Where Malaysia's Workforce Actually Is — and What It Costs by State

The first idea every founder brings to Malaysian site selection is that moving out of the Klang Valley cuts the wage bill. The data says it does — and the law says most of the saving is not yours to take, because the floor and every contribution above it are set federally.

Both halves of that are true at once. Here is where the line falls.

What does labour actually cost by state?

DOSM’s Salaries and Wages Survey for reference year 2024, covering Malaysian citizens only, is the only official state-level wage series that exists:

Read the column definition before you use it. This is salaries and wages — employees only. It excludes the self-employed, employers and unpaid family workers, so it is not household income and not a measure of local prosperity. For Malaysia the median monthly salary is RM2,793 while median household income is RM7,017; the two answer different questions and must never be put in one table.

State	Median monthly salary	Mean	Recipients (‘000)
W.P. Putrajaya	RM4,598	RM5,091	38.5
W.P. Kuala Lumpur	RM3,687	RM4,782	752.9
Selangor	RM3,350	RM4,052	2,610.5
Pulau Pinang	RM2,934	RM3,787	642.2
W.P. Labuan	RM2,931	RM3,812	32.3
Johor	RM2,582	RM3,414	1,388.0
Melaka	RM2,528	RM3,469	350.9
Negeri Sembilan	RM2,503	RM3,626	356.6
Pahang	RM2,491	RM3,337	465.6
Sarawak	RM2,490	RM3,424	771.7
Sabah	RM2,236	RM3,389	730.0
Kedah	RM2,051	RM3,053	569.3
Perak	RM2,050	RM3,172	735.7
Perlis	RM2,038	RM3,257	80.0
Terengganu	RM2,021	RM3,107	320.3
Kelantan	RM1,914	RM3,091	393.8
Malaysia	RM2,793	RM3,652	10,238.4

Putrajaya’s median is 2.4 times Kelantan’s. That is a real spread and it is worth understanding — but read the mean column beside it. Kelantan’s mean of RM3,091 is 61 per cent above its median, which is the signature of a small professional layer sitting on top of a large low-wage base. You are not hiring the median. You are hiring from the thin top of a thin distribution.

Penang sits at RM2,934, roughly 14 per cent below Selangor and 5 per cent above Labuan — a much narrower gap than the cluster’s reputation suggests.

Where are the people?

Absolute labour force by state, from DOSM’s quarterly report, persons in thousands:

State	Labour force Q1 2026	Employed Q1 2026	Unemployment rate Q1 2026
Selangor	4,136.4	4,053.9	2.0%
Johor	2,185.0	2,132.9	2.4%
Sabah	1,885.5	1,782.2	5.5%
Sarawak	1,262.1	1,219.8	3.4%
Perak	1,233.0	1,190.6	3.4%
W.P. Kuala Lumpur	1,210.4	1,169.6	3.4%
Kedah	999.0	974.9	2.4%
Pulau Pinang	947.8	922.3	2.7%
Malaysia	17,232.0	16,725.6	2.9%

Selangor and Kuala Lumpur together hold 5.35 million workers — 31 per cent of the national labour force in two adjoining jurisdictions. Penang, for all its industrial weight, has 948,000, smaller than Kedah next door.

On the annual measure, participation was highest in Putrajaya at 79.0 per cent, Selangor 78.4, Kuala Lumpur 75.8, Penang 72.2 and Johor 72.0, against a national 70.8 per cent — the highest DOSM has ever recorded. Unemployment was lowest in Putrajaya at 1.1 per cent, Melaka 1.6 and Pahang 1.8, and highest in Sabah at 7.2 per cent, Labuan 6.3, Perlis 3.9, and Kelantan and Perak both at 3.5.

There is no single 2025 state unemployment rate, and you should stop looking for one. DOSM’s annual report and its own revised quarterly series do not reconcile: Sabah is 7.2 per cent on the annual report against 6.3 per cent averaged from the revised quarters, Kelantan 3.5 against 4.1, Selangor’s participation 78.4 against 77.4. Two official DOSM publications, two answers, no explanation in either. Always name the vintage — 7.2 per cent, Labour Force Survey Report 2025 or 5.5 per cent, Q1 2026 — and never multiply an annual rate by a quarterly level.

Two further traps if you go to the source yourself. There is no state-level annual open dataset at all — the annual state figures exist only inside chart images in the report PDF. And DOSM’s open-data CSVs are not safe substitutes for the publications: labour/lfs_qtr_state.csv is a stale unrevised vintage that stops at Q3 2025 and contradicts the published revised series, and it carries an unflagged 2024 Q1 break from the Census 2020 rebasing. Cite the PDF or the XLSX.

Why is the statutory saving so much smaller than the wage gap?

Because the floor and the on-costs are federal.

The Minimum Wages Order 2024, P.U.(A) 376, sets RM1,700 a month and RM8.72 an hour in every state. The only carve-out that ever existed, a deferment to RM1,500 for employers with fewer than five employees, ran from 1 February to 31 July 2025 and expired; paragraph 5 brought everyone to RM1,700 from 1 August 2025. No 2026 Order has been made.

Above the floor, the add-ons are identical everywhere: EPF at 13 per cent of wages up to RM5,000 and 12 per cent above, SOCSO at 1.75 per cent to a RM6,000 ceiling, EIS at 0.2 per cent on the same ceiling, and the HRD Corp levy at 1 per cent for employers with ten or more staff.

One federal rate does discriminate — by citizenship, not geography. The HRD Corp levy is payable on Malaysian citizens only, and EPF for non-citizens registered from 1 August 1998 is 2 per cent rather than 13. A team’s blended statutory cost moves far more with its citizenship mix than with its postcode.

So the arithmetic is this. In Kelantan the median wage is RM879 below Selangor’s, but a large share of your roles will sit at or near a floor that is the same in both, and the entire on-cost stack is the same in both. The differential is real in the middle of the distribution and vanishes at the bottom.

Which clusters actually exist?

Penang and Kulim: electrical and electronics. Manufacturing is 47.3 per cent of Penang’s economy — the highest manufacturing share of any Malaysian state — and the state’s economy grew 7.3 per cent in 2025 to RM130.3 billion, second-fastest nationally. MIDA recorded RM11.3 billion of approved E&E investment in Penang across 83 projects in 2025, carrying 12,798 employment opportunities. Nationally the electrical, electronic and optical products sector holds 658,500 jobs with 34,900 vacancies unfilled. No official figure exists for the stock of E&E workers in Penang, which is a striking gap given how central the claim is to every Penang pitch.

Klang Valley: services and head offices. Selangor contributed 26.5 per cent of national GDP in 2025 at RM460.1 billion, Kuala Lumpur RM265.1 billion. Services are 66.0 per cent of national employment. Kuala Lumpur also has the country’s lowest skills-related underemployment at 22.8 per cent against a national 35.6 — meaning graduates there are more likely to be doing graduate work.

Johor: capital without commensurate headcount. Johor grew 8.0 per cent in 2025, fastest in the country, and took RM110.03 billion of approved investment — about 26 per cent of the national total, and first place. But the same MIDA appendix records only 24,584 potential jobs from those projects, roughly 10 per cent of the national 244,902. Kuala Lumpur, with RM63.3 billion, generated 82,985. That gap is the data-centre and heavy-capital profile of the JS-SEZ, which alone accounted for RM77.0 billion across 393 projects and 18,406 jobs in 2025. Johor is winning capital faster than it is winning employment, and a labour plan built on the investment headlines will overestimate the pool being created.

What about graduates and foreign workers?

Malaysia produced 332,871 graduates in 2025 — but that headline includes certificates and foundation completions. The degree figure is 143,241 bachelors, or 187,301 at bachelor level and above. By field: business, administration and law 94,891; engineering, manufacturing and construction 54,729; ICT 30,685. Employability is 92.3 per cent overall and 95.7 per cent for engineering, the highest of any field.

Output by state is not published. Enrolment by campus state is, and it is concentrated: Selangor 479,992, Kuala Lumpur 200,840, Perak 90,969, Johor 89,360, Penang 63,622.

Non-citizens were 3.38 million, or 9.9 per cent of the population, in 2025, and the distribution is not what most people expect: Sabah 1.04 million, Selangor 727,900, Johor 385,100, Kuala Lumpur 270,700, Penang 179,900. Note this counts non-citizens, not permit holders — dependants, students and permanent residents are inside it — and no official count of documented foreign workers by state exists.

What changes in Sabah and Sarawak?

Separate law, not a variation. Sabah applies the Labour Ordinance Cap. 67 and Sarawak Cap. 76, both amended by Acts A1753 and A1754 with effect from 1 May 2025 except Part IVA. Sarawak’s replacement First Schedule keeps the RM4,000 threshold but disapplies the section 2 definitions of normal hours of work and overtime along with several other provisions, and adds new Second and Third Schedules with no Peninsular equivalent. There is no section-for-section mapping.

Sarawak also runs its own work authorisations through GENESIS, covering non-Sarawakians including Malaysians from the Peninsula. And for foreign hires, Sabah keeps sector-based MYFutureJobs advertising periods — 7 days for agriculture, plantation and construction, 14 for services, 30 for manufacturing — where the expatriate period elsewhere has been 7 days since 1 January 2025. Exemption from advertising is never exemption from approval: s.60K of the Employment Act 1955 requires the Director General’s prior consent for every foreign employee, on penalty of RM100,000 or five years.

Common mistakes

Treating the median salary gap as a statutory saving. The floor is RM1,700 everywhere and every on-cost percentage is federal.

Mixing DOSM’s annual and quarterly 2025 series. They disagree materially on Sabah, Labuan, Kelantan, Putrajaya and Selangor. Applying an annual rate to a quarterly level produces a number neither publication supports.

Pulling state comparisons from DOSM’s open-data CSVs. They parse cleanly and return plausible numbers, which is exactly the problem. The quarterly state labour file is a stale unrevised vintage, and the household income file’s 2024 Gini column is wrong for five states including Sabah, which it shows as 0.405 against a published 0.389. Use the publication PDF or XLSX.

Treating median salary as household income. Salaries and wages covers employees only. Malaysia’s median monthly salary is RM2,793 and its median household income is RM7,017 — a table that mixes the two is measuring nothing.

Reading Johor’s investment lead as a hiring lead. Twenty-six per cent of national capital, ten per cent of national potential employment.

Assuming a Peninsular HR policy travels east. Sarawak’s amended First Schedule removes definitions the Employment Act depends on, and a transfer to Kuching goes through GENESIS.

Quoting an E&E headcount for Penang. No government body publishes one. What is published is investment, exports and approved-project employment — all flows, none of them a stock.

What’s next

Shortlist on the pool, not the price: Penang and Kulim for electronics engineering, the Klang Valley for services depth and the lowest graduate underemployment in the country, Johor for manufacturing with an international wage anchor. Then price the site with business costs by region, check you can move goods with infrastructure and connectivity by region, and find out who can actually approve an incentive in state investment agencies.