Country-by-Country Reporting (CbCR) has become one of the most important new sources of information on the global activities of multinational enterprises (MNEs). Originally introduced within the OECD/G20 BEPS framework for tax transparency and risk assessment purposes, micro-level CbCR data now offer unprecedented opportunities for both academic research and tax administrations to analyze the international organization of multinational firms, the location of profits and economic activity, and the functioning of the global tax system.
However, despite the increasing analytical importance of these data, firm-level CbCR datasets remain difficult to use in practice. Raw files frequently contain inconsistencies in identifiers, duplicates, reporting errors, sign reversals, unit inconsistencies, currency problems, and heterogeneous reporting practices across firms and jurisdictions. Moreover, the structure of the CbCR itself — based on jurisdiction-level aggregation of unconsolidated entity accounts — raises additional challenges related to intra-group transactions, permanent establishments, and the potential double counting of dividends.
This Note provides a practical methodological guide for cleaning and processing firm-level (“micro”) CbCR data. It proposes a set of harmonized procedures aimed at improving the internal consistency, longitudinal comparability, and analytical usability of the data. The document covers all major stages of the cleaning process, including:
The guide is intended both for researchers working with micro-CbCR data and for statisticians or analysts within tax administrations responsible for processing exchanged CbCR files and producing aggregated statistics. Particular emphasis is placed on reproducibility, documentation of corrections, and the importance of maintaining clear audit trails throughout the cleaning process.
A central message of this guide is that careful cleaning is a prerequisite for any meaningful use of firmlevel CbCR data. At the same time, the document shows that, once properly processed, micro-CbCR data appear broadly consistent with other major sources of information on multinational enterprises. This confirms the considerable potential of CbCR data as a rich source for tax analysis, international macroeconomic statistics, and research on multinational firms.
Finally, the Note argues that the exchange of best practices across researchers and tax administrations is essential. Greater convergence in cleaning methodologies would improve the comparability of empirical studies, strengthen the quality of OECD aggregated CbCR statistics, and contribute to a more reliable use of these data for both analytical and administrative purposes.