Modélisation de la Volatilité des recettes mensuelles de la Direction Générale des Douanes et Accises (DGDA ex-OFIDA) en RDC de janvier 1982 à décembre 2005

For a few years the revenue services of the DGDA have increased in a spectacular way in Democratic Republic of Congo. Thus, the objective of this paper is to empirically examine the evolution of these monthly receipts of 1982 to 2005. The Heteroskedastic Conditional Autoregressive model (ARCH) initiated by Engle (1982) was chosen to take into account the conditional variance of the error depending on time. The econometric literature informs us that if we generalize the model ARCH(p), while adding of the explanatory variables who represent variances of the shifted errors of q period, we obtain the Generalized Autoregressive Conditional Heteroskedastic GARCH(p,q) model. The empirical results of our study show that the receipts DGDA are represented by a model GARCH(1,1). This means that the heteroskedastic conditional variance of the errors of the monthly revenue services of the DGDA from January 1982 to December 2005 depends on its lag value and on the lag values of the error squared.