Every four years, during the US presidential elections, discussions on combating inequality take center stage among economists, politicians, and journalists. Proposed measures often include wealth taxes, income tax reforms, and changes to exemptions. The success of such policies hinges on accurate inequality estimates, as flawed data can lead to ineffective or harmful policies. This underscores the need for robust methodologies to measure wealth inequality trends. This letter investigates two primary measures of wealth inequality: one based on the Survey of Consumer Finances (SCF) and the other derived from individual income tax returns (PUF) under various modelling assumptions. The analysis incorporates sampling and nonresponse errors in SCF estimates and introduces a bootstrapping procedure to quantify sampling errors in PUF estimates. The findings reveal how neglecting uncertainty in these estimates can lead to conflicting interpretations of long-term trends in top wealth inequality. Additionally, the research demonstrates how different datasets and modelling assumptions can significantly alter conclusions about the Great Recession’s impact on the top 10 percent wealth shares – ranging from rapid increases to slower growth to stagnation.
Marta Boczoń (2025) studied this question.