Apprenticeship Grants Exceeded $1.2 Billion. ETA's Workforce Fell to 894.

Federal apprenticeship grants have become a billion-dollar administrative system. USAspending records $1.221 billion in obligations under the Apprenticeship USA Grants assistance listing from fiscal 2018 through fiscal 2026, across 440 awards. The Department of Labor says its Office of Apprenticeship now manages more than $830 million in active grants, cooperative agreements and contracts.

The federal workforce responsible for that system is moving in the other direction. Labor's latest budget reports 143 actual full-time-equivalent positions in Apprenticeship Services in fiscal 2025 and 120 enacted FTE in fiscal 2026. Across the larger Employment and Training Administration, covered employee headcount fell from 1,202 in September 2023 to 894 in June 2026.

Those numbers do not prove that apprenticeship oversight is understaffed. They do not establish how many employees monitor grants, how contractor support changed or how much labor one award requires. FTE, employees, award obligations and active portfolio value are different measures.

But an Aug. 4 Labor Department inspector general report shows why the capacity question matters. Auditors found that the Employment and Training Administration did not effectively administer a completed $183.9 million apprenticeship grant initiative. In a sample of 195 participant files, 130 lacked required eligibility documentation. Recipients reported meeting one of six tracked outcome targets; a seventh measure was not tracked. The inspector general issued six recommendations covering documentation, questioned costs, data validation, monitoring, risk assessment and analytics.

The audited grants were awarded in 2019 and ended by 2024, before the largest recent workforce reduction. It would be wrong to say that the later decline caused the earlier failures. The more defensible finding is institutional: Labor is expanding apprenticeship investments while recent audits show how much federal work is required to validate results, resolve findings and distinguish a participant count from a verified outcome.

A 23-grant test of federal oversight

The Scaling Apprenticeship through Sector-Based Strategies initiative awarded $183,883,271 to 23 recipients, largely community colleges. The grants began July 15, 2019, and ended between July 2023 and July 2024. Recipients spent $164,837,656; Labor de-obligated the remaining $19 million during closeout.

The program was not the same as the current Office of Apprenticeship grant portfolio. ETA's Office of Workforce Investment managed the initiative, which used H-1B employer-fee funding and sought to expand apprenticeship into industries that rely on H-1B workers. The distinction matters because an audit of this cohort cannot be generalized automatically to every registered-apprenticeship award.

Auditors examined reported performance for all 23 recipients, reviewed 22 monitoring reports and tested 195 participant files. They found that 130 sampled files lacked documentation required to support eligibility. The report estimated $255,710 in questioned costs and $64,796,914 in funds that could be put to better use in a future program if documentation controls improved.

Those estimates need careful language. The inspector general did not identify $64.8 million already recoverable from recipients. It projected a future opportunity for better use based on the sampled documentation rate and an average cost per participant. The $255,710 figure was also an estimate; ETA agreed to review the cases and use its audit-resolution process to determine whether any costs should be disallowed and repaid.

The sampled files did not merely lack a preferred form. ETA guidance and federal award rules required recipients to retain source documents supporting participant eligibility and reported outcomes. Auditors said some recipients relied on employers to hold proof or used inconsistent practices, leaving ETA unable to verify the record centrally.

The program reported 83,785 participants served against a target of 79,380. Yet ETA's system showed 16,968 employed participants whose eligible status at enrollment, such as underemployed or incumbent, could not be determined because required logic was absent. Auditors also found 4,766 reported participants who received neither training nor any of five identified grant services.

Those problems do not establish that every affected person was ineligible. They establish that the program could not demonstrate eligibility and service delivery at the level federal oversight required.

Participant counts were stronger than completion outcomes

The program exceeded its top-line target for apprentices served by 6%. Results weakened at later steps.

Recipients reported 51,459 participants hired by an employer and enrolled in an apprenticeship education or training program, 93% of the 55,131 target. They reported 27,351 completions, 54% of target, and 25,000 participants who completed and received a degree or other credential, 52% of target. The reported incumbent-worker advancement measure reached 66% of target.

The table reproduces the final inspector general report. It should not be treated as an uncontested administrative record.

ETA disputed the fifth outcome. In its written response, the agency said wage-record matching showed 18,643 formerly unemployed or underemployed apprentices completed training and maintained employment, equal to 86% of the target. ETA argued that the data it initially gave auditors in February 2025 were incomplete and that updated March 2026 data captured employment more accurately.

The inspector general did not accept that total. It said only 11,706 unemployed and underemployed participants were shown as beginning training, so 18,643 could not represent that population alone. Auditors also said ETA supplied aggregate totals without participant-level documentation needed for independent verification. The final report revised its original figure downward to 2,275 and retained its conclusion.

This disagreement is not a footnote to be discarded. It demonstrates the central oversight problem: if program definitions, system logic and source records do not line up, even a seemingly simple result such as “completed and employed” can produce incompatible official answers.

The audit identified another break in the path. Thirty-nine percent of participants, or 32,326 people, received initial services but did not advance to training. Recipients were not required to record why. Ten of 22 monitoring reports discussed pandemic disruption, but ETA did not perform a program-wide analysis of how COVID-19 affected outcomes while the grants were active. Later independent evaluations found greater access to training and employment but could not determine the program's overall causal effect.

Tracking findings is work, not paperwork theater

ETA used GrantSolutions and a spreadsheet to track monitoring findings and corrective-action plans. Auditors said the spreadsheet could not show which findings were resolved or closed, whether corrective actions had been implemented, how long issues remained open, whether problems recurred or whether multiple recipients shared a systemic weakness.

That missing structure changes management decisions. A late quarterly report and an unsupported participant file are not the same risk. A finding open for a week and one open for a year should not look identical. A documentation problem repeated across ten colleges calls for a different response than a one-time error at one recipient.

The inspector general recommended six actions:

ETA said it had already changed later H-1B grant administration. The agency described data-validation procedures for the Industry-Driven Skills Training Fund awards made in September 2025, including source-document guidance and training for recipients. It also said its regional offices continued reviewing monitoring reports and corrective actions in GrantSolutions.

Those are meaningful corrective steps. They are not the same as a closed recommendation. The report does not provide a final closure date for all six actions, and FederalHiringData found no basis to label them completed.

A broader award stream reached $1.2 billion

Scaling Apprenticeship was one grant cohort. The larger Registered Apprenticeship funding system continued after it ended.

FederalHiringData queried the official USAspending API for Assistance Listing 17.285, Apprenticeship USA Grants. From fiscal 2018 through Aug. 26, 2026, the transaction series records $1.221 billion in obligations. Fiscal 2024 was the largest year in that interval at $272.2 million. Fiscal 2026 stood at $207.3 million through the query date.

These are obligations, not a simple sum of announcement press releases. Obligations reflect legally recorded federal commitments and can include later increases, decreases and corrections. They are not appropriations, outlays or the value of awards still active on a particular day. The 440-award count is also not a count of unique recipients, because one recipient can hold multiple awards.

The series excludes the audited Scaling Apprenticeship grants, which used a different H-1B assistance listing. Combining the two as if they were one program would erase the administrative boundaries the audit itself emphasizes.

Current expansion is substantial. In July 2026, Labor announced nearly $162 million through five cooperative agreements tied to performance incentives in shipbuilding, the defense industrial base and emerging technology. At least 85% of each award is intended to flow to eligible apprenticeship sponsors. Those agreements are new activity, not evidence about whether the 2019 cohort succeeded.

The Department is also pursuing a goal to reach or surpass one million active apprentices. Its fiscal 2025 performance report lists 702,466 active apprentices and targets of 850,000 for fiscal 2026 and 900,000 for fiscal 2027. Targets communicate ambition; they are not achieved counts.

The Office of Apprenticeship budget shows fewer FTE

Labor's fiscal 2027 Program Administration budget gives the most direct current staffing denominator for the federal apprenticeship function. Apprenticeship Services used 143 actual FTE in fiscal 2025. Congress enacted a level associated with 120 FTE for fiscal 2026, 23 fewer, or 16.1% below the actual prior-year labor total.

FTE measures labor over a year. It is not the number of people on payroll at one instant. The fiscal 2026 value is enacted, not an actual year-end result. It can therefore show the planned resource level, not exactly how many employees worked or for how long.

The 120 FTE also cannot be assigned entirely to grants. Office of Apprenticeship employees register and review programs, assist employers and sponsors, support state apprenticeship agencies, maintain standards and systems, perform quality reviews, develop policy and oversee grants and contracts.

The workload table in the same budget makes that breadth visible. For fiscal 2025, it lists more than 30 apprenticeship contracts; 217 apprenticeship grants and cooperative agreements; 145 extended program reviews; 2,785 quality and provisional program reviews; 26,342 registered programs; 334 national programs; 158 approved new or revised standards and frameworks; and 14 new apprenticeable occupations.

Dividing the $830 million active portfolio by 120 FTE produces about $6.9 million per FTE, but that ratio is not a caseload or productivity measure. The numerator includes grants, cooperative agreements and contracts. The denominator includes people performing many non-grant duties. Some monitoring is conducted by ETA regional Federal Project Officers outside the narrow apprenticeship budget activity, while contractors and state staff add capacity that federal FTE does not capture.

The fiscal 2027 request merges Apprenticeship Services into a consolidated Training and Employment activity. That may give managers flexibility, but it also ends the comparable stand-alone apprenticeship FTE line. The requested consolidated account contains work transferred from several programs and cannot be read as a sudden increase in apprenticeship staff.

ETA's larger workforce also contracted

Office of Personnel Management records supply a different denominator: covered employees across all of ETA. The series includes apprenticeship, workforce programs, unemployment insurance, foreign labor certification, grants management, regional operations and administration.

ETA had 1,380 covered employees in September 2000. It fell below 1,000 in 2006, rose to 1,265 in 2010, and reached 1,202 in September 2023. Headcount then declined to 1,141 in September 2024, 1,113 in September 2025 and 894 in June 2026.

The 308-employee decline from September 2023 to June 2026 was 25.6%. The sharper monthly change came late in 2025: headcount moved from 1,113 in September to 987 in December and 879 in January 2026.

Personnel-action records show 32 ETA accessions and 314 separations during calendar 2025, a net flow of minus 282. In the first half of 2026, the records show 37 accessions and 17 separations, a positive net flow of 20. The later result is evidence of stabilization and hiring, but six months should not be compared casually with a full calendar year.

This workforce series does not isolate the Office of Apprenticeship. It also excludes contractors. It should not be used to claim that apprenticeship oversight lost 308 people.

It does establish the broader environment in which apprenticeship work is performed. Regional grant officers, program analysts, budget staff, data specialists and managers can support multiple ETA programs. A reduction elsewhere in the agency may affect shared services even when the narrow 1109 grants-management count changes little.

The skill mix changed unevenly

Between September 2024 and June 2026, ETA's 0142 Workforce Development series fell from 231 employees to 152, a decline of 79. Management and Program Analysis fell by 28, Unemployment Insurance by 28, Program Management by 20, and Workforce Research and Analysis by 19.

The 1109 Grants Management series fell only from 51 to 49. Miscellaneous Administration and Program rose by one, from 86 to 87.

The relative stability of grants-management specialists is important counterevidence to a simple claim that grant capacity collapsed. It may indicate that ETA protected a core occupation during the wider contraction.

It does not prove that monitoring capacity was unchanged. Federal Project Officers may be classified as workforce-development specialists, program analysts or other administrative series. Financial review, data validation and technical assistance require additional skills. OPM records show occupational series, not the program or portfolio assigned to each employee.

The data-science series contained seven ETA employees in September 2024 and three in June 2026. That small count should not be treated as the agency's entire analytics capability; program analysts, statisticians, economists, IT staff and contractors can also perform data work. It does reinforce the need for agencies to publish function-level capability rather than asking occupation codes to answer every question.

Public recruiting returned, but at a lower level

FederalHiringData counted distinct ETA USAJOBS announcements opened from Jan. 1 through Aug. 26 in each year. Using the same partial-year window avoids comparing eight months of 2026 with full prior years.

Related research

Put this finding in context

All articles

Data investigations/August 23, 2026

ICE Got 150,000 Applications. Its Workforce Grew by About 9,600 at the Peak.

OPM recorded 10,030 new-hire actions during ICE's surge, but tentative offers were not final hires and covered employment had receded by June.

Data investigations/August 24, 2026

FAA Received 23 Electric-Propulsion Projects. Its Certification Skill-Gap Review Is on Hold.

FAA says current electric-aircraft projects remain supported, but two certification branches reported an eight-position specialty gap as a selected technical workforce fell 7.3%.

The archive contains 315 ETA announcements in the 2022 window, 187 in 2023, 106 in 2024, five in 2025 and 26 in 2026.

The 2026 total is more than five times the 2025 low, so the public channel had not stopped. Four 2026 announcements had “grant” in the title, including Grants Management Specialist and Supervisory Grants Management Specialist roles. The strict title rule found no 2026 announcement containing “apprenticeship” through Aug. 26, although generic titles can support apprenticeship work and current announcements may use a broader Labor agency code.

An announcement is not a vacancy, application, selection or hire. Some announcements cover multiple openings; others produce none. A title keyword is not an organization assignment. The chart measures public recruiting activity, not replacement of every employee who left.

What changed after the audit period

ETA's response describes controls that were not in place for the 2019 awards. The agency said it began requiring stronger data-validation procedures for its September 2025 Industry-Driven Skills Training Fund grants. It reviewed source-document requirements, prepared program guidance and trained recipients on performance reporting.

Labor's current budget also reports that regional offices monitored 23% of their assigned grants in fiscal 2025, above a 20% target, and completed 96% of initial risk assessments for new discretionary grants within 60 days, above a 95% target. Those figures cover the wider ETA grant portfolio, not just apprenticeship awards, and say nothing by themselves about the depth of each review.

The budget says ETA had 3,258 active grants worth $21.86 billion across its programs at the end of 2025. Regional and national Federal Project Officers use onboarding, risk assessments, enhanced desk monitoring and on-site reviews. ETA aims to monitor 20% of active grants annually. A comprehensive on-site review can take one to three weeks and may require a team.

That workload description helps explain why “number of grants per employee” is inadequate. Risk, recipient capacity, award terms, reporting quality and participant services determine effort. Monitoring a formula grant with mature systems is not equivalent to reconstructing eligibility across multiple colleges with inconsistent records.

Labor also issued new apprenticeship registration guidance in March 2026, including a commitment to make final determinations within 30 days of receiving registrations. The fiscal 2027 budget says only about half of programs submitted through Standards Builder were visible in that system during fiscal 2025, and average approval times differed between federally managed and state-agency programs. This is another reminder that federal statistics cover only the systems collecting them.

What the public still cannot see

The federal record now supports a strong description of scale. It does not provide a complete operating model.

The public cannot determine how many Office of Apprenticeship employees are Federal Project Officers, how many awards each officer monitors, how portfolios are weighted by risk, how many contractor hours supplement federal staff or how much time is spent resolving a finding. It cannot link each of the six OIG recommendations to a public completion date. It cannot compare current apprenticeship staffing with an independently validated requirement.

Those gaps prevent two opposite claims. The data cannot prove that 120 FTE is too few. It also cannot prove that an $830 million active portfolio, 217 grants and cooperative agreements, thousands of program reviews and a million-apprentice goal are adequately staffed.

A useful public dashboard would report actual apprenticeship FTE by major function; vacancies and time to fill; contractor support separately; active awards and weighted risk; monitoring coverage; median age of open findings; repeat findings across recipients; corrective-action closure; and participant outcomes that can be reproduced from validated records. It would distinguish registrations, active apprentices, participants served, completions and employment outcomes rather than treating them as one growth number.

The strongest conclusion is therefore not a staffing verdict. It is a measurement standard. A larger apprenticeship system requires more than announcing awards and counting participants. It requires federal capacity to test who was eligible, verify what happened, identify patterns while a grant is still operating and show that corrective actions closed.

The Scaling Apprenticeship audit found that ETA could not do all of those things reliably for one completed cohort. Labor says it has strengthened later grant controls. Meanwhile, its apprenticeship budget activity moved from 143 actual FTE to 120 enacted, ETA-wide headcount fell to 894 and the stand-alone apprenticeship staffing line is being merged into a broader account.

That combination does not prove failure. It makes transparent, function-level oversight data more important.

Methodology and limitations

FederalHiringData analyzed the final DOL OIG report and ETA's included response for the Scaling Apprenticeship cohort. Audit estimates, recipient-reported results and ETA's disputed Measure 5 total are labeled separately. The analysis does not extend audit findings to current recipients without evidence.

The USAspending analysis uses the official spending-over-time API for Assistance Listing 17.285 and assistance award types. It sums transaction obligations from fiscal 2018 through Aug. 26, 2026. Obligations can be revised and do not equal appropriations, outlays or active portfolio value. The ended H-1B Scaling Apprenticeship cohort is not folded into that assistance-listing series.

OPM headcount uses ETA agency subelement DLET. The long-run series uses September observations through 2025 and June for the partial 2026 point. Occupation comparisons use September 2024 and June 2026. Covered employee headcount excludes contractors and cannot identify Office of Apprenticeship assignments.

USAJOBS counts are distinct archived control numbers associated with DLET or the canonical ETA name, opened Jan. 1 through Aug. 26 in each year. Announcements do not equal vacancies, applications, selections or hires. Title keywords are examples, not a complete staffing map.

Sources and related reading

• DOL OIG report 02-26-001-03-375
• DOL fiscal 2027 Program Administration budget
• DOL fiscal 2025 performance report and fiscal 2027 plan
• Apprenticeship.gov data and statistics
• Apprenticeship Grants Performance Dashboard
• July 2026 apprenticeship cooperative-agreement announcement
• USAspending API
• Browse federal jobs
• Explore workforce statistics
• Read more data investigations