GAO Could Verify Agency Ethics Training for 49 of 206 Identified DOGE Personnel—but the Denominator Is Incomplete

The cleanest percentage is also the most misleading one

GAO identified 206 people who held DOGE-related positions within the Executive Office of the President from January 2025 through January 2026. Nine outside agencies supplied activity information for 64 of them: 49 completed ethics training at those agencies, 18 completed records-management training and 38 filed financial disclosures.

Those figures equal 76.6%, 28.1% and 59.4% of the 64-person agency subset. They must not be presented as completion rates for all 206. Some detailees were expected to complete requirements at home agencies, categories can overlap, and the Executive Office did not provide the central records GAO requested.

Turnover exceeded three-fifths

At least 128 of the 206 identified personnel had separated from their Executive Office positions by January 31, 2026—62.1%. Twenty-seven were described as special government employees, or 13.1%, and four were characterized as leaders. The 206 count is itself a lower bound because it excludes people not identified in court filings or agency responses by GAO’s cutoff.

A position table totals 210 rather than 206 because four people held positions in both the U.S. DOGE Service and White House Office. Treating the table total as unique headcount would double-count them.

Missing documentation is not proof of noncompliance

Executive Office officials said DOGE personnel received the same ethics and records training as other staff, but did not provide records that would let GAO verify that statement. Ten agencies also did not provide requested information on time or at all. The audit therefore could not determine complete training and disclosure status.

That is an oversight failure, not proof that every unverified person skipped training or concealed a conflict. The distinction is central when interpreting percentages from the nine-agency subset.

Method and original contribution

What’s the Scoop With Broach recomputed subset shares, the 62.1% departure share and the 13.1% special-government-employee share, then reconciled the 210 position rows to 206 unique people using GAO’s footnote. The result shows why the public record supports a coverage-gap finding—not a universal violation claim.