monetary base summarizes the money supply effects of all central bank actions. We shall implement the model of equation (16), using both monthly changes in interest rates and quarterly percent changes in the monetary bases.

There are differences in the specification of equation (16) for the two policy variables, imposed by data availability and the time series properties of the left-hand side variables. In the interest rate version of equation (16), DX consists of domestic inflation rates (measured in terms of the consumer price index) and the growth rate of output (measured in terms of the index of industrial production). Changes in the U.S. money market rate of interest represent policy actions outside the EMS.7 To reduce the number of free parameters in the system, the following additional restrictions are imposed on A(L):

where ϕj are country j’s share in regional GNP and the group of countries 4,…,7 consists of Belgium, the Netherlands, Ireland, and Denmark. Conditions (17) state that in the equations for Germany, France, and Italy we use the weighted average of the interest rates of these four countries, rather than each interest rate individually. Within that group, we use the weighted sum of the remaining three countries. The German, French, and Italian interest rates are left unrestricted.

In the monetary base version of equation (16), DX consists of a trend, with seasonal dummies and government budget deficits a proportion of the lagged monetary base. DW consists of a weighted average of the monetary base growth of the United States, the United Kingdom, Canada, and Japan.8 Here as well, we reduce the number of free parameters in the system by imposing the following, additional restrictions on A(L):

Conditions (18) restrict the off-diagonal polynomials of the matrix A(L) so that Germany reacts to a weighted average of the monetary base growth of the other six countries, whereas each of the six countries reacts to Germany and a weighted average of the remaining five countries.

Empirical Results

Our empirical strategy was first to estimate a “broad” specification—allowing for several lags—and later to eliminate variables that had contributed “little” to the explanatory power of the model. The selection process had to satisfy the twin criteria that the reduction would neither lower the information value of the model—as measured by the Akaike’s AIC statistic—nor generate residual serial correlation. The shrunken version of the model was then estimated with 3SLS and used for testing the German dominance hypothesis. The interest rate model was tested and, where appropriate, adjusted for structural breaks in all parameters after March 1983, using interactive dummy variables. This model therefore has two parameterizations, one for 1979 to the third quarter of 1983 and one for the period thereafter. For the quarterly monetary base model, this procedure was not possible because of the more limited number of degrees of freedom. All tests were conducted as F-tests, imposing the relevant restrictions on the entire system. The restrictions of H1 through H3 (and H1W through H3W) were imposed alternatively at the country level—that is, leaving the other equations of the system unaltered—and at the system level.

The test results are reported in Tables 5, 6, and 7. For both policy variables, the strong form of German dominance is clearly rejected. Germany, to be sure, is a significant player in the EMS, but so are other central banks in the system, particularly the Banque de France and the Banca d’Italia. German interest rates and monetary base growth rates respond significantly to monetary policy actions in other EMS countries, and other countries interact and react to the rest of the world independently. From this point of view, the EMS is interactive and not hierarchical.

Table 5.Test of German Dominance Monthly onshore interest rates(1979M3 to 1983M3)
Strong formWeak form
H1H2H3H4H1H2H3H4
Germany,3.9**3.1*
Fed. Rep. of
(8)(3)
Belgium1.02.7**0.91.31.70.1
(2)(7)(3)(1)(3)(1)
Denmark5.2**9.1**3.2*10.2**2.9*0.2
(2)(5)(3)(1)(3)(1)
France7.4**3.7*7.9**0.40.220.4**
(2)(2)(4)(1)(1)(1)
Ireland3.1*6.6**2.6*0.95.1**3.8*
(3)(6)(3)(1)(3)(1)
Italy2.04.8*2.30.37.2**3.0
(3)(3)(2)(1)(2)(1)
Netherlands2.55.8**5.0**4.8*1.113.5**
(2)(7)(3)(1)(3)(1)
System3.5**5.6**4.1*3.43.0**6.8**
(14)(30)(18)(6)(15)(6)
Note: See note to Table 6.
Note: See note to Table 6.
Table 6.Test of German Dominance Monthly onshore interest rates(1983M4 to 1988M4)
Strong formWeak form
H1H2H3H4H1H2H3H4
Germany,4.31**0.383.03*
Fed. Rep. of
(8)(3)
Belgium0.992.58**4.41**1.312.335.47**
(2)(8)(3)(1)(3)(1)
Denmark0.632.24*1.120.592.61*1.10
(2)(6)(3)(1)(3)(1)
France7.42**3.70*7.92**0.740.2020.4**
(2)(2)(4)(1)(1)(1)
Ireland3.94**5.94**4.26**1.534.89**0.63
(3)(7)(3)(1)(3)(1)
Italy2.44*4.48**2.186.51**8.61**0.12
(4)(4)(3)(1)(2)(1)
Netherlands4.97**4.78**5.04**1.231.5213.47**
(3)(8)(3)(1)(1)(1)
System3.24**3.94**4.39**2.063.32**7.00**
(14)(35)(19)(6)(15)(6)
Note: H1 refers to world insularity; H2 to EMS insularity; H3 to independence from German policy; H4 to German policy independence. Entries are values of the F-tests based on 3SLS estimates. Numbers in parentheses are numerator degrees of freedom. Denominator degrees of freedom are 612 for all tests.* and ** denote significance at the 5 and 1 percent level, respectively.
Note: H1 refers to world insularity; H2 to EMS insularity; H3 to independence from German policy; H4 to German policy independence. Entries are values of the F-tests based on 3SLS estimates. Numbers in parentheses are numerator degrees of freedom. Denominator degrees of freedom are 612 for all tests.* and ** denote significance at the 5 and 1 percent level, respectively.
Table 7.Test of German Dominance Quarterly percent changes of the monetary base(1979QII to 1988QI)
Strong formWeak form
H1H2H3H4H1H2H3H4
Germany,
Fed. Rep. of2.9*3.50
(2)(1)
Belgium0.903.1*0.900.206.2*0.00
(2)(2)(5)(1)(1)(1)
Denmark12.4**0.300.702.600.601.00
(2)(2)(5)(1)(1)(1)
France12.1**0.203.4**12.1**0.300.00
(2)(2)(5)(1)(1)(1)
Ireland6.4**2.8 +4.9**9.7**2.104.0*
(3)(2)(5)(1)(1)(1)
Italy6.88**9.3**9.7**4.9*20.7**19.5**
(2)(3)(5)(1)(1)(1)
Netherlands6.5**3.0*2.6*12.9**5.9*1.10
(3)(2)(5)(1)(1)(1)
System8.7**3.9**4.7**7.1**5.9**5.0**
(13)(13)(30)(6)(6)(6)
Note: H1 refers to world insularity; H2 to EMS insularity; H3 to independence from German policy; H4 to German policy independence. Entries are values of the F-tests based on 3SLS estimates. Numbers in parentheses are numerator degrees of freedom. Denominator degrees of freedom are 174 for all tests.+, *, and ** denote significance at the 5 and 1 percent level, respectively.
Note: H1 refers to world insularity; H2 to EMS insularity; H3 to independence from German policy; H4 to German policy independence. Entries are values of the F-tests based on 3SLS estimates. Numbers in parentheses are numerator degrees of freedom. Denominator degrees of freedom are 174 for all tests.+, *, and ** denote significance at the 5 and 1 percent level, respectively.

The results for the weak form of German dominance are more interesting. They are summarized in Table 3, where the focus is on the three large countries in the system. Here, a “Y” means that the test result is consistent with the German dominance hypothesis, whereas “N” denotes the opposite. The overall result of the weak dominance tests is, again, that German dominance is rejected. But the table reveals several interesting characteristics of the structure of the EMS. Consider the interest rate model first. Comparing the two sub-periods, EMS reactions to the outside world show more conformity in the later period. During the second half of the EMS period, world insularity is accepted for all countries except Italy, and for the system as a whole. By itself, this finding could be interpreted as a sign of German leadership in the system, with Germany preventing other countries from reacting independently to interest rate movements outside the EMS. However, the results for the remaining hypotheses reveal the limits of that interpretation.

Table 3.Summary of Weak German Dominance Tests
Policy Variable: Money Market Rates, 1979M4 to 1983M3
GFISmaller EMS CountriesEMS
H1YY2YN
H2YN2YN
H3YN2YN
H4N
Policy Variable: Money Market Rates, 1983M4 to 1988M4
H1YN4YY
H2YN2YN
H3YN2YN
H4N
Policy Variable: Monetary Base Growth Rates, 1979QII to 1987QIV
H1NN2YN
H2YN2YN
H3NN1YN
H4Y
Note: “Y” and “N” mean result consistent and not consistent with German dominance, where consistent results are: accept H1, H2, H4 and reject H3. “EMS” reports the result for imposing the restrictions on the entire system.
Note: “Y” and “N” mean result consistent and not consistent with German dominance, where consistent results are: accept H1, H2, H4 and reject H3. “EMS” reports the result for imposing the restrictions on the entire system.

Furthermore, the results point to a striking difference between the French and the Italian positions in the EMS. French interest rate responses accord well with all three restrictions posed by the German dominance hypothesis, while the Italian responses do not. French interest rates are much more closely bound to German interest rates than are Italian rates.

Next, consider the longer-run-oriented results of the monetary base growth rates. A first important finding is that German money growth does not respond to policies in other EMS countries in the long run. This is consistent with the view that the Bundesbank pursues an independent policy in the EMS. But German policy independence does not prevent the other countries from reacting to one another and to the rest of the world independently. German policy independence is therefore not the same as German dominance.

Both Italy and France show significant long-run reactions to the rest of the world; Italy in addition reacts significantly to other EMS countries. But the more striking difference between the French and the Italian position in the system is that Italy’s monetary growth path depends importantly on Germany’s in the long run, while the French does not.

Table 3 thus raises two puzzles. Why does the close link between German and French short-run policy actions, expressed by interest rates, not show up in the monetary base growth rates; and why are Italian short-run movements more disconnected from Germany’s despite its apparently closer long-run connection? The answer lies in the institutional arrangements of the EMS. Although monetary policies are intertwined in the short run, realignments give member countries the ability to choose their own trend monetary growth and inflation rates. The stark contrast between the interest rate and the monetary base results, therefore, is resolved by the fact that France has used realignments more often and effectively than other countries to set its own monetary trend. As for Italy, the EMS has granted this country more short-run flexibility than France because of the wider exchange rate band (12 percent versus 4.5 percent). Consequently, Italy’s need to align with other EMS countries in the short run is weaker.

To summarize, our tests reject German dominance in the EMS both in its strong and weak form. Overall, the system appears to be more interactive than hierarchical. But the tests indicate that other asymmetries exist in the system. One is German long-run monetary policy independence. Another is that the Banque de France, most likely because of the narrower exchange rate band, is more restricted than the Banca d’Italia by the short-run constraints of the system.

Other Evidence on German Dominance

Cohen and Wyplosz (1989) test the German dominance hypothesis for a sub-group of EMS countries using vector autoregressions of domestic interest rates and monetary base growth rates. They find that German interest rates and base growth have significant impacts on the same variables in other EMS countries, but the opposite is also true. A similar outcome for base growth is reported by Mastropasqua et al. (1988). Again, German dominance fails the empirical test.

De Grauwe (1988) has recently presented an alternative approach to testing German dominance. His test asks whether expected exchange rate devaluations of an EMS member country against the deutsche mark affect short-term interest rates only in this country—as the German dominance hypothesis would suggest—or both in Germany and the depreciating country. He finds that the German interest rate does not respond to expected devaluations of the Belgian franc, the French franc, and the Italian lira. However, interest rates in these three countries do not fully adjust either. It follows that capital controls and dual exchange rate systems were critical in isolating domestic money markets from international forces. For the Netherlands, where there are no capital controls, De Grauwe finds that interest rate adjustments are shared between Germany and that country, a result that is consistent with ours. De Grauwe concludes that his findings provide no support for German dominance.

IV. Asymmetries in the EMS: Another Look

The rejection of the German dominance hypothesis does not imply that the EMS has performed symmetrically in the sense that each country has contributed its “proper” share to the EMS policy game. In this section, we explore the issue of asymmetries by relaxing the strict constraints implied by German dominance. In particular, we investigate the interest rate responses in the EMS to interest rate innovations in the three large countries, Germany, Italy, and France.

Let Δij(k) be the change in the level of country j’s interest rate due to a unit innovation in country i’s rate k months ago.9 Estimates of Δij(k) are reported in Table 6 for j = Germany, France, and Italy; i = Germany, France, Italy, and the group of the smaller EMS countries; and lags k = 3, 6. Not surprisingly, the effects of interest rate innovations in the three large countries are quite different. But this may simply reflect the different relative sizes of their economies, and not necessarily asymmetries.

To ascertain asymmetries, we compare the relative strength of the interest rate responses to a simple measure of the countries’ relative size, namely their weights in the ECU, wi.10 We define the following indicator of asymmetry:

If Aij(k) < 0, the ratio of country i’s effect on j to j’s effect on i is smaller than the ratio of the two weights in the ECU. Assuming that the ECU weights correctly reflect the relative economic size of the countries, this indicates an asymmetry in that country i contributes too little to the EMS policy game. The opposite holds if Aij(k) > 0.

These indicators are reported in Table 4. Two observations are noteworthy.11 First, Germany has consistently responded too little to French interest rate innovations and too much to innovations from the group of the smaller countries. Similarly, the small countries have consistently responded too little to French innovations. Rather than German dominance, the asymmetry revealed by this pattern is that the French position in the system has been weaker than the relative size of the French economy would lead one to expect. Second, a similar observation holds for Italy in the first EMS period, when, in addition, the Italian reaction to French interest rate innovations was too strong compared with the Italian impact on the French interest rate. But Italy has improved her position in the second half of the EMS period, reversing the sign of the indicator with respect to France and Germany. The small Italian response to French interest rate changes in this period strengthens the impression that France has a relatively weak position in the EMS.

Table 4.Indicators of Asymmetry
1979M4-1983M31983M4-1988M4
Aij(k)Aij(k)
Germany,Germany,
country jFed. Rep. ofFranceItalyQi(k)Fed. Rep. ofFranceItalyQi(k)
country ik = 3k = 3
Germany,
Fed. Rep. of.39.31
France.20.09
Italy.10++.23
Smaller
Countries+++.31+++.37
k = 6k = 6
Germany,
Fed. Rep. of.41.32
France.20.10
Italy.08+.20
Smaller
Countries+++.31+++.38
Note: Aij(k) and Qi(k) are the indicators defined in equations (19) and (20). Aij(k) = −Aij(k), so that we report only the latter.
Note: Aij(k) and Qi(k) are the indicators defined in equations (19) and (20). Aij(k) = −Aij(k), so that we report only the latter.

Next, we define the average ECU money market rate RE = ΣiwiRi, where the wi’s are again the ECU weights of each currency i = 1,…,7. We define the following measure of country i’s relative strength in the EMS:

which is the relative contribution of an innovation in country i’s interest rate k months ago to the change in RE resulting from simultaneous innovations of equal size in all seven countries. By construction, 0 ≤ Qi(k) ≤ 1. These measures are again reported in Table 4. In the first EMS period, Qi(k) is very close to each country’s ECU weight. But in the second period, the French relative contribution falls much below its benchmark, while the contributions of Italy and the small countries exceed their ECU weights. Again, the asymmetry is not German dominance, but France’s relatively weak position in the system.

V. Conclusions

In this paper, we have presented a critical assessment of the credibility hypothesis of the EMS. Regarding the theoretical argument as it stands in the literature, our main points can be summarized as follows.

  • Borrowing central bank credibility through a fixed exchange rate weakens the credibility of the low-inflation country and, hence, creates a conflict between this and the high-inflation country. Credibility models of the EMS need to explain German membership in the system convincingly by resolving this conflict. One way of doing so would be to allow for reputational considerations in addition to mere central bank credibility, in the analysis of the EMS game (von Hagen, 1989).

  • A high and rising degree of economic integration among the members of the EMS may reduce the high-inflation countries’ benefits from borrowing Bundesbank credibility.

  • The benefits from borrowing credibility in the end can only stem from political pressures that keep countries in the EMS and prevent them from adopting discretionary monetary policies. The empirical evidence on the importance of such political pressures is at best inconclusive.

Our empirical analysis of the EMS policy game leads us to reject the structural implication of the credibility hypothesis, German dominance. The German position in the EMS is best described by long-run independence, not dominance. Realignments and capital and exchange controls explain why German long-run independence is not equivalent to German dominance. We find that the EMS has constrained interest rate policies in France more strongly than in Italy. The relative weakness of France in the system is the most striking asymmetry in the EMS. French under representation and the stronger constraints on French interest rate movements may well explain why the complaints about German dominance have come most forcefully from the French side. We conclude that many observers have mistaken German long-run independence and French under representation in the EMS for German dominance.

Indirectly our findings corroborate the alternative hypothesis, which holds that the general consensus in the region to reduce inflation from the high levels of the 1970s—and the resulting willingness of policymakers to bear the political cost of restrictive domestic policies—have been far more important for European disinflation in the 1980s than EMS membership and the borrowing of Bundesbank credibility. The Bundesbank’s strong commitment to price stability has likely strengthened this consensus and favored disinflation in the region by serving as a focal point for monetary policies in the other EMS countries and facilitating their turning to low-inflation strategies. Eichengreen (1989) argues that in the classical gold standard and the Bretton Woods system the United Kingdom and the United States, as the largest countries, respectively, performed a similar coordinating function rather than being hegemons. But this is a much weaker, yet more plausible, role than the Bundesbank’s coercive dominance in the EMS claimed by the credibility hypothesis.

Yet, the credibility argument carries a lesson for the future of the EMS. The construction of more formal institutions, proposed in the Delors Report to achieve European Monetary Union, will inevitably increase the political cost of discretionary monetary policies in all member countries. From this perspective, the EMS should enter “Stage Two” of the Delors proposal fairly soon, so as to lock in the disinflation gains of the 1980s, and raise the cost of a policy reversal, should the consensus on low-inflation policies disappear in the future.

Table 8.Interest Rate Responses Interest rate innovation k months ago in country
1979M4-1983M31983M4-1988M4
Germany,FranceItalyGermany,FranceItaly
Fed. Rep. ofFed. Rep. of
Interest Ratek = 3k = 3
Response in:
Germany,
Fed. Rep. of.395.33.00.00.49
France.99.221.73.80.57
Italy.96.40.110.36.00
Small Countries.31.73.16.49.19.36
k = 6k = 6
Germany,
Fed. Rep. of.38.00.08.51
France1.06.23.75.23
Italy1.11.43.23.00
Small Countries.43.25.13.55.20.34
Note: Weighted averages for small countries, using ECU weights. Numbers on the diagonal are ECU weights.
Note: Weighted averages for small countries, using ECU weights. Numbers on the diagonal are ECU weights.

More generally, our review suggests that the interpretation of the EMS as a shock absorber is more robust than the credibility hypothesis. Empirical evidence (Artis and Taylor, 1988; and Fratianni and von Hagen, 1990) shows that the EMS has significantly reduced exchange rate uncertainty (both nominal and real) among its members. But in doing so, it has increased exchange rate uncertainty with outside currencies, most notably the U.S. dollar. In Fratianni and von Hagen (1990), we have also shown that the EMS has weakened the coherence between inflation surprises inside and outside the system, while strengthening the coherence of inflation surprises among the member countries. This is consistent with the view that monetary policy coordination in the EMS has partly disconnected the system from external shocks and has equalized their effects in the region. In contrast to the credibility hypothesis, this means that the welfare implications of exchange rate policies in the EMS, both at the group and the world level, are ambiguous and need further analysis.

Comment

Stefan Schonberg

The main conclusion of the paper by Fratianni and von Hagen is that the European Monetary System is not “dominated” by the deutsche mark (or by the monetary policy of the Deutsche Bundesbank), in the same sense that other EMS countries have to borrow credibility from the Bundesbank, and that the EMS itself is characterized by strong “asymmetries.” The EMS countries’ success in reducing inflation is attributed to the willingness of policymakers to bear the political cost of restrictive domestic policies, rather than to EMS membership.

In my view the two aspects are intertwined.1 There are certainly no built-in features that would make “dominance” of one or more currencies inherent to the system. The authors’ view is also supported by the fact that the success of the EMS countries in reducing inflation in the 1980s is not fundamentally superior to that of other non-EMS countries during the same period. Disinflation and convergence have not been unique to the EMS.

However, EMS membership appears to have facilitated necessary policy decisions in member countries. The willingness of policymakers in EMS countries to bear the political costs of restrictive domestic policies cannot be dissociated from EMS membership and from the opportunity of borrowing Bundesbank credibility. The theory that the EMS countries other than Germany pursue “autonomous” policies to combat inflation is not entirely consistent with the fact that most of these countries have followed regularly the interest rate moves of the Bundesbank. Indeed, some countries abstained from any interest rate action when their own economic development would have made such action desirable, waiting instead until the Bundesbank eventually took such action. From this point of view one could say that the “leading role” of the deutsche mark is characterized by the fact that currency competition in Europe encourages central banks to pursue less inflationary policies than they would otherwise.

This conclusion is consistent with the assessment of the Delors Report (“Report on Economic and Monetary Union in the European Community”) that the EMS “has benefited from the role played by the deutsche mark as an anchor for participants’ monetary and intervention policies.” The express recognition of the usefulness of the deutsche mark as a standard of stability contrasts with the criticism often heard in the past that the Bundesbank is forcing its EMS partners into a deflationary policy that prevents them from fully exploiting their economic growth potential.

The deutsche mark has assumed its leading role for lack of a better alternative. The factors that have made the deutsche mark the anchor of the system are the same as those that have made it the second largest reserve currency worldwide (a 21 percent share in total foreign exchange reserves, which is more than double the share of Japan, whose economy is double that of the Federal Republic of Germany): German price stability and the absence of exchange controls. The Bundesbank has not sought the role of anchor. It is rather the result of policy decisions made by other countries. The decisive signal was sent in March 1983 when the French Government, in the midst of a deep crisis, decided to shift its economic and fiscal policies to give priority to the defense of a fixed exchange rate vis-à-vis the deutsche mark.

The EMS has developed into a zone of relative stability. Exchange rate fluctuations have remained within the limits set under the agreed margins of fluctuations since January 1987. However, the average inflation rate of EMS members (except for Germany) is still about 4.5 percent, that is, approximately 1.5 percentage points higher than the rate in Germany. Under these conditions, as a matter of principle, member countries cannot dispense with the option of adjusting exchange rates in the EMS. This would only shift the adjustment pressure to such other areas as wage policy, fiscal policy, and so on.

At present, we are experiencing the paradoxical situation that the countries with the highest inflation rates and the largest current account deficits are revaluing in real terms vis-à-vis the deutsche mark. This would not appear to correspond to the underlying fundamentals and to the wish of member countries (including Germany) to see German surpluses reduced. In view of the rising German external surpluses and strong domestic growth, a revaluation of the deutsche mark—in particular against the currencies of the countries with considerably higher inflation—would be an appropriate reaction.

Germany is at present exporting stability to its European trading partners and importing inflation. This demonstrates that Bundesbank policies, for their part, are influenced by the monetary and exchange rate policies of its partner countries. The policy mix in Germany is different, at present, from what it would be without the exchange rate constraint. From this point of view, the EMS is indeed, as the authors claim, interactive and not hierarchical.

The major EMS partners have, however, excluded exchange rate changes as an adjustment instrument for the foreseeable future. For all practical purposes, they are behaving as though a currency union—defined as a zone of fixed exchange rates and free movement of capital—already existed. The inevitable consequence of such behavior is, of course, a certain loss of policy autonomy. This is, by the way, the main reason the United Kingdom has not yet joined the EMS. Under these conditions any monetary policy decision taken by the Bundesbank to promote monetary stability in Germany tends to imply at the same time greater monetary stability for Europe.

The Bundesbank has to take this into account when evaluating current plans for a European monetary union. There is a real risk that the cost of an early monetary union, which would end the existing competition among the different European currencies, would be a higher inflation rate. Therefore, in our view the anchor function of the deutsche mark in the EMS should be abandoned only when there is a guarantee that something equally good, or better, will take its place.

References

    ArtisMichaelP. andMarkJ. Taylor“What has the European Monetary System Achieved?” in FrancescoGiavazziStefanoMicossi andMarcusMillereds.The European Monetary System (Cambridge, England: Cambridge University Press1988).

    ArtusPatrick“The European Monetary System, Exchange Rate Expectations and the Reputation of the Authorities,”presented at the Conference on International Economic Policy Coordination Aixen ProvenceJune 24–251988.

    Banca d’ItaliaAnnual Report 1973.

    BarroRobert andDavidGordon“Rules, Discretion and Reputation in a Model of Monetary Policy,”Journal of Monetary EconomicsVol. 12 (Amsterdam: North-Holland1983) pp. 10121.

    CanzoneriMatthew andJoA. Gray“Monetary Policy Games and the Consequences of Non-Cooperative Behavior,”International Economic ReviewVol. 26 (Philadelphia: University of Pennsylvania1985) pp. 54764.

    CanzoneriMatthew andDaleHenderson“Is Sovereign Policymaking Bad?” in KarlBrunnerand andAllanH. Meltzereds.Carnegie-Rochester Conference Series on Public PolicyVol. 28 (Amsterdam: North-Holland1988).

    ChouraquiJ.C. andR.W.R.Price“Medium-time Financial Strategy: The Coordination of Fiscal and Monetary Policies,”OECD Economic Studies No. 2 (Paris: Organization for Economic Cooperation and Development1984).

    CohenDaniel andCharlesWyplosz“The European Monetary Union: An Agnostic Evaluation,”Macroeconomic Policies in an Interdependent World (Washington: International Monetary Fund1989).

    CollinsSusan“Inflation and the European Monetary System,” in FrancescoGiavazziStefanoMicossi andMarcusMillereds.The European Monetary System (Cambridge, England: Cambridge University Press1988).

    DeGrauwePaul“Is the European Monetary System a DM-Zone?” (unpublishedOctober1988).

    DeGrauwePaul“The Cost of Disinflation and the European Monetary System,”Centre for Economic Policy Research Discussion Paper No. 326 (London: CEPR1989).

    DelorsReportReport on Economic and Monetary Union in the European Community (Luxembourg: European CommunityApril 121989).

    DornbuschRudiger“Credibility, Debt and Unemployment: Ireland’s Failed Stabilization,”Economic Policy (Cambridge, England: 1989) pp. 174208.

    EichengreenBarry“Hegenomic Stability Theories of the International Monetary System,” in RichardN. Cooperet al.eds.Can Nations Agree? (Washington: The Brookings Institution1989).

    FratianniMichele“The European Monetary System: How Well Has it Worked?”Cato Journal (San Francisco: Cato Institute1988) pp. 477501.

    FratianniMichele andJürgenvon Hagen“Asymmetries and Realignments in the EMS,”for the Conference on the EMS in the 1990sCEPS and Central Bank of Greece (AthensSeptember 11989).

    FratianniMichele“The European Monetary System Ten Years After,”AllanH. Meltzer andCharlesPlossereds.Carnegie-Rochester Conference Series on Public PolicyVol. 32 (Amsterdam: North-Holland1990).

    GiavazziFrancesco andMarcoPagano“Capital Controls and the European Monetary System,” in Capital Controls and Foreign Exchange Legislation Euromobiliare Occasional Paper 21June1985.

    GiavazziFrancesco“The Advantage of Tying One’s Hands: EMS Discipline and Central Bank Credibility,”European Economic ReviewVol. 32 (Amsterdam1988) pp. 105582.

    GiavazziFrancesco andAlbertoGiovannini“Models of the EMS: Is Europe a Greater Deutschmark Area?” in RalphC. Bryant andRichardPorteseds.Global Macroeconomics (New York: St. Martin’s Press1987).

    GiavazziFrancesco“The Role of the Exchange-Rate Regime in a Disinflation: Empirical Evidence on the European Monetary System,” in FrancescoGiavazziStefanoMicossi andMarcusMillereds.The European Monetary System (Cambridge, England: Cambridge University Press1988).

    GiavazziFrancescoLimiting Exchange Rate Flexibility: The European Monetary System (Cambridge, Massachusetts: MIT Press1989).

    GiavazziFrancesco andLuigiSpaventa“Italy: The Real Effects of Inflation and Disinflation,”Economic Policy (Cambridge, England1989) pp. 13363.

    International Monetary FundInternational Financial Statisticsvarious issuesWashington.

    LaskarDaniel“International Coordination and Exchange Rate Stabilization,”Journal of International EconomicsVol. 21 (Amsterdam: North-Holland1986).

    MastropasquaCristinaStefanoMicossi andRobertoRinaldi“Interventions, Sterilization and Monetary Policy in the EMS Countries, 1979–87,” in FrancescoGiavazziStefanoMicossi andMarcusMillereds.The European Monetary System (Cambridge, England: Cambridge University Press1988).

    MelitzJacques“The Welfare Case for the European Monetary System,”Journal of International Money and FinanceVol. 4 (Guilford, England: 1985) pp. 485506.

    MelitzJacques“Monetary Discipline, Germany and the European Monetary System: A Synthesis,” in FrancescoGiavazziStefanoMicossi andMarcusMillereds.The European Monetary System (Cambridge, England: Cambridge University Press1988).

    Organization for Economic Cooperation and DevelopmentWhy Economic Policies Change Course: Eleven Case Studies (Paris: OECD1988).

    RogoffKenneth“The Optimal Degree of Commitment to an Intermediate Monetary Target,”Quarterly Journal of EconomicsVol. 100 (Cambridge, Massachusetts1985) pp. 116989.

    SachsJeffrey andCharlesWyplosz“The Economic Consequences of President Mitterrand,”Economic Policy (Cambridge, England1986) pp. 262322.

    TabelliniGuido“Monetary and Fiscal Policy Coordination with a High Public Debt,” in FrancescoGiavazzi andLuigiSpaventaeds.High Public Debt: The Italian Experience (Cambridge, England: Cambridge University Press1988).

    ThompsonEarlA.“Who Should Control the Money Supply?”American Economic Review Papers and ProceedingsVol. 71 (Nashville, Tennessee1981) pp. 35661.

    UngererHorstet al.The European Monetary System: Recent Developments (Washington: International Monetary FundDecember1986).

    VaubelRoland“The Return to the New EMS,” in KarlBrunner andAllanH. Meltzereds.Monetary Institutions and the Policy Process Carnegie-Rochester ConferenceVol. 13 (Amsterdam: North-Holland1980) pp. 173221.

    vonHagenJurgen“Monetary Policy Delegation and Fixed Exchange Rates: Credibility, Reputation and Inflation in the EMS,”Working Paper Indiana University School of Business1989.

    vonHagen andMicheleFratianni“German Dominance in the EMS: Evidence from Interest Rates,”Working Paper Federal Reserve Bank of St. Louis and Italian International Economic Center1989.

    WyploszCharles“Capital Controls and Balance of Payments Crises,”Journal of International Money and FinanceVol. 5 (Guilford, England: 1986) pp. 16779.

For a review of the debate surrounding the creation of the EMS, see Fratianni and von Hagen (1990).

The restriction to static, simple-shot games implies that there are no reputational considerations in the formulation of central bank strategies such as in Barro and Gordon (1983), Melitz (1988) analyzes a repeated game of policymaking in the EMS, where, however, only the high-inflation countries’ strategies are considered. Von Hagen (1989) presents a repeated game version with both countries endogenous. Note that Giavazzi and Pagano’s (1988) model is of a static nature although their game has several stages called “periods.”

Additionally, Giavazzi and Giovannini (1987) argue that intervention practices in the EMS free the Bundesbank from the need to adjust to EMS portfolio shocks, a point, however, that holds strictly only for nonobligatory, intramarginal interventions.

furthermore, and in contrast to what the credibility hypothesis implies, there is no evidence that EMS countries have experienced a lower cost of disinflation than non-EMS countries. On this point refer to Dornbusch (1989), De Grauwe (1989), and Fratianni and von Hagen (1990).

A more detailed explanation of the test can be found in Fratianni and von Hagen (1990) and von Hagen and Fratianni (1989).

We define ci = 0 and αij = 0 to mean ci(L) ≡ 0 and αij(L) ≡ 0

The data come from the International Monetary Fund’s International Financial Statistics. Money market interest rates are measured by line 60b, the index of industrial production by line 66c, the index of consumer prices by line 64. With the exception of domestic output, all data are seasonally unadjusted.

The monetary base is measured by line 14 and the budget deficit by line 80 in the International Financial Statistics. The weights underlying the construction of the world monetary base are based on GNP: 0.626 for the United States, 0.097 for the United Kingdom, 0.06 for Canada, and 0.215 for Japan.

Empirically, Δij(k) can be computed by summing over k lags of the appropriate impulse response function obtained from the estimates of the matrices A0, A1,…, in equation (16).

We use the following weights: Belgium 10.4 percent, Denmark 3.2 percent, France 22.1 percent, Germany 39.5 percent, Ireland 1.3 percent, Italy 11.0 percent, and the Netherlands 12.5 percent.

Note that the indicator is symmetric in the sense that Aij(k) = −Aij(k).

The authors’ argument that “realignments and capital and exchange controls explain why German long-term independence is not equivalent to German dominance” is not convincing per se; there have been no realignments for almost three years, and member states are committed to abolishing all exchange controls by mid-1990.

    Other Resources Citing This Publication