The Last Thing on Your Mind: Recall Bias in EU Income Measurement
通过比较欧盟家庭调查EU-SILC中不同时间收集的收入数据，发现受访者回忆前一年收入时存在显著偏差，其报告值偏向当前收入，导致官方统计失真，呼吁统一数据收集时间并更多使用行政记录。
ABSTRACT This paper studies the soundness and consistency of income measurement in the EU‐SILC, the main household survey of the European Union. By comparing income data collected at different times of the year, I show that households exhibit strong recall bias when asked to report their income for the previous calendar year. In particular, retrospective valuations are skewed towards respondents' contemporaneous income at the time of the interview. Given large heterogeneity in data collection schedules, this present bias has serious consequences for the stability, reliability, and international comparability of income statistics across the EU. Microdata analysis from 2007 to 2021 shows that income self‐reports diverge by more than 10% across sampling quarters on average, which seriously distorts official statistics, income trends, and poverty estimates in several member states. These findings call for the harmonization of EU‐SILC data collection periods, more extensive use of registers, and a thorough review of related EU social statistics.