Wealth surveys suffer from the 'missing rich' problem. This paper proposes a new method to improve representativeness by calibrating the survey’s income distribution using income tax data.