<![CDATA[This study aims to compare tax revenue forecasting model on nett tax revenue in contrast to its tax baseline component. We identifies tax baseline component that reflecting natural economic growth through interviews with Indonesia's Directorate General of Taxes (DGT). We employ ARIMA, ETS, linear model, and forecast combination to forecast both the baseline and nett tax revenue data using monthly national time series data from 2021-2023. By comparing Mean Absolute Percentage Error (MAPE), we determine the most accurate model and dataset combination for tax revenue forecasting. We finds that forecast from linear model in baseline tax revenue has the best MAPE of 5,17% and perform better than forecast combination as the best model from nett tax revenue with 8,30% MAPE. This study offers a novel perspective on tax revenue forecasting by employing a micro approach that focused on identifying baseline component from overall tax revenue. It has the potential to more comprehensive understanding of tax revenue behavior and lead to more improved fiscal control in Indonesia.]]>
Copyrights © 2025
