Univariate Time Series Forecasting: A Study on Monthly Tax Revenue of Bangladesh

Authors

  • Biplob Kumar Nandi East West University
  • Muntasir Chaudhury East West University
  • Gazi Quamrul Hasan East West University

Keywords:

Holt-Winters seasonal additive approach, SARIMA, ARIMA, Box-Jenkins Method, Tax revenue forecasting

Abstract

In recent years, Bangladesh has taken various steps to modernize tax system in order to enhance tax effort. Due to subsequent increase in financial constraints globally, economy's reliance on domestic resource mobilization continues to intensify. As a result, tax revenue target for every forthcoming budget appears to be buoyant albeit the prevalence of domestic constraints, i.e. inefficiencies in tax system, narrower tax base along with numerous exemptions and political instability. To enhance tax effort to reduce fiscal vulnerability, a esoteric revenue forecasting procedure is necessary. But, in Bangladesh, during the budget preparation, the method to target tax revenue is based on the growth rate extended with discretionary adjustments for a number of updated assumptions and personal judgments, which can lead to huge forecast error. This exercise attempts to identify an appropriate model by scrutinizing three approaches - ARIMA SARIMA multiplicative approach, Holt-Winters seasonal multiplicative approach and Holt-Winters seasonal additive approach - to forecast monthly tax revenue of Bangladesh and finds that, Holt-Winter seasonal multiplicative approach is the most appropriate method with minimum forecast error.

Downloads

Published

2015-07-28

How to Cite

Nandi, B. K. ., Chaudhury, M. ., & Hasan, G. Q. . (2015). Univariate Time Series Forecasting: A Study on Monthly Tax Revenue of Bangladesh . EAST WEST JOURNAL OF BUSINESS AND SOCIAL STUDIES, 4, 1–28. Retrieved from https://ojs.rsi-lab.com/index.php/ewjbss/article/view/113