An Anatomy of a GAO Audit: What the Process Actually Looks Like From the Inside

WASHINGTON - The Government Accountability Office gets invoked in floor speeches, appropriations battles, and agency press releases with enough frequency that its name functions almost as shorthand for oversight. What gets far less attention is the actual work: the months of structured inquiry that precede the reports that land on committee desks.

A GAO engagement typically begins with a congressional request. A committee chairman, a ranking member, or a group of members can formally ask the office to examine a program, an agency, or a pattern of spending. The office also conducts some work on its own authority, but the majority of its roughly 900 reports produced in a given year trace back to Capitol Hill.

Once a request is accepted, a team of analysts - usually between four and eight people for a mid-size engagement - is assigned. They spend the first several weeks doing what the office calls "scoping," a process of defining the precise questions the audit will try to answer. The questions matter enormously. A poorly scoped audit produces findings that agencies can dismiss as irrelevant to what they actually do.

Data collection runs parallel to a document review. Analysts send formal requests to the agency under review, requesting internal reports, budget justifications, policy memos, and inspector general findings. Federal agencies are legally obligated to cooperate, though cooperation in practice ranges from prompt and complete to slow and partial. Former GAO staff describe a recurring pattern: agencies that have something to hide tend to produce documents in formats that are difficult to analyze, or they interpret requests narrowly and wait to be asked again.

Interviews follow the document phase. GAO analysts are trained to conduct structured interviews - not the open-ended conversations of journalism, but a consistent set of questions asked of multiple officials at multiple levels. The goal is to find where accounts diverge. A senior official's description of a program and a field-level employee's description of the same program frequently do not match.

The draft report goes to the agency before publication. This comment period, typically 30 to 60 days, is where the most contested work happens. Agencies can formally agree with findings, disagree, or offer what the office calls "technical comments" - corrections of fact without disputing the larger conclusion. When an agency's written response disagrees with a finding, GAO includes that response in the final report, alongside its own assessment of whether the disagreement has merit.

The resulting document is written to a house style that prioritizes clarity over nuance. Each recommendation is numbered. Each finding is tied to a specific standard - usually criteria drawn from the agency's own regulations, from statute, or from GAO's own internal control standards, a framework called the Green Book that sets baseline expectations for how federal programs should be managed.

What the reports cannot do is compel action. GAO has no enforcement authority. Its power is reputational and political. When an agency ignores a recommendation, the office tracks that refusal and reports it to Congress in an annual digest of unimplemented findings. At any given time, several hundred recommendations from prior years remain open.

For reporters covering federal agencies, the open-recommendations tracker is one of the more reliable guides to where management problems persist. An agency that has accumulated a large backlog of unaddressed findings is an agency that has decided, for whatever reason, that the cost of inaction is lower than the cost of reform.

That calculation, repeated across enough programs over enough years, is what GAO audits are ultimately designed to disturb.