THE GREAT SUCKING SOUND IN ENGINEERING • Part 4 of 4

H-1Bs: When the Control Group Told the Story

Mar 12, 2026

H-1Bs: When the Control Group Told the Story

Herbert Roberts, PE

Two Hallways, Two Economies

I had the privilege of working on classified defense programs during my career in aviation R&D. I cannot discuss the work itself. What I can discuss is a workforce observation that requires no classification to understand, because it reveals something fundamental about how engineering labor is valued in the United States.

Classified programs carried a strict requirement: U.S. born, U.S. citizenship, full security clearance. No exceptions. Offshore engineers were not an option. H-1B visa holders were not an option. The work could not leave the building, and the people performing it could not be substituted with lower-cost labor from any source. The government mandated a closed labor pool.

And within that closed pool, a predictable thing happened. Salaries were higher. Annual raises were meaningfully larger. Compensation reflected what basic economics would predict when supply is genuinely scarce and demand cannot be met by importing alternatives. The market, constrained to operate on actual supply and demand, worked exactly as the textbooks describe.

Walk down one hallway and you found engineers working on classified programs, compensated in proportion to their rarity. Walk down another hallway in the same building, at the same OEM, and you found engineers of equal skill working on commercial programs, compensated at rates suppressed by the existence of offshore teams and visa-dependent workers who could be hired for less. Same company. Same discipline. Same tools. Two different labor markets, separated by nothing more than a security door and a government policy.

The Stated Purpose and the Arithmetic That Disproves It

The H-1B visa program was designed, in theory, to fill gaps in the domestic labor supply. The premise was that certain specialized roles could not be staffed with American workers, and that allowing employers to sponsor qualified foreign nationals would benefit the economy by ensuring critical work could proceed. The program carried provisions requiring employers to pay prevailing wages and to demonstrate that domestic candidates were not available for the positions being filled.

The arithmetic tells a different story. According to the National Center for Education Statistics, U.S. universities have awarded between fifteen thousand and forty-five thousand bachelor’s degrees in mechanical engineering per year over the last four decades, with the number climbing steadily from the mid-1980s onward. Aerospace engineering added another fifteen hundred to nearly nine thousand per year over the same period, depending on the era. Combined, the domestic pipeline of mechanical and aerospace engineering graduates has averaged roughly twenty thousand or more per year—the overwhelming majority of whom are U.S. citizens and permanent residents, since temporary visa holders account for only about eleven percent of engineering bachelor’s degrees.

Now recall the demand side. The entire U.S. gas turbine structural design workforce across all major OEMs numbers roughly ten thousand engineers. To fully staff every structural design position in the industry—including replacement of retirees, normal attrition, and growth—you would need to capture two to five percent of each year’s domestic graduating class. Even accounting for competition from automotive, oil and gas, defense systems, HVAC, and every other industry that recruits mechanical and aerospace engineers, the domestic supply was never the bottleneck.

The bottleneck was compensation. The OEMs did not face a shortage of qualified American engineers. They faced a shortage of qualified American engineers willing to accept salaries that had been suppressed by the very mechanisms described throughout this series. The “skills gap” that justified H-1B hiring was, in practice, a wage gap dressed in workforce development language. Had the OEMs offered compensation that reflected the rarity and consequence of the work—the rate that classified programs proved the market would bear—they would have had no difficulty filling every position from the domestic pipeline.

The H-1B program provided an alternative to letting the market clear at its natural price. Rather than competing for domestic engineers by offering salaries commensurate with the scarcity of the role, OEMs could sponsor visa holders at compensation levels that undercut the domestic market. The “prevailing wage” determination—the mechanism supposedly ensuring fair pay—was itself influenced by the wages that visa-dependent hiring had already compressed. The benchmark was contaminated by the practice it was meant to regulate.

Credential Versus Capability

This is not an argument against the talent of foreign-born engineers. Many H-1B holders I worked alongside were technically gifted, academically accomplished, and deeply committed to the work. Some became among the best analysts in their groups. The issue was never individual competence. The issue was systemic: what happens when you bring engineers into a design organization that has already lost its institutional knowledge transfer mechanisms?

As described throughout this series, the tribal knowledge that sustained gas turbine design was transmitted through mentorship—years of sitting next to a senior engineer who had watched hardware succeed and fail, who could explain why a particular boundary condition mattered, who had seen the gap between analysis and test and knew how to close it. That mentorship pipeline was already fraying by the time H-1B hiring scaled up. Senior engineers were retiring. The documentation effort had captured steps but not judgment. The feedback loops connecting design to manufacturing and test were weakening.

Into this degraded environment came engineers who possessed excellent academic credentials but had no access to the institutional context that credentials could not provide. They arrived with strong fundamentals in finite element analysis, solid mechanics, and materials science. What they could not arrive with was a decade of watching their designs get built, tested, and returned from service. That context was not something a university could teach or a hiring process could evaluate. It was something only a functioning mentorship pipeline and an intact feedback loop could provide—and both were disappearing.

The confusion of credential with capability was not the fault of the engineers hired under the program. It was the fault of the system that treated a master’s degree and FEA proficiency as interchangeable with the calibrated judgment that took a decade to develop inside a functioning design organization. The credential got people through the door. The capability required an infrastructure that the corporation was simultaneously dismantling.

The Mentorship Collapse

The compounding effect of outsourcing and H-1B hiring on the mentorship pipeline was devastating in a way that headcount numbers could not reveal. A healthy engineering organization maintains a ratio of senior to junior engineers that allows knowledge to transfer across generations. The senior engineers carry the judgment. The junior engineers carry the computational speed. The overlap—the years during which a junior engineer works under senior supervision on real hardware programs—is where institutional knowledge is reproduced.

Outsourcing removed work from the domestic organization, which reduced the need for domestic junior engineers, which eliminated the entry point of the mentorship pipeline. H-1B hiring filled some of those domestic positions with engineers who were qualified on paper but entered an organization where the senior mentors were being pushed toward early retirement to reduce cost. The pipeline was being drained from both ends: fewer experienced mentors at the top, fewer long-tenure domestic engineers developing at the bottom, and a middle increasingly populated by analysts who could execute procedures but had never been taught to question them.

The H-1B structure itself contributed to the instability. Engineers on employer-sponsored visas were, by the nature of the program, dependent on their employer for their immigration status. This dependency created a workforce that was less likely to push back on unreasonable schedules, less likely to challenge analytical assumptions that seemed questionable, and less likely to raise concerns that might jeopardize their position. It was not a workforce that could afford to say, “This boundary condition doesn’t look right, and I need three months to investigate before I sign this analysis.” The incentive structure rewarded compliance, not inquiry.

The Same Coin

Outsourcing and H-1B hiring were not separate strategies. They were two faces of the same coin—complementary mechanisms for achieving the same objective: decoupling engineering compensation from the actual scarcity and consequence of engineering work. Outsourcing established that the work could be done elsewhere for less. H-1B hiring established that cheaper labor could be brought here to do it. Together, they created a pincer that compressed domestic engineering wages from both sides.

The classified programs proved, by exception, that this compression was a policy choice and not an economic inevitability. When the government closed the labor pool, compensation responded to actual scarcity. When the commercial side opened the pool, compensation responded to artificial abundance. The engineers were the same. The skills were the same. The tools were the same. The only variable that changed was whether the corporation was permitted to substitute cheaper labor—and that single variable determined whether an engineer earned what the work was worth or what the market’s most affordable participant would accept.

The same dynamic that suppressed wages in gas turbine design has played out across software engineering, hardware engineering, and other highly skilled technical fields. The specific technologies differ. The mechanism is identical: document the work, demonstrate it can be performed at lower cost elsewhere or by imported labor, and use that demonstration to cap domestic compensation regardless of the rarity or importance of the role.

What Rarity Should Have Meant

Ten thousand engineers in a nation of millions, performing work that determines whether jet engines stay on wings at forty thousand feet. A domestic university system producing more than twenty thousand mechanical and aerospace engineering graduates every year—more than enough to staff the entire industry twice over. A classified defense program proving that when substitution is prohibited, the market pays what the work is worth.

There was no skills shortage. There was never a skills shortage. There was a compensation shortage—a deliberate refusal to let the market clear at its natural price, enabled by offshore outsourcing and visa programs that created the appearance of a labor supply problem to justify what was, at its core, a labor cost strategy.

The classified programs demonstrated this principle. The commercial programs violated it. The difference was not talent, not productivity, not market forces in any honest sense of the term. The difference was that one side of the building operated under rules that respected the actual labor supply, and the other side operated under rules designed to circumvent it.

Ross Perot’s great sucking sound was not just the sound of jobs leaving. It was the sound of value being extracted from the people who create the most consequential engineering in the world—the engineers who keep us in the air. The sound did not stop when the factory floor went quiet. It moved upstairs, into the design offices, and it has been running ever since.

This is the final installment of The Great Sucking Sound in Engineering, a four-part series examining how cost pressure, tool transitions, outsourcing, and labor policy reshaped the engineering workforce that designs the gas turbines powering commercial and military aviation.

The series doesn’t offer solutions, and that’s appropriate. The purpose was diagnosis, not prescription. But the question I’m left with is whether this is reversible, or whether the industry is now permanently operating on a lower base of institutional knowledge. The classified programs prove that when you close the pool, the market pays. But the classified programs are a small island. The commercial side is the ocean, and the ocean has been diluted.

The Great Sucking Sound series is essentially the post-mortem on what happens when you run a knowledge-intensive industry on cost-arbitrage logic. Outsourcing, at its core, is a lazy business idea — it’s the C-suite reaching for the easiest lever in the room and calling it strategy. If CEOs want to be known as true leaders, they need to bring the same creativity and innovation to their business solutions that they demand from their engineers on products and services. You cannot stand in front of an engineering team and ask for breakthrough thinking, then walk back to the boardroom and greenlight the most unimaginative decision in the portfolio. That contradiction is what the series documents. Low science. Long consequences.

One question before you go.

If you work in engineering — any field, any country — answer this honestly: Do you believe the people around you are paid what the work is actually worth? Not what HR says the market bears. What the work is worth.

Leave your answer in the comments. Especially if you’re outside the U.S. The classified hallway was American. I want to know what the other hallways look like.

Thank you for following me on this journey. I would be very interested in reading about your experiences, opinions and feedback, good and bad. Please leave a comment, especially if you are outside the US or work in a different engineering field.

The Inventor's Mind Blog's Substack

Discussion about this post

Ready for more?