Reason #25: Pipeline Mismatch Is Built In

On campus at 1 a.m., the EE is tracing a jittery signal on a scope. The SWE just pushed a tiny script that saves a teacher an hour a week. The Aero kid glues a cracked airframe, rebalances the CG, tries again. You printed a pristine CAD bracket and told yourself you were done with grease. In EE, Aero, and SWE, the hobby becomes the job. In ME, the hobby is what you hoped to leave behind.

Those other pipelines filter for patience early. Debugging code, chasing noise, and fixing airframes are slow loops, and the people who stick with them already like the rhythm they will live inside later. Mechanical engineering takes two kinds of students, hands-on tinkerers and problem-set specialists who want hardware without touch time. That split makes expectations fuzzy. You picture systems thinking, not stripped threads. You imagine invention, not fixture buy-offs. Then the semester count creeps and you are still orbiting labs, CAD, and theory while the patience you were trying to avoid waits for you at the door.

Recruiters can read this mismatch. A SWE shows a GitHub with working commits. An EE brings schematics with oscilloscope traces and notes. An Aero shows flight logs, repairs, and performance. Your portfolio is a team CAD file, a campus machine shop part you could not personally fabricate, and a simulation snapshot. None of that proves you can live in the slow loop that physical products demand. So internships, the few that exist see Reason #5, slip away to the classmates who already like the loop, and the first real offer comes from a place that needs bodies near the line, see Reason #11 and Reason #20.

Day one on the job explains the fine print. You do not hold the socket, but the socket still decides your calendar. You move holes on drawings, call out threads, argue over a torque table, and update test plans so the unit survives vibration. Suppliers were locked before you arrived, so your clever redesign becomes a washer stack note and a test fixture tweak. The technician next to you learned the tools your work actually needs, see Reason #10 and you learn what that means for status the first time a build slips tolerance, and when your borrowed authority fails there is no shield around the title, see Reason #13.

You do not turn the wrench; you answer to it.

Reason #24: Your Applicant Pool Is Global

The hiring funnel for mechanical engineers in the United States is structurally crowded, see Reason #1. Labor-market totals show fewer projected openings than new graduates, and that baseline oversupply is further amplified by a steady inflow of H-1B workers in mechanical engineering occupations. Together, these streams create a persistently deep applicant pool for entry-level and early-career ME roles (Bureau of Labor Statistics [BLS], 2025; National Center for Education Statistics [NCES], 2022; U.S. Citizenship and Immigration Services [USCIS], 2025). 

Quantitatively, the BLS projects about 18,100 mechanical-engineer openings per year on average over 2024–2034 (BLS, 2025). In the latest fully consolidated NCES year, U.S. institutions awarded 36,224 bachelor’s degrees in mechanical engineering (AY 2020–2021), roughly two graduates for every projected opening (NCES, 2022). This two-to-one ratio exists before adding experienced candidates, internal transfers, or those re-entering from graduate programs: factors that further intensify competition (BLS, 2025; NCES, 2022). 

Global labor supply compounds the squeeze. In FY 2024, USCIS approved 8,010 H-1B petitions in Mechanical Engineering Occupations alone (2,714 initial and 5,296 continuing. The practical result is that the same requisitions attract domestic new grads, experienced MEs, and returning H-1B candidates who already know the employer’s systems, fixtures, routings, and paperwork. 

These conditions are visible on the shop-floor side of ME work. Day to day, junior engineers are evaluated on risk reduction rather than invention, moving holes and thread callouts to match supplier revisions, rewriting DFMEAs when a casting tolerance shifts, shimming test fixtures to preserve repeatability, and pushing ECOs through signatures. In an oversupplied market, managers can wait for applicants who already did these tasks in the same plant. As if not bad enough, internally, the pressure is reinforced by peer competition and by the lack of a protective professional guild, see Reason #6 and Reason #13.

You were never outmatched, only outnumbered.

References

Bureau of Labor Statistics, U.S. Department of Labor. (2025). Mechanical engineers (Occupational Outlook Handbook). https://www.bls.gov/ooh/architecture-and-engineering/mechanical-engineers.htm 

National Center for Education Statistics. (2022). Table 325.47: Degrees in chemical, civil, electrical, and mechanical engineering conferred by postsecondary institutions, by level of degree: Academic years 1959–60 through 2020–21. In Digest of Education Statistics. https://nces.ed.gov/programs/digest/d22/tables/dt22_325.47.asp 

U.S. Citizenship and Immigration Services. (2025, April 29). Characteristics of H-1B specialty occupation workers: Fiscal year 2024 (Annual report to Congress). https://www.uscis.gov/sites/default/files/document/reports/ola_signed_h1b_characteristics_congressional_report_FY24.pdf 

U.S. Citizenship and Immigration Services. (2025, July 18). H-1B specialty occupations. https://www.uscis.gov/working-in-the-united-states/h-1b-specialty-occupations

Reason #23: You Are a Cost Center, Not a Contributor

Your badge says engineer, your cost code says overhead. The first-time finance walks the floor you learn the hierarchy that matters. Sales is revenue, operations is throughput, you are an expense to be managed. You propose a better bracket; they ask about unit cost and cycle time. You save a line from slipping schedule, the thank you is a reminder to hold the tooling budget flat next quarter.

This is not personal; it is how mechanical engineering is positioned. You live in validation, fixtures, packaging, and release, work that companies file under cost containment rather than value creation. The board is picked, the software is chosen, the suppliers are locked, and you inherit a pile of drawings that need holes moved, threads called out, and a torque table that nobody agrees on. The more disciplined you are, the more invisible you feel, because a perfect day has nothing to show but a green dashboard and a smaller variance. What does that do to your raise conversation.

Being labeled a cost center shapes everything downstream. Your projects are approved when they reduce scrap, shorten test time, or make the same thing cheaper, not when they make a new thing possible. When a fixture slips out of tolerance, a technician shims it and keeps the cell alive, you route the ECO and update the model later, see Reason #16. 

When the plant calls at 5 a.m., you drive in because the product exists where you live, not in a slide deck, see Reason #20.

Cost center status also drags your calendar toward coordination. You sit in three standups to defend capacity and two reviews to defend tolerances. You ship more slides than designs, see Reason #9.

Meanwhile, the programs that earn strategic credit live elsewhere. Budgets and headlines migrate to batteries, chips, and code, see in Reason #7. 

You will work hard, reduce risk, and keep the operation steady. The spreadsheet will still call you a cost to be minimized.

Reason #22: You Probably Won’t Work in the Field

Your first badge may say Engineer. Your second one often does not. In the National Survey of College Graduates, among employed workers whose highest degree is a bachelor's in mechanical engineering, about half say their job is closely related to the degree, roughly four in ten say somewhat related, and about one in ten say not related at all (National Center for Science and Engineering Statistics [NCSES], 2025). The Census reports that the most common occupation group for mechanical-engineering majors is Computer, Engineering, and Science, yet it covers only 48 percent of degree holders, with the remainder spread across management, business, and other groups (U.S. Census Bureau, 2025). Broad snapshots from the Bureau of Labor Statistics show engineering degree holders dispersed across architecture and engineering, management, and computer occupations, not just engineer titles (U.S. Bureau of Labor Statistics, 2025).

This drift is ordinary, not dramatic. You start by covering a launch, you sit in on supplier calls, you own a cost tracker for a quarter that never ends. Soon your calendar measures your value better than your drawings do, and people introduce you as the point of contact, not the designer. See Reason #9 for how PowerPoint replaces design as your actual output.

Sales wants a technical voice, so you quote from vendor selectors and learn the discount ladder. Operations wants throughput, so you chase OTIF, expedite a liner that missed a truck, and move a build because the shaker is only free on Sunday. Quality wants a narrative, so you write CAPAs, tidy control plans, and paste three clean screenshots from a test that passed on the third try. You are near the product, see Reason #20, but you are not really doing mechanical engineering.

You saw the contours on day one. The technician hears the bad bearing first and gets the cell running, while you document why the deviation is acceptable so production can keep the build alive. See Reason #16 for why the hands-on work belongs to someone else. That is why the exits are already mapped. If the company rewards schedule, cost, and customer emails, how long before your title follows the work? See Reason #14 for what custodial work looks like when the org chart still calls you an engineer.

Your cost-center status seals it. Finance does not measure your contribution by what you design. It measures you by variance and schedule adherence. See Reason #23 for why that framing turns design into overhead. When the real raises arrive, they come with new badges. Program manager. Product manager. Operations lead. Those roles still lean on your ME training, but they live in different departments with different metrics. See Reason #28 for why climbing means walking away from the work you trained for.

You wanted to design. You got to coordinate.

References

NCSES. (2025). National Survey of College Graduates: 2023 (Table 1-3). National Science Foundation. https://ncses.nsf.gov/pubs/nsf25322/assets/nsf25322.pdf

U.S. Bureau of Labor Statistics. (2025, August 28). Field of degree: Engineering (Occupational Outlook Handbook). https://www.bls.gov/ooh/field-of-degree/engineering/engineering-field-of-degree.htm

U.S. Census Bureau. (2025, July 9). Field of Bachelor’s Degree in the United States: 2022 (Table 7). https://www2.census.gov/library/publications/2025/demo/acs-59.pdf

Reason #21: Cost Down Is the Job

Your first performance goal is not invent something, it is remove dollars. You get a number that looks small on paper and huge in tooling, a cost-down target to hit before year end. You change a fastener to a cheaper grade, you shave thickness and promise the test will still pass, you swap a supplier the buyers can process in two hours. The part survives, the margin smiles, the word innovation stays in the slide template.

Most mechanical work is value engineering in plain clothes. You trade stainless for zinc-plated steel and attach a salt-spray chart. You drop an ABEC rating and accept a bushing where a bearing lived. You consolidate fastener lengths so the kit has one size instead of five, then switch to flange bolts to kill the washers. You relax a flatness from 0.05 to 0.10 so grinding disappears, you bump a surface finish from Ra 0.8 to 1.6 so a polishing step goes away, you trim weld lengths and thin a gusset because FEA says it still clears fatigue. You replace a machined spacer with a laser-cut shim stack, you change FKM to NBR and add a line in the temperature table. None of this is glamorous, all of it moves the costed BOM.

What counts as innovation when the goal is pennies? You write the ECO, update the control plan, and paste the before-after rollup so Finance can see the delta. The architecture does not change, the interfaces get cheaper. You want invention, but instead you will find yourself packaging other people’s breakthroughs see Reason #7 and Reason #14

You hit the target, then you get a new target, and that is the plan.

Reason #20: The Plant Picks Your Zip Code

Your specialization chose the metro. See Reason #11. Now the plant chooses the rest. You drove past cornfields, smokestacks, and a rail spur, then parked beside a building that smells like coolant and cardboard. Before lunch, a supervisor asks if you can swing by bay three. This is mechanical engineering. You are hired to be near the thing as it becomes the product.

Other engineers work on-site too. But civil engineers build in every city. Their jobsites are wherever people live. Chemical engineers run processes in refineries and pharma plants spread across dozens of metros. Electrical engineers increasingly do board layout, firmware, and signal work that lives in a laptop. Mechanical engineering gets the worst of both: the work is physical and the industries are clustered. A civil engineer can take a structural role in Portland or Pittsburgh or Pensacola. If your experience is in NVH for powertrain, you go to southeast Michigan, or you do not do NVH. The Bureau of Labor Statistics confirms the broader pattern. In the first quarter of 2024, workers in production, transportation, and material moving occupations teleworked at a rate of 3.2 percent. Management and professional occupations teleworked at 37.9 percent (BLS CPS, 2024). The BLS Occupational Outlook Handbook describes mechanical engineers as people who "generally work in offices" and "may occasionally visit worksites." Anyone who has worked in a plant knows what "occasionally" means. It means the phone rings at 2 a.m. because a test fixture drifted, a fitting is weeping, or the line stopped and nobody can find the revision that explains why.

When the vendor's motor shows up with a new flange, you adapt the mounting, update the keyway, move holes on drawings, call out threads, and rewrite the torque table so the build keeps moving. The shop does the hands-on work while you route ECOs and CAPA rationales, which is exactly why you must be present. See Reason #16. You are not building the thing. You are standing next to the people who are, writing the paperwork that lets them continue.

Oversupply makes the geography harsher. There are more mechanical engineers than attractive roles, so openings are where the plant is, not where you want to be. You take the job in a town chosen by rail access, utility capacity, and tax abatements because you cannot negotiate for remote when ten qualified applicants will move tomorrow. See Reason #1. Plants do not sit downtown. They sit in industrial parks, along freight corridors, in exurban strips zoned for noise and emissions. Your apartment options are whatever is within commuting distance of that zoning. Managers will say "we are flexible," then schedule standups around the production shift.

The irony is that the closer you are to shipping, the further you are from freedom. Your impact is real, very real, but the badge opens doors to the plant, not to a coastal remote policy. See Reason #30. A civil engineer can relocate and still be a civil engineer. You relocate and start over. The plant chooses where you live.

References:

U.S. Bureau of Labor Statistics. (2025). Telework trends in 2024. Beyond the Numbers, 14(2). https://www.bls.gov/opub/btn/volume-14/telework-trends-in-2024.htm

U.S. Bureau of Labor Statistics. (2025). Occupational Outlook Handbook: Mechanical Engineers. https://www.bls.gov/ooh/architecture-and-engineering/mechanical-engineers.htm

Reason #19: Grad School Doesn’t Help

You run out of momentum after the bachelor's. The job market is crowded (See Reason #1), the entry-level postings want three years you do not have (See Reason #12), and a professor suggests you stay for a master's. The postings say "MS preferred." The logic sounds clean. Two more years, a stronger resume, a better slot in the queue. So you stay.

Here is what happens to that investment. The Bureau of Labor Statistics published a cross-discipline comparison of MS-versus-BS wages for engineers and found that the mechanical engineering master's premium is 9 to 13 percent over the bachelor's median (BLS, 2015). At ME's current median of $102,320, that is roughly $9,000 to $13,000 a year. The opportunity cost of two years out of the workforce at ME's early-career median of $80,000 is $160,000 in forgone earnings, plus $40,000 to $75,000 in tuition. The total investment is north of $200,000. At a $9,000 to $13,000 annual return, the break-even is 15 to 26 years. That is the entire mid-career. You will be paying off the decision to stay in school until you are old enough to wonder whether you should have retired.

The premium is not the real problem. The real problem is that the degree disappears. Mechanical engineering graduates pursue master's degrees at rates comparable to chemical engineering, roughly 36 to 39 master's degrees for every 100 bachelor's degrees in both fields (NCES, 2022). But the ME workforce carries the lowest share of advanced degree holders among named engineering disciplines. Bankrate's 2026 analysis of Census data shows 38.9 percent of ME workers hold an advanced degree, the lowest of any engineering major. Chemical engineering: 46.7 percent. Aerospace: 48.9 percent. Electrical: 47.6 percent (Bankrate, 2026). ChemE produces master's degrees at the same rate as ME and retains them at a dramatically higher rate. The degrees are not failing everywhere. They are failing here.

The structural reason is the same one that runs through this entire blog. Chemical engineering's MS unlocks pharma R&D, process development, and specialty chemicals, work that a ChemE bachelor's cannot easily access. Electrical engineering's MS opens chip design, signal processing, and ML-adjacent hardware roles. Computer science's MS is a gatekeeper for AI and machine learning positions that explicitly require it. In each case, the credential opens a door to a different tier of work. In ME, the BLS Occupational Requirements Survey reports that 97.4 percent of mechanical engineering positions require only a bachelor's degree (BLS, 2025). On-the-job training is required for 62.3 percent. The MS does not unlock a different tier because the tier does not exist for 93 percent of the workforce. The 7 percent who work in the dedicated R&D industry earn a genuine premium (See Reason #7). The rest are in manufacturing, engineering services, and compliance, where the currency is experience, not letters (See Reason #14).

Where do the missing degrees go? ASEE and NSF workforce data show that 44 percent of engineering master's graduates work outside engineering entirely, with management being the most common destination (ASEE, 2019). The ME master's does not function as a deeper investment in mechanical work. It functions as an exit ramp into a management role that did not require the degree in the first place. You could have reached that role with two years of plant experience and a PMP, and you would have arrived $200,000 richer (See Reason #28).

The professor who suggested you stay did not show you the break-even math. The department that accepted your tuition did not show you the workforce retention data. You invested two years and six figures into a credential that your own field discards at a higher rate than any other engineering discipline. The line you left is still there when you get back. It is just two years longer.

References:

ASEE. (2019). A snapshot of engineering degree holders in the U.S. workforce. https://ira.asee.org/a-snapshot-of-engineering-degree-holders-in-the-u-s-workforce/

Bankrate. (2026, February). 2026 college majors data study. https://www.bankrate.com/loans/student-loans/college-majors-data-study/

Bureau of Labor Statistics. (2015, September). Should I get a master's degree? Career Outlook. https://www.bls.gov/careeroutlook/2015/article/should-i-get-a-masters-degree.htm

Bureau of Labor Statistics. (2025). Occupational requirements survey: Mechanical engineers. https://www.bls.gov/ors/factsheet/mechanical-engineers.htm

National Center for Education Statistics. (2022). Table 325.47: Degrees in chemical, civil, electrical, and mechanical engineering. Digest of Education Statistics. https://nces.ed.gov/programs/digest/d22/tables/dt22_325.47.asp

Reason #18: You're Paid Less Than Your Peers

The first thing you notice when you compare offer letters is that your friends in other branches of engineering make more. Electrical, chemical, computer, even civil, their starting salaries pull ahead of yours. You thought mechanical would be "the broadest," which meant "the safest." Instead, it meant you were slotted into the lowest-paying tier of the engineering ladder. See Reason #1 and Reason #34.

This is not an accident. Table 1 lays it out. As of May 2024, the Bureau of Labor Statistics reports a median annual wage of $102,320 for mechanical engineers. Electrical engineers earn $111,910. Chemical engineers earn $121,860. Aerospace engineers earn $134,830. Software developers earn $133,080. ME sits near the bottom of every named engineering discipline. The gap between you and a chemical engineer is nearly $20,000 a year at the median. The gap between you and a software developer is $30,760. And the 90th percentile column shows where the ceiling is. The best-paid 10% of mechanical engineers clear $161,240. The best-paid 10% of aerospace engineers clear $205,850. The best-paid 10% of software developers clear $211,000. Your ceiling is their midpoint. These are not cherry-picked comparisons. They are the same BLS survey, the same reference period, the same methodology.

Table 1. Annual Wages by Engineering Discipline, May 2024

Occupation	Median	90th Percentile	ME Deficit (Median)
Aerospace Engineers	$134,830	$205,850+	-$32,510
Software Developers	$133,080	$211,000+	-$30,760
Electronics Engineers (exc. computer)	$127,590	$199,060+	-$25,270
Chemical Engineers	$121,860	$182,150+	-$19,540
Electrical Engineers	$111,910	$175,460+	-$9,590
Mechanical Engineers	$102,320	$161,240+	---
Industrial Engineers	$101,140	$157,140+	+$1,180
Civil Engineers	$99,590	$160,990+	+$2,730

Source: U.S. Bureau of Labor Statistics, Occupational Outlook Handbook (May 2024 OEWS data). Median = 50th percentile. 90th percentile = the wage above which the top 10% of earners in that occupation are paid, a proxy for the late-career ceiling. "+" indicates BLS reports "earned more than" this amount. Software developers includes QA analysts and testers (SOC 15-1256). The top 10% of mechanical engineers earn above $161,240. The top 10% of aerospace engineers earn above $205,850. That is a $44,610 gap at the ceiling.

The gap does not close with experience. It widens. The Federal Reserve Bank of New York tracks median wages for recent college graduates (ages 22-27) and mid-career workers (ages 35-45) by major, using U.S. Census data. Table 2 shows the numbers for the 2024 American Community Survey. ME starts at $80,000 in early career and reaches $120,000 by mid-career. Chemical engineering starts at $85,000 and reaches $135,000. Computer engineering starts at $90,000 and reaches $131,000. Aerospace starts at $85,000 and reaches $130,000. The discipline that supposedly keeps your options open leaves you $15,000 behind chemical engineering at mid-career, $11,000 behind computer engineering, and $10,000 behind aerospace. Over a twenty-year mid-career window, the chemical engineering gap alone is $300,000 in lost earnings before compounding. See Reason #63.

Table 2. Median Wages by Engineering Major, Early Career and Mid-Career (2024 ACS)

Major	Early Career (22-27)	Mid-Career (35-45)	ME Deficit (Mid)
Computer Engineering	$90,000	$131,000	-$11,000
Chemical Engineering	$85,000	$135,000	-$15,000
Aerospace Engineering	$85,000	$130,000	-$10,000
Electrical Engineering	$82,000	$123,000	-$3,000
Mechanical Engineering	$80,000	$120,000	---
Civil Engineering	$75,000	$115,000	+$5,000
Industrial Engineering	$83,000	$100,000	+$20,000

Source: Federal Reserve Bank of New York, The Labor Market for Recent College Graduates, February 2026 (2024 ACS data). Early career = ages 22-27. Mid-career = ages 35-45.

Day to day you will see the gap widen. Electrical engineers at your company sit in fewer meetings and cash bigger checks. Chemical engineers in process industries receive bonuses tied to output, while your role is considered overhead (see Reason #23). Software engineers at startups you never heard of jump jobs every two years and double their salaries. You chase a 3 percent raise that is eaten alive by insurance premiums. By mid-career the discrepancy is not a feeling. It is $15,000 a year between you and a chemical engineer who took the same number of credits, sat through the same thermodynamics sequence, and graduated in the same four years.

Table 3 puts a dollar figure on the cost. The mid-career deficit against each discipline, projected over a twenty-year window from age 35 to 55. This is not a model. It is simple multiplication. The actual cost is higher once you account for compounding, investment returns on the difference, and the fact that raises in higher-paying fields tend to be larger in absolute terms.

Table 3. Cumulative Mid-Career Pay Deficit: ME vs. Other Engineering Disciplines

Compared to...	Annual Gap (Mid-Career)	20-Year Cost
Chemical Engineering	-$15,000/yr	-$300,000
Computer Engineering	-$11,000/yr	-$220,000
Aerospace Engineering	-$10,000/yr	-$200,000
Electrical Engineering	-$3,000/yr	-$60,000

Source: Derived from NY Fed mid-career median wages (Table 2). Gap = difference between ME mid-career median ($120,000) and comparison discipline. 20-year cost = annual gap x 20. Does not account for compounding, investment returns, or divergent raise trajectories. The actual lifetime cost is higher.

You will find yourself explaining to family members why you are still stuck near the bottom of the engineering pay scale, even after years of experience. And because your title becomes your label (see Reason #15), those same relatives will assume that "engineer" means prestige and prosperity. You will correct them, awkwardly, while your cousin in software drives off in a new car.

It is not that mechanical work has no value. It is that the market has decided it is cheap. And in this field, the market always wins.

You will be an engineer, but you will not be paid like one.

References

Bureau of Labor Statistics. (2025). Aerospace engineers. Occupational Outlook Handbook. https://www.bls.gov/ooh/architecture-and-engineering/aerospace-engineers.htm

Bureau of Labor Statistics. (2025). Chemical engineers. Occupational Outlook Handbook. https://www.bls.gov/ooh/architecture-and-engineering/chemical-engineers.htm

Bureau of Labor Statistics. (2025). Civil engineers. Occupational Outlook Handbook. https://www.bls.gov/ooh/architecture-and-engineering/civil-engineers.htm

Bureau of Labor Statistics. (2025). Electrical and electronics engineers. Occupational Outlook Handbook. https://www.bls.gov/ooh/architecture-and-engineering/electrical-and-electronics-engineers.htm

Bureau of Labor Statistics. (2025). Industrial engineers. Occupational Outlook Handbook. https://www.bls.gov/ooh/architecture-and-engineering/industrial-engineers.htm

Bureau of Labor Statistics. (2025). Mechanical engineers. Occupational Outlook Handbook. https://www.bls.gov/ooh/architecture-and-engineering/mechanical-engineers.htm

Bureau of Labor Statistics. (2025). Software developers, quality assurance analysts, and testers. Occupational Outlook Handbook. https://www.bls.gov/ooh/computer-and-information-technology/software-developers.htm

Federal Reserve Bank of New York. (2026). The labor market for recent college graduates. https://www.newyorkfed.org/research/college-labor-market

Reason #17: Professional Licensure Rarely Pays

You are told the Professional Engineer license is the brass ring, the line that separates you from the crowd. Then you look around mechanical engineering and find there is almost nowhere to use it. Civil has stamps baked into bridges and buildings (National Society of Professional Engineers [NSPE], 2024). Electrical engineers must seal power distribution and protection system designs in most states (NSPE, 2024). Medicine, law, and accounting enforce similar gates: doctors cannot practice without a medical license, lawyers cannot practice without passing the bar, and CPAs must sign financial audits (National Conference of State Legislatures [NCSL], 2023; American Institute of CPAs [AICPA], 2024). Those licenses sit at the center of the work and the pay reflects it. In mechanical, most of the work lives inside product companies and factories where no one asks for a stamp. The credential mostly decorates your HR file while your day stays the same.

The few mechanical niches that truly demand a PE are thin, mostly HVAC and building services, pressure vessels, public sector compliance, and a small slice of consulting. That is not where most MEs are hired. In product design and manufacturing, liability sits with a senior reviewer or an outside firm, and many managers prefer that you never seal anything. You wanted leverage. You got continuing education. Meanwhile you inherit vendor-chosen parts and locked suppliers, then you move holes on drawings, call out threads, argue over a torque table, and tweak fixtures so a test barely passes.

The path does not match the payoff. You pass the FE, hunt for a PE supervisor in an industry where few exist, log the hours, pay the fees, and chase PDHs. Your reward is the same job description and the same PLM clicks. The raise, if it appears, is modest. The work does not change. You still rewrite test plans and route ECOs. You still collect signatures so CAPA can close. The license does not open new rooms, it decorates the one you are already in. See Reason #27 for how your salary flattens regardless of credentials.

Worse, the supply is upside down. There are far more MEs with or chasing PEs than roles that require stamps, which is why you see licensed MEs applying to jobs that never mention licensure. In a crowded field, credentials become a way to feel less interchangeable, not a way to change what you do. See Reason #1 for how oversupply turns credentials into theater. You have already seen why this field has no guild to turn that license into protection. See Reason #13 for how ABET audits courses, not careers.

The pattern repeats with every credential you chase. PMP nudges you toward project management. Lean and Six Sigma sound impressive until you see how many green belts a single plant can mint without changing anything on the floor. At review time the bullet reads well, your salary still sits near the same plateau. See Reason #48 for why most badges serve vendors and HR, not you.

You hang letters after your name. The job stays the same.

References

American Institute of CPAs. (2024). Becoming a CPA. https://www.aicpa.org/becomeacpa

National Conference of State Legislatures. (2023). Professional licensure. https://www.ncsl.org/research/labor-and-employment/occupational-licensing.aspx

National Society of Professional Engineers. (2024). Licensure. https://www.nspe.org/resources/licensure

Reason #16: Technicians Do the Real Work, You Do the Paperwork

Your first week on the line, a technician teaches you the machine by sound. He hears the bad bearing before you can find the panel latch. You are in clean PPE with a notebook. He is three steps ahead with a nut driver. Everyone looks at you for the decision anyway, because you are the mechanical engineer. See Reason #3. You cost more. Your employer pays a 49 percent wage premium to have you there instead of a mechanical engineering technician. You earned a bachelor's degree in five or six years. See Reason #2. The tech earned an associate's in two. The Bureau of Labor Statistics puts your median at $102,320. The tech's median is $68,730 (BLS, 2024). The question is what the premium buys.

It does not buy you a day spent on engineering. Trevelyan, studying practicing engineers through ethnographic observation, found they spend approximately 60 percent of their time communicating with others and less than 30 percent on solitary technical work (Trevelyan, 2007). An earlier time study at an aerospace company found design engineers spent only 28 percent of tracked hours on problem solving, with the remaining 72 percent consumed by documentation, coordination, information gathering, meetings, and negotiation (Crabtree, Baid, & Fox, 1993). A more rigorous follow-up using 11,137 data points from 78 engineers over 20 working days found a higher share of technical work, around 63 percent, but also found that 56 percent of all time involved some form of information behavior: seeking, receiving, or providing information rather than producing it (Robinson, 2012; Robinson, 2010). The estimates vary. The pattern does not. You spent five years learning thermodynamics, machine design, and materials science. A substantial share of your day goes to routing, documenting, and coordinating instead.

The technician's day does not look like yours. ONET lists five physical work activities on the technician profile that do not appear on the engineer profile at all: repairing and maintaining mechanical equipment, controlling machines, operating mechanized devices, performing general physical activities, and handling objects. Seventy percent of technicians wear safety equipment every day. Thirty-two percent of engineers wear it at all, and only occasionally (ONET, 2024). A fixture slips tolerance, the tech shims it and keeps the cell running. A vendor swaps a motor, the tech adapts the mounting in an hour. You chase shaft fits, update the keyway drawing, move holes, argue over a torque table in a release meeting, and rewrite the test plan so the unit survives vibration. See Reason #9. When a build goes sideways, you sign off on a quick deviation. Then you own the paper trail when the field return arrives. The tech fixed the problem. You documented the fix. Sociologist Beth Bechky, studying a semiconductor equipment manufacturer, found that engineering drawings function as jurisdictional artifacts: they enable engineers to preserve occupational boundaries over technicians not through superior technical knowledge but through control over the documentation that defines what gets built (Bechky, 2003). The 49 percent premium bought your employer documentation authority. See Reason #33.

The workforce structure makes this systemic. The BLS counts 293,100 mechanical engineers and 38,300 mechanical engineering technicians and technologists in the United States (BLS, 2024). That is a ratio of nearly 8 to 1, the most lopsided of any major engineering discipline. Electrical engineering runs 2 to 1. Civil runs about 6 to 1. Industrial runs about 5 to 1 (BLS, 2024). Twenty years ago, the mechanical ratio was closer to 5 to 1. It has widened because ME employment grew 30 percent while MET employment declined 14 percent. The BLS projects zero growth for mechanical engineering technicians through 2034, noting that automation of routine tasks "may reduce the need for workers in this occupation" (BLS, 2024). In most professions, the ratio of support staff to professionals runs the other way. Medicine has more nurses than doctors. Construction has more tradespeople than architects. Mechanical engineering has replaced its hands-on tier with specification-writers and is still hiring. See Reason #72.

You wear the title. They do the work. You sign the blame.

References:

Bechky, B. A. (2003). Object lessons: Workplace artifacts as representations of occupational jurisdiction. American Journal of Sociology, 109(3), 720-752. https://doi.org/10.1086/379527

Bureau of Labor Statistics. (2024). Occupational outlook handbook: Mechanical engineers. https://www.bls.gov/ooh/architecture-and-engineering/mechanical-engineers.htm

Bureau of Labor Statistics. (2024). Occupational outlook handbook: Mechanical engineering technologists and technicians. https://www.bls.gov/ooh/architecture-and-engineering/mechanical-engineering-technicians.htm

Crabtree, R. A., Baid, N. K., & Fox, M. S. (1993). Where design engineers spend/waste their time. AAAI Technical Report WS-93-07, AI in Collaborative Design Workshop, 209-219. https://cdn.aaai.org/Workshops/1993/WS-93-07/WS93-07-018.pdf

O*NET OnLine. (2024). Summary reports: 17-2141.00 Mechanical Engineers; 17-3027.00 Mechanical Engineering Technologists and Technicians. https://www.onetonline.org

Robinson, M. A. (2010). An empirical analysis of engineers' information behaviors. Journal of the American Society for Information Science and Technology, 61(4), 640-658. https://doi.org/10.1002/asi.21290

Robinson, M. A. (2012). How design engineers spend their time: Job content and task satisfaction. Design Studies, 33(4), 391-425. https://doi.org/10.1016/j.destud.2012.03.002

Trevelyan, J. P. (2007). Technical coordination in engineering practice. Journal of Engineering Education, 96(3), 191-204. https://doi.org/10.1002/j.2168-9830.2007.tb00929.x

Reason #15: It Becomes Your Identity, For Better or Worse

You introduce yourself and the degree comes first. Not your name, not your interests, not anything you chose on purpose. The degree. It walked in ahead of you the day you graduated and it has not stepped aside since. Strangers hear "mechanical engineer" and assume competence, money, and machines. See Reason #3. People inside the field hear it and think: sustaining, ECOs, and a salary that plateaued three years ago. You spent five or six years earning the label. See Reason #2. The label now earns you. Engineering education researchers have documented how the curriculum builds this identity systematically, through lab culture, design projects, problem sets, and socialization, so that by graduation students do not merely have engineering skills; they have engineer self-concepts (Stevens, O'Connor, Garrison, Jocuns, & Amos, 2008). Erin Cech's longitudinal research found the process goes further: engineering education narrows students' broader identities, stripping away civic engagement and social breadth until what remains is almost purely technical (Cech, 2014). The narrower the identity, the more it depends on the profession delivering what it promised.

The problem is not that ME becomes your identity. Every profession does that to some degree. The problem is what the identity buys you compared to what it costs. Medicine extracts total commitment and repays it with a $250,000 median salary, legal scope protection, and near-universal employment in the field. Law extracts total commitment and repays it with title protection, bar-required practice, and a professional guild. Computer science extracts less commitment and repays it with higher wages, geographic freedom, and a job market nearly six times the size of yours. See Reason #75. Mechanical engineering extracts maximum commitment, the hardest undergraduate curriculum in engineering, the longest time-to-degree, physical work that chains you to a plant in a zip code you did not choose, and repays it with the lowest job satisfaction of any engineering discipline. The NSF's National Survey of College Graduates finds ME degree holders report the lowest share of "very satisfied" (41.2 percent) and the highest share of dissatisfaction (9.5 percent) of any named engineering branch (author's analysis of NSCG 2021 public-use microdata; NSF, 2023). You gave more. You got less. That is not an identity crisis. That is a transaction. Sociologists have a name for it. Gerhard Lenski identified "status inconsistency" in 1954: when your education rank says professional and your income rank says something lower, the mismatch generates measurable stress and dissatisfaction (Lenski, 1954; Vatter & Meuleman, 2022). Eliot Freidson spent a career studying what separates real professions from aspirational ones. His conclusion: without control over your own work, over who gets trained, who gets hired, and what counts as acceptable practice, the professional identity is ideology without a structural foundation (Freidson, 2001). You carry the ideology. See Reason #13 for what happened to the foundation.

The transaction gets worse when you measure how many people carry the identity without holding the corresponding job. There are 1,014,000 people in the United States whose highest degree is in mechanical engineering (NSF, 2023). The Bureau of Labor Statistics counts approximately 281,000 people employed as mechanical engineers (BLS, 2024). That is a titled-occupation utilization rate of roughly 28 percent. Fewer than one in three ME degree holders actually works as a mechanical engineer. Compare that to civil engineering: 663,000 degree holders, approximately 328,000 employed as civil engineers, a utilization rate near 49 percent (NSF, 2023; BLS, 2024). Civil places about half its degree holders into the job the degree names. ME places about a quarter. The rest carry the identity into management, sales, operations, adjacent technical roles, or out of engineering entirely. See Reason #22 and Reason #63. You are statistically more likely to call yourself a mechanical engineer than to work as one.

The financial penalty for staying committed makes the trap tighter with every year. Labor economists have measured what happens when long-tenured workers in manufacturing are displaced: earnings losses of 15 to 30 percent that persist for a decade or more, with each additional year of prior tenure deepening the scar (Jacobson, LaLonde, & Sullivan, 1993). Industry-specific human capital, the knowledge you build inside one sector's fixtures, standards, suppliers, and regulatory systems, is a primary determinant of wages for manufacturing workers, and switching industries forces you to forfeit those returns (Neal, 1995). ME is almost uniquely concentrated in durable manufacturing, the sector where these penalties are largest. Your two years learning the tolerance stack on a diesel fuel rail do not transfer to medical device packaging. Your five years qualifying castings for an OEM do not translate to semiconductor fixture design. Each year you stay committed, the exit gets more expensive. Psychologist Barry Staw demonstrated in 1976 that people who are personally responsible for a failing investment commit more resources to it, not fewer, because quitting means admitting the original decision was wrong (Staw, 1976). Staw and Ross extended the finding in 1987: the more factors lock you in, years of tenure, social identity, relocation, sunk cost of the degree, the less likely you are to exit even when exit is the rational choice (Staw & Ross, 1987). Staw was studying financial decisions. He might as well have been studying engineers in their forties. See Reason #46 and Reason #71.

And the field knows this about itself. Architecture and engineering occupations carry a median tenure of 4.9 years, the highest of any professional occupation group in the country (BLS, 2024). You stay longer than software engineers, longer than financial analysts, longer than managers. Not because the work is rewarding. Because leaving costs more than staying. See Reason #6. The people around you have made the same calculation. They stay. You stay. The identity calcifies into a career because the career has made every alternative more expensive than itself.

Other fields let you outgrow the label. A software engineer who moves into product management is not abandoning a credential. An EE who pivots into systems architecture is building on a foundation that compounds. See Reason #71. A civil engineer who gets a PE and opens a firm turns the identity into an asset with a retail value. See Reason #58. In ME, the identity does not compound. It confines. Recruiters filter your resume by the keyword and nothing else. Hiring managers read ten years of "Mechanical Engineer" titles and pattern-match you into more of the same. The label that was supposed to open every door opens one door, repeatedly, into the same room.

You committed to the identity. The occupation did not commit back.

References:

Bureau of Labor Statistics. (2024). Employee tenure in 2024. https://www.bls.gov/news.release/tenure.t06.htm

Bureau of Labor Statistics. (2024). Occupational employment and wage statistics, May 2023: Mechanical engineers (17-2141). https://www.bls.gov/oes/2023/may/oes172141.htm

Bureau of Labor Statistics. (2024). Occupational employment and wage statistics, May 2023: Civil engineers (17-2051). https://www.bls.gov/oes/2023/may/oes172051.htm

Cech, E. A. (2014). Culture of disengagement in engineering education? Science, Technology, & Human Values, 39(1), 42-72.

Freidson, E. (2001). Professionalism: The third logic. University of Chicago Press.

Jacobson, L. S., LaLonde, R. J., & Sullivan, D. G. (1993). Earnings losses of displaced workers. American Economic Review, 83(4), 685-709.

Lenski, G. E. (1954). Status crystallization: A non-vertical dimension of social status. American Sociological Review, 19(4), 405-413.

National Science Foundation. (2023). National Survey of College Graduates, 2021 (NSF 23-306), Table 1-1. https://ncses.nsf.gov/pubs/nsf23306

Neal, D. (1995). Industry-specific human capital: Evidence from displaced workers. Journal of Labor Economics, 13(4), 653-677.

Staw, B. M. (1976). Knee-deep in the big muddy: A study of escalating commitment to a chosen course of action. Organizational Behavior and Human Performance, 16, 27-44.

Staw, B. M., & Ross, J. (1987). Behavior in escalation situations: Antecedents, prototypes, and solutions. Research in Organizational Behavior, 9, 39-78.

Stevens, R., O'Connor, K., Garrison, L., Jocuns, A., & Amos, D. M. (2008). Becoming an engineer: Toward a three dimensional view of engineering learning. Journal of Engineering Education, 97(3), 355-368.

Vatter, J., & Meuleman, B. (2022). Status inconsistency and subjective social status. Social Forces, 101(1), 150-179.

Reason #14: You're a Custodian, Not an Innovator

You will size the fitting on the heat exchanger. You will not design the heat exchanger. By the time you show up, the motor frame is fixed, the impeller diameter is set, the pipe schedule is locked, and the vendor is chosen. You inherit an assembly that needs bolt patterns shifted, thread callouts corrected, clearance cuts added, and a torque table that nobody agrees on. You will spend a week arguing about a gasket while the thing you are gasketing was designed without you. See Reason #33. This is not a failure of ambition. It is a description of the job.

The money confirms it. The National Science Foundation tracks R&D spending by industry. In 2022, machinery manufacturing, the sector that employs more mechanical engineers than any other manufacturing subsector, spent 3.7 percent of its revenue on research and development. That is below the all-industry average of 4.9 percent, which includes retail, food service, and hospitality. Semiconductors invested at 25.8 percent. Pharmaceuticals at 16.9 percent. Computer and electronic products at 14.3 percent (NSF NCSES, 2022). In total dollars, machinery manufacturing accounts for 3 percent of all U.S. business R&D. Information and software publishing accounts for 25 percent. The industry you trained for invests less in innovation per dollar of revenue than the national average across every industry in the country. See Reason #60.

The people who study engineering services have a word for this. Bain & Company surveyed over 500 senior engineering executives in 2023 and found that companies are "outsourcing legacy disciplines that are often less strategically relevant" for manufacturers. The legacy disciplines they named: mechanical engineering, testing, simulation, and compliance. Service providers have "significantly optimized" these operations. Digital engineering, the territory of electrical and computer science, commands a dominant share of the outsourcing market and is growing at 19 percent annually, nearly double the overall rate (Bain, 2023). The consulting firms that study your field do not call it traditional. They do not call it foundational. They call it legacy. See Reason #40. Patent filings tell the same story. Electrical engineering accounts for roughly 40 percent of all international patent applications filed under the Patent Cooperation Treaty. Computer technology alone is the single largest technology field in global patenting. Mechanical engineering's share has been declining for two decades (WIPO, 2024). The innovation output of your discipline is shrinking in the one metric that measures it directly. See Reason #7.

This is not random. Economists James Utterback and William Abernathy showed in 1975 that once a product reaches a "dominant design," innovation shifts from product creation to process optimization, and engineering work becomes incremental (Utterback & Abernathy, 1975). ME's core technologies, engines, pumps, compressors, heat exchangers, bearings, gears, hydraulic circuits, reached dominant design decades ago. Some reached it centuries ago. The S-curve flattened. Semiconductor devices, software architectures, wireless protocols, and AI models are still climbing. Their engineers create new products. You maintain existing ones. Civil engineers stamp buildings under PE requirements that give them design ownership of every project. See Reason #58. Chemical engineers run processes where optimization IS the innovation, because yield improvements have direct economic value. Your optimization is a cost reduction exercise on a product someone else conceived. See Reason #42 and Reason #51. The compliance layer you manage exists to protect the architecture you did not build. See Reason #65.

You are not a custodian because you chose the wrong employer. You are a custodian because the industry reclassified your entire discipline as custodial.

References:

Bain & Company. (2023). The digital shift fuels outsourcing in engineering and R&D. https://www.bain.com/insights/the-digital-shift-fuels-outsourcing-engineering-r-and-d-report-2023/

National Center for Science and Engineering Statistics. (2022). Business Enterprise Research and Development Survey, Table 3: Sales, R&D, R&D intensity, and employment. NSF 24-334. https://ncses.nsf.gov/pubs/nsf24334

Utterback, J. M., & Abernathy, W. J. (1975). A dynamic model of process and product innovation. Omega, 3(6), 639-656. https://doi.org/10.1016/0305-0483(75)90068-7

World Intellectual Property Organization. (2024). PCT yearly review 2024: The international patent system. https://www.wipo.int/edocs/pubdocs/en/wipo-pub-901-2024-en-patent-cooperation-treaty-yearly-review-2024.pdf

Reason #13: No Guild, No Protection

Degreed Engineers brag about ABET accreditation, then wonder why nothing about their jobs feels protected. ABET audits syllabi, not careers. It inspects capstone rubrics and faculty CVs. It does not police titles in industry, it does not limit program seats, it does not lobby to restrict who does mechanical work inside corporations. Medicine and law fuse accreditation to licensure and scope of practice, so the pipeline narrows on purpose. Engineering hands you an accredited diploma, then sends you into a market that treats you like a replaceable cost.

The problem is bigger than campus. ABET accredits far beyond the United States. That includes massive producer countries where graduating classes dwarf the output of American programs. Your “gold seal” is not a moat, it is a universal stamp that expands the pool everywhere, then invites everyone to the same party. Managers chase lower costs, and your bargaining power sinks.

ABET also accredits non-engineering programs, including mechanical engineering technology at the bachelor and associate levels, often inside the very same institutions that confer mechanical engineering degrees. Employers blur the distinction because their needs are practical: prints, GD&T, ERP, testing, schedule. The technician looks useful on day one. You can guess who gets hired when a team is trying to hit a deadline, see Reason 10.

The result is predictable. Universities keep growing enrollment, because tuition is revenue and accreditation signals legitimacy. Employers avoid training, then demand experience. Entry-level requires three to five years doing the job already, see Reason 12. Wages flatten under a permanent surplus of candidates. When supply rises, your leverage falls, and no guild shows up to say otherwise.

If ABET behaved like a guild, it would tie accreditation to workforce outcomes, limit the pipeline where demand is weak, and enforce scope of practice in the private sector. It does none of that. It checks whether your heat transfer class had measurable learning outcomes, then it walks away. You are left to compete with everyone who has that same stamp, including graduates produced at a scale your market cannot absorb, see Reason 1.

Mechanical engineering feels “open.” In practice, it is unprotected. Your credential certifies that your program existed. It does not certify that your job will.

Reason #12: Entry-Level Requires Experience You Do Not Have

"Entry-level" in mechanical engineering means something the Bureau of Labor Statistics cannot agree with itself about. The BLS Employment Projections table classifies mechanical engineering as requiring no prior work experience and no on-the-job training to enter the occupation (BLS, 2024). The BLS Occupational Requirements Survey, which asks what real employers actually require, found that 54.7 percent of mechanical engineering positions require prior work experience and 78.3 percent require on-the-job training (BLS, 2023). Same agency, two surveys, opposite answers. The projection table describes the credential you need on paper. The employer survey describes what happens when you hand that paper to a hiring manager. See Reason #1.

The gap is not unique to engineering. An analysis of 126 million job postings found that 35 percent of all roles labeled "entry-level" now require three or more years of prior experience (Metaintro/Burning Glass, 2026). But the trend is worse in fields where the work is physical and the training is expensive. A software company can onboard a junior developer with a laptop and access to a repository. If the hire struggles, the cost is recoverable. Your employer needs you on a test rig that costs a half-million dollars, running DV validation on a timeline locked to a supplier's launch schedule. If you struggle, the cost is a missed gate, a delayed launch, and a line manager who never hires unproven again. The infrastructure makes tolerance for ramp-up failure low, which makes demanding prior experience rational from the employer's side and punitive from yours.

That punishment has no workaround. A computer science graduate who cannot get hired can build a GitHub portfolio, contribute to open-source projects, complete a bootcamp, deploy a personal project to the public, and accumulate measurable skill signals that substitute for formal experience. A mechanical engineer who cannot get hired cannot run a PPAP from a laptop. You cannot execute DV/PV testing without calibrated equipment. You cannot demonstrate DFMEA competency without access to a system design under NDA. You cannot route an ECO without a PLM system and a live BOM. The credentialing environment is the physical work environment, and both are controlled by the employer you are trying to convince. There is no side door. See Reason #8.

The entry-level market is contracting, and the Burning Glass Institute's 2025 analysis identifies four structural forces driving the elimination: AI handling entry-level tasks, lean staffing models that persisted after the pandemic, AI amplifying senior worker productivity and reducing the need for junior learning curves, and graduate oversupply saturating the applicant pool (Burning Glass, 2025). The symptom shows up in hiring data: junior-level postings fell 7 percent from August 2024 to August 2025, while senior-level postings edged up (Indeed, 2025). Meanwhile, entry-level software developer postings grew 47 percent over a single year from October 2023 to November 2024 (Lightcast, 2024). The junior tier is not disappearing everywhere. It is disappearing where you are.

The internships that were supposed to solve this are scarce, geographically captive, and excluded from the one growth channel in the market (See Reason #5). The graduate degree that was supposed to substitute for experience costs $200,000 and takes 15 to 26 years to break even (See Reason #19). The industry that demands you arrive trained is also the industry that invests the least in training you. Small manufacturing plants, where 75 percent of ME worksites operate, train less than large ones. Plants with cyclical demand and high contractor use train even less (BLS, 1998). The employers who need you to already know the fixtures, the test rigs, and the ECO workflow are the same employers who will not teach you any of it.

By the time you finally collect the experience that entry-level demanded, you trained yourself on your own time and your own dime. The industry did not invest in you. It filtered you. That is the real function of "entry-level" in mechanical engineering.

References:

Bureau of Labor Statistics. (2023). Occupational requirements survey: 2023 occupational profiles. https://www.bls.gov/ors/factsheet/2023/orsprofiles.htm

Bureau of Labor Statistics. (2024). Table 5.4: Education and training assignments by detailed occupation. https://www.bls.gov/emp/tables/education-and-training-by-occupation.htm

Bureau of Labor Statistics. (1998, June). Results from the 1995 Survey of Employer-Provided Training. Monthly Labor Review. https://www.bls.gov/opub/mlr/1998/06/art1full.pdf

Burning Glass Institute. (2025, July). No country for young grads. https://static1.squarespace.com/static/6197797102be715f55c0e0a1/t/68fa5bb9f727046443900bc4/1761237945960/No+Country+for+Young+Grads+V_Final7.29.25+(1).pdf

Indeed Hiring Lab. (2025, September). The labor market squeeze on new entrants. https://www.hiringlab.org/2025/09/25/september-labor-market-squeeze-on-new-entrants/

Lightcast/Tech Elevator. (2024, December). 2024 job growth: Rising demand grows for software developers. https://www.techelevator.com/2024-job-growth-rising-demand-grows-for-software-developers/

Metaintro. (2026, January). Gen Z faces entry-level job paradox. https://www.metaintro.com/blog/gen-z-faces-entry-level-job-paradox