기업 재무자료를 활용한 법인세수 추계모형

초록

Estimation of corporate tax payment is important to the government’s budget operation. This is because most government’s budget depends on national income tax. A study by the National Assembly Budget Office shows that two estimating methods are in competition：one from estimation of macroeconomic variables such as GDP (traditional approach) and the other from estimation at firm level data (another approach). The flaws of the traditional approach are that it does not take into account company and industry level factors and that government agencies overestimate or underestimate economic fluctuations. However, if the new approach fails to take into account the appropriate company characteristics and tax－related factors, then incorrect estimation results will occur. Therefore, the researcher should consider financial data related to past, present and future corporate tax in order to reduce estimation error of corporate tax payment. The purpose of this study is to estimate the amount of corporate tax payments that show the greatest difference between the tax budget and the actual levy using firm－level financial data. First, we extract the financial variables related to corporate tax payment and develop an estimation model based on the relation between these variables and the actual tax payment amount. The validity of this estimation model is verified by comparing the estimated tax payment with the actual tax payment. The estimation results are as follows. First, we find that past amount of corporate tax, interest earnings, tax payables, and EBT are important factors in estimating the corporate tax payment. Second, we find estimation results are stable regardless of industry and firm size. However, we find that the estimation results are unstable if we include small firm size, small industries and specific year sample such as 2011 in which the amount of corporate tax payment fluctuated structurally. Also we find that short－term estimation provides better results because it minimizes estimation error and noise derived from historical tax data. The contribution of this paper is to find a new model that estimates corporate tax payment using firm－level financial data and reflecting corporate characteristics.

키워드