Audit Selection under Weak Fiscal Capacity

Policy Context

Well-targeted tax audits are critical to deter tax evasion. In many high-income economies, the selection of which firm to audit is based on risk-based analysis systems. These systems are often beyond the capacity of tax administrations in many low- and middle-income countries. As a result, tax administration in these contexts often relies on the manual screening of tax returns by tax inspectors. The individual perspectives of these officials can help leverage local expertise, but also may subject taxpayers to discrimination.

Study Design

Should low-income countries leave discretion to auditors to select firms to audit or should selection be rule based? In partnership with the tax administration of Senegal, this pilot project aimed to understand how auditors select cases under discretion and compare these discretionary audits to audits selected by a risk-scoring algorithm. The intervention changed the selection method for audits at the tax authority by selecting part of the audit program with the risk-based algorithm.

Results and Policy Lessons

Initial results found that the inspector-selected audits are more likely to be conducted and lead to larger amounts of evasion recovered compared to algorithm-selected audits. There is suggestive evidence, however, that the evasion rate (recovery/turnover) is higher for algorithm-selected cases. The research team did not find any evidence that the provision of third-party data or the provision of third-party data and risk flags impacts the likelihood that an audit is started or the audit outcome. However, as inspectors are selective in their reporting of audit outcomes, these results may well change once we can conduct the analysis on the digitized data.

For more information, read the working paper here.