Methodology

The federal dollar, from appropriation to award.

Public Money Tracker follows the federal dollar from congressional appropriation to individual contract or grant award. Every figure published on this platform traces to a primary federal data source — a government-operated database, a statutory disclosure filing, or an agency bulk file — and every record entering the database carries source provenance metadata: the URL or endpoint it was retrieved from and the date of retrieval. PMT does not aggregate secondary sources, does not license data from commercial vendors, and does not infer figures from proxies when direct data is available.

The platform is non-partisan and observational. Public Money Tracker does not assert causal relationships, assign institutional blame, or editorialize about the significance of any individual transaction or pattern. The data is presented as reported by federal agencies. Where estimation methods are applied — for tax contribution projections, district demographic attribution, or place-of-performance mapping — those methods are described at a high level in this document and in the sources file attached to published analytical reports.

Eleven primary public sources. No aggregators.

Public Money Tracker does not use OpenSecrets, ProPublica, or any secondary aggregator as a primary data source. When those organizations are referenced for context in reporting, it is noted explicitly alongside the citation.

Collection, processing, and validation on a weekly cycle.

Collection

Automated collectors query federal APIs and download bulk data files on a weekly schedule, running each Friday. Every record is stored alongside its source URL, the API endpoint or file path it originated from, and a retrieval timestamp. No figure enters the database without provenance metadata.

The collection layer is intentionally narrow in scope: it pulls exactly the fields required for analysis and discards fields that are not used. This keeps the pipeline auditable — a reviewer can trace any published figure back to a specific row in a specific source file retrieved on a specific date.

Processing

Federal data is inconsistent across agencies. Vendor names vary between SAM.gov registrations and USASpending award records. Congressional district assignments shift after redistricting. Dollar figures posted to USASpending can be revised months after an award is initially reported, as agencies submit modifications to existing contracts. The processing layer normalizes these fields — standardizing entity names against SAM.gov canonical registrations, resolving duplicate records created by agency re-submissions, and mapping awards to current congressional district boundaries.

Where a direct match is not available — for example, when a ZIP code straddles two congressional districts — the processing layer applies a documented attribution rule and records the method used. Reports generated from estimated or partially-attributed data include explicit notation in their sources file.

Validation

Before data appears in any published report, automated checks flag records that fall outside expected ranges — spending figures that represent a substantial departure from a vendor's or district's historical baseline, vendor names that do not resolve to an active SAM.gov registration, and district assignments that conflict between two data sources covering the same award. Flagged records are held for additional review or excluded from published output until the discrepancy is resolved.

Federal spending received versus federal taxes contributed.

The District Return on Investment (ROI) metric compares federal spending received by a congressional district against the estimated federal tax contributions originating from that district. The spending figure draws from USASpending.gov district-level obligation totals — the same data the federal government uses for public accountability reporting — covering contracts and grants with a place-of-performance address in the district. The tax contribution figure draws from the IRS Statistics of Income ZIP code data, which reports individual income tax liability at the ZIP level, mapped to congressional districts using Census Bureau boundary files.

The ROI ratio is a comparative indicator, not a measure of fiscal efficiency or political performance. It is a single-number summary: federal dollars received per dollar contributed in individual income taxes. A ratio above 1.0 indicates the district receives more in federal obligations than its residents contribute in individual income tax. A ratio below 1.0 indicates the reverse. The ranking is refreshed weekly as USASpending.gov posts new obligation data. The tax and demographic figures are updated annually when new IRS and Census vintages are released. The specific fiscal years and tax years used in any published ranking are disclosed in the accompanying data file at publicmoneytracker.com/districts.

What this data does not capture.

Federal spending data is broad but not comprehensive. The limitations below reflect deliberate scope decisions, not gaps to be filled — each is a consequence of how federal agencies collect and publish their own data.

Spending and tax figures cover different time windows. USASpending records obligations by federal fiscal year (October 1 through September 30). IRS Statistics of Income data is reported by calendar tax year. The two figures are not directly comparable on an annual basis; multi-year aggregation reduces but does not eliminate this mismatch. The data window used in any specific analysis is disclosed in that report.

ZIP-to-district mapping is imprecise at district boundary ZIP codes where a single ZIP code straddles two congressional districts. In those cases, PMT applies a consistent geographic attribution rule, which introduces a modest margin of error in districts with many boundary ZIPs.

USASpending.gov tracks discretionary federal spending — contracts, grants, direct payments, and loans. Mandatory programs including Social Security, Medicare, and Medicaid disbursements are not reflected in the spending figures. The ROI metric therefore captures the discretionary portion of the federal budget only; mandatory transfer payments flow through separate agency reporting systems and are not included in the current pipeline.

The IRS Statistics of Income captures individual income tax liability. It does not include corporate income tax, payroll taxes, excise taxes, or estate taxes. The tax contribution figure used in District ROI calculations represents individual income tax only.

Place-of-performance attribution means a contract performed at a federal facility — a military installation, a research laboratory, or a federal building — is credited to the district where that facility is located, not to the district where the contractor's workforce resides or where the contractor is headquartered. Districts with large federal facilities will show higher obligation totals than districts where the economic activity from those contracts actually occurs.

FY2026 spending data is partial year-to-date and will continue to increase as agencies post additional obligation records through September 30, 2026. Rankings that include FY2026 will shift with each weekly refresh.

Lobbying disclosure data is reported quarterly to the Senate Office of Public Records under the Lobbying Disclosure Act. There is an inherent lag between lobbying activity and its appearance in the database. Individual contacts, meetings, and communications are not itemized in LDA filings; only aggregate semi-annual spending and general issue areas are reported.

When each dataset refreshes.

How to cite this data.

Public Money Tracker is a publication of Sievela LLC. When citing PMT data in research, journalism, or policy analysis, please reference publicmoneytracker.com/methodology and the specific report or data page from which the figure was drawn. A machine-readable sources file accompanies published analytical reports. It lists, for each figure: the exact primary public source, the time period covered, the estimation method applied if any, and the value as stored in the PMT database at time of publication. PMT is designed to preserve the provenance chain from source record to published figure.