Identifying Best Method for Forecasting Tax Income using Time Series Analysis

Fitri Pebriani Wahyu (Departement of Public Administration, Faculty of Social and Political Sciences, UIN Sunan Gunung Djati Bandung, Indonesia)
Indriyani Dwi Rahmawati (Departement of Public Administration, Faculty of Social and Political Sciences, UIN Sunan Gunung Djati Bandung, Indonesia)
Khaerul Umam (Departement of Public Administration, Faculty of Social and Political Sciences, UIN Sunan Gunung Djati Bandung, Indonesia)

Tax Literacy and Leadership Agility: Indigenous Peoples Awareness in Paying Taxes
2022: Proceedings IAPA International Conference and International Indonesia Conference on Interdisci

Regional independence can be seen from the high or low local indigenous income. In doing planning, forecasting is needed as consideration for policy making. For economic development planning, accurate predictions of regional income are needed. Majalengka Regency as one of the districts included in the national legislation program Segitiga Rebana area is projected as the driving force for the economic growth of Java Province Barat. The research method uses secondary data on regional tax revenue receipts of Majalengka Regency obtained from the Central Statistics Agency. Data analysis using time series with models tested including Single Exponential Smoothing, Double Exponential Smoothing, Winters Method Additive, and Winters Method Multiplicative. The study aimed to find the best forecasting methods for receiving local tax incomes. The results indicated the Winters Method Additive is the best forecasting method that can be used to forecast local tax incomes. The Mean Absolute Percentage Error of Winters Method Additive reaches the accurate category with a value of 14% when level is 0.1, trend is 0.2, and seasonal is 0.1.