USDA has issued management directed reassignment letters to 725 employees, part of a move to five regional hubs that is expected to reach about 2,500 staff. The department says 64 percent of those 725 have accepted.
Employees and their union call that figure highly inflated. They are right, and the reason is structural rather than a matter of anyone's honesty.
The form only has one useful answer
An employee who wants a hardship or medical exemption must first accept the reassignment. The exemption is requested from inside the acceptance, not instead of it.
An employee who declines loses the position. As one Natural Resources Conservation Service employee put it: say no to the letter and the job you are assigned to goes away.
So the choices on offer are: accept, and preserve the ability to argue you cannot go; or decline, and be removed. Under those terms, "accepted" does not mean "will relocate." It means "has not yet quit, and has kept open the only door through which not going might be permitted."
Keywan Johnson, a labour relations specialist at the American Federation of Government Employees, made the operative point: many of the employees who accepted will not be able to relocate if their exemption requests are denied. The acceptance is provisional in a way the percentage does not record.
The department's own number contradicts its number
This is the part that settles it without needing to adjudicate anyone's motives.
USDA expects to retain no more than 2,000 of 4,600 employees in the national capital region. That is a planning assumption of losing well over half.
You cannot simultaneously hold that assumption and present 64 percent acceptance as evidence that the move is going smoothly. One of those figures describes what the department expects to happen and the other describes what a form recorded. They are not in tension because someone is lying; they are in tension because they are measuring different things, and only one of them is a forecast.
The 2019 precedent points the same way. When USDA moved the Economic Research Service and the National Institute of Food and Agriculture to Kansas City, the GAO found the vast majority of affected staff quit rather than go. That is the base rate for this exercise, established by this department, within living memory of the people now receiving letters.
Why relocations lose people even when they work
The mechanism is not resistance to the destination. It is that a federal scientist or economist in their forties has a spouse with a career, children in schools, and often a mortgage in a market they cannot replicate elsewhere. The decision is not about Kansas City or the hubs. It is about whether a household can move.
That population skews toward exactly the employees an agency least wants to lose: the experienced ones, who are senior enough to have accumulated the commitments that make moving hard, and employable enough to have alternatives. Relocation is therefore a selection mechanism as well as a logistical one, and it selects against tenure.
Which is the same finding this desk reached about FEMA this week: saving an agency is not the same as keeping its capability, because capability lives in people who have done the work before, and those people are the ones with the most to lose by moving or leaving.
The measurement problem, a third time this week
This is now the third story in five days in which the reported number is an artefact of how it was collected.
Childcare waiting lists count only the states that choose to keep a list, so the jurisdictions that record their shortfall look worse than the ones that turn people away at the door. Local "budget gaps" bundle three different problems under one word. And an acceptance rate gathered on a form that punishes refusal counts something other than acceptance.
None of these are frauds. Each is a number that answers the question its collection process was shaped to ask, presented as an answer to a question somebody else is asking.
The figure that will settle it
Not acceptances. Count separations and vacancies twelve months after the hub moves complete, by job series, against the same series in agencies that did not move.
That is the only measurement here that cannot be produced by the design of a form, and it is the one the GAO produced in 2019. It will be available eventually. The decision it bears on is being made now.
USDA's statement that 64 percent of the 725 employees issued management directed reassignment letters have accepted relocation; the characterisation of that figure as highly inflated by employees and by the American Federation of Government Employees; the requirement that an employee accept the reassignment in order to request a hardship or medical exemption; the consequence of removal for declining; the quoted remarks of an NRCS employee and of AFGE labor relations specialist Keywan Johnson; the expectation of about 2,500 notices in total and of no more than 2,000 of 4,600 national capital region employees remaining; the establishment of five regional hubs; and the reference to Mary Pletcher Rice, USDA's acting principal deputy assistant secretary for administration, are as reported by Government Executive on 3 September 2026. The finding that the vast majority of Economic Research Service and National Institute of Food and Agriculture staff quit rather than relocate to Kansas City in 2019 is from Government Accountability Office work as cited in that reporting. The analysis is our own.
Topics nationalfederal workforcestate and local
