Author Description

Assessing tax gaps-the difference between the potential and actual taxes raised-plays a vital role in achieving positive domestic revenue objectives through improved and reformed taxation. This is particularly pertinent for growth outcomes in developing countries. This study uses a bottom-up approach based on micro-level audit information to estimate the extent of tax misreporting in Zambia. Our methods predict the extent of tax evasion using a regression and a machine learning algorithm based on a sample of audited firms, after which we estimate tax gaps using a standard approach. We estimate total tax gaps as 56 per cent and 47 per cent for the two approaches, respectively. These gaps are mainly driven by corporate taxes. Applying our gap to key industries shows that the extractives sector in Zambia records the highest gaps in terms of CIT and one of the lowest gaps in terms of VAT.