Has Exchange Rate Uncertainty Increased in the EMS Since September 1992?
Eric Perée, Alfred Steinherr *

Since the Summer of 1992 the European Monetary System has been rocked by several waves of turbulence. In September 1992, the Italian lira and the British pound left the exchange rate mechanism of the EMS. While remaining in the ERM, the Spanish peseta, Portuguese escudo and Irish punt have been devalued. At the beginning of August 1993, the permitted fluctuation margins within the ERM have been widened to 15 percent. This has renewed the fear that increased exchange rate flexibility could have a negative effect on trade flows within the European Community. Indeed, variability of nominal exchange rates within the EMS has increased substantially in the last eighteen months.

We have argued elsewhere (see Perée and Steinherr, 1989) that short-term exchange rate volatility may entail only small costs to international traders as this type of risk can usually be covered in financial markets. However, it is much more difficult to hedge risk beyond a one year horizon both because financial markets are much less complete for periods beyond one year and because exchange needs are not known without precision. Furthermore, from a resource allocation perspective, it is not clear that nominal exchange rate volatility is of crucial importance. High short-term exchange rate volatility at times does not create large and protracted changes in real exchange rates, which are likely to be more costly in term of welfare and resource allocation.

In the reminder of this paper, we introduce two measures of medium-term exchange rate uncertainty (Section 1) and apply them to the ITL/DEM bilateral rate (Section 2). In Section 3, we apply our measures of exchange rate uncertainty to large trade zones in the world economy before providing a few figures of the estimated export loss induced by recent exchange rate uncertainty in Section 4.

1. Measuring exchange rate uncertainty

For several years, the European Monetary System has provided an island of relative exchange rate stability. Nominal exchange rate fluctuations were constrained by the ERM arrangements and no adjustment of central rates took place in the period between February 1987 and August 1992. While this suggests that, until the end of August 1992, the EMS arrangement successfully limited nominal exchange uncertainty, it did not stabilise real exchange rates due to diverging inflation performances. Hence large shifts in competitive conditions resulted and economic agents, while mostly free from nominal exchange rate risk were confronted with increased real exchange rate uncertainty.

We consider two alternative measures of medium-term uncertainty which capture simultaneously nominal and real exchange rate developments. Consider first the measure:

where Xt is the nominal exchange rate at time t, the max and min operator refer to maximum and minimum values over a past period of time of length k up to time t and   is the "equilibrium" exchange rate as measured with relative purchasing power parity based in 1960. V1 captures nominal exchange rate variations over a given past period. For example, if the exchange rate in the period t-1 remains constant but had achieved in previous years a record fall or rise then we postulate that agents will not have high confidence in future exchange rate stability. To nominal fluctuations, the term V2 adds the effect of deviations from equilibrium and postulates that as misalignment grows linearly uncertainty increases exponentially.

2. An application to the ITL/DEM bilateral rate

To illustrate the effects of exchange rate misalignments and of exchange rate uncertainty, we focus on the Italian Lira (ITL). Since September 1992 the value of the ITL has declined by over 20 per cent in terms of DEM. This decline was accompanied by a sharp increase in volatility which is the standard measure of risk. According to Figure 1 the ITL became undervalued on a PPP basis in 1991 and on the basis of a Williamson-type equilibrium rate in 1992. The depreciation of the ITL seems therefore to overshoot any possible correction for overvaluation.

The degree of misalignment is depicted in Figure 2, ranging at present from 15 to 25 per cent.

Figures 3 to 5 illustrate the increase in exchange rate uncertainty based on the proposed measures. In Figure 3 the maximum movement of the ITL rate over a five year period (with decay) is the measure of uncertainty. Clearly, the perceived ITL/DEM uncertainty increases substantially in 1993, but remains well below those levels observed for the last years before creation of the EMS. Similar results are obtained for measures V and U, respectively graphed in Figures 4 and 5.

Thus, for the ITL/DEM rate - as for most rates in the EMS - exchange rate uncertainty has increased dramatically in 1992.





3. Exchange rate uncertainty in a multilateral perspective

Figures 6 and 7 depict the two uncertainty variables V and U for the DEM. In the European Union exchange rate uncertainty reached its maximum in the late 1970s before the creation of the EMS. It then declined substantially until 1992 and increased again since then on the basis of measure V, but not yet on the basis of U (which is a smoothed measure and requires more time for a reversal).

The question then is: has this increased uncertainty had any effects on foreign trade? The next section outlines an answer.

4. The effect of exchange rate uncertainty on exports

In Perée and Steinherr (1989) we have estimated export equations in which the variables U and V were statistically significant. We now use these estimated coefficients to compute the effects of increased U and V on export volumes. The results in Table 1 are therefore only due to changes in uncertainty and do not include the effects of changes in real exchange rates.

In 1993 increased uncertainty indeed depressed export volumes of the various countries considered. The reduction varies between 1 percentage point in the UK to 2.5 per cent in Germany and Japan. Belgium - a small and very open economy - suffers much more and its exports may be depressed by as much as 5 per cent due to higher exchange rate uncertainty.

Table 1: Export Volume Losses Caused by Exchange Rate Uncertainty in 1993

(% of 1980 export volume)

Exports to the US
Measure Japan Germany  UK Belgium
V 1.7395 1.9324 1.0248 1.6671
U - 1.4594 - 1.3831

Aggregate Exports
Measure US Japan Germany  UK Belgium
V - - - 1.1022 5.0506
U - 2.6907 2.5302 1.1376 3.4681


- = not significant at 5% level. A positive change is a reduction in exports.

5. Conclusion

The EMS upheavals have contributed to higher exchange rate uncertainty which in turn has depressed foreign trade in 1993. Exchange rate stability in real terms seems to matter and this empirical finding should not be forgotten by policy-makers.


PEREE Eric and Alfred STEINHERR, 1989, Exchange Rate Uncertainty and Foreign Trade, European Economic Review (33) 1241-1264.