How do we improve the collection and use of administrative tax data to better serve both men and women? This question was at the heart of the recent Community of Practice on Gender and Tax (CoPGT) meeting, where researchers and practitioners explored how administrative data can help uncover gendered patterns in taxation and compliance. The discussion, featuring presentations from experts at the World Bank, ICTD and UNU-WIDER, underscored both the opportunities and challenges in leveraging administrative tax data for gender analysis.
What is Administrative Tax Data and Why Does it Matter?
Administrative tax data refers to information collected by tax authorities in the process of carrying out their functions, including taxpayer registrations, income declarations, withholding tax reports, customs data, audit records, and payment transactions.
Sex-disaggregated administrative tax data is a powerful tool for analysing disparities in tax and other policy outcomes. Unlike traditional surveys, administrative data covers the entire taxpayer population, allowing researchers to, among other things:
- Analyse gender gaps in income, business ownership, wealth distribution, etc through tax records;
- Assess differences in tax compliance behaviour by comparing filing and payment rates for men- and women-owned businesses;
- Identify potential biases in tax enforcement such as whether women-owned businesses are audited at the same rate as those owned by men;
- Measure the impact of tax reforms on different genders by tracking changes in tax variables over time.
Tax administrative data helps to understand gender disparities but presents accessibility limitations
Administrative tax data has both huge potential and several limitations, as outlined by Giulia Mascagni, ICTD Executive Director and Co-Convener of the CoP, in her introduction to the topic. She emphasised its importance in understanding gender disparities and explained that, while surveys have long been used to analyse gender gaps, administrative data offers a more systematic and large-scale approach.
However, she noted that gaps persist in the availability, accessibility, and quality of sex-disaggregated data across many tax jurisdictions. Some examples include the exclusion of the informal sector which has a resultant effect on women who form the majority of workers in this sector; data accuracy concerns where information in tax registers may be outdated or incorrect, particularly regarding gender markers; limited gender variables especially when gender data is collected at registration but is not always linked to tax returns, audits, or compliance records.
Mascagni also discussed the challenges of data access, noting that privacy laws and inter-agency restrictions often make it difficult for researchers to improve gender data integration and expand research efforts. By leveraging administrative tax data alongside qualitative research, policymakers and researchers can develop more gender-responsive tax policies, ensuring that tax systems are both fair and effective for all citizens.
A Latin American perspective: sex-disaggregated data for tax policy analysis
Hitomi Komatsu, tax economist at the World Bank, built on this introduction, discussing the importance of sex-disaggregated data for tax policy analysis, using case studies from Colombia and Argentina. She highlighted that disparities in tax compliance between men and women remain poorly understood due to a lack of reliable sex-disaggregated data. Even in high-income countries, sex-disaggregated tax compliance data is not systematically collected or analysed, with only 56% of OECD countries having sex-disaggregated tax return data available for policy analysis. In low- and middle-income countries, the availability of such data are even rarer, limiting insights into gender differences in tax burdens, compliance, and wealth distribution.
In Colombia, the Tax and Customs National Authority (DIAN) undertook an initiative to analyse sex-disaggregated tax return data, driven by a 2022 tax reform that mandated the collection of taxpayer’s gender. DIAN faced however faced significant barriers as only 18.5% of taxpayers provided the relevant information in their tax returns. The tax authority therefore relied on alternative strategies to infer sex, including:
- Merging tax and pension data (matched 42.8%);
- Using ID number rules (ID numbers worked for pre-2000 births, matching 34.6%);
- Applying name-based identification algorithms to identify male or female taxpayers (matched 22.3% of taxpayers).
This process allowed DIAN to publish data on gender disparity at the top of the income distribution, showcasing that women make up only 28.2% of the top 0.1% of income earners.
Komatsu also shared some high-level results from a World Bank study where Flores, Cruces, Bermúdez, Scot, Schiavoni, and Tortarolo analysed gender disparities in property ownership and tax compliance in the municipality of Tres de Febrero, Buenos Aires, Argentina. The study finds:
- a sharp decline in women’s share of property ownership after the 40th percentile of the value distribution, with women’s share at 20% in the top 1%;
- women facing slightly higher effective tax rates because of their ownership of lower-valued properties and a regressive tax schedule;
- a significant increase in timely payments, with equal responses by men and women, following a randomised communication campaign, although men responded more quickly.
- women are less likely to use electronic payments than cash payment.
These insights underscore how sex-disaggregated data can provide information on gender disparities in income, property ownership, tax burdens, and tax compliance responses, and help develop equitable policies.
Gender-based tax compliance behaviour differences: Evidence from 5 African countries
For her part, Sripriya Iyengar Srivatsa, ICTD’s gender and tax focal point, presented findings from a forthcoming multi-country project which explores gendered tax compliance attitudes and behaviours in Eswatini, Ethiopia, Nigeria, Rwanda and Sierra Leone using administrative data merged with survey responses.
Some of the key findings were:
- No consistent pattern of gender differences in tax compliance behaviour across countries, although attitudinal differences were present. This suggests that although compliance outcomes may not be dissimilar for men and women, women’s behaviours are influenced by factors different to men.
- Women participate in the tax system with lower knowledge about it than men.
- In Rwanda, women were more likely to file tax returns
- In Eswatini, however, they were more likely to be non-filers
- Women property owners in Sierra Leone were more compliant with tax payments than men
Focus group discussions also introduced questions about how women’s experiences vary based on where they interact with tax authorities. Larger, centralised tax offices may be more structured, whereas smaller branch offices may have less formalised processes, potentially leading to gendered differences in taxpayer experiences. This is an area that requires further research.
Taxpayer registration represents a big challenge in Africa
The meeting concluded with a presentation by Maria Jouste, Research Associate at UNU-WIDER, outlining current research on gender and taxation using administrative data. Despite the rarity of such studies, a few have been conducted in select countries: South African research has focused on wage gaps, gender disparities in income, and tax incentives; while in Ecuador studies have examined gender gaps in top income earners. In Uganda, Zambia, and Rwanda, some sex-disaggregated tax data is available, but research using these datasets is still in its early stages.
Jouste stressed that one of researchers’ biggest challenge is incomplete gender coverage in tax records due to limited registration of individuals as taxpayers. In Uganda, for example, employers were not historically required to enforce the registration of employees with the tax authority, obstructing collection of employees’ gender in tax filings. This in turn limited the ability to analyse gendered tax compliance trends. Recent reforms introduced mandatory taxpayer registration and, thus, gender reporting, but older records remain incomplete. Another challenge is restricted access to individual-level tax data. In some cases, only aggregated data is available, preventing researchers from analysing individual taxpayer behaviour by gender. Privacy regulations and administrative barriers further complicate access to tax microdata for gender research. In many low-income countries, tax data is not fully digitised, inhibiting the possibility to track gender disparities over time. Even where digital records exist, different tax datasets are often siloed, limiting their usefulness for gender research.
Finally, Jouste introduced UNU-WIDER’s work with the Uganda Revenue Authority (URA) and Southern African Social Policy Research Institute (SASPRI) to build a microsimulation model, UGAMOD-TAX. The model can be used to analyse the distributional and budgetary effects of various tax policies and policy reforms.
Moving Forward: Bridging Data Gaps for Policy Impact
Discussions reinforced the need for strengthening sex-disaggregated data collection and integration across tax systems. While some countries are making strides in this regard, widespread challenges such as data availability, privacy restrictions, and inter-agency coordination remain. Going forward, collaborative efforts between researchers, policymakers, and tax authorities will be essential in ensuring that tax policies are inclusive and equitable. By investing in better data collection, increasing transparency, and expanding access to sex-disaggregated data, countries can create tax systems that are more responsive to the diverse needs of their populations. As the CoPGT continues its work, fostering partnerships and supporting research will be key to driving impactful reforms that bridge existing gender gaps in taxation.
