For decades, global anti-corruption efforts have been based largely on composite indicators of national level aggregate data such as the Transparency International Corruption Perceptions Index, the World Worldwide Governance Indicators - Control of Corruption index, and several others, which provide important information on corruption prevalence at country level. Such aggregate corruption scores make headlines, but they hide the real story of who is most affected by corruption, where, and how. To design anti-corruption policies that truly leave no one behind, we need disaggregated data: granular statistics that reveal how corruption impacts different groups based on gender, age, income, disability, geographical location, and more.

Global norms and policy commitments increasingly promote this kind of disaggregated data:

• On corruption, UNCAC-COSP Resolution 10/10 (2023) (the Conference of the States Parties to the United Nations Convention against Corruption) explicitly encourages States Parties to collect corruption data disaggregated by gender and age, and to understand how corruption specifically affects women, men, girls, and boys.
• The UN Statistical Commission principles require that SDG indicators be disaggregated ‘by income, sex, age, race, ethnicity, migratory status, disability and geographic location’.
• The UNODC Manual on corruption surveys (the United Nations Office on Drugs and Crime) mandates that national surveys should at minimum disaggregate bribery prevalence by type of public official, sex, age, income, and education level.
• The Asian Development Bank (ADB) practical guidebook on data disaggregation frames disaggregation as essential for identifying groups ‘left furthest behind’.
• The UN Women – ‘Women count’ initiative aims to close global gender data gaps, helping countries improve the production and use of gender-disaggregated statistics across the Sustainable Development Goals.
• The Praia City Group on Governance Statistics, created in March 2015 at the 46th session of the United Nations Statistical Commission (UNSC), led to the 2023 revision of the UN Classification of Statistical Activities (CSA 2.0). Governance statistics on non-discrimination, equality, participation, trust, and prevalence of corruption are now formally recognised as an official statistical domain on par with economic and social statistics.

This blog refers to several surveys and approaches to data gathering from a number of sources. These examples are illustrative rather than exhaustive, drawing on a selection of global, regional, and national datasets to illustrate both the insights and blind spots that emerge when corruption data is (or is not) disaggregated.

What corruption surveys reveal – and what they still miss

When corruption data is not disaggregated, it conceals the diverse and unequal ways people experience corruption. It masks how gender roles, social norms, and structural inequalities intersect with corruption risks and impacts. Moreover, disaggregated data helps uncover the role of discrimination in fuelling corruption. Research by Transparency International (TI) and the Equal Rights Trust has shown that corruption often acts as a vehicle for discrimination and oppression, with marginalised groups such as persons with disabilities, migrants, or ethnic minorities being more frequently harassed, denied services, or extorted.

A forthcoming analysis by the UNCAC Coalition’s Gender and Corruption Data Taskforce shows that the architecture of global corruption data is both broad and shallow. Across 18 datasets examined, they found that gender-disaggregated corruption data is rarely analysed or published, even where it already exists. Only a handful of international surveys – such as Afrobarometer, or recent waves of the Global Corruption Barometer (GCB) – enable some level of analysis on gendered aspects of corruption. Even then, the detail often varies by country and round.

Yet even when surveys do collect sex-disaggregated data, there is an additional layer that is often missed. Gendered abuses such as sexual corruption, defined as the abuse of power to demand or obtain sex or acts of sexual nature, may remain hidden if surveys do not intentionally seek to uncover it. When TI first included questions on sexual corruption for some regions of the 2019 GCB, the results were striking: in Latin America, up to 20% of respondents said they had either experienced it themselves or knew someone who had. This demonstrates how powerful survey questions can be in revealing patterns that would otherwise stay invisible. The 2019 GCB for Africa, in contrast, did not include questions on sexual corruption, yet a national study by TI Zimbabwe conducted the same year found that 57.5% of women had personally experienced it. This hints at the scale of the problem that is missed by failing to ask the right questions.

The 2019 GCB Africa data further highlight differences in how corruption is experienced across population groups. It shows that men report paying bribes more often than women (32% of men compared to 25% of women in the previous 12 months), probably because of frequent interaction with public institutions rather than any indication that women face fewer corruption risks. The data also reveals that the poorest citizens are nearly twice as likely to pay bribes for accessing services as the wealthiest. However, the lack of further disaggregation, such as by education level, disability status, or sub-regional location, means that other inequalities may remain hidden. Furthermore, according to the forthcoming UNCAC Coalition study, sector-specific and qualitative evidence, especially in health, education, and law enforcement, remains scarce, which makes it difficult to pinpoint where the likelihood of corruption is highest.

Afrobarometer surveys collect data that can be disaggregated by gender and rural/urban residence for questions on bribery for services such as healthcare and identity documents. Although the data collected is openly available, only a subset is routinely published. Often, the data is not fully analysed nor included in reports: a missed opportunity for uncovering insights into differential experiences of corruption.

From broad indicators to granularity: emerging insights from national data

Fortunately, at the national level, some countries have started reshaping what high-resolution governance data can look like. Ghana, Nigeria and Kenya offer examples of this shift. Vietnam’s Provincial Governance and Public Administration Performance Index (PAPI) also collects gender disaggregated data on experiences of corruption although these are not yet analysed or reported on.

• In Ghana, a governance and corruption survey gathers evidence from households, enterprises, and public officials. It integrates granular demographic detail that enables analytical depth on corruption’s societal impact. It disaggregates corruption incidence and reporting patterns by age, sex, region, urban/rural location, education, and employment status – revealing, for example, that young urban men report the highest frequency of bribe requests in interactions with police, while rural women encounter distinctive obstacles in accessing justice and grievance mechanisms.
• Nigeria’s Corruption in Nigeria: Patterns and Trends – produced jointly by UNODC and the National Bureau of Statistics – provides a granular corruption dataset that disaggregates results by gender, income quintile, education, employment, geopolitical zone, and type of official involved. It also includes data on sexual corruption. This approach shows how corruption affects different socio-economic groups in distinct ways.
• Kenya’s National Ethics and Corruption Survey (NECS) by the Ethics and Anti-Corruption Commission breaks down corruption prevalence, reporting behaviour, bribe requests, and trust in public institutions by county, gender, income, education, and type of service. The latest edition continues to refine its county-level indicators and includes detailed analysis of which groups face the highest exposure to bribery across service points such as health, policing, local administration, and licensing.

Taken together, these national examples show what is possible when governments commit to gathering and using higher-resolution data.

Disaggregate, cross-reference, and connect the dots to ensure we ‘leave no one behind’

Disaggregated corruption data shapes what policymakers can see, and therefore what they can act on. Where granular evidence does exist, it can drive meaningful change. For example, civil society (in the form of TI) helped secure recognition of sexual corruption in UNCAC-COSP Resolution 10/10 by providing credible, gender-disaggregated evidence showing that sexual corruption was widespread yet invisible in mainstream governance indicators.

Disaggregated corruption data shapes what policymakers can see, and therefore what they can act on. Where granular evidence does exist, it can drive meaningful change. For example, civil society (in the form of TI) helped secure recognition of sexual corruption in UNCAC-COSP Resolution 10/10 by providing credible, gender-disaggregated evidence showing that sexual corruption was widespread yet invisible in mainstream governance indicators.

Granular, gender-disaggregated evidence can be even more insightful when combined with information on inequality, public service performance, and patterns of exclusion, by cross-referencing corruption data with other governance and inclusion datasets on public service satisfaction, poverty, and inequality. This would help reveal the full picture of how corruption, discrimination and poor service delivery reinforce each other.

In this regard, ongoing efforts by the Praia Group on Governance Statistics to standardise data on corruption, discrimination, and other governance dimensions are important. Such a holistic approach will make it possible to see where disadvantage accumulates, which services are most affected, and who is left behind, providing the foundation for better policy, improved public services, fairer governance, and more equitable outcomes for all.