Replication Package for "Reaching for Yield: Evidence from Households"

This replication package includes the data and code necessary to replicate the tables and figures in Gomes, Peng, Smirnova & Zhu (2025). It contains raw data files obtained from public sources that were used in our analysis. For the data derived from licensed or proprietary sources, the package is based on a sampled dataset (consisting of 50k investors as opposed to over 1m in the original data) and provides pre-processed data set with hidden initial identifiers and randomized variables.
The master file executes all the code used to process and compile the data, and runs the analysis generating the 15 figures and 3 tables in the paper. While the full dataset code takes approximately 14 hours to run on a modern computer, the code provided for the subsample should complete in about 1 hour.