Working Paper 71 - Exchange Rate Policy and Currency Substitution: The Case of Africa’s Emerging Economies

This paper uses an error-correction model to examine the dynamic of the currencysubstitution phenomenon (CS) in two of Africa’s emerging economies: Egypt andSouth Africa. The study also assesses the causal relationships of this phenomenon.There are three main CS- related differences between Egypt and South Africa.These are: (i) the orientation of economic policy and the instrument used, (ii) thedegree and level of CS, and (iii) the direction or trend of CS. During the studyperiod 1991-2001, Egypt used the exchange rate as an anchor to its economicstabilization program. While in the case of South Africa, the monetary authoritiesdirectly targeted inflation by controlling the growth of money. During the sameperiod, CS in Egypt started with a very high level (51% of M2 was in the form offoreign currency deposits) and experienced a steady decline till it reached less than20% in 1999. On the contrary, CS in South Africa was less than half a percentagepoint in 1991 and observed a significant and uninterrupted increase till it passed the6% mark in 2001. The results of the error-correction model suggest that the elasticityof CS, with respect to exchange rate, of South Africa is 2.3 times that of Egypt,and that the speed of adjustment in South Africa is 5 times faster than in Egypt.The results of the Granger-causality tests between the exchange rate and CS, inboth Egypt and South Africa, agree that there is a unidirectional Granger-causalrelationship from the exchange rate to CS. However, the tests between the interestrate differential and CS indicate that while the causality runs in one direction fromthe interest rate differential to CS in the case of South Africa, it runs in the oppositedirection in the case of Egypt. The implication of these results is that the policy ofexchange rate anchoring is more suitable to a high CS environment, and moreeffective in reducing the rate of substitution. However, this policy is not free ofcosts. As for the inflation targeting policy, the implication is that the policy can beeffective in achieving its objective as long as the CS is insignificant. In an environmentof High CS, monetary authorities should consider switching to the exchange rateanchoring policy.