The Balassa - Samuelson Hypothesis: Turkey and foreign trade partners application

This paper aims to investigate validity of "Balassa-Samuelson Hypothesis" for Turkey and 10 member countries of the European Union (Germany, France, Britain, Italy, Netherlands, Spain, Belgium, Greece, Poland, Romania) and NAFTA countries (Canada, Mexico, United States) which Turkey's major trading partners using annual data in period from 1991 to 2013. Hypothesis argues that differential between the tradable and non-tradable sector productivity among countries determine the real exchange rate. In this study, first of all Levin Lin Chu; Hadri; Breitung; Im Pesaran Shin; Fisher ADF; Fisher PP unit root tests were performed to investigate stationary of the series (the real effective exchange rate index and relative productivity index), then The existence of long-run relationships between variables were investigated by Pedroni Panel cointegration test ve Kao Panel cointegration test, and finally the long-run coefficient between variables were estimated by Fully Modified Ordinary Least Squares (FMOLS) and Dynamic Ordinary Least Squares (DOLS). Findings can be summarized as follows: a) According to unit root tests the series are not stationary at level but in the first difference they are stationary at 1% significance level. This finding also proves the "Purchasing Power Parity" does not holds in the these countries; b) According to Pedroni Panel cointegration test ve Kao Panel cointegration test long-run cointegrating relationship between variables has determined at 5% significance level; c) FMOLS and DOLS methods prove "Balassa-Samuelson Hypothesis" in both terms of panels and for all countries at 1% significance level. According to FMOLS and DOLS methods a 1 unit increase in the relative productivity index increases the real effective exchange rate index approximately 0.96 unit in the long-run. Key Words: Balassa-Samuelson Hypothesis,Real Exchange Rate, Tradable and Non-Tradable Sector Productivity, Fully Modified Ordinary Least Squares (FMOLS), Dynamic Ordinary Least Squares (DOLS)