The Revenue Administration Gap Analysis Program
Technical Notes and Manuals · 2026-11 · 2026 (014) · 1
DOI: 10.5089/9798229060639.005 ↗Kurz gefasst
Fachartikel in der Zeitschrift Technical Notes and Manuals, 2026-11. Verfasst von Shan Zhong und 3 weiteren Autor:innen.
Abstract (Originalsprache)
This technical note provides practical guidance on using machine learning to support bottom-up tax gap estimation within revenue administrations. It explains how taxpayer-level administrative and audit data can be used to estimate the assessment component of the compliance gap, while emphasizing the need to address sample selection bias arising from non-random operational audits. The note discusses data readiness, audit-selection correction, model selection, validation, interpretation, and uncertainty, and compares machine learning with traditional econometric approaches. It also highlights how granular estimates can inform compliance risk management by identifying taxpayer segments, sectors, and behaviors associated with higher predicted non-compliance. Throughout, the note stresses that machine learning should complement—not replace—sound methodology, professional judgment, and existing RA-GAP frameworks. The focus is on using advanced analytics responsibly to translate administrative data into credible, actionable insights for tax administration.
Zitieren
APA
Zhong, S., Julian, L., De, S., & Loro, P. D. (2026). The Revenue Administration Gap Analysis Program. Technical Notes and Manuals, 2026(014), 1. https://doi.org/10.5089/9798229060639.005
Vancouver
Zhong S, Julian L, De S, Loro PD. The Revenue Administration Gap Analysis Program. Technical Notes and Manuals. 2026;2026(014):1. doi:10.5089/9798229060639.005
BibTeX
@article{Zhong2026,
author = {Zhong, Shan and Julian, Lucienne and De, Supriyo and Loro, Pierfrancesco Di},
title = {The Revenue Administration Gap Analysis Program},
journal = {Technical Notes and Manuals},
year = {2026},
volume = {2026},
number = {014},
pages = {1},
doi = {10.5089/9798229060639.005}
}
EUVIMED