“This paper looks at the functioning of the parallel market and discusses macroeconomic policies in the country to shed light on the context within which the market evolved. Analysis of time series properties of the variables used in an empirical model covering 1970–1998 reveals that the premium is stationary, guiding the decision to estimate a stationary model of the premium. Econometric results show that the premium is determined by the expected rate of devaluation, trade policy variables and GDP growth. This conforms with empirical studies on other African countries, although the Burundian case is particular in one respect:
The market is characterized by a relatively low premium, which, as a stationary variable,
is also generated by stationary fundamentals, in contrast with findings for other African countries. This particularity reflects a relatively more flexible exchange rate policy over the sample period. In terms of policy, the paper argues that political stability and more fundamental changes in Burundi’s economy and its management will be needed to ensure the success and sustainability of the current foreign exchange reforms.”
