Analysis of official statistics · Edition 1 · August 2026
The underemployment paradox: 1.96 million graduates, working below their qualifications.
Malaysia has near-record-low graduate unemployment and record graduate underemployment, at the same time. Employers report they cannot fill digital roles while a third of employed graduates work beneath their qualifications. Both things are true, and understanding why they coexist changes what an individual graduate should actually do about it.
Every figure below comes from an official or reported source, numbered and linked. Free to read, cite and republish under CC BY 4.0.
1.96m
Malaysian graduates working below their qualifications (Q3 2025) - 35.5% of employed degree and diploma holders2
3.2%
Graduate unemployment rate (2024) - near a decade low. This is not a jobs shortage1
66.0%
Skill-related underemployment among graduates aged 24 and under (2024)1
21,815
High-value digital jobs approved in a single quarter (Q3 2025), 38% of them AI-related3
Finding 01
It is not an unemployment problem.
The most common framing of the Malaysian graduate situation - that graduates cannot find work - is contradicted by the official figures. Graduate unemployment fell to 3.2% in 2024, down from 3.4%, with 165,900 unemployed graduates out of a graduate labour force of 5.14 million.1 That is close to a decade low, and the graduate labour force participation rate is high at 86.0%.1
What has risen instead is skill-related underemployment: graduates who are employed, but in low- or semi-skilled roles that do not require the qualification they hold. In 2024 that was 1.60 million people, or 32.2% of employed graduates1, and by Q3 2025 the figure had risen to 1.96 million, or 35.5% of employed degree and diploma holders.2
The distinction matters because it changes the prescription. If the problem were a shortage of jobs, the answer would be to apply more widely and wait for the cycle to turn. It is not a shortage of jobs. Roughly one in three employed Malaysian graduates is in a role their qualification did not require - which is a mismatch between what people were trained in and what the economy is paying for.
Finding 02
Two-thirds of graduates under 25 are working below their qualification.
Skill-related underemployment rate by age group, 2024.
- Aged 24 and under66%
- Aged 45 and above18.6%
66.0% of employed graduates aged 24 and under were working below their qualifications in 2024, against 18.6% of those aged 45 and above.1 DOSM attributes the young-graduate figure to recent graduates accepting work below their qualifications in order to gain experience and strengthen their position in the labour market.1
There is an optimistic reading of that gradient - it resolves with age, and the 45-plus figure suggests most people do eventually reach roles that match their training. There is also a less comfortable one: two-thirds of young Malaysian graduates begin their working lives in jobs their degree did not require, and early-career position compounds. Someone who spends three years in a role beneath their qualification is not only earning less during those years; they are accumulating experience that the market prices accordingly.
Finding 03
Meanwhile, 21,815 high-value digital jobs were approved in one quarter.
In the third quarter of 2025 alone, MDEC approved RM54.13 billion of Malaysia Digital investments, projected to create 21,815 high-value jobs across 402 digital companies.3 Of those roles, 8,328 - some 38% - were AI-related, spanning digital professionals, data scientists, AI engineers and specialised service talent.3 Across the full year, approved digital investments reached roughly US$22 billion.4
The concentration is worth noting for anyone planning around these figures: Kuala Lumpur and Selangor accounted for 88% of that investment (RM47.8 billion) and close to 90% of the digital jobs.3 This is a Klang Valley phenomenon first, which matters given that much of the underemployed graduate population is not there - and which is part of why remote work features so heavily in what career-changers say they want.
A caveat worth stating plainly, because investment headlines routinely overstate employment: data centres in particular are capital-intensive rather than labour-intensive. The SCMP has reported concerns that the promise of job creation in that sector may be illusory, and the Johor Talent Development Council’s data centre technician programme launched to fill 200 vacancies at RM3,500 to RM4,000 a month.5 The larger and better-paid employment effect from digital investment shows up in software, AI and services roles rather than in the facilities themselves - which is also where the skills bar is highest.
Finding 04
Malaysia sits mid-table internationally - better than the region, behind advanced economies.
Graduate skill-related underemployment rate, as compiled by DOSM (2024).
- United States27.3%
- Australia27.5%
- United Kingdom28.5%
- Malaysia32.2%
- South Korea36.8%
- Indonesia42.2%
- Philippines56.7%
At 32.2%, Malaysia sits ahead of South Korea (36.8%), Indonesia (42.2%) and the Philippines (56.7%), and behind the United States (27.3%), Australia (27.5%) and the United Kingdom (28.5%).1 DOSM frames the position as evidence of progress with room to improve.
The comparison is genuinely useful context - this is a regional pattern rather than a uniquely Malaysian failure, and the gap to advanced economies is around five percentage points rather than a chasm. It is also, for an individual, beside the point. Whether a third of your peers are underemployed because of a national structural issue or a regional one does not change what you personally can do about your own position.
Interpretation
What the paradox actually tells an individual.
Put the two halves together. Nearly two million Malaysian graduates are working below their qualifications, while tens of thousands of high-value digital roles are being approved every quarter. Those are not contradictory facts about one labour market; they are facts about two different populations that the education pipeline has not connected.
The practical implication is narrower than the scale of the numbers suggests. If unemployment were the constraint, more applications would be the answer. Since the constraint is mismatch, the lever is which skills you hold, not how hard you look. And the figures point at where the unmet demand is concentrated: software, AI and data roles account for the bulk of the high-value digital employment being created, with AI alone at 38% of projected roles in the most recent quarter.3
Our interest here is obvious and worth stating: we are a school that teaches exactly those skills, so treat our framing accordingly. What is not our framing is the underlying data - the underemployment figures are DOSM’s, the investment and job figures are MDEC’s via BERNAMA, and the caveat about data centre employment is the SCMP’s. Every one of them is linked below so you can read them without us in the middle.
If you are in the 1.96 million, the companion pieces are our 90-day plan for graduates and how to check whether you are underpaid. For what the roles on the other side of the gap pay, see our hiring and salary report; for who is actually making the move across it, see our composition study.
Sources
Method and references.
Method. This report contains no primary data collection. Every figure is drawn from official statistics or reported coverage of them, and each is numbered inline and listed below. Where two sources give different figures for the same measure - as with the 2024 annual underemployment figure and the Q3 2025 quarterly one - both are given with their period stated, rather than blended. Percentages are reproduced as published.
Definitions.“Skill-related underemployment” is DOSM’s term for people employed in low- or semi-skilled occupations despite holding a degree or diploma; it is distinct from unemployment and from time-related underemployment (working fewer than 30 hours for want of available work, which stood at 41.7 thousand graduates, or 0.8%, in 2024).1 Graduates are defined by DOSM as holders of a certificate from a university, college, polytechnic or recognised body with a study duration of at least two years.1
Licence and corrections. Our analysis is released under CC BY 4.0 ; the underlying statistics remain the property of their publishers and should be cited to them directly. If you find an error, write to support@sigmaschool.co and a dated correction will be published here.
- 1
Department of Statistics Malaysia (DOSM)
Graduates Statistics 2024 · Released 31 October 2025
View source - 2
DOSM, via The Edge Malaysia
Malaysia’s unemployment steady at 3% in 3Q2025 but over one-third of graduates remain underemployed · Q3 2025 labour force data
View source - 3
MDEC, via BERNAMA
Malaysia Secures RM54.13 Bln Approved Digital Investments In 3Q 2025 · Q3 2025
View source - 4
Digital News Asia
MDEC secures US$22 billion in digital investments in 2025, driven by AI · Full-year 2025
View source - 5
South China Morning Post
Data centres are booming in Malaysia’s Johor. Now comes the hard part: job creation · 2025
View source
Questions
Frequently asked.
How many Malaysian graduates are underemployed?
According to the Department of Statistics Malaysia, 1.60 million graduates were employed below their qualifications in 2024 - a skill-related underemployment rate of 32.2%. More recent labour force data for Q3 2025 puts the figure at 1.96 million, or 35.5% of employed degree and diploma holders. Both figures count people who are employed, but in low- or semi-skilled positions that do not match the qualification they hold.
Is graduate unemployment a big problem in Malaysia?
No, and this is the most misread part of the picture. Graduate unemployment fell to 3.2% in 2024, down from 3.4% - near a decade low, with 165,900 unemployed graduates. Malaysia does not have a shortage of jobs for graduates. What it has is a mismatch: roughly a third of employed graduates are in roles beneath their qualification level. Framing this as an unemployment crisis leads to the wrong response, which is usually "apply to more jobs" rather than "become qualified for different ones".
Why is graduate underemployment so much worse for young Malaysians?
The age gradient is stark. In 2024, graduates aged 24 and under recorded a skill-related underemployment rate of 66.0%, against 18.6% for those aged 45 and above. DOSM attributes the young-graduate figure to recent graduates accepting work below their qualifications in order to gain experience. The optimistic reading is that it resolves with time. The less comfortable reading is that two-thirds of young Malaysian graduates begin their careers in roles their degree did not require, which is a slow start to compound from.
If graduates are underemployed, why do Malaysian employers say they cannot find talent?
Because both are true at once, and they are describing different populations. In Q3 2025 alone, MDEC approved RM54.13 billion of Malaysia Digital investments projected to create 21,815 high-value jobs across 402 companies, with 8,328 of those roles - 38% - AI-related. Across the full year, approved digital investments reached roughly US$22 billion. The vacancies are real and the underemployed graduates are real; what connects them is a skills gap rather than a jobs gap. A graduate qualified in a field with no shortage cannot fill a vacancy requiring skills they were never taught.
Are data centre and AI investments actually creating jobs for ordinary Malaysians?
Partly, and it is worth being sceptical of the headline numbers. Data centres are capital-intensive rather than labour-intensive: the SCMP has reported concerns that the promise of job creation in the sector "may be illusory", and the Johor Talent Development Council's data centre technician programme launched to fill 200 vacancies at RM3,500 to RM4,000 a month. The larger employment effect from digital investment appears in software, AI and services roles rather than in the facilities themselves - which is also where the skills requirement is highest.
What actually closes the gap for an individual graduate?
The data does not prescribe a solution, but it does narrow the question. Since unemployment is low and underemployment is high, the binding constraint is not access to work - it is holding skills that a shortage exists for. That points away from generic further study and toward demonstrable capability in a field with documented unmet demand, of which software, AI and data are the clearest current examples in the Malaysian figures. Our companion report on developer salaries covers what those roles pay, and our composition study covers who is actually making that move.
The constraint is which skills, not how hard you look.
That is what the two halves of this data mean together.
If you are one of the 1.96 million, the cheapest possible test of the other side is a week of actually building - six real projects, a live instructor session, one signup, no card.