Testing the purchasing power parity theory for Pula/Rand and Pula/US dollar exchange rates in Botswana

Abstract

This paper tested the Purchasing Power Parity (PPP) theory between Pula/Rand and Pula/US dollar exchange rates in Botswana. The cointegration method, error correction model (ECM) and autoregressive distribution lag model (ARDL) were estimated to determine the validity of the theory between Botswana and South Africa and between Botswana and United States of America. The Engle-Granger cointegration method did not establish a relationship between the two exchange rates and prices and agreed with the results of studies by Tshipinare (2006), Rapelana (2014) and Sinha, Rapelana and Motlaleng (2018) that the PPP does not hold for Pula/Rand exchange rate. However, the estimated error correction model (ECM) showed rapid deviation of the variables to the long-run equilibrium, indicating a short-run relationship for Pula/Rand and Pula/US dollar exchange rates in Botswana. This contradicted the results of Rapelana (2014) and Sinha, Rapelana and Motlaleng (2018) and corroborated with Atta, Jefferis and Monnathoko (1996) for the Pula/Rand exchange rate and corroborated with Paul and Motlaleng (2008) in the case of Pula/US dollar exchange rates. A further investigation of a long-run PPP was conducted using the autoregressive distribution lag model (ARDL). The ARDL bound tests were conducted and they showed that the variables were cointegrated with each other for both Botswana and South Africa and between Botswana and United States of America. This indicated a long-run association between the variables and validated the long-run PPP theory between Botswana and South Africa and between Botswana and United States of America. The validity of the Pula/Rand and Pula/US dollar exchange rates indicates that Botswana has strong trade relations with the two countries. Hence, it is recommended that monetary authorities should try to balance the weights in the Pula basket to promote both the importing and exporting sectors.