The impact of measurement error on trends in earnings inequality in the USA

Abstract Measurement error is pervasive in self-reported income and has likely evolved over time. As such, the extent of and trends in income inequality are difficult to discern. While administrative data are a possible solution, they face their own measurement issues and are frequently unavailable. Here, we propose a new method based on stochastic frontier analysis to correct self-reported income for measurement error throughout the distribution. We then apply this method to repeated cross-sectional data in the USA spanning 1960–2023 to assess trends in earnings inequality using ‘error-free’ data on earnings.