EMU and the International Monetary System

Comments on Cohen

Thomas Krueger, Paul Masson, and Bart Turtelboom
Published Date:
September 1997
  • ShareShare
Show Summary Details
Joly Dixon

This is an ambitious paper dealing with an interesting and relevant question. It develops a model of the relationship between the real exchange rate and monetary and fiscal policy instruments, uses this to investigate the change brought about by EMU, and then provides some empirical results. It gives a provocative answer to one of the key issues by claiming that the euro under certain conditions will be more volatile than the weighted average of the existing currencies.

It should first be noted that volatility in this paper has a very special meaning. It is defined as the failure to restore output and competitiveness or real exchange rates to previous levels after a price or demand shock. This is not what is generally meant by exchange rate volatility, and Bryant made a similar point about the paper by Bénassy-Quéré, Mojon, and Pisani-Ferry in this volume.

Second, it should be noted that even on its own terms the paper's conclusion only holds in a limited set of cases. It holds only with respect to “symmetric” shocks and only in the case in which price shocks are involved. In cases in which demand shocks are involved, the author's arguments conclude that EMU will dampen the volatility of the exchange rate.

I want to discuss two main issues: (1) the nature of the experiment; and (2) the applicability of the model. I will then add a short comment on the empirical results.

The experiment starts by looking at how France and Germany react to a symmetric price shock, that is, a shock that affects them each equally, but which has no effect on the rest of the world. It is assumed in that baseline case that there is no cooperation between the policymakers in each of the countries. From the perspective of each policymaker, competitiveness will be threatened (through higher prices for the representative good) and output could be reduced. Given the absence of coordination, each country will overstate the negative trade impact of the shock and monetary policy will be overly active. Because the overall trade impact is overstated, the policy reaction is overstated and the correction on the exchange rate with the rest of the world is higher than optimal—hence, lower volatility. As I mentioned before, volatility is defined in this specific way: real exchange rate volatility is low if the situation after policy measures have been implemented is close to the situation before the price shock occurred. The dynamics of exchange rate swings in between are not taken into account. That is the baseline against which EMU is compared.

In EMU the single monetary policymaker will in contrast correctly assess the trade impact and apply policy in line with the objective (or loss) function. The erroneous component due to lack of coordination will be absent—hence, in the language of the paper, a less accommodating monetary policy and a higher exchange rate volatility.

The high volatility argument thus boils down to a comparison between two situations in which policymakers have exactly the same set of objectives with respect to output effects of trade distortions and inflation, but in the baseline scenario the result is less optimal than anticipated because of the lack of coordination.

It thus seems erroneous to claim that in EMU volatility will be higher because policymakers care less about trade; they care the same, but now act on the basis of a correct assessment of the situation. Furthermore, I doubt that the baseline as specified in the paper is a good point of comparison. The existing situation in the EU should not be characterized as one of a total lack of cooperation between policymakers in the various countries. In fact, on the fiscal side there are many provisions for Community-level surveillance; and on the monetary side actions are clearly not independent. Some have indeed suggested that a more adequate representation of reality is that German monetary policy is made on the basis of German conditions, and then this policy is applied to the whole area. If this were indeed the case, the model here would give exactly the same results before and after EMU because we are only considering symmetric shocks.

Now I would like to turn to the model itself. As usual, a lot of assumptions have to be made when formalizing economic behavior; some are explicit, some are well hidden in the model. I want to discuss just three of these assumptions. First, the importance, or unimportance, of inflation. It seems that in this model inflation is a nonissue. In the model, real wages are simply assumed to be constant, and inflation has no real impact on the economy. Also, because the model is a static one-period model (which reduces in my view its usefulness as a tool for the analysis of exchange rate volatility, which is essentially a time-dependent concept), any dynamic relationship between an increase in the present price level (caused by the initial price shock) and expected inflation is explicitly assumed to be nonexistent. This is a strong assumption for a model that discusses monetary policy and exchange rate responses to price shocks.

Second, the objective or loss function. The paper constructs a composite loss function. Fiscal and monetary policymakers together aim at minimizing the weighted average of deviations of output from its trend, the fiscal stimulus parameter, and the inflation rate. I will come back to the arguments in this function, but a more fundamental issue is the composite nature of the function. I very much doubt that it can do justice to reality, especially when the main issue being investigated is exchange rate stability, which is often highly influenced by tensions between policymakers. The Maastricht Treaty is clear on the primary responsibility of the ECB toward price stability. Article 109 also lays out the framework within which exchange rate policy will be conducted, and recently a great deal of progress has been made on fiscal policymaking in EMU, with the agreement on the details of the Stability and Growth Pact. But the arrangements for the interactions between policymakers remain to be fully determined and tested. It seems to me that a loss function that treats them all together may well be insufficient. I also doubt that the weights attached to various arguments will be stable over time.

On the issue of the arguments in the loss function, we saw yesterday from the paper by Bénassy-Quéré, Mojon, and Pisani-Ferry that even if only the monetary authority's loss function is considered, results change depending on whether the relevant variable is the real exchange rate or the current account. I imagine that this model is equally sensitive to the specification of the arguments chosen.

Third, the paper assumes that the rest of the world is totally passive. In reality, the stability or instability of the euro will, I think, depend not so much on the European policy mix as on the relative policy mix. As Mr. Szász said yesterday in the roundtable discussion, “exchange rate stability cannot be attained unilaterally.“Nor can instability. One of the key issues for the stability of the euro/dollar exchange rate is the effect that EMU may eventually have on international monetary arrangements generally. My own opinion is that it has great potential.

Let me conclude by adding just one comment on the empirical results. A number of regressions have been run, using what is called a sample of European currencies. It is not at all clear why or how those currencies were chosen. It is, however, rather noticeable that a number of currencies that have been stable vis-é-vis the deutsche mark, for example, the Dutch guilder, are not included, and I suspect that this might well have a substantial impact on some of the empirical results obtained.

    Other Resources Citing This Publication