Automation and Inequality in Wealth Management

Abstract

We examine how access to automated wealth managers affects households’ investment in financial markets and welfare across the wealth distribution. Our setting features novel microdata from a major U.S. robo advisor and a quasi-experiment in which the advisor reduces its account minimum by 90%. Based on a difference-in-difference estimator, the reduction increases middle-class households’ participation by 110% but does not affect wealthier or poorer households. We rationalize this behavior with a life cycle model calibrated using portfolio-level data. Our calibration suggests that the reduction significantly raises middle-class households’ welfare, and 65% of this gain reflects improved diversification.