How ImpactLine reads a budget

Every figure on ImpactLine originates in a record a government published itself. Nothing is scraped from secondary commentary, and no figure is sourced from another aggregator.

• U.S. Treasury Monthly Treasury Statement (MTS)Federal receipts, outlays and the monthly deficit or surplus.
• USAspending.govFederal awards, contracts and grants by agency and recipient.
• Federal Reserve Economic Data (FRED)Economic context — GDP, employment and price indices.
• U.S. Census Bureau Annual Survey of State and Local Government FinancesComparable state and local revenue and expenditure actuals.
• State enacted budget documentsBudget authority as passed by each state legislature and signed by its governor.
• City and county adopted budgetsDepartment-level appropriations from each city's adopted budget.

Where a jurisdiction publishes both a machine-readable data service and a budget document, the data service is preferred and the document is used to check it. Where only a document exists, the figures are entered from that document and the page says so.

The single most common misreading of a government budget is treating the enacted figure as the amount spent. An enacted budget is permission to spend, granted for a fiscal year — federally the same measure is called budget authority. Outlays are money that actually left the treasury. The two rarely match, and the gap can run to tens of billions of dollars at federal scale.

ImpactLine labels which of the two a view is showing rather than blending them. Where both are available for the same jurisdiction and year, both are shown, because the difference between them is often the substantive finding.

Jurisdictions are reorganised into a common structure so that comparable things sit next to each other — a department in one state lines up with the equivalent department in another. That reorganisation is presentational. It changes how a figure is grouped, never the figure itself.

Amounts are never adjusted, interpolated, deflated or projected. If a government has not published a value, the page shows it as unavailable rather than filling the gap with an estimate. If a government revises a figure, the revised figure replaces the original at the next refresh.

Where a comparison is drawn between jurisdictions, the page states which funds and which fiscal years are in scope, because a comparison between two governments that account for the same service differently is not a like-for-like comparison and should not be read as one.

Individual spending analysis v5 applies across federal, state and local spending. Every service receives a distinct cost critique, proposed mechanism and success measure. Status is based first on a linked, jurisdiction- and period-specific program finding, then on qualifying financial signals and source quality. Missing outcome data alone never creates an adverse rating. Outcome unverified is a neutral label when the accounts do not establish delivery; missing evidence must not produce a green Stable rating. A positive result requires a matched commitment and delivery measure.

Leak identifies documented or explicitly estimated improper-payment exposure. Estimates can include underpayments and documentation failures; they are not automatically fraud, recoverable cash losses or the whole function’s spending. Critical identifies a serious sourced delivery or contract-performance failure and explains why it is a priority. Warning identifies a concrete program risk, cost pressure, allocation opacity or source limitation. These are ImpactLine assessments, not labels issued by the source auditor. A finding about one program is not applied to every dollar in its parent function.

Comparable state spending growth triggers a cost-pressure Warning when the nominal increase is at least 20%, the prior amount is at least $1 million and the increase is at least $1 million. A new allocation of at least $1 million after a comparable zero also triggers review. Residual spending of at least 10% of a reconciled total and at least $1 million triggers an allocation-visibility Warning. These are disclosed review thresholds, not validated efficiency measures: inflation, service volume, construction timing and classification changes can explain growth. Cuts alone are not treated as success or failure.

When a published funded target, matching result and deadline are supplied, a missed target after its deadline is Warning, or Critical at a shortfall of at least 50%. Before the deadline it is In progress; a reached target is Target met. Reported zeros show No reported spending and negative net amounts show Accounting offset. State source items and direct-spending/transfer components must reconcile exactly; unreconciled amounts are rejected.

Solutions describe changes in procurement, incentives, service design or budget choices rather than a universal request for more data or a universal AI recommendation. Housing and homelessness proposals can tie part of a contract to verified permanent placement and sustained housing at 6 and 12 months, while retaining upfront essential care and adjusting for client needs. Proposals and success measures are editorial judgments, not proven savings or claims that the government already uses the proposed contract.

Local spending critiques preserve city source flags: imputed (I), unknown classification (M), analyst correction (A), alternative source (S) and unrecognized flags can produce source-quality Warning. Reported (R) does not override a linked local finding; inapplicable (N) entries remain excluded. Revenue, debt and trust entries retain their financial meaning and are not treated as service spending. Individual snapshots are not fabricated into a complete city total or a validated time series. Linked audits disclose their own review periods and publication dates; a historical finding does not assert that later corrective action was absent.

Every Details panel separates the source amount, what the entry funds or measures, why its status applies, a proposed corrective or verification plan, and the source evidence. Aggregate finance records do not identify every invoice, final recipient or proof of delivery. Official auditor reports require a separate review and are not inferred from a status badge.

The Impact Score is a 0–100 signal supplied by the ImpactLine analytical service, not a measurement published by a government. The current API does not publish a reproducible score calculation or validation date, so its precise meaning cannot be independently verified here.

Treat the score as an application assessment, not an official audit rating, fraud finding or verified performance measure. Source-based state and city views omit unsupported quality scores. Reported source flags describe the financial observation, not service performance.

It depends on the government. Federal Treasury statements publish monthly. Most states publish once a budget cycle. Census state and local finance actuals lag by roughly two years, which is why they sit alongside enacted figures rather than replacing them.

Every page states the fiscal year each of its datasets belongs to. If a page shows an FY2025 enacted budget next to FY2023 reported actuals, that is deliberate and both years are labelled — an enacted budget, actual spending and federal awards are published on three separate cycles and are never forced onto one year here.

ImpactLine is independent and is not affiliated with, endorsed by, or funded by any government or agency. It reads records that governments publish and presents them in one place.

Reading budgets on ImpactLine is free and does not require an account.

If a figure here does not match the published record it cites, that is a defect and we want to know. Corrections are made against the source document, and a figure is changed only when the source supports the change.