Arima Models For Forecasting Own Source Revenue For The County Of Machakos

Abstract

Fiscal stresses have created a need for County Governments in Kenya to use accurate Own Source Revenue figures in budgeting for good economic planning. This is because they are faced with unlimited demands from taxpayers coupled with limited tax or revenue resources. Counties have not adopted a formal any quantitative forecasting technique leading to forecasting errors and disruption in government services. Therefore, forecasting is becoming increasingly relevant and essential in the context of county governments in Kenya. Forecasting is not only a legal requirement, but an essential tool for fiscal planning. This study therefore aims at developing an autoregressive integrated moving average model and using it to forecast own source revenue for the County Government of Machakos and then comparing the forecast results with prediction generated using the expert judgment approach which is the current in-house forecasting technique in the county. The findings show that the auto-regressive integrated moving average method generated forecasts with a higher level of accuracy than those generate through the expert judgment approach. This study uses own source revenue data from the financial year 2013/2014 to 2019/2020 while the mean absolute percentage error is used as the measure of accuracy. This study recommends adoption the Auto-regressive Integrated Moving Average models for forecasting of Own Source Revenue for Machakos County.