Academic journal article Seoul Journal of Economics

Dynamic Analysis of Trade Balance and Real Exchange Rate: A Stationary VAR Form of Error Correction Model Approach

Academic journal article Seoul Journal of Economics

Dynamic Analysis of Trade Balance and Real Exchange Rate: A Stationary VAR Form of Error Correction Model Approach

Article excerpt

This paper analyzes the dynamics of trade balance and real exchange rate based on the elasticity and purchasing power parity (PPP) approaches. Here, a stationary vector autoregressive model with cointegration error, transformed from the error correction model in Kim (2012), is employed. Trade balance and PPP are jointly considered as the two long-run cointegration relationships that represent external economy equilibria. The model was applied to the dynamic analyses of Korea's trade balance using monthly data from 1990, where model variables from the elasticity and PPP approaches were selected. Based on the estimation, we first confirmed die finding of Cheung et al. (2004), whereas trade balance is additionally considered. The nominal exchange rate adjustment not the price adjustment, is the key engine that governs the speed of PPP convergence, and the nominal exchange rates were found to converge much more slowly than the prices. The nominal exchange shock did not significantly affect trade balance, whereas the price shocks did. Therefore, manipulation of the nominal exchange rate through intervention to improve trade balance might not be an effective policy tool.

Keywords: Trade balance. Real exchange rate, Error correction model. Stationary VAR

JEL Classification: F3

I. Introduction

Trade balance is mainly determined by real exchange rate based on elasticity approach, which reflects the relative price levels between two transacting countries. In particular, at trade equilibrium, the necessary condition for improvement of trade balance after an increase in real exchange rate is the well-known Marshall-Lerner condition. The dynamic relationship between real exchange rate and trade balance is substantiated by the J-curve effect, which predicts a trade deficit in the shortrun and a trade surplus in the long-run after a depreciation shock.

This kind of dynamic analysis of real exchange rate and trade balance involves the vector autoregressive (VAR) model that links directly the real exchange rate and trade balance. For instance, the traditional impulseresponse analysis from real exchange rate shock to trade balance was conducted using such VAR models (Goldstein and Kahn 1985; Rose and Yellen 1989; Moura and Silva 2005; Hsing 2008).

In another perspective, we can consider the J-curve effect and the convergence of real exchange rate from tíiat in a convergence process to cointegration equilibrium because economic theories generally do not know about the specific convergence process in equilibrium. In this respect, Pesaran and Shin (1996, 1998) proposed the persistence profiles and impulse-response analyses as indicators of the adjustment speed in a cointegrated VAR model. Another interesting application is the purchasing power parity (PPP). Engel and Morley (2001) showed that nominal exchange rates converge slowly, whereas prices converge relatively fast, by formulating the adjustment equations as a state-space model. Using the error correction model (ECM), the persistence profiles, and the generalized impulse-response analyses by Pesaran and Shin (1996, 1998), Cheung et oL (2004) discovered that the nominal exchange rate adjustment (not the price adjustment) is the key engine that governs the speed of PPP convergence, and the nominal exchange rates are found to converge much more slowly than prices. l This finding is very important in international or open macro-economic theory because it challenges conventional price-stickiness theories and raises new problems in modeling of PPP disequilibrium dynamics, essentially driven by the nominal exchange rate.

However, the aforementioned persistence profile approach mainly focused on the impulse-response (of cointegration error) analysis based on the Beveridge-Nelson decomposition. In addition, the Granger causality test and optimal forecasting of cointegration error are useful tool kits for the dynamic analyses of co-integration error dynamics. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.