MODELING MONEY DEMAND COMPONENTS IN LEBANON USING AUTOREGRESSIVE MODELS

34212

This paper analyses monetary aggregate in Lebanon and its different components using the methodology of AR model. Thirteen variables in monthly data have been studied for the period January 1990 through December 2005. Using the Augmented Dickey-Fuller(ADF) procedure, twelve variables are integrated at order 1, thus they need the filter(1- B) to become stationary, however the variable X 13,t (claims on private sector) becomes stationary with the filter (1- B)(1- B^12 ). The ex-post forecasts have been calculated for twelve horizons and for one horizon (one-step ahead forecast). The quality of forecasts has been measured using the MAPE criterion for which the forecasts are good because the MAPE values are lower. Finally, a pursuit of this research using the cointegration approach is proposed.

105

monetary aggregate , autoregressive model , stationarity , forecasting

Lebanese Science Journal

119

Lebanese Science Journal