In response to Congressional interest and as part of our ongoing efforts to enhance transparency, the Congressional Budget Office plans to publish additional computer code on our GitHub page throughout the year showing details of the modeling underlying CBO's analyses.
Last week, we published several new code repositories for CBO's models, including the following:
- Conventional Tariff Analysis Model: The code used to project the budgetary effects of tariffs, which CBO described in a blog post last November.
- Discount Factors Model: The code used to calculate discount factors from CBO's forecasts of Treasury yields, which are used to estimate the budgetary costs associated with a variety of federal loan and loan guarantee programs.
- Permitting Model: The code used to model the effects of changes in permitting requirements on private investment, which informed our economic and budgetary analysis of the House-passed version of the 2025 reconciliation act (H.R. 1) published last June.
- State Unemployment Rate Model: The code used to model, forecast, and analyze state-level quarterly unemployment rates and their implications for Unemployment Insurance Extended Benefits eligibility.
CBO detailed its plans to post additional modeling resources in a blog post last November, which explained how we develop and share models—including by posting data, computer code, and documentation on GitHub—to provide greater insight into CBO's analytical methods.
In addition to posting advanced modeling resources, CBO publishes lists of available data, easy-to-use modeling tools, and detailed explanations of our analytical methods on our Topics and Transparency pages.
CBO will continue to assess opportunities to post additional modeling resources and develop new tools, while adhering to strict federal standards to ensure that sensitive information is not released.
Kevin Perese is a Senior Adviser for Data Science and Transparency at CBO.