Causality between Devaluation and Trade Balance: Evidence from Portugal and Spain
Vamvoukas, George A., Indian Journal of Economics and Business
The Marshall-Lerner (ML) condition has been proposed as a theory to examine the links between devaluation and the trade balance. Since devaluation is a monetary phenomenon which affects several macroeconomic variables, its anticipated effects depend on the validity of the ML condition. This paper provides additional evidence of the effects of devaluation on the trade balance, using data from Portugal and Spain. The paper makes an attempt to standardize for very important variables which determine the behaviour of both the real effective exchange rate and the trade balance. Testing for causality is based on the procedure of Toda and Yamamoto (1995), who developed a technique which guarantees the asymptotic distribution of Wald tests. The paper incorporates in the causality procedure the information provided by the cointegration vectors in order to evaluate properly the long-run dynamic properties of the four-variable systems. Toda-Yamamoto causality results are not in line with the ML condition for Portugal and Spain.
Keywords:Marshall-Lerner condition; cointegration analysis; Toda-Yamamoto causality tests.
JEL classification:F31, F14, P22.
In the international trade literature three main theoretical approaches have been proposed, each having its own set of arguments (1). First, the elasticities approach which represents the basis of the Marshall-Lerner (ML) condition. Second, the absorption approach analysing haw devaluations affect the terms of trade, switch expenditure from foreign to domestic goods, raise total production and decrease absorption relative to total production, so improving the trade balance. Third, the monetarist approach supporting that devaluations change the relative price of traded and non-traded commodities and decrease the real value of cash balances, and, hence, the picture of the trade balance and that of the balance of payments. According to Miles (1979, p. 601), "if devaluation causes a significant improvement in the trade balance, this improvement should be statistically observable regardless of which theoretical approach is used".
The concept of real exchange rate constitutes the central variable in all three approaches. The nominal devaluation is considered successful when the real exchange rate is altered to its equilibrium value, whereupon the trade balance or the current account will improve provided that the demand elasticities of imports and exports are high enough. Most of the available empirical work shows that the total effects of devaluation will spread out over a period of more than one year. These effects will rarely take place within a month, a quarter or even a year. Many studies have attempted to establish a pattern of causality between devaluation and the trade balance. The ulterior purpose of a number of these studies is to explore the validity of the Marshall-Lerner (ML) condition.
Empirical studies presented by Maffett (1989), Bahmani-Oskooee (1991), Kulkarni (1996), Demeulemeester and Rochat (1995), Bughin (1996), Zhang (1998), Wang and Dunne (2000), Tang and Nair (2002), Vamvoukas (2003) and based on current time series techniques, have led to the conclusions that in a considerable number of countries (i) their trade balance and their effective exchange rate follow a long run cointegration relationship, (ii) the devaluation of the domestic currency serves to improve the trade balance, and (iii) the devaluation does not have any improvement on trade balance. In order to assess the relative validity of the ML condition, this paper applies a methodology which adds two basic points to the empirical literature. First, using data from two European Union countries such as Portugal and Spain, the paper investigates the causal channels between the real effective exchange rate and the trade balance within the framework of multivariate systems. The purpose of this methodological procedure is to explore the direction of causality between the real effective exchange rate (R) and the trade balance (T), introducing other important variables such as the real output and the rate of interest which influence the behaviour of both R and T. …