Development of audit selection strategy for Delhi VAT

Abstract

Value Added Tax (VAT) has been implemented recently in India. One of the important aspects of VAT administration is Audits. Auditing of taxpayers on one hand improves compliance, but on the other hand, it leads to costs to the tax department as well as to the taxpayers. Every year, a large number of taxpayers fail to correctly declare their tax liabilities, and the Tax Administration is forced to tackle a tough task to detect them (and enforce compliance from them) without increasing the compliance costs of the tax compliant taxpayers. Selecting returns for Audits is like looking for a needle in a haystack. It is not possible to identify the likely taxevaders by simple query and reporting tools. Data mining techniques have been widely used in learning the patterns in the past data and these learnt patterns have been used to predict the behaviour of the current data. Knowledge Discovery in Databases (KDD) is a hot and emerging area of research as well as applications. An Audit Strategy based on KDD can meet the conflicting issues of audit planning, viz., trade off between maximizing audit benefits and minimizing audit costs. There is a possibility of application of data mining algorithm to meet the objective of an Audit selection strategy. This paper thus seeks to develop a model through application of data mining techniques for developing an Audit Selection Strategy for Delhi VAT.