The Comparison Malaysia Isn’t Part Of

labour market
malaysia
southeast asia
education
Author

the haddock

Published

September 2, 2026

DOSM has tracked skill-related underemployment (SRU); tertiary-educated workers stuck in semi-skilled or low-skilled jobs, every quarter since 2020. It’s a solid, regularly updated number. What it doesn’t come with is any easy way to see where Malaysia sits next to its neighbours.

That comparison should exist. The International Labour Organization runs a database built for exactly this: the Education and Mismatch Indicators (EMI) framework, which classifies workers as over-educated for their job using the same normative logic comparing a worker’s education against what their occupation actually requires that DOSM itself uses for SRU. If two countries both feed this framework, the numbers are supposed to mean the same thing.

Vietnam, Indonesia, and the Philippines feed it. Malaysia doesn’t.

What the data shows

Using ILOSTAT’s EMI indicator, 2023 the most recent year all three countries report the share of tertiary-educated workers over-qualified for their jobs:

data = FileAttachment("data/clean.csv").csv({typed: true})

bars = data.filter(d => d.role === "bar").sort((a, b) => d3.ascending(a.rate, b.rate))
malaysia2023 = data.find(d => d.role === "callout")
malaysiaCurrent = data.find(d => d.role === "callout_current")
Plot.plot({
  marginLeft: 100,
  width: 640,
  height: 260,
  x: { label: "Over-educated, tertiary-educated workers (%)", domain: [0, 80] },
  y: { label: null },
  color: { range: ["#14264A"] },
  marks: [
    Plot.barX(bars, {
      x: "rate",
      y: "country",
      fill: "#14264A",
      sort: { y: "x", reverse: true }
    }),
    Plot.text(bars, {
      x: "rate",
      y: "country",
      text: (d) => `${d.rate}%`,
      dx: 8,
      textAnchor: "start",
      fontFamily: "IBM Plex Mono",
      fontWeight: 600
    })
  ]
})
html`<div style="
  border: 2px solid #C1401C;
  background: #FBF9F3;
  border-radius: 8px;
  padding: 1.25rem 1.5rem;
  margin: 1.5rem 0;
  font-family: Inter, sans-serif;
">
  <div style="font-family: 'IBM Plex Mono', monospace; font-size: 0.8rem; letter-spacing: 0.05em; color: #C1401C; text-transform: uppercase; margin-bottom: 0.4rem;">
    Not in the ILOSTAT comparison
  </div>
  <div style="font-size: 2rem; font-weight: 700; color: #14264A; font-family: 'IBM Plex Mono', monospace;">
    ${malaysia2023.rate}%
  </div>
  <div style="font-size: 0.9rem; color: #14264A;">
    Malaysia, skill-related underemployment, 2023 — DOSM, own methodology
  </div>
  <div style="font-size: 0.8rem; color: #14264A; opacity: 0.7; margin-top: 0.5rem;">
    Most recent available quarter: ${malaysiaCurrent.rate}% (${malaysiaCurrent.note.match(/\((.*?)\)/)[1]})
  </div>
</div>`

Read the four numbers side by side and Malaysia’s SRU rate is the lowest of the four but that comparison comes with an asterisk large enough to be the actual point of this piece, which the next two sections are about.

Why Malaysia isn’t a fourth bar

DOSM’s SRU methodology is conceptually the same approach ILO uses for the EMI indicator: both compare a worker’s education level against what their occupation is classified as requiring, and flag the ones sitting above that line. The difference is in who does the classifying. ILOSTAT’s figures for Vietnam, Indonesia, and the Philippines run through ILO’s own harmonised ISCO occupational mapping, applied consistently across every country in the database. Malaysia’s SRU runs through MASCO — DOSM’s national occupational classification — processed entirely within Malaysia’s own statistical system, never submitted into ILO’s cross-country pipeline.

Same underlying logic. Different classification system, different institution doing the coding, no shared audit trail between them. That’s not the same as saying the numbers are meaningless next to each other — but it’s also not nothing, and folding Malaysia into the same bar chart as the other three would quietly erase a real methodological gap that a reader has no way to see on their own.

The comparison itself is uneven

Even setting Malaysia aside, the three countries that are in ILOSTAT’s framework don’t report on the same schedule. Vietnam’s data runs almost continuously from 2016 to 2024. Indonesia has 2016, 2017, 2019, 2022, and 2023; gaps at 2018, 2020, and 2021. The Philippines has only 2017, 2021, 2022, and 2023, and its 2021 figure carries ILO’s own “break in series” flag, meaning something changed in the underlying survey methodology that year that ILO doesn’t consider fully comparable to the years around it.

2023 is simply the most recent year all three happen to share. It’s not the most recent year for any of them individually, and it’s not a year chosen because it tells a particular story; it’s the year the data allowed.

What this adds up to

An international framework exists that’s specifically designed to make this comparison possible. Three of Malaysia’s neighbours use it. Malaysia doesn’t; not because the underlying concept doesn’t apply to it (DOSM’s own SRU series shows Malaysia measures essentially the same thing, has for years, and updates it quarterly) but because that measurement was never routed into the system that would let it sit next to anyone else’s.

Whether that’s a data-sharing gap worth closing is a separate question from what this piece can answer. What the data does show is where the comparison stands today: three countries in one framework, one country tracking the same idea on its own.

Data & sources

  • ILOSTAT, Education and Mismatch Indicators (EMI) database — indicator “Employment by sex, educational mismatch - normative approach and education,” annual, accessed via the ILOSTAT Data Explorer on 28 August 2026. Covers Viet Nam, Indonesia, and the Philippines; Malaysia is not present in this indicator.
  • DOSM, Quarterly Skills-Related Underemployment by Sex (lfs_qtr_sru_sex) — OpenDOSM data catalogue, licensed CC BY 4.0. Used for Malaysia’s figure.