Every economist is responsible for their own data. Here is mine.
I am moving data hosting off Google Drive and onto self-hosted mirrors at files.heterodata.org. Each dataset below lists the self-hosted mirror as the primary location, with the existing Google Drive link preserved as a legacy fallback. Datasets under 1 GB are hosted whole; larger ones ship final outputs plus exact instructions for pulling the raw data from the original source. Programmatic access via the Robin API is planned for a future release.
The HMDA Explore dashboard at hmdaexplore.com is the public-facing version of this work. For HMDA, only the final aggregates are mirrored — the raw Loan/Application Register (LAR) is too large to host whole and must be pulled directly from the CFPB.
For programmatic access to bank regulatory data spanning 1863 to the present:
The replicated and extended data series behind the Anwar Shaikh Capitalism (2016) replication and the Shaikh & Tonak national-accounts work are published, with per-series provenance and downloads, at:
International-accounts data — balance of payments, flow of funds, and cross-border value transfers across 258 countries — is explorable at:
Bank-panel data on U.S. bank failures and balance sheets, 1865–2026, is explorable at:
The U.S. BEA Input-Output accounts and the Leontief inverse across sectors, 1997–2024:
35 boxes / 77,452 pages / 532 GB of digital content at the Levy Economics Institute of Bard College. Public archive: