Analisis Kebijakan Compliance Risk Management Berbasis Machine Learning Pada Direktorat Jenderal Pajak

Abstract

Implementation of tax self-assessment system gives full trust to taxpayers to calculate, pay, and report their tax themselves. To get the optimum result, the implementation of this system is determined by the level of compliance of the taxpayers. It is affected by internal and external factors such as technology, resources, legislation where the tax authority operating, organization’s aim and strategy, and public general tax conformity. This study aim to analyze taxpayer conformity level with machine based Compliance Risk Management (CRM) policy. This study is using qualitative approach through interview with people who have roles in implementing tax policy in Indonesia. This study founds the importance of machine learning based CRM policy, in which the tax authority cannot apply the same treatment to all taxpayers, so it needs to decide which taxpayer needs to be investigated with rational justification based on risk level. Tax authority needs to focus on implementing big data analytics with machine learning algorithm as an important source of information in decision making process, and helps predict taxpayers with potential fraud, so it can be used to reduce task risk before it happens.