Labor Cut $100 Million of Apprenticeship Red Tape. The Race Goals Were 1% of It.

On Friday the Labor Department published a final rule that strips the affirmative action machinery out of the federal Registered Apprenticeship system, the program that trains more than 700,000 electricians, pipefitters, sheet metal workers and nurses across 26,628 programs.¹ The coverage has been about race and sex goals. I read the agency's own cost tables, and that is not where the money is. The utilization analyses and goals that drew every headline account for about 1% of the $100 million a year the rule saves. The other 99% is generic paperwork that every sponsor, large or small, had to carry just to stay registered. That matters more than the politics, because the country is short of exactly the workers these programs train, and the companies building data centers and factories are rationing work by how many of those people they can find.

Key Highlights

• The action: DOL's final rule, RIN 1205-AC21, rescinds the affirmative action requirements in 29 CFR part 30 and replaces them with a single duty to follow existing federal and state anti-discrimination law, effective Jan. 5, 2027.²
• The savings: $100.18 million a year at a 7% discount rate and $1.016 billion over 10 years undiscounted, against $9.87 million of one-time costs.³
• The quiet number: Race and sex workforce and utilization analyses save about $986,000 a year combined. Orientation sessions and mandatory anti-harassment training save $56.31 million.⁴
• The shortage: Construction alone has nearly 270,000 registered apprentices, and the rule cites South Dakota at 41 available workers for every 100 open construction jobs, with only 10 states reporting a surplus.⁵
• The idea: Comfort Systems USA (FIX), with a record $17.81 billion backlog, up 90% in a year, closed Friday at $1,017.50, about half its 52-week high of $2,073.99.^{6, 7}

What The Rule Actually Does

Part 30 is the rulebook for equal opportunity in registered apprenticeship. Under the version DOL just replaced, every sponsor had a general duty to engage in affirmative action. Every sponsor had to name an equal opportunity coordinator, post an equal opportunity pledge, run universal outreach to recruitment sources, and give anti-harassment training to apprentices and to the journeyworkers who supervise them.⁸ Sponsors with five or more apprentices also had to write an affirmative action plan, compare the race, sex and ethnicity of their apprentices to the local labor pool, set a utilization goal for any group that came in low, aim for a 7% disability utilization goal, and invite every applicant and apprentice to self-identify a disability.⁹

The final rule removes all of that. What survives is one sentence: sponsors must comply with all applicable federal and state laws prohibiting discrimination on the basis of race, color, religion, national origin, sex, age (40 or older), genetic information or disability.¹⁰ Registration agencies stop investigating discrimination complaints themselves and instead refer apprentices to the EEOC, the Attorney General or a state fair employment agency. A sponsor that loses a final, non-appealable discrimination case must disclose it to the registration agency within 60 days, and the agency can then suspend new registrations or start deregistration if the sponsor does not carry out the remedy.¹¹

Trade and HR press read it the same way. WorldatWork called it a complete revamp that eliminates the written plan, the demographic analysis and the mandatory training¹²,, and the International Foundation of Employee Benefit Plans listed the same removals.¹³ Bloomberg Law's headline said Labor "guts" the discrimination rules.¹⁴ The proposal came out on July 2, 2025, so this has been a year and four months in the making, and OMB designated it deregulatory and economically significant.¹⁵

Where The Money Actually Is

Here is the part I have not seen anyone write. DOL broke its savings estimate into line items, and the line items tell a different story from the coverage.

The largest single item is the annual orientation session and anti-harassment training, at $56.31 million a year.¹⁶ Next is the disability self-identification invitation, at $19.37 million, then the affirmative action program reviews and plan updates at $12.71 million, then universal outreach at $10.58 million.¹⁷ The workforce analysis by race, sex and ethnicity saves $876,680 a year and the utilization analysis $109,585.¹⁸ Targeted outreach for groups that came in under goal saves $137,135.¹⁸ Add the three goal-related items together and you get about $1.1 million, roughly 1.1% of the total, by my arithmetic.

So the honest description of this rule is not "Labor ends race quotas in apprenticeship." The old rule itself said quotas were "expressly forbidden".¹⁹ The honest description is that Labor took a stack of process obligations, most of which applied to every sponsor regardless of size, and replaced them with the law that already applied to every employer. For a sheet metal contractor with eight apprentices, the binding cost was never the goal. It was the plan, the annual 45-minute session, the disability forms and the records.¹⁷

The per-sponsor number is small. DOL projects 26,810 active sponsors in 2026²⁰, which works out to roughly $3,700 a year each, by my calculation. Nobody builds a training program over $3,700. But that average hides the real effect, which is on the decision to register at all.

The Front Door Problem

Registration is voluntary. An employer can train its own people any way it likes. Registration buys the apprentice a portable credential. It costs the sponsor the compliance stack. DOL quotes its own earlier task force: "many employers choose to establish apprenticeship programs outside of the registered apprenticeship program, in part because of the paperwork and process involved in registering a program".²¹

That is the market failure a free-market reader should care about. Part 30 front-loaded its requirements onto prospective sponsors before they had registered a single apprentice: name a coordinator, write a policy, post a pledge, document selection procedures, plan for the rest.²² Fixed costs at the door fall hardest on the smallest entrants. A 30-person mechanical shop looks at the form and trains informally instead, and the apprentice ends up with skills but no credential that travels.

The numbers on growth are worth stating plainly. Active apprentices rose 77% from 395,764 in fiscal 2016 to more than 700,000 in fiscal 2025, growing about 7% a year against 2.3% a year in the decade before.²³ An executive order now asks for 1,000,000 active apprentices.²⁴ DOL's own projection has sponsors rising from 26,810 in 2026 to 30,114 by 2035.²⁰ That is not fast enough to close a shortage the rule itself sizes at roughly 5.25 million workers needing training beyond high school through 2032.⁵

One more detail tells you who cared. On Oct. 2, the Sheet Metal and Air Conditioning Contractors' National Association and the SMART union met jointly with OMB about this rule before it was finalized.²⁵ When the contractors' association and the union show up together, the issue is the pipeline of trained workers into union shops, and those are the shops wiring and cooling the data center buildout.

The Free-Market Read

I want to steelman the other side first, because it is not frivolous. Supporters of part 30 point out that women apprentices grew 12.3% a year and Black apprentices 7.6% a year from fiscal 2016 to 2025, faster than total enrollment.²⁶ Commenters cited women at 14.3% of apprentices against 47.1% of the labor force.²⁷ They argue the outreach and training requirements are why those numbers moved, and that routing complaints to the EEOC takes away a faster, apprenticeship-specific forum.

The evidence does not carry that argument as far as its advocates want. DOL says it found no direct causal link between the part 30 requirements and the gains in representation.²⁸ To its credit, the department is equally candid about its own side, conceding it cannot show causally that the old rules held growth back either.²⁹ When neither side can show the rule moved the outcome, the default should be the cheaper rule, and the cheaper rule here still keeps discrimination illegal under Title VII and every state law that applies.

The deeper point is about what the government is selling. A registered apprenticeship is a credential the state certifies. Every condition attached to that credential is a tax on producing it. If you want more of something, you lower the tax. A contractor desperate for workers recruits from everywhere. A labor market with 41 workers for every 100 openings does more for outreach than any pledge on a breakroom wall.

Where I part company with the cheerleaders is the size claim. This rule alone will not get the country to a million apprentices. The savings are real but modest. The bigger barriers are wages, the multiyear length of many programs and the willingness of firms to invest in workers who can leave. Deregulation removes a cost at the front door. It does not build the house.

Winners And Losers

Winners: Small and midsize sponsors with five or more apprentices, which DOL says may lack the staff to run demographic analyses.³⁰ Construction, which DOL calls the largest concentration of programs, at 33.5% of programs in its 2024 data.³¹ Newer fields such as IT and health care, where DOL says the old requirements weighed more heavily on new entrants.³² And the large mechanical and electrical contractors who live or die on craft labor.

Losers: The compliance consulting work that built affirmative action plans for sponsors. Apprentices who would have used an apprenticeship-specific complaint process and will now go to the EEOC or a state agency. And, in a sense, the state apprenticeship agencies, 32 recognized ones plus California's, which DOL estimates will each spend about a year of staff time rewriting their state plans, at $6.10 million in total.³³

The Bear Case

Start with the states. SAAs must submit plans within a year of the effective date, and the rule requires state rules to "conform to" part 30, not "conform only to" it.³⁴ That wording leaves room for a state to keep its own requirements as long as they do not demand anything illegal. If the largest states with their own agencies keep most of the old stack, the national savings shrink and sponsors in those states see little change.

Second, litigation. A rule this visible will draw challenges, and even a failed challenge can delay state implementation. Third, the savings are small relative to the problem, roughly $3,700 a sponsor a year. If you buy anything on this rule alone, you are buying a headline.

For the investment idea below, the bear case is different and bigger: the AI construction cycle. Last week's selloff hit the companies building data centers hard, and contractors with large technology backlogs fell with them.

The Idea: Comfort Systems USA (FIX)

If any listed company runs on the supply of trained craft workers, it is Comfort Systems. The mechanical and electrical contractor reported third-quarter revenue of $3.77 billion, up from $2.45 billion a year earlier, earnings of $13.91 a share against $8.25, and operating cash flow of $1.01 billion in a single quarter.⁶ Backlog hit a record $17.81 billion, up from $14.06 billion at the end of June and $9.38 billion a year earlier.⁶ Revenue beat the consensus estimate of $3.40 billion and earnings beat $12.71.³⁵

Management is explicit that labor sets the speed limit. On the call, President Trent McKenna said the company employs over 25,000 craft workers and that "if we've taken work, it means labor planned against every dollar in that backlog".³⁶ Coverage of the call noted that labor availability remains a key capacity constraint³⁷, and the company's quarterly filing says it expects labor-cost pressures to persist over the next several quarters.³⁸ "We pay well, we train relentlessly," McKenna said.³⁶

And yet the stock closed Friday at $1,017.50, against a 52-week high of $2,073.99, at 21.97 times trailing earnings of $46.32 a share.⁷ It fell about 9% the day after that blowout quarter and is down 49% from its June 30 close of $1,981.95, by my calculation from month-end prices, while its backlog rose 27% between June 30 and Sept. 30.^{6, 39}

What the market may be missing. My view is that investors are pricing Comfort Systems as an AI derivative and ignoring what actually limits its earnings. The limit is not demand, which is contracted 18 to 24 months out according to CEO Brian Lane.³⁷ The limit is how fast the industry can turn a new hire into a journeyworker. Anything that widens that pipe, including a rule that makes it cheaper for every contractor in America to register and run an apprenticeship program, helps the companies that are rationing work by labor. This rule is a small push. It is a push in the right direction for a stock the market just cut in half.

Risks. Technology work is 59% of revenue, up from 42% a year earlier, and industrial customers were 76% of revenue in the first nine months.³⁷ A pullback in data center spending would hit bookings. Management itself raised customers' ability to pay as a factor in what work it accepts.³⁶ Wage inflation can squeeze fixed-price contracts. And the apprenticeship rule's effect on labor supply will take years to show up, if it shows up at all.

What would prove me wrong. A sequential decline in backlog when the company reports fourth-quarter results, scheduled for Feb. 19, 2027⁴⁰, or a gross margin that falls meaningfully below the 24.8% it reported in the third quarter.³⁷ Either would say the cycle, not the labor supply, is now the constraint. I keep score on every idea here, winners and losers: the FMR Idea Scorecard.

What To Watch Next

• Jan. 5, 2027: The rule takes effect. Sponsors must meet the streamlined standard and registration agencies start assessing compliance under it.⁴¹
• Jan. 5, 2028: Deadline for State Apprenticeship Agencies to submit conforming plans, unless DOL grants an extension. Watch whether the big state agencies keep their own requirements.
• Feb. 19, 2027: Comfort Systems' fourth-quarter report and its first backlog figure after the AI selloff.
• DOL's next registration data: The count of new sponsors in 2027 against DOL's projection of 344 newly registering sponsors in 2026.²⁰

I track dated catalysts like these on the FMR Catalyst Calendar. For the last Labor Department rule I covered, see A Judge Certified an 80/20 Tip Case. Then Labor Buried the Rule., and for another credential gatekeeper losing its grip, Colleges Can Now Fire Their Accreditor.

Related analysis

• A Judge Certified an 80/20 Tip Case. Then Labor Buried the Rule.
• Colleges Can Now Fire Their Accreditor. Watch the Transfer Rule.
• The SEC Redefined Small. Three in Four Advisers Now Qualify.

Know someone who runs a contracting business or invests in the trades? Forward this one along.

The Free Markets Report is provided by Lead-Lag Publishing, LLC. All opinions and views mentioned in this report constitute our judgments as of the date of writing and are subject to change at any time. Information within this material is not intended to be used as a primary basis for investment decisions and should also not be construed as advice meeting the particular investment needs of any individual investor. Trading signals produced by the Free Markets Report are independent of other services provided by Lead-Lag Publishing, LLC or its affiliates, and positioning of accounts under their management may differ. Please remember that investing involves risk, including loss of principal, and past performance may not be indicative of future results. Lead-Lag Publishing, LLC, its members, officers, directors and employees expressly disclaim all liability in respect to actions taken based on any or all of the information on this writing. Nothing in this report shall be construed as an offer to provide investment advice or services. Nothing in this report shall be construed to be an offer to provide investment advisory services. Nothing in this report shall be construed as an offer or solicitation to buy or sell any security. The Free Markets Report is for informational purposes only and does not constitute a complete description of our investment services or performance. THE INFORMATION IN THIS REPORT HAS BEEN PREPARED WITHOUT REGARD TO THE INDIVIDUAL CIRCUMSTANCES OF INDIVIDUALS WHO MIGHT RECEIVE IT. THE INVESTMENTS DISCUSSED MAY NOT BE SUITABLE FOR ALL INVESTORS. INVESTORS MUST MAKE THEIR OWN INVESTMENT DECISIONS BASED ON THEIR OWN PARTICULAR FINANCIAL SITUATIONS AND OBJECTIVES. This report is not an offer to provide investment advice or services. Nothing in this report shall be construed as an offer to provide investment advice or services. Nothing in this report shall be construed to be an offer to provide investment advisory services in any jurisdiction in which Lead-Lag Publishing, LLC is not authorized to provide such services. All investments involve risk, including the risk of loss of principal. No representation is being made that any security, account, or strategy is or is likely to achieve results similar to those shown. Investing in securities involves risks, and there is always the potential of losing money when you invest in securities. Before investing, consider your investment objectives, and Lead-Lag Publishing, LLC's charges and expenses. This report is not a prospectus and does not contain all of the information necessary to fully evaluate any security or strategy discussed, including, but not limited to, the risk of adverse effects of related market developments. You should not rely solely on the information contained in this report, and should seek advice based on your situation from a licensed investment professional. Carefully consider the investment objectives, risks, charges, and expenses of any investment product before investing. This report may not be copied or distributed without the express written consent of Lead-Lag Publishing, LLC. Any decisions made based on any or all of the information in this writing are the sole responsibility of the user, and the user should exercise caution and take appropriate advice before acting.