This thesis investigates the monetary sector in Cameroon within an open economy framework. Two main hypotheses: money demand and Granger-causality are investigated. The data used are found to be non-stationary. Consequently, the money demand relationship is tested for the null hypothesis that it is spurious or not co-integrated. This is rejected in all the models put forward. The models are estimated and found to exhibit elasticities that are not unusual. Price homogeneity is found to be data incompatible. Income elasticities are generally found to be significantly less than unity suggesting economies of scale in money holdings. Corresponding dynamic models in the form of error correction are constructed using the familiar general to specific methodology and generally found to exhibit desirable statistical properties. Model preference is in terms of the narrow Ml definition of money with explanatory variables which include a foreign interest rate. For Granger-causality, the non-stationary data are transformed into stationarity where the null hypothesis of noncausality is tested in bivariate and multivariate contexts. Lag length selection is by the Final Prediction Error statistic. Results are mixed but two appear striking: domestic money and prices are found to be independent while domestic prices are Granger-caused by foreign variables but not by domestic ones.