M-17-22 lifted the 2017 hiring freeze and asked agencies to justify staffing from their own workload data

OMB’s April 12, 2017 memo ended the federal hiring freeze and told agencies to set staffing baselines from agency data, not prior budgets. In our reading, that made workforce planning a records problem, and the 2025 orders asked the same kind of question.

Looking Back revisits past weeks one at a time, up to the most recent, and asks what it meant for the banks, health systems and public agencies we serve. The series was written in 2026, with the benefit of hindsight, and each article says which week it looks back on.

Two-minute summaryThe dek and the first sentence of each section, quoted from the article

OMB’s April 12, 2017 memo ended the federal hiring freeze and told agencies to set staffing baselines from agency data, not prior budgets. In our reading, that made workforce planning a records problem, and the 2025 orders asked the same kind of question.

  1. The week The freeze had been in place since January.
  2. The baseline: staffing measured from the work A budget-based baseline needs one number: the FTE ceiling in last year’s allocation.
  3. The cost: total personnel cost, grades and layers Two units with the same headcount can cost very different amounts.
  4. The roles: positions that technology or shared services might change The third set of instructions concerned what jobs should exist at all.
  5. The record: performance management and the data behind it The fourth set of instructions concerned the people agencies kept.
  6. How the memo reached different kinds of agencies The large departments were to meet with OMB in July 2017 in place of their usual summer reviews. [1] They had chief human capital officers, workforce planning staff and personnel data systems.
  7. Objections, in order Its stated purpose was to reduce the size and cost of the workforce and reform programs. [1] Its method for deciding where to cut was data: FTE baselines from agency data, total personnel cost, vacancy reviews and performance records. [1] An agency could disagree with the purpose and still need the data.
  8. What agency leaders could check The four parts translate into checks that apply under any workforce policy.
  9. With hindsight The process M-17-22 began reached its end point in June 2018.

The week

On April 12, 2017, the Director of the Office of Management and Budget issued Memorandum M-17-22, Comprehensive Plan for Reforming the Federal Government and Reducing the Federal Civilian Workforce. [1] The memo said it fulfilled a requirement in the January hiring freeze memorandum for OMB to prepare a long-term plan to reduce the size of the federal workforce. [1] As a result, it said, the government-wide hiring freeze was lifted upon its issuance. [1]

The freeze had been in place since January. A presidential memorandum dated January 23, 2017 ordered a freeze on hiring federal civilian employees across the executive branch. [2] No vacant position existing at noon on January 22, 2017 could be filled and no new positions created, except in limited circumstances. [2] Military personnel were excluded, and agency heads could exempt positions needed for national security or public safety. [2]

The January memorandum set the clock for what followed. Within 90 days, the OMB Director, consulting the Director of the Office of Personnel Management, was to recommend a long-term plan to reduce the size of the federal workforce through attrition. [2] The freeze would expire when the OMB plan was implemented. [2] It also said that contracting outside the government to circumvent the memorandum’s intent was not permitted. [2]

A second instruction had come on March 13, 2017. Executive Order 13781 directed each agency head to submit, within 180 days, a proposed plan to reorganize the agency, if appropriate, to improve efficiency, effectiveness and accountability. [3] It directed the OMB Director to invite public suggestions through a Federal Register notice and then to propose a government-wide reorganization plan. [3] The order listed factors to consider, including whether functions would be better left to state or local governments or the private sector, whether they were redundant, and whether their costs were justified. [3]

M-17-22 tied those threads together. It required all agencies to begin immediate actions toward near-term workforce reductions and cost savings, including planning for the funding levels in the President’s FY 2018 Budget Blueprint. [1] Agencies were to develop a plan to maximize employee performance by June 30, 2017. [1] They were to submit an Agency Reform Plan to OMB in September 2017 with their FY 2019 budget submissions, including long-term workforce reductions, with a high-level draft due June 30, 2017. [1]

The memo set a calendar beyond that. In July 2017, in place of the FedStat and Strategic Reviews that normally occur in the summer, OMB would meet with the Chief Financial Officers Act agencies and a limited number of others. [1] OMB would publish a government-wide reform plan as part of the President’s FY 2019 Budget. [1] Starting in February 2018, OMB would track progress through the federal performance framework, including public reporting of workforce reductions in all major agencies. [1]

The memo defined agency by reference to section 551(1) of title 5, consistent with the executive order. [1] [3] All agencies, boards and commissions were to submit plans in September 2017 unless OMB granted an exception. [1] Detailed timelines, templates and examples were placed on a MAX site available only to executive branch agencies. [1]

Its analytical framework asked agencies to sort proposals into four categories: eliminate activities, restructure or merge, improve organizational efficiency and effectiveness, and workforce management. [1] A table of factors set out when to explore each option, from duplication and non-essential functions to cost-benefit and customer service. [1]

The workforce section held the memo’s instructions for anyone running an agency’s people, budget or technology. It is the part this article examines, because it turned the end of a freeze into a set of questions that an agency could answer only from its own records. [1]

A hiring freeze is simple to administer. Nobody is hired, with exceptions. Lifting one is harder, because something has to decide which positions are filled next. M-17-22 answered that question with data. It told agencies to set staffing levels from evidence about the work, not from last year’s budget allocation. [1]

Our argument is that M-17-22 made workforce planning a records problem. Each of its workforce requirements asked for a fact that lived in a personnel, payroll, time or case system, and each could be answered only as well as those systems allowed. [1] The same kind of request returned in 2025, with tighter terms and a new set of approvers. [5] [6]

We take the memo’s workforce requirements in four parts:

  1. The baseline: how many people a program needs, measured from the work.
  2. The cost: what the workforce costs in full, and how its grades and layers are arranged.
  3. The roles: which positions technology or shared services might change or remove.
  4. The record: how managers handle performance, and what data shows it.

The first part concerns size. The second concerns price and shape. The third concerns which jobs exist. The fourth concerns how well the people in them are managed. Together they cover the questions any staffing decision must answer. After the four parts we look at how the memo reached different kinds of agencies, take objections in order, and list what an agency leader could check.

A note on sources and our reading. We describe the memo, the hiring freeze and Executive Order 13781 from the documents themselves. [1] [2] [3] We describe later events from OMB’s 2018 reform plan and two 2025 executive orders. [4] [5] [6] What the memo implied for an agency’s systems and records is our reading. The memo itself said that agency heads kept approval authority for their final workforce plans. [1]

The baseline: staffing measured from the work

The memo told agencies to use agency data to determine appropriate full-time equivalent baselines. [1] It named data sources: career field benchmarking and time studies, among others. [1] Instead of relying on previous budget allocations that set FTE levels, agencies should better examine how many people are required to perform tasks at the level required, the memo said. [1]

What a work-based baseline requires

A budget-based baseline needs one number: the FTE ceiling in last year’s allocation. A work-based baseline needs at least three. It needs the volume of work, such as claims processed, inspections done or calls answered. It needs the time each unit of work takes. And it needs the level of service the program is expected to deliver.

The budget office holds none of those numbers. Work volume sits in case management systems. Time sits in time and attendance systems, if anywhere. Service levels sit in program performance plans. In our reading, M-17-22 asked agencies to join those sources for each program, in a few months, as the basis for decisions about people. [1]

Benchmarking and time studies

The memo’s two named methods differ. Benchmarking compares staffing in one career field with similar work elsewhere. Time studies measure how long tasks actually take. Each produces a defensible number only if the comparison or the measurement is sound. [1]

Benchmarking is faster and depends on a good comparator. A field office of a benefits agency can be compared with another field office doing similar work. A unique program has no comparator. Time studies work for any program but take time and invite resistance from the people being measured. An agency had to choose, program by program, which method its data would support before its long-term staffing plan went to OMB with its FY 2019 budget submission in September. [1]

Vacancies as decision points

The memo also told agencies to review positions as they became vacant. [1] Each vacancy was a chance to ask whether the duties, qualifications and placement still matched current mission needs. [1] It was also a chance to ask whether the work could be done at a lower level, and how any change could be made quickly. [1] It said that eliminating unnecessary vacant positions could begin immediately. [1]

In our reading, this turned the vacancy list into the first management report of the post-freeze period. An agency that knew which positions were vacant, how long they had been vacant and what work they covered could act at once. One that tracked vacancies only through hiring requests could not tell an unneeded position from an unfunded one.

Who owns the numbers

A work-based baseline crosses four offices. The program office knows the work. The chief human capital officer’s staff hold the position and personnel records. The chief financial officer’s staff hold payroll and budget. The chief information officer runs, or buys, the systems that store all of it. The memo gave the plan to the agency head and did not assign the data to any one of them. [1]

That left a governance gap to close in weeks. Someone had to decide which count of employees was authoritative, which time data could be trusted, and who would reconcile them when they disagreed. An agency that had settled those questions before April 2017 could spend its time on analysis. One that had not would spend it on definitions.

The risk of measuring what is easy to count

A work-based baseline has a known weakness. It favors work that produces countable units. Claims, inspections and transactions are easy to count. Policy analysis, oversight, investigations and relationship management are not. A program whose value lies in judgment can look overstaffed when measured by volume.

The memo partly addressed this. Its baseline instruction spoke of the people needed to perform tasks at the level required, and elsewhere it asked agencies to build a portfolio of evidence from program monitoring, evaluations and performance measures. [1] In our reading, an agency that wanted to protect judgment-based work needed to define its outputs in advance, in terms an examiner of the plan would accept. Leaving them undefined invited a count of whatever was easy to count.

The January memorandum had frozen every position vacant at noon on January 22. [2] By April, those vacancies had been frozen since January. M-17-22 lifted the freeze but asked agencies to decide each vacancy against mission needs, and to plan reductions over a period it described as FY 2018 through 2022. [1]

The freeze therefore left a backlog of decisions, and the memo told agencies to make them with evidence. An agency with clean position data could clear the backlog in an orderly way. In our reading, one without it was more likely to fill positions in the order hiring managers asked, which would not meet the memo’s instruction to review each vacancy against mission needs.

The cost: total personnel cost, grades and layers

The memo’s second instruction was to examine the total cost of personnel as well as the number of employees. [1] It listed what that cost included: salary and benefits, overtime, training, awards and career ladder progression, plus employee services and office expenses. [1]

Why headcount was not enough

Two units with the same headcount can cost very different amounts. One may be staffed at higher grades, with more overtime, more awards and more space. The memo gave an example: restructuring duties so that lower-graded employees do lower-level work previously assigned to higher-graded positions, and consolidating the higher-graded work into fewer positions. [1]

That example is a data question. To find it, an agency must know what work each position actually does, beyond its title and grade. Position descriptions are written when a job is classified and may not be revisited after that. In our reading, the memo’s example could be acted on only where an agency had current descriptions or some other record of real duties.

Spans of control and layers

The memo asked agencies to review organizational design and position structures. [1] Spans of control and delegations of authority were to be set so the work is done with the fewest management layers needed for appropriate risk management, oversight and accountability. [1] It named deputy positions, lower-level chief of staff positions, special projects and management analysts as places to look. [1]

Spans and layers come from the organization chart, and the chart lives in the personnel system’s supervisory codes. Those codes are only as accurate as the last reorganization recorded in them. An agency that wanted to report its spans of control had to trust its codes or rebuild the chart by hand.

Policy written twice

The memo also asked agencies to stop recrafting agency-wide policy at the component or regional level. [1] It described bureaus whose human resources and financial management staff customize agency-wide policies, and field locations that write local policy where no unique local expertise is needed. [1]

In our reading, this was an instruction about records as much as staffing. Every local policy is a document to maintain, train on and audit. An agency that consolidated policies would reduce the staff time spent on them and the number of versions an inspector general might find out of date.

Two counts of the same workforce

Agencies count their people in more than one way. The personnel system counts positions and employees on board. The payroll system counts who was paid. The budget counts full-time equivalents, a measure of hours worked rather than people. The memo asked for FTE baselines and total personnel cost, which draw on all three. [1]

The three counts disagree for ordinary reasons: part-time staff, people on leave without pay, detailees, seasonal workers and positions funded from more than one account. In our reading, a plan that mixed counts without reconciling them would not survive OMB’s review or a later inspector general audit. The reconciliation was unglamorous work, and it decided whether the cost analysis could be trusted.

The cost of reporting itself

One cost the memo addressed directly was reporting. Each government-wide management council was to identify policy and regulatory reporting requirements that were low-value, duplicative or no longer necessary. [1] Within 60 days of the memo, OMB, working with agencies that impose reporting on other agencies, would identify reporting that could be stopped or modified at once. [1] Agencies were also to list statutorily required reports they believed Congress should eliminate or modify. [1]

For a technology organization, reporting is a hidden workload. Data calls, compliance reports and recurring metrics each take staff time to assemble. The memo’s attention to reporting burden was an opening for CIOs to name the reports that consumed their staff, at the same moment they were asked to justify that staff.

The budget connection

The cost analysis fed the budget. Agencies were to plan for the FY 2018 budget reductions where applicable and to include long-term staffing plans in their FY 2019 budget submissions. [1] OPM would provide streamlined templates for voluntary early retirement and voluntary separation incentive requests and promised expedited reviews. [1]

In our reading, any request for those authorities would have had to say which positions and organizations it covered. The templates could speed up the request; they did not remove the need for the position data behind it.

The roles: positions that technology or shared services might change

The third set of instructions concerned what jobs should exist at all. The memo told agencies to keep positions current by assessing how technology may have changed or eliminated the need for some of them. [1] It said agencies should build in flexibility to adapt to technological advances while offering separation incentives as needed to create openings. [1]

The named fields

The memo named fields undergoing rapid transformation or affected by the availability of shared services. They were database administration, invoice processing, human resources transactional services, financial management and management analysts. [1]

Two of those fields are technology jobs or close to them. Database administration is the work of running the systems that hold an agency’s data. Invoice processing and human resources transactions are high-volume back-office tasks that a shared service provider or automation can absorb. The memo did not say these jobs should go. It said agencies should assess how technology may have changed the need for them. [1]

Shared services and mission support

Separately, the memo asked agencies to streamline mission-support functions such as IT, acquisition, financial management, human resources and real estate. [1] It listed options: intra- and inter-agency shared services, lines of business or shared IT infrastructure, external service providers including best-in-class contracts, and outsourcing where the total cost would be lower. [1] It also mentioned insourcing a function where a contract could be eliminated or scaled back. [1]

These options trade one kind of risk for another. A shared service provider can process transactions at lower unit cost. It also becomes a dependency for the agency’s payroll, payments or personnel actions. An agency that moved invoice processing to a provider would need to keep the staff who manage that provider, review its controls and handle its exceptions.

The technology leader’s exposure

For a chief information officer, this part of the memo cut both ways. It named database administration as a field to reassess, which touched the CIO’s own staff. [1] It also asked agencies to adopt technology that automates processes, to use shared IT infrastructure and to provide online service delivery, which in most agencies the CIO’s organization would have to supply. [1]

Doing both at once is difficult. The people who would build automation are the people whose positions might be reassessed. A CIO who read the memo in April 2017 had to decide early which technical skills were core and which could be bought. The June 30 draft plan was the place to say so.

Online service and the people behind it

Among alternative service delivery models, the memo listed providing online service delivery, alongside co-locating offices and aligning field staffing across agencies. [1] Its table of factors included customer service, asking whether a program could be redesigned to meet the public’s needs in a more accessible and effective manner. [1]

Online service moves work rather than removing it. Fewer people may staff counters and phones, while more maintain the site, support identity checks, handle exceptions and answer the people who cannot use it. A staffing plan that counted the first effect and not the second would overstate the savings. The memo’s request for total cost was the check against that mistake. [1]

The skills an agency keeps

Moving work to a shared service or a contractor changes the skills an agency needs. It needs fewer people doing the transactions and more people writing requirements, monitoring service levels, reviewing controls and handling escalations. The memo’s list of options did not spell this out, but each option implied it. [1]

For a technology organization, that shift has a specific shape. Database administration moved to a provider still needs someone in the agency who understands the data, approves changes and checks backups and access. An agency that cut those oversight roles along with the transactional ones would save money until the first serious incident, and then lack the people to respond.

The contracting limit

The January memorandum had barred contracting outside the government to circumvent the freeze. [2] M-17-22 lifted the freeze and allowed outsourcing where the total cost would be lower. [1] [2] Those two statements fit together only if an agency could show the cost comparison.

That comparison needs the total personnel cost from the second part and the full cost of the contract, including oversight. In our reading, any shift from staff to contractors in this period would be judged on whether that comparison was on file.

The record: performance management and the data behind it

The fourth set of instructions concerned the people agencies kept. By June 30, 2017, every agency had to develop a plan to maximize employee performance, with a timeline for five actions. [1]

  1. Review and update formal agency policy on addressing poor performance and conduct, removing steps not required by statute or regulation.
  2. Provide transparency around the performance improvement plan process, and maintain data on how many employees are placed on such plans and how many improve.
  3. Ensure that senior executives, supervisors, managers, team leads and employee relations staff complete training on managing performance and conduct.
  4. Ensure that supervisors’ and managers’ own performance plans hold them accountable for managing employee performance and conduct.
  5. Establish real-time support for managers, such as a Manager Support Board of employee and labor relations experts.

The list comes from the memo. [1] Its second item is the clearest data requirement in the document: agencies were to maintain data on performance improvement plans, including the number of employees placed on them and the number who successfully improve. [1]

Where performance data lives

Performance improvement plans are personnel actions with a paper trail. In our reading, an agency that tracked them in a case management tool for employee relations could report the counts quickly. One that kept them in supervisors’ files could not, and would have to build a tracking process before it could report.

The memo also linked the plan to the Federal Employee Viewpoint Survey, naming improvement on its questions about addressing employee performance as one sign of success. [1] That gave agencies a second, outside measure. A plan could look good on paper and still show no change in what employees reported.

Administrative leave

The memo asked agencies to review all employees on administrative leave because of performance deficiencies or misconduct. [1] Each was to be returned to work with alternative duties, or subjected to other appropriate action up to and including removal, consistent with current law. [1]

That review needed a list. Administrative leave is coded in time and attendance systems, but the reason for it may not be. An agency had to join leave codes with employee relations case files to know who was on leave for performance or misconduct. This small instruction tested whether two separate systems could be joined in a hurry.

Training and support records

The memo’s third and fifth actions also produced records. Training for executives, supervisors, team leads and employee relations staff had to be scheduled and tracked. [1] The memo pointed agencies to OPM courses and to the training regulation at 5 C.F.R. 412.202. [1] A Manager Support Board, if an agency chose that model, would hold regular check-ins with managers working on a case. [1]

Each of these is easy to describe in a plan and harder to evidence. Training completion lives in a learning management system. Check-ins live in calendars and notes unless someone records them. An agency that wanted to show progress in February 2018, when OMB tracking was due to begin, needed to decide in 2017 where those records would be kept. [1]

What the June 30 deadline tested

The June 30 performance plan was not a list of finished changes. It was a timeline for five actions. [1] That design was realistic. It also meant the first evidence of progress would come later, in agency strategic plans and human capital operating plans that the memo expected to be published in February 2018. [1]

In our reading, the deadline tested whether an agency knew its own starting point. A timeline is credible only if it begins from an accurate account of current policy, current training and current case data. An agency that could not describe its starting point in June would struggle to show progress against it later.

Collective bargaining

The memo said its guidance must be implemented consistent with applicable collective bargaining obligations, and that agencies must meet any lawful bargaining obligations related to accountability and performance management. [1] The January memorandum had said the freeze did not abrogate any collective bargaining agreement in effect. [2]

For an agency with bargaining units, those obligations shaped the timeline. A change to performance policy can require notice to a union and, depending on the agreement, negotiation over its impact. The June 30 deadline was for a plan with a timeline, not for the changes themselves. [1] That distinction mattered for agencies whose agreements required bargaining before implementation.

How the memo reached different kinds of agencies

The memo applied to all agencies under the title 5 definition, but its process was built around the larger ones. [1]

Chief Financial Officers Act agencies

The large departments were to meet with OMB in July 2017 in place of their usual summer reviews. [1] They had chief human capital officers, workforce planning staff and personnel data systems. Their difficulty was scale. A department with several components had to aggregate component plans into one agency plan, using data defined differently in each component.

Small agencies, boards and commissions

Small agencies were also required to submit plans in September unless OMB granted an exception. [1] Some share personnel and payroll services with a larger provider. For them, the data to answer the memo’s questions sat with the provider. A small commission’s first task was to find out what reports its service provider could produce, and how quickly.

Components and field offices

For bureaus and field offices, the memo’s instruction on local policy and its attention to management layers fell directly on their structures. [1] A regional office that wrote its own human resources guidance, or kept its own deputy layer, would be asked to justify both. Its answer depended on whether headquarters could see what the region actually did.

Agencies with exempt missions

The January memorandum let agency heads exempt positions needed for national security or public safety. [2] M-17-22 did not repeat that exemption in its workforce sections, but it left the agency head with final approval of the workforce plan. [1] An agency with a large public safety workforce therefore had room to protect it, and needed data to show which positions fell within that mission.

State, local and tribal governments

The memo was not addressed to states and localities. Its reform options, though, included delegating responsibilities to state, local and tribal governments, and its factors asked whether functions could be better performed by another level of government. [1] Executive Order 13781 had listed the same consideration. [3] For a state agency that ran a federal program, this was a signal to watch the federal reform plan for duties that might move its way, with or without funding.

Agencies on older personnel systems

Consider an agency whose personnel, payroll and time systems do not share identifiers or reporting tools. For it, every question in the memo becomes a manual data call. The memo assumed that agencies could use agency data to set baselines, without saying how good that data had to be. [1]

The burden of the memo is heaviest for such an agency, because it has to build reports before it can analyze anything. Its plan is likely to rest on estimates. An estimate is acceptable if it is labeled as one. The risk was an estimate presented as a measurement.

Inspectors general and auditors

The memo pointed agencies to evidence sources including GAO reports, inspector general reports and evaluations. [1] That cut both ways. Agencies could cite oversight findings to justify reforms. Oversight bodies could later test whether the data behind a reform was sound. In our reading, any staffing decision made under the memo became a candidate for audit, and the working papers behind it were the agency’s defense.

Contractors

For firms supplying staff augmentation to agencies, the two memoranda read together set a boundary. [1] [2] Work could move to contractors where the cost comparison supported it. It could not move simply to avoid a freeze. A contractor that wanted to win such work needed a client that had documented its comparison.

Objections, in order

1. The memo was about cutting, not data

Its stated purpose was to reduce the size and cost of the workforce and reform programs. [1] Its method for deciding where to cut was data: FTE baselines from agency data, total personnel cost, vacancy reviews and performance records. [1] An agency could disagree with the purpose and still need the data.

2. Agencies already had workforce plans

Agencies had human capital operating plans under 5 C.F.R. Part 250, which the memo referenced. [1] The memo asked for something narrower and harder: program-level staffing levels justified by evidence about the work. [1] In our reading, an existing plan helped only if it already rested on that evidence.

3. Time studies are impractical in a few months

They are slow. The memo named them as one of several data sources, alongside career field benchmarking. [1] An agency could use the faster method where it had comparators and plan studies for the rest. What it could not do, under the memo, was default to the prior budget allocation. [1]

4. The memo’s details were on a restricted site

Detailed timelines, templates and examples were on the MAX site, available to executive branch agencies only. [1] That limited what outsiders could see. It did not change what agencies had to produce, which the public memo set out. [1] An outside reader can still see the deliverables, the dates and the data the memo asked for. What cannot be seen from outside is how each agency answered, which is why the answers mattered to the agencies themselves.

5. Later administrations changed course, so the memo did not matter

The process it began produced OMB’s June 2018 reform plan, which said agencies had submitted initial reform ideas in June 2017 in response to the April 12 memo. [4] The kind of request it made, for staffing justified by data, returned in 2025 executive orders. [5] [6] The approach outlasted the specific deadlines.

6. Performance management is an HR matter, not a technology one

The memo asked for counts of performance improvement plans and outcomes, and a review of who was on administrative leave and why. [1] Those answers come from systems that a CIO’s organization runs or buys. Whether they can be produced on request is a technology question.

7. Data-driven staffing ignores mission judgment

It can, if the data is limited to what is easy to count, as the baseline section noted. The memo asked agencies to consider evidence and the level of service required, and left the final plan with the agency head. [1] Judgment remained in the process. The memo asked that it be written down next to the data.

What agency leaders could check

The four parts translate into checks that apply under any workforce policy.

  1. For each major program, can we state the work volume, the time per unit and the service level that justify its staffing?
  2. Can we report total personnel cost by organization, including overtime, awards, training and space as well as salary and headcount?
  3. Do our supervisory codes and position descriptions match the work people actually do, closely enough to report spans of control?
  4. Can we list every vacant position with its age, its funded status and the work it covers?
  5. Can we count performance improvement plans and their outcomes without asking supervisors to search their files?

The checks depend on each other. The first needs case and time data. The second needs payroll and facilities data. The third and fourth need clean personnel records. The fifth needs an employee relations case system. An agency that can answer all five has the evidence the memo asked for, and the evidence any later policy will ask for.

The order of work matters. Start with the fourth check, because a vacancy list is the quickest to build and the first decisions any staffing policy forces are about vacancies. Then the third, because spans of control and position data underpin both cost and baseline. Then the second and the first, which need data from outside the personnel office. Leave the fifth to run in parallel, since it depends on a different team and a different system.

Each check also has a natural owner. Vacancies and positions belong with the chief human capital officer. Total cost belongs with the chief financial officer. Work volume belongs with program offices. The systems that hold all of it belong with the chief information officer. Naming those owners in writing is the cheapest control an agency can put in place before the next staffing review arrives.

A short test shows where the gaps are. Pick one program and try to answer the first check in a day. If the answer requires a data call to the field, the agency has found the work that would decide its next staffing review.

With hindsight

The process M-17-22 began reached its end point in June 2018. OMB published Delivering Government Solutions in the 21st Century: Reform Plan and Reorganization Recommendations. [4] The plan said the assessment began in June 2017, when agencies submitted initial reform ideas in response to an April 12, 2017 OMB memo, with public comments shared with them. [4]

The plan described the steps between. Agencies refined their ideas with OMB over the summer and fall of 2017 and submitted proposals with their FY 2019 budget requests. [4] Some agency-specific improvements went into the FY 2019 Budget released in February 2018 or were adopted under existing authorities. [4] The plan itself focused on cross-agency reorganization. [4]

Its proposals included reorganizing the Office of Personnel Management. The plan proposed moving OPM’s policy function into the Executive Office of the President and moving certain operational activities, including fee-for-service human resources and IT services, to other federal entities. [4] It proposed transferring the National Background Investigations Bureau from OPM to the Department of Defense. [4] Those were proposals, and we do not trace their outcomes here.

The data-based staffing request returned in 2025. A presidential memorandum of January 20, 2025 imposed a new hiring freeze, which Executive Order 14210 of February 11, 2025 cites. [5] That order told the OMB Director to submit a plan under which each agency would hire no more than one employee for every four who depart. [5] The ratio did not apply to functions related to public safety, immigration enforcement or law enforcement. [5]

Executive Order 14210 also required each agency head to develop a data-driven hiring plan, in consultation with the agency’s Department of Government Efficiency team lead, so that new career hires would be in highest-need areas. [5] It directed agencies to prepare for large-scale reductions in force, with priority given to functions not mandated by statute. [5] The language of data-driven planning was close to M-17-22’s instruction to use agency data for FTE baselines. [1] [5]

Executive Order 14356, signed October 15, 2025, went further. It stated that the administration had surpassed the four-to-one ratio of departures to hires set in Executive Order 14210. [6] It required each agency to establish a Strategic Hiring Committee to approve the creation or filling of each vacancy. [6] Each agency was also to prepare an Annual Staffing Plan, with quarterly updates to OPM and OMB beginning in the second quarter of fiscal year 2026. [6] In those plans, agencies were to seek to eliminate duplicative or unnecessary functions and positions and to reduce unnecessary or low-value contractor positions. [6]

One clause survived almost word for word. The January 2017 memorandum said contracting outside the government to circumvent its intent shall not be permitted. [2] Executive Order 14356 said contracting outside the federal government to circumvent the intent of the order is prohibited. [6] In both cases the clause depended on someone being able to see the contracts and the positions side by side.

What has held up since April 2017 is the idea that staffing decisions would be asked to rest on agency data. M-17-22 asked for FTE baselines from workload evidence; the 2025 orders asked for data-driven hiring plans and annual staffing plans. [1] [5] [6] In our view, the agencies better placed in 2025 were those that had built the records M-17-22 asked for and kept them current.

What changed was who decided. In 2017, the memo left agency heads with approval authority for their final workforce plans. [1] In 2025, new career hires were to be decided in consultation with a team lead reporting monthly to the U.S. DOGE Service Administrator, and later approved by a Strategic Hiring Committee. [5] [6] The data requirement stayed; more approvers stood between a vacancy and a hire.

Spotted an error? Tell us and we will check it and correct the article where needed.

  • OMB
  • M-17-22
  • hiring freeze
  • workforce planning
  • agency reform
  • performance management