현행 세수추계 방식의 개선방안 연구

Abstract

This study examines the Korean government's national tax revenue forecasting method, evaluates its performance, and derives its improvement measures. Like domestic and foreign tax revenue forecasting institutions, the Korean government forecast tax revenue using a bottom-up approach, where individual heads of revenues are forecast and then aggregated to produce a total revenue forecast. As a result of evaluating the government's budget-year revenue forecasting, the average error rate was －1.44% for 2000～2009, －1.62% for 2010～19, and －9.62% for 2020－22. Thus, the bias of underestimation has recently expanded. The fact that the average absolute error was 4.00% in 2000～2009, 4.78% in 2010～19, and 11.17% in 2020～22, suggests that the prediction accuracy has gradually decreased. The fact that the two-thirds spread of errors was 15.2% in 2000～2009 and 17.5% in 2010～22 implies that the uncertainty of forecasting has also gradually increased. Such the government’s performance for forecasts of tax revenue is mainly derived from corporate tax and asset taxation such as capital gains tax, inheritance and gift taxes, and securities transaction tax. Therefore, this study presents some recommendations that would improve the government's tax revenue forecasting performance. First, it is proposed to reduce the forecast errors of economic variables used as explanatory variables for revenue forecasting models, and to estimate and provide variables in the national account corresponding to the tax base for each tax item through the macroeconomic models. Next, it is suggested to break down revenue forecasts by sector and/or types of each tax item, making use of available sector-specific information, and to build a micro simulation model of income tax and corporate tax. In the stage of correcting non-model factors, this study recommends that it is necessary to accurately estimate and disclose the tax revenue effect of each government policy, to establish and operate communication channels with large companies to identify their latest business conditions, and to strengthen the expertise of tax revenue forecasters. Lastly, in the result evaluation and feedback phase, it is suggested that the public-private joint revenue forecasting committee should be permanently established to analyze and update the forecasting models, improve the contents of the 「Revenue Budget Estimation Analysis Report」, and prepare and disclose the tax revenue forecasting performance evaluation report.