A ribbon cutting is not a completed project. A press release is not proof that a policy took effect. And a mayor’s claim that an administration has “delivered” on a promise may be technically true, materially incomplete, or flatly unsupported. A guide to municipal fact checking starts with that distinction: government produces a lot of statements, but public accountability depends on evidence.
Municipal fact checking is harder than checking a national political quote. The relevant record may be spread across a mayoral announcement, a City Council bill, an agency rule, a budget schedule, a procurement notice, a court filing, and a dataset that has not been updated in months. The work is not just finding a document. It is determining what that document proves, what it does not prove, and whether the underlying government action changed anything residents can actually measure.
Start With a Checkable Claim
The first task is to reduce a broad political statement to a claim that can be tested. “We made the city safer” is a campaign message, not yet a fact-checkable proposition. “Major crime fell 8% in the first six months of the administration compared with the same period last year” is a claim with a timeframe, a metric, and a comparison point.
A useful municipal claim has four parts: an actor, an action or outcome, a timeframe, and a measurable standard. If any part is missing, do not fill the gap with assumptions. Identify the ambiguity instead.
Consider the statement, “We funded 500 new affordable homes.” Does “funded” mean money was appropriated, contracts were executed, construction began, or units opened? Are these new units, preserved units, or a mixture? Is the 500-unit figure citywide, in one district, or connected to a specific program? Each version carries a different evidentiary burden.
The strongest fact checks quote the claim exactly, then define the narrow question being evaluated. This keeps the analysis from drifting into a verdict on the speaker’s politics, intentions, or popularity.
Build an Evidence Chain, Not a Clip File
News coverage can help establish context, but it should rarely be the final stop. Municipal government is documented through primary records, and those records should form the backbone of the review.
For an executive action, begin with the order, directive, memorandum, or agency notice itself. For a spending claim, inspect the adopted budget, budget modifications, financial plan, contract registration, and actual spending data where available. For a legislative claim, read the bill text, legislative history, vote record, and implementation timetable. For claims about service delivery, look for agency performance reports, public dashboards, audits, inspection records, or independently collected data.
The point is not to collect the most documents. It is to establish a chain from announcement to authority to implementation to outcome. A mayor may announce a new initiative on Monday. On Tuesday, an agency may issue guidance. Months later, the relevant question becomes whether money moved, staff were hired, rules changed, and residents received the promised service.
That chain also exposes where a claim is stalled. An item can be authorized but unfunded, funded but unspent, launched but undersubscribed, or fully implemented without producing the stated result. Those are not interchangeable conditions.
Know What Each Record Can Prove
Primary documents are not automatically decisive. A budget line proves that money was allocated, not that it was spent. A signed executive order proves that an order was issued, not that agencies complied. A press release may accurately describe a policy goal while overstating its reach.
Treat every source as evidence with limits. An agency’s own performance report may be the best available record, but it can contain changed definitions, delayed reporting, or favorable framing. A comptroller audit may offer a more independent assessment, but it may cover a period that predates a current policy. Court records can establish legal status, but not necessarily operational impact.
The central question is simple: what does this document establish directly? State that answer plainly. Avoid asking a source to carry more weight than it can support.
Separate Inputs, Outputs, and Outcomes
Municipal claims often move between three levels of performance without warning.
An input is a resource: dollars appropriated, staff authorized, contracts awarded. An output is what government did: inspections completed, shelter beds opened, trees planted, applications processed. An outcome is the real-world condition the government says it wants to improve: fewer people sleeping outdoors, shorter permit wait times, lower traffic deaths, better student attendance.
All three matter. But they require different ratings. A city can keep a promise to hire 100 inspectors and still fail to reduce illegal dumping. It can announce a record capital commitment while completing few projects. Conversely, an agency might improve a service outcome through changes that were not part of a mayoral headline initiative.
This distinction is especially important when assessing campaign promises. A promise to “end street homelessness” should not be marked kept because a new outreach program launched. The program may be a relevant action, but the promised outcome demands outcome-level evidence. Where the administration has taken meaningful steps but results are incomplete or unavailable, “in progress” or “stalled” is more accurate than either victory or failure.
Use the Right Baseline and Timeframe
Percentages are among the easiest municipal figures to misuse. A claim that a program “doubled” may describe an increase from two sites to four. A claim that a metric is at a historic low may depend on a narrow time series or a newly changed reporting method.
Always establish the baseline before accepting the change. Compare like with like: the same months, geography, definitions, and data collection method. Seasonal services, crime categories, school attendance, and emergency response times can all shift sharply during the year. A single month may be newsworthy, but it is rarely enough to prove a durable trend.
Also distinguish between fiscal years and calendar years. In New York City, the fiscal year begins July 1. A claim tied to a budget may use a fiscal-year figure, while a public safety claim may use calendar-year data. Mixing them can create a misleading picture of timing and responsibility.
When definitions change, say so. If an agency revised how it counts placements, inspections, or eligible households, a before-and-after comparison may not be clean. That does not make the data useless. It means the conclusion must be narrower.
Assign a Verdict That Matches the Record
A verdict should describe the evidence, not reward rhetorical force. The most useful ratings are specific enough to show where a claim stands.
“Supported” fits when primary records substantiate the central claim. “Partly supported” fits when a claim contains a verified element but leaves out a major limitation, such as partial implementation or a narrower population served. “Unsupported” fits when the available record does not establish the assertion. “False” should be reserved for claims directly contradicted by reliable evidence. “Unverifiable” is appropriate when key data are unavailable, incomplete, or too delayed to support a confident conclusion.
For performance tracking, status labels can add needed precision: kept, broken, stalled, in progress, or pending evidence. A promise may be kept procedurally but unresolved substantively. For example, a mayor could keep a promise to introduce legislation, even if the legislation later fails. The fact check should state both facts rather than compressing them into one misleading score.
Show Your Work So Readers Can Audit It
A municipal fact check earns trust when the reader can see the path to the finding. Lead with the claim, verdict, and core evidence. Then explain the limits: missing data, changed definitions, implementation delays, disputed figures, or unresolved legal questions.
Use exact dates. Identify the responsible agency. Name the document type and reporting period. If the evidence changed after publication, update the finding with a clear note rather than silently replacing the old conclusion. Government is ongoing, and a fair record must be able to change when the facts do.
This approach is central to public-interest accountability work, including the dashboard model used by ReviewMamdani.com: not treating every announcement as a final result, and not treating every setback as the whole story. The record should remain open to new evidence while staying firm about what is known now.
The Discipline Is the Point
Municipal fact checking is not a contest to catch officials in a mistake. It is a method for making public claims legible. The most valuable findings may be mundane: a deadline slipped, a budget allocation did not become spending, an agency reported progress using a narrower metric than the public was led to believe.
Those details are where accountability lives. When residents can distinguish an announcement from an appropriation, an appropriation from implementation, and implementation from results, government becomes harder to spin and easier to judge.