Do Fiscal Convergence Indicators Survive Verification? Tax Burden Dispersion and Sample Integrity in the EU-27, 2011–2023
Do Fiscal Convergence Indicators Survive Verification? Tax Burden Dispersion and Sample Integrity in the EU-27, 2011–2023

Summary/Abstract: Convergence indicators built on ratios of tax revenue to gross domestic product are used to monitor fiscal performance across the European Union and read as evidence that member states are becoming more alike. This paper applies to one such indicator the verification an auditor would apply to any reported figure, through three checks: the composition of the sample, the level of government at which the measure is aggregated, and the strength of the inference. The data are Eurostat tax aggregates by receiving level of government for 2011–2023. The extract as delivered contains 403 rows for a panel of 27 member states; 52 rows belong to two statistical aggregates and two non-member states, and removing them leaves 351 observations for 27 countries over 13 years. Dispersion is measured by the coefficient of variation; catch-up by absolute beta convergence. On the corrected sample the coefficient of variation of total tax revenue falls from 0.1691 to 0.1587, a trend significant at p = 0.035. The levels of government disagree: central government revenue converges strongly (p < 0.001), social security funds only marginally (p = 0.058), and local government revenue, starting from a dispersion more than six times that of the total, drifts apart (p = 0.070). Catch-up has the expected sign but a confidence interval that includes zero. Romania moves against the aggregate: its gap from the mean widens from 7.24 to 9.23 percentage points. A convergence claim is only as good as its sample and its level of aggregation.

• Details
• Contents