How Much Should We Trust the 30–40 Percent? Financial Knowledge and Retirement Wealth Inequality
Description
Replication package for "How Much Should We Trust the 30–40 Percent? Financial Knowledge and Retirement Wealth Inequality". This package contains the code, outputs, and documentation needed to reproduce every number, table, and figure in the manuscript. The paper audits the flagship claim that financial knowledge explains 30–40 percent of retirement wealth inequality (Lusardi, Michaud, and Mitchell 2017, Journal of Political Economy). Contents: (1) a Python port of the LMM structural model plus LMM's official Fortran code for provenance; (2) NFCS 2009–2021 financial literacy and behavior pipelines; (3) HRS 2004 wealth regressions with corrected education coding; (4) PSID moment construction and formal likelihood-ratio inference for the knowledge return gradient; (5) the manuscript's 13 tables extracted verbatim and 3 figures. To replicate: unzip, place the external public-use survey files in data/raw/ following data/codebook/source-register.md, install the pinned Python 3.12 requirements (code/requirements.txt), and run "bash code/run_all.sh". Core pipeline runtime is about 2.5 hours; the PSID section requires the PSID_DATA_DIR environment variable and adds about 3 hours. Raw survey microdata are not bundled; all sources are public or registered public-use.
Files
Steps to reproduce
1) Download the replication package file in this dataset and unzip. 2) Obtain the external public-use survey files (NFCS 2009–2021, HRS 2004 experimental module + RAND HRS, SCF 2004, PSID family files 2001/2003/2005) following data/codebook/source-register.md and place them in data/raw/; point the PSID section to its folder with PSID_DATA_DIR. 3) Create a Python 3.12 environment and install code/requirements.txt. 4) Run "bash code/run_all.sh". The core pipeline takes about 2.5 hours; the PSID/beq3v2 section adds about 3 hours and runs automatically when PSID_DATA_DIR is set. 5) All tables and figures are written to output/tables/ and output/figures/. Full instructions are in the README.md inside the package.