Impact of real exchange rate fluctuations on aggregate cocoa and coffee exports in Sierra Leone

Abstract

The regression and the vector autoregressive VAR models have been employed in this analysis. I use the autodistributed lag regression model to estimate both the short and the long-run impacts. In the VAR model, orthogonalized impulse response functions are employed to estimate the short-run. The regression result shows that while depreciation of the RER increases aggregate cocoa and coffee exports AGX in the current year, this variable is not significant in determining AGX in Sierra Leone. This is due to the fact that AGX have long gestation periods and until this period is over, suppliers cannot actually raise their output and hence exports. The negative effect of the one period lag of RER variable on AGX can be attributed to the fact that in the long run, depreciation in the nominal exchange rate leads to real exchange rate depreciation. This will lead to increase in cost of imported farming inputs in domestic currency terms. The reduction in imports that follows decreases the output and hence cocoa and coffee exports. However, this variable is not significant in determining AGX in Sierra Leone. An increase in the orthogonalized shock to the first difference of log RER causes a short series of increases in first difference of log AGX followed by a decrease, followed by an increase that dies out after four periods. The null hypothesis that the lag of first difference of log RER does not Granger-cause the lag of first difference of AGX cannot be rejected. The paper concluded that in the short and long-run, the RER should not be taken as policy variable to influence AGX in Sierra Leone.