The bill directs the Congressional Budget Office to disclose the modeling tools, data-preparation routines, and the computational details underlying its cost estimates. It adds a new publication requirement to the Congressional Budget Act of 1974 so outside parties can see how CBO reached its numbers.
This matters because the CBO’s score carries heavy weight in congressional debate; opening its methods aims to improve reproducibility, public trust, and external review but also raises questions about data confidentiality, proprietary inputs, and agency workload.
At a Glance
What It Does
The Act amends section 402 of the Congressional Budget Act to require the Director of the CBO to make available on the agency website (and to Members of Congress) each fiscal model, policy model, and data-preparation routine used for estimates, and to publish any updates to those models. For each estimate the CBO must also disclose the data, programs, models, assumptions, and computation details in a form sufficient to permit replication by people not employed by the CBO. When statutory restrictions bar disclosure of particular data, the bill requires CBO to publish a list of variables, descriptive statistics, the governing statute, and contact information for those with access.
Who It Affects
Directly affects the Congressional Budget Office and its modelers, congressional offices that rely on CBO scoring, external researchers and academic modelers who replicate or critique scores, and executive-branch agencies or private entities that supply sensitive data to CBO. Data vendors and holders of proprietary datasets used in scoring will also be implicated.
Why It Matters
By requiring publication of code, models, and computation details, the bill changes the default transparency standard for federal budget scoring and creates a new expectation of reproducibility. That can accelerate external validation and policy analysis but will force CBO and data providers to sort out confidentiality, licensing, and technical-disclosure logistics.
More articles like this one.
A weekly email with all the latest developments on this topic.
What This Bill Actually Does
The bill adds a new subsection to the statutory provision that governs CBO’s operations, creating a presumption in favor of disclosure: CBO must put the models it uses — fiscal, policy, and the routines that prepare data for those models — on its public website and make them available to Members of Congress. ‘‘Models’’ in practice will include the code, scripts, or executable programs, the data inputs and metadata, and the assumptions and parameter values used in running those models.
For each estimate the CBO issues, the agency must publish enough detail about the computational steps so an external analyst can re-run the calculations and reach the same—or closely comparable—results. The bill builds in a process for handling information CBO cannot legally release: instead of raw restricted files, CBO must publish variable lists, summary statistics (means, standard deviations, counts, and correlations where disclosure rules allow), cite the nondisclosure statute, and say who may be contacted for access.
Those disclosure-replacement elements are intended to preserve analytical transparency even when raw data remain protected.The law also requires CBO to publish updates when it changes a model or a data-preparation routine, which creates an audit trail of model evolution. Practically, that means CBO will need publication workflows, version control, documentation standards, and likely new legal and IT reviews for each release.
The statute requires public posting on the CBO website and availability to Members, but it does not specify file formats, licensing terms, or procedures for third parties to request access to restricted datasets — leaving operational detail to CBO implementation.Finally, the bill sets a short compliance window: the amendment takes effect six months after enactment. That timetable will determine which estimates and models are subject to immediate publication and will push CBO to prioritize which legacy models and datasets to document and post first.
The Five Things You Need to Know
The bill amends section 402 of the Congressional Budget Act of 1974 (2 U.S.C. 653) by adding a new subsection (b) specifically directing publication obligations.
CBO must publish each fiscal model, policy model, and data-preparation routine it uses, plus any subsequent updates to those models or routines.
For every cost estimate, CBO must disclose the data, programs, models, assumptions, and computation details in a way that is 'sufficient to permit replication by individuals not employed by' the CBO.
When law prevents disclosure of particular datasets, CBO must instead publish a complete variable list, descriptive statistics (including averages, standard deviations, counts, and correlations where permitted), the statutory citation blocking disclosure, and contact information for the party with access.
The new publication requirements apply beginning six months after the Act’s enactment; materials must be posted on the CBO website and made available to Members of Congress.
Section-by-Section Breakdown
Every bill we cover gets an analysis of its key sections.
Short title: 'CBO Show Your Work Act'
A single line establishes the Act’s short title. That’s the bill’s public handle used in implementing documents and guidance; it signals the statute’s transparency focus but creates no legal obligations by itself.
Statutory amendment to 2 U.S.C. 653
This provision modifies the existing opening sentence of section 402 by inserting a subsection structure (shifting the current language into subsection (a)). Mechanically, that change creates room to add new duties without rewriting existing CBO authorities. The move is administrative but necessary to graft the disclosure obligations onto current statutory text.
Publish models and updates
Paragraphs (1) and (2) require CBO to make each fiscal model, policy model, and data-preparation routine available publicly and to Members of Congress, and to publish updates. The practical effect is a continuous disclosure requirement for model artifacts and versioning: any substantive change must be documented and posted. That raises questions about how CBO will define 'update,' how it will document version differences, and whether minor maintenance changes trigger publication.
Replication standard and handling restricted data
Paragraph (3) sets a replication standard: disclose the data, programs, models, assumptions, and computation details so external actors can replicate CBO estimates. Paragraph (4) creates a compliance path for legally nondisclosable data by requiring variable lists, descriptive statistics, citation of the nondisclosure statute, and contact details for authorized access. Practically, paragraph (4) is an attempt to balance reproducibility with legal constraints, but it leaves operational choices—file formats, aggregation levels for summary statistics, and mechanisms for third-party access—undefined.
Six-month compliance window
The Act becomes operative six months after enactment. That deadline compresses CBO’s implementation planning, requiring the agency to prioritize model documentation, establish publication processes, and negotiate data-sharing arrangements with suppliers in a short timeframe.
This bill is one of many.
Codify tracks hundreds of bills on Government across all five countries.
Explore Government in Codify Search →Who Benefits and Who Bears the Cost
Every bill creates winners and losers. Here's who stands to gain and who bears the cost.
Who Benefits
- Members of Congress — gain direct access to the models and computational details behind scores, improving their ability to scrutinize and question assumptions during legislative drafting and negotiation.
- Independent researchers and academic economists — can reproduce CBO estimates, run sensitivity checks, and propose alternative specifications, which strengthens external validation and methodology debate.
- Policy and advocacy organizations — obtain material needed to produce counter-estimates or to explain scoring choices to stakeholders, improving transparency in policy advocacy and public communications.
Who Bears the Cost
- Congressional Budget Office — will face increased staff time, documentation, legal review, IT hosting costs, and new processes for publishing code, data, and version histories; these are concrete resource and project-management burdens.
- Federal agencies and other data providers — may need to renegotiate data-sharing agreements, implement anonymization or aggregation, or refuse to supply data that conflicts with proprietary or privacy constraints.
- Private data vendors and licensors — risk disclosure of inputs or derived variables used in scoring, which may prompt licensing renegotiations or higher charges for data access to preserve proprietary protections.
Key Issues
The Core Tension
The central dilemma is between maximizing reproducibility and protecting legally or commercially sensitive inputs: full methodological transparency advances accountability and independent verification, but it can expose personal data, proprietary inputs, or model architectures that stakeholders want to keep private — and complying with both goals will require costly, judgment-filled implementation choices.
The bill sets a strong transparency objective but leaves multiple implementation choices to CBO, which creates both risk and flexibility. The statute prescribes what must be published but not how: it does not specify file formats, documentation standards, licensing terms, timelines for posting after an estimate, or quality-control processes for released models.
Those gaps will force CBO to develop policies that balance reproducibility against technical feasibility and legal limits.
A second practical tension concerns data that CBO uses but does not own or cannot disclose. Paragraph (4) requires descriptive statistics and variable lists, but summary measures can be insufficient for certain replication tasks and could still permit re-identification in small samples.
CBO will need to decide acceptable aggregation or masking approaches, negotiate access paths for vetted researchers, and navigate conflicts with statutes that bar sharing. Those negotiations could delay releases, reduce analytical utility, or trigger litigation if data providers or other agencies resist disclosure.
Try it yourself.
Ask a question in plain English, or pick a topic below. Results in seconds.