The Agency That Fired the People Who Knew Where the Files Were

A stark, documentary-style wide shot of a mostly empty open-plan government office — rows of vacant desks, cleared of personal items, a few remaining monitors dark and unplugged, desk chairs pushed in neatly as if cleared out in a hurry. Filing cabinets line one wall, some drawers open and empty. A single surviving plant sits on a windowsill. Fluorescent overhead lighting. The aesthetic is institutional hollowing-out — not chaos, but the eerie orderliness of a workplace after a mass departure. Muted tones of beige, institutional gray, and pale office-window light.

When GAO auditors asked the General Services Administration for basic records on which federal buildings it had sold, and when, and how long each sale took, the agency handed over a dataset in which nearly 45 percent of the values were missing across the two fields that mattered most. Not obscure fields. The Acceptance Date and the Completed Disposal Date, the start and finish lines of the entire disposal timeline GAO was asked by statute to evaluate. GSA runs the federal government’s primary real estate disposal program. It could not, on the first attempt, produce a clean accounting of its own recent transactions.

Half the Staff, All of the Institutional Memory

The explanation, once GAO pressed for it, tracks a specific and traceable event. In March 2025, GSA’s Public Buildings Service began a major reorganization that cut staffing levels by roughly 50 percent. GSA officials told auditors that even after producing a revised dataset, still missing about 6 percent of values in the same two fields, they could not fill in the remaining gaps because the staff who had worked on those specific disposal projects were no longer employed by the agency.

Sit with the shape of that admission. This isn’t a case of records being lost to a server failure or a filing error. The people were the record. Once they left, the knowledge of what happened on individual property sales left with them, because nobody had apparently built a system robust enough to survive the departure of the humans who ran it. A disposal program depends on knowing when a building was accepted for sale and when the sale closed. GSA cut the workforce that tracked those two dates by half and then discovered, when someone finally asked, that it could no longer fully answer the question.

Not a New Complaint, Just a Newly Confirmed One

This is where the story stops being a one-off staffing hiccup and starts looking like a pattern GAO has been circling for years. In April 2026, a separate GAO review concluded that GSA was not using quality disposal data and recommended the agency establish a formal plan to ensure the reliability of its sales and disposal records going forward. GSA agreed with that recommendation. As of August 2026, it remained open, meaning the agency has acknowledged the problem in writing and has not yet fixed it.

The Public Buildings Reform Board, for its part, told GAO it has its own version of the same complaint. Data in the Federal Real Property Profile, the government’s central database for tracking what it owns, was described by Board officials as unreliable and potentially inaccurate on something as basic as a building’s deferred maintenance needs. The Board also noted that OMB’s last formal call for agency disposal candidate data before January 2025 went out in October 2020, a gap of nearly five years during which the underlying inventory the entire disposal apparatus runs on simply wasn’t being refreshed.

A Program Built on Records It Can’t Verify

None of this data infrastructure is incidental to FASTA’s mission. The entire premise of the program, that the government can systematically identify unneeded buildings and dispose of them faster than the traditional process allows, depends on knowing what the government owns, what condition it’s in, and how long past disposals actually took. GAO’s own statistical analysis of disposal timelines, the one that found FASTA properties sold roughly five months faster than non-FASTA ones, had to be built on a dataset with known, acknowledged gaps, gaps the agency running the program could not close because it no longer employed the people who could have closed them.

The irony writes itself without much help. A law designed to speed up how quickly the government sheds property it doesn’t need is administered by an agency that, mid-implementation, shed roughly half its own staff and then couldn’t fully reconstruct the paper trail on the properties it had already sold. GSA says it’s working on it. The recommendation to fix the underlying data quality problem has been open since 2022, reaffirmed in 2026, and remains open still. Somewhere in that gap between acknowledgment and correction is where a federal real estate program has been quietly running for the better part of a decade, on records nobody can fully vouch for.


Advertisements

Discover more from Bureaucracy Times

Subscribe to get the latest posts sent to your email.