Wealth survey calibration : imposing consistency with income tax data

Wealth surveys tend to underestimate wealth concentration at the top due to the "missing rich" problem. We propose a new way of improving the credibility of wealth surveys: We make them consistent with tabulated income tax data. This is possible with the Household Finance and Consumption Survey (HFCS), which takes place in most European countries every three years and collects data on both income and wealth. Consistency is achieved by calibrating survey weights using the income part of HFCS. We apply the calibration method of Blanchet, Flores and Morgan (2022), but propose a new way to determine the merging point where the calibration starts. Calibrated weights are then used with HFCS wealth values. We test the method on Austrian data and find that calibration increases the top 1 % wealth share from 26 % to 37 % in 2014 and from 23 % to 27 % in 2017. The effect is small and negative in the 2011 HFCS wave, even though the net worth of the top 1 % increases. We also highlight a strong downward bias in the Austrian HFCS income distribution, which begins even before the 80th percentile. Following the calibration, we test other top tail adjustments: replacing the survey top tail with a Pareto distribution and combining the data with a magazine rich list.