|
Title
|
|
|
|
Tax disclosure : firms’ governance characteristics and utilization by tax authorities
| |
|
Author
|
|
|
|
| |
|
Abstract
|
|
|
|
| Through two archival studies, this dissertation explores the relationship between firm-level governance characteristics and tax disclosures. In addition, a third methodological study provides a framework to value tax disclosures based on a machine learning approach, which subsequently applies this framework to a real-world case. The findings demonstrate that increasing family ownership in a firm is significantly negatively associated with the level of tax disclosure, but do not demonstrate robust evidence of a relationship between family involvement in management and tax disclosures, or between family involvement on the board of directors and tax disclosures. Furthermore, the dissertation shows a positive relationship between employee representation on corporate boards and a firm’s level of tax disclosure and a positive relationship between the existence of a CSR subcommittee and a firm’s level of tax disclosure. We find no robust evidence of a relationship between board gender diversity and the firm’s level of tax disclosure. Additionally, the consequences of governance mechanisms are shown to be influenced by a country’s formal and informal institutions. Finally, this dissertation develops and applies a machine learning approach to value data based on the theoretically sound Shapley value to assist the Belgian Federal Tax Authorities in managing large quantities of data received under various exchange of information agreements. This valuation technique determines which data is valuable for predicting successful tax audits. This valuation technique does not rely on the assumption that all features in the machine learning model must be independent like the often-applied SHAP and SAGE methods, enhancing the validity of its results in a real-world setting. This dissertation intends to inform regulators and other stakeholders of the pitfalls and effectiveness of governance mechanisms that could monitor a firm’s tax management and payments. In addition, this thesis provides a framework to value tax disclosures that can effectively assist tax authorities in filtering out useful tax disclosures. |
| |
|
Language
|
|
|
|
English
| |
|
Publication
|
|
|
|
Antwerp
:
University of Antwerp, Faculty of Business and Economics, Department of Accountancy and Finance
,
2024
| |
|
DOI
|
|
|
|
10.63028/10067/2079140151162165141
| |
|
Volume/pages
|
|
|
|
xviii, 216 p.
| |
|
Note
|
|
|
|
:
Jorissen, Ann [Supervisor]
:
Van de Vijver, Anne [Supervisor]
:
Martens, David [Supervisor]
| |
|
Full text (open access)
|
|
|
|
| |
|