The Effects of exchange rates on bilateral trade balances of SACU members states with their trading partners

The fluctuations of exchange rates prevent countries from achieving stability in their external account records. Appreciation or depreciation has effects on international trade. This thesis examines the relationship between exchange rate fluctuations on bilateral trade balances focusing on the SACU region. There are several theories made to explain the relationship between exchange rate and trade balances. In examining this phenomenon, this thesis will unveil if the purchasing power parity theory, the Marshall-Lerner condition and the J-curve effect holds in the Southern African Customs Union (SACU) countries. This analysis is divided into three parts. The first part examines the stability of the exchange rate in the SACU countries in the long run as given by the purchasing power parity. To test for the Purchasing Power Parity theory, the recently developed powerful unit root test was applied with multiple smooth structural breaks of Omay (2015), based on a Fractional Frequency Flexible Fourier Form (FFFFF) on unique data of SACU countries covering the monthly period of 1995M01-2017M11. The Purchasing Power Parity (PPP) results show that the nominal effective exchange rate (NEER) of all SACU members does not provide evidence for PPP theory. In terms of the real effective exchange rate (REER), the PPP condition holds in the case of South Africa only. Further unit root investigations were carried out using the panel data for all SACU members, NEER and REER. The FFFFF test results for panel data shows strong evidence of the PPP while the standard DF test rejects PPP theory in the SACU’s NEER. Both the standard DF and the FFFFF tests show strong evidence of PPP theory in the case of SACU’s REER. The second section of the analysis examines the Marshall-Lerner condition employing annual data from the period of 1980-2017. The import and export model were examined firstly in a time series format and then in a panel data format. The time series data was examined using the ARDL (PMG) model while the panel data used the panel ARDL, fully modified OLS (FMOLS) method and the Dynamic OLS (DOLS) method of estimation. The PMG/ARDL model shows no evidence to support the existence of the Marshall-Lerner condition in the short run for all SACU members. However, only two out of five countries show evidence of the Marshall-Lerner condition in the long run. There is strong evidence of the Marshall-Lerner condition in Namibia and Botswana in the long run using the PMG/ARDL model