When the "At-risk-of-poverty" rate is at risk - Data quality issues in Hungarian EU-SILC income data

This presentation was delivered at the United Nations Economic Commission for Europe (UNECE) – Conference of European Statisticians, Group of Experts on Measuring Poverty and Inequality, held on April 16, 2026, in Rome.

The presentation examines data quality issues in Hungarian EU-SILC income data and their implications for the at-risk-of-poverty (AROP) indicator.

While income data from 2005–2013 appear broadly consistent, multiple anomalies emerge from 2014 onwards. One major issue is the occurrence of cases where net income exceeds gross income in a substantial share of households (reaching up to 4–28% in some years), which is not compatible with the Hungarian tax system.

For the period 2018–2021, the analysis identifies strong and systematic “bunching” in income distributions just above the poverty threshold, including identical reported income values across large numbers of households. These patterns are statistically implausible and are observed across multiple population groups, including single-person households, households with children, and the full sample.

The presentation also discusses the 2019–2023 data revision. Although the revision eliminated the observed clustering, changes in income components frequently follow a limited set of deterministic multipliers, raising concerns about the underlying data generation process.

Overall, the findings suggest that Hungarian EU-SILC income data between 2014 and 2024 are not reliable for poverty analysis. Given that EU-SILC is the primary source of income data in Hungary, this has significant implications for both national and EU-level poverty monitoring.

The presentation also summarizes the authors’ communication with the Hungarian Central Statistical Office and Eurostat, and outlines recommendations for a comprehensive review of the income data and the introduction of more systematic pre-publication quality checks.