Tucked into Australia’s 2020 university funding reforms was something of an absurdity. The government decided to make humanities degrees 113 per cent more expensive for students whilst cutting fees for nursing, teaching and STEM subjects by up to 62 per cent. The rationale was to steer youngsters toward fields with “the best employment outcomes”. Never mind that the data underpinning this logic was rubbish. Three years after graduation, humanities graduates in Australia earned more (70,300 Australian dolllars) than maths and science graduates (68,900 Australian dolllars) and enjoyed a 91.1 per cent employment rate—higher than their science and maths colleagues (90.1 per cent).
Back in Europe, the received wisdom about university degrees—whether to pursue STEM (science, technology, engineering, and mathematics) versus humanities or practical versus perceived frivolous subjects—has become widely accepted. Choosing the wrong degree can lead to a lifetime of underemployment, while selecting a STEM field, especially one involving computers, seems to promise financial security.
However, evidence suggests that this perspective is not only overly simplistic but also potentially misleading.The real trap facing students isn’t choosing the ‘wrong’ degree. It’s believing that any single credential will insulate them from a labour market that increasingly values what degrees cannot teach.
The non-existent STEM shortage
Start with the supposed STEM shortage. Industry leaders have spent decades lamenting insufficient graduates in science, technology, engineering and mathematics. The European Commission commissioned studies in 2015 to investigate whether Europe needed more STEM graduates. The answer? At an aggregate EU level, no. Projections indicated that yearly STEM graduate supply would meet overall expansion and replacement demand through 2025. Regional shortages existed, certainly, particularly in engineering and ICT. But these stemmed not from low graduation rates but from “under-employment of mobile STEM graduates, barriers to transition of recent graduates to the labour market, and a large share of graduates ending up in non-core STEM jobs.”
That final point bears emphasis. According to US Department of Commerce data, there were 11.9 million workers with STEM degrees but only nine million in STEM occupations—meaning roughly two-thirds of STEM degree holders work in non-STEM jobs. A 2018 study by the Universities of Leicester and Warwick found that most science graduates choose not to—or cannot—work in highly skilled STEM occupations at any point in their careers. “While the majority of engineering graduates worked in these kinds of occupations, a relatively small number of biological science graduates were employed in these roles,” the researchers noted.
The pattern repeats across disciplines. In the United States, six years after finishing their degrees, 57 per cent of computer science graduates work as programmers; at 20 years, when most are in their early 40s, the figure falls sharply to 19 per cent. Civil engineers fare better, with 52 per cent still in their field after two decades. This suggests that for many STEM fields, the premium students pay in effort during university buys them perhaps a decade of directly relevant work before they pivot elsewhere.
Oh the humanities!
Meanwhile, the supposedly unemployable humanities graduate soldiers on rather better than advertised. The unemployment rate for humanities majors in 2021 stood at 5.2 per cent for terminal bachelor’s degree holders—higher than the 4.3 per cent average for all degree holders, but considerably better than the 7.1 per cent for those who completed only secondary education. For humanities graduates with advanced degrees, unemployment fell to 3.3 per cent.
American research shows English majors earning 53,000 US dollars annually, only 1,000 US dollars less than molecular biology majors and 5,000 US dollars more than neuroscience majors. Unsurprisingly, the highest salaries remain in engineering. But European data from the 2018 Eurograduate survey reveals humanities graduates contributing substantially to legal, social and cultural occupations whilst demonstrating higher rates of civic engagement and volunteering than their peers.
The employment gap narrows further over time. Australian census data shows 94.7 per cent of humanities and social science graduates aged 25-34 employed, essentially identical to the 94.4 per cent for science graduates. Graduate outcomes data indicate that full-time employment rates for humanities graduates align with other disciplines three years post-graduation, “suggesting that initial disparities ease with time spent in the workforce.”
The false dichotomy
The obsession with STEM versus humanities rests on a category error. It assumes that students acquire specific knowledge in university and then deploy it throughout their careers. But modern careers rarely work that way. McKinsey’s research on generative AI’s impact finds that automation will affect up to 30 per cent of hours currently worked across the US economy by 2030. Yet “we see generative AI enhancing the way STEM, creative, and business and legal professionals work rather than eliminating a significant number of jobs outright.”
Automation’s biggest effects will hit routine roles: office support, customer service, food service. The positions with staying power? Those requiring what researchers increasingly call ‘durable skills’—communication, critical thinking, collaboration, ethical judgment. Eight of the top ten most-requested skills in US job postings are these human capabilities. Sixty-six per cent of all tasks in 2030 will still require human skills or human-technology combinations. Thirty-nine per cent of key job skills are expected to change by 2030, with 59 per cent of workers requiring upskilling or reskilling.
This undermines the premise that choosing the ‘right’ degree provides insurance. No credential protects against obsolescence when the skills landscape shifts so rapidly. What matters is meta-competence: the ability to learn, adapt, synthesise information across domains, and communicate complex ideas to diverse audiences. Humanities degrees explicitly train these capabilities. STEM degrees do only sometimes.
The Turkish warning
Over the past two decades Türkiye has expanded the number of public universities in from 53 in 2003 to 204. The result has made Türkiye the only European country where university graduates face higher unemployment (9.2 per cent) than the overall population (8.8 per cent). “The quantity of tertiary-level graduates has grown faster than the number of jobs that require a degree,” says the Organisation for Economic Co-operation and Development (OECD). “This has eroded the previous benefits of having a university degree.”
Europe is edging toward a similar trap. Ireland’s 2023 graduate outcomes show arts and humanities graduates with the lowest employment rate (64.5 per cent) nine months post-graduation—but also the highest proportion (21 per cent) pursuing further study. Education graduates top employment at 90.3 per cent. Yet 53.8 per cent of graduates willing to share salary data earnt under 40,000 euros. The system produces degree holders faster than it creates graduate-level work.
What, then, should students study?
Any honest answer depends less on subject than on how it is studied. Engineering graduates do well because they typically learn structured problem-solving, project management and collaboration alongside technical knowledge. Philosophy graduates who can write clearly and construct rigorous arguments fare better than those who merely regurgitate theory. The Harvard career service advises humanities students to frame their skills explicitly: “Critical thinking and analytical reasoning—draw parallels between your studies, projects, and research papers with job responsibilities and qualifications.”
The ideal probably combines elements of both. Technology graduates who can write persuasively possess rare advantages. Humanities graduates with quantitative literacy stand out. The most valuable students master both technical competence and human judgment—what one analyst calls “the ability to solve technical problems in human ways.”
Students should certainly consider employment prospects. But fixating on first-year placement rates misses the point. Research comparing 18 countries found that “wider economic issues, such as the employment rate of graduates in subjects like humanities, are key predictors” of labour market outcomes. The broader economy matters more than credentials.
Perhaps the better question isn’t ‘What should I study?’ but ‘How do I cultivate adaptability?’ The answer involves intellectual curiosity, comfort with ambiguity, and practice synthesising information across domains—qualities neither STEM nor humanities monopolise. Universities teaching students to think rather than merely training them for specific jobs serve them better over careers spanning 40-50 years.
Governments might note that when they manipulate prices to engineer student choices, they often engineer the wrong choices. Australia’s 2020 reforms assumed humanities graduates underperform economically. Data showed the opposite. The market for skills evolves too rapidly for central planning. Better to trust students to balance passion, aptitude and opportunity—whilst ensuring universities actually teach the adaptability everyone claims to value.
Photo: Dreamstime.
