Ertan ERSOY

The Determination of Long-Run Real Exchange Rates Under an Alternative Theory: The Case of Turkey

2008

(Supervisor: Nedim Süalp)

ABSTRACT

The exchange rate determination has become one of the most important fields of research in the studies of international economics during the last three decades. Although there has been an extensive literature about this subject over the same period, orthodox economists have expressed an increasing disappointment over their failure to explain the exchange rate movements.

Nevertheless there has been a pretentious approach developed by Anwar Shaikh (1980, 1991 and 1995) as an alternative theoretical model to the ones prevailing in the literature. Shaikh demonstrated, in his studies, that the long run real exchange rates move in such a direction to equalize profit rates, rather than price levels among trading nations. The main hypothesis of the model is that the long run real exchange rates are determined by the real unit labor cost ratios of the trading partners. The long run real exchange rates moving in line with the equalization of the profit rates may also be consistent with the chronic trade imbalances covered by endogenously generated capital flows among trading partners.

Besides, since, there is no presumption of labor mobility in the international scale, yet financial and non-financial capital and goods move freely across countries in the search for higher returns, real wage differences between countries can be greater and persistent. In addition, levels of technology can differ across countries.

This hypothesis empirically necessitates establishing a cointegration relation among long run real exchange rates, real unit labor cost ratios and real interest rate differentials. The findings of the empirical studies, performed by Shaikh (1998) for the data of the USA and Japan, by Roman (1997) for the data of Spain and by Antonopoulos (1997, 1999) for the data of Greece and her trading partners, showed that this model provides a sufficient theoretical and empirical framework for the explanation of movements in real exchange rates in the long run.

In the literature various orthodox models have also been applied to the data of the Turkish economy none of which are able to explain the determination of the long run real exchange rates. Therefore, it has been absolutely necessary to apply this alternative model for the data of Turkey. In order to investigate the theoretical hypothesis under consideration for the case of Turkey, we applied unit root tests, cointegration and error correction techniques to the quarterly data of Turkey and her main trade partners spanning from 1970 to 2004. In our empirical study we found that each of the long run real effective exchange rates, real unit labor cost ratios and real interest rate differentials are non-stationary in levels and stationary in first differences. In addition there is a quite significant cointegration relation between the variables and error correction mechanism works for the variables.