Government Revenue Forecasting in Nepal

This paper attempts to identify appropriate methods for government revenues forecasting based on time series forecasting. I have utilized level data of monthly revenue series including 192 observations starting from 1997 to 2012 for the analysis. Among the five competitive methods under scrutiny, Winter method and Seasonal ARIMA method are found in tracking the actual Data Generating Process (DGP) of monthly revenue series of the government of Nepal. Out of two selected methods, seasonal ARIMA method albeit superior in terms of minimum MPE and MAPE criteria. However, the results of forecasted revenues in this paper may vary depending on the application of more sophisticated methods of forecasting which capture cyclical components of the revenue series. The prevailing forecasting method based particularly on growth rate method extended with discretionary adjustment of a number of updated assumptions and personal judgment can create uncertainty in revenue forecasting practice. Therefore, the methods recommended here in this paper help in reducing forecasting error of the government revenue in Nepal.