Contenu du sommaire : Économie des taux de change

The Real Equilibrium Exchange Rate: An Introduction by Hervé Joly, Céline Prigent and Nicolas Sobczak This article presents the two main real equilibrium exchange rate theories. The first approach is based on a two-sector general equilibrium model. It defines the real exchange rate as the relative price of tradeable goods compared with goods. It is the most complete as a theory, but its empirical application is complicated. The second approach is "macroeconomic". The real exchange rate is defined here as a competitiveness indicator. Its equilibrium level enables equilibrium to be attained whilst remaining compatible with internal equilibrium. The empirical calculation is easier and uses most robust mechanisms in macroeconometric models. These two approaches have certain points in common, notably intertemporal definition of external equilibrium. They moreover appear to be largely complementary insofar as they implicitly to different forecasting periods.

Heterogeneity, Long Memory and the Dynamics of the Real Exchange Rate by Patrick Fève The purchasing power parity assumption could be considered to be one of the most influential equilibrium exchange rate in open economy macroeconomics. It is empirically one of the most studied assumptions, but also one of the most disputed. identify the main opportunities and threats of a quantitative evaluation of the PPP and propose testing this assumption by fractional processes. This type of modeling partially addresses the problem of heterogeneity, but above all provides a more representation of the dynamics of the real exchange rate. Long memory tests are applied to the real exchange rate for the franc against the deutschemark, the US dollar, the Italian lira and the pound sterling. Yet the scope of the results derived from modeling remains somewhat limited, as it is difficult to be conclusive about the empirical pertinence of the long-run PPP for the deutschemark and the lira. Moreover, the unit root assumption cannot be rejected for the dollar and the pound sterling.

Forecasting the Canadian-US Dollar Exchange Rate: An Approach in Terms of "Fundamentals" by Mark P. Taylor We develop and estimate a forecasting equation for the Canadian-US dollar exchange rate by conditioning on short-run and determinants of the exchange rate. The equation performs well compared with an equation developed by the Bank of is statistically well-determined. The model does, however, have certain Keynesian features which, at the long-run intriguing.

Monetary Exchange Rate Models: A New Empirical Study by Éric Jondeau This paper makes an empirical study of the two main monetary exchange rate models, which assume perfectly flexible or short-run prices. In view of the non-stationarity of exchange rates and their determinants, the testable specifications normally from these models should be interpreted as long-run relations and the tests carried out using a multivariate error correction The monetary models are therefore estimated for the yen-dollar and the deutschemark-dollar over the 1973-1994 period. The induce a virtually systematic rejection of the flexible-prices model and appear considerably more favourable to the rigid-prices model.

Government Debt and Exchange Rate Dynamics: Empirical Evidence for G7 Countries over the Last Two Decades by Alain Paraponaris Empirical exchange rate studies can be roughly divided into univariate models and open economy macroeconomic models. The former have no particular grounding in macroeconomic theory, but have the best predictive short-run performance. The latter perform less well in terms of the short run, but help explain long-run changes in exchange rates. This paper presents an overview of some stylised facts for G7 countries and endeavours to link the recent behaviour of exchange rates to changes in government debt. The empirical study is based on cointegration tests, causality tests within error-correction models, simulations of shocks on variables and the breakdown of forecasting error variance. This study distinguishes the countries in which the causal relation runs from government debt to exchange rates (Canada and the United States) from those in which the relation is reversed (France and Germany). The relation is found to be bidirectional for Great Britain and Italy. Results for Japan are inconclusive.

The Temporal Consistency of Exchange Rate Expectations: A Study Using Survey Data by Agnès Bénassy-Quéré and Hélène Raymond The consistency of forecasts concerning different dates is a less stringent assumption than that of rational expectations, frequently rejected by survey data. However, temporal consistency tests are conditional on the forecast modeling adopted. The temporal consistency assumption is tested here for expectations relating to four exchange rates conditional on four forecast models. Three different surveys are processed and several models estimated to guarantee reasonably general results. Nevertheless, these results need to be made relative to the explanatory capacity of each model. The three-month exchange rate forecasts appear to comply with a naive model, while the longer run forecasts are more in line with a mixed extrapolative-adjustive-regressive model in which the regressive component is often preponderant. This variety of models used depending on the future time periods explains why the temporal consistency assumption is often rejected and, consequently, strengthens the assumption of irrational expectations.

EMS Credibility and Macroeconomic Fundamentals: An Empirical Study by Benoît Cœuré and Antoine Magnier The connection between the credibility of the EMS parities and changes in macroeconomic fundamentals is at the core of the economic debate that started following the 1992 and 1993 exchange rate crises. This study proposes an empirical evaluation of this link over the 1983-1993 period. We first put together a measure of devaluation expectations for the French franc/deutschemark and lira/deutschemark parities using the method of exchange rate "drift correction" in a target-zone arrangement. We then use non-stationary VAR modeling, Granger causality tests and the calculation of response-to-macroeconomic-shock functions to evaluate the role of the fundamentals in the formation of devaluation expectations. Certain macroeconomic variables seem to affect the credibility of the parities studied over the 1983-1993 period, but these links do not generally appear to be particularly intuitive or significant from an economic point of view. Nevertheless, the effect of exogenous shocks on the credibility of the French franc/deutschemark parity tends to ease off fairly quickly, while this same effect appears highly persistent in the case of the lira/deutschemark parity. The results are not really conclusive as regards the nature of the 1992 and 1993 exchange rate crises. However, they do suggest that the theoretical explanation for the crises in terms of multiple equilibria could be justified in part. More importantly, the results underscore the need for a more structural econometric approach.

Exchange Rate Crisis Models: A Summary Focusing on the 1992-1993 French franc Crisis by Olivier Jeanne The purpose of this article is to provide a summary presentation and comparison of the exchange rate crisis mechanisms posited by economic theory; to identify the assumptions suggested by these mechanisms as to the origin of the 1992-1993 French franc crisis; and to compare them with the "stylised facts". We distinguish two theoretical approaches: speculative attack models and the "opt-out clause" approach. The speculative attack models focus the analysis on the phenomenon of the drain on exchange reserves when crises occur. The opt-out clause approach looks more at the reasons why monetary authorities choose to defend or not defend the currency when it is attacked. The opt-out clause approach makes the connection between the French franc crisis and certain fundamentals such as French unemployment and Germany's monetary policy. Moreover, the two approaches concur with the assumption that the French franc crisis was not caused by fundamentals, but was self-fulfilling.

The Vulnerability of Fixed Exchange Rate Systems and Capital Control by Bernard Bensaïd and Olivier Jeanne In this paper, we study the dynamics of currency crises under the following two hypotheses. First, it is costly for the to raise the nominal interest rate above a certain level in order to defend the currency. Second, the interest rate level that government is willing to bear before it decides opt out is not perfectly known by the foreign exchange market players. We that, given these conditions, the fixed exchange rate system is vulnerable to self-fulfilling currency crises unconnected with fundamentals and that such crises could force the government to opt out of the fixed exchange rate system. We then specify conditions under which capital controls might deter this type of currency crisis.

Is the New EMS More Asymmetric than the Old? by Christian Bordes, Éric Girardin and Velayoudom M arimoutou The purpose of this paper is to make a new study of the asymmetric relations between the interest rates of the EMS member and the German interest rates. The approach is twofold. First, we select a formulation in terms of the response function of monetary authorities. In addition to the American rates, which explain the external influences on the EMS, the VAR model also includes the rates of inflation for an EMS member country and Germany. Secondly, we extend the estimation period to expansion of the margins and carry out stability tests to show that 1987 was a major break point. This break can be associated the relaxation of capital controls. Over these two sub-periods, the EMS appears as an asymmetric system due to the fact that influence of the German rate innovations on the forecasting errors for the other European rates is three times greater overall the influence in the reverse direction. The new EMS generally looks at least as asymmetric as the old and probably more so in French case.

A Return to Average Rates in the European Exchange Rate Mechanism: 1987-1995 by Jean Sébastien Pentecôte and Thierry Roncalli This paper studies the effect of the return to average daily rates for the European Exchange Rate Mechanism currencies and the dollar against the deutschemark over the 1987-1995 period. The error correction term appears to fluctuate over this interval. However, it is rarely significant. Our study does not conclusively uphold the findings of the theoretical target zone models. With the exception of the Belgian franc and the guilder, the estimated exchange rate appeal value differs from the official parity. As regards the dollar, the Louvre Agreements would appear to have been brought into question following the October 1987 stock exchange crash.

Conditional Volatility, Trading Signals and Risk Perception by Giampiero M. Gallo and Barbara Pacini In this paper, we study the role of a conditional volatility term in modeling the presence of risk in the relationship between monthly spot and forward rates. We comment on the disappointing performance of the Garch-M model when used to study this relationship. This performance is due mainly to the high degree of persistence implied by the parametric estimation. In our view, a more flexible formulation should be adopted to model the risk premium and a distinction should be made between the alternating periods of weakness (when the risk premium is positive) and strength (when the risk premium is negative and becomes a risk discount). Taking three currencies against the deutschemark (French franc, Italian lira and pound sterling), we suggest a semiparametric estimator of the relationship, adopting a nonparametric measure of the conditional volatility. This enables us to consider the currency purchase and sale signals derived from technical analysis filters.

Uncertainty, Trade-Off and Exchange Rates by Jean Laurent Viviani Although we do not deny the pertinence of the explanations put forward for the existence of price brackets on the financial markets, we show that they could also be due to agents' rational behaviour in an environment to which probability cannot be applied. In this context, the aversion of agents to uncertainty is behind a difference between the purchase and sale price, which is compatible with the absence of trade-off opportunities.