Liabooks Home|PRISM News
Editorial cartoon. A heap of faceless wooden peg figurines sits on a wooden table while two hands in navy suit sleeves sweep them into two piles. The left pile is all cream figurines; in the middle of the right pile stands one cluster of red figurines. From the bottom, a hand in a blue sleeve holds a magnifying glass over the red cluster, and a factory topped by a large round wafer is faintly sketched in the distance.
EconomyAI Analysis

Korea Youth Jobs (15–29) Down 196K; Population ~100K

19 min read

Korea's employed youth (15–29) fell 196,000 in October. PRISM's split: ~100,000 population decline, ~90,000 lower employment ratios, mostly ages 20–24.

South Korea had 196,000 fewer employed young people (ages 15–29) this October than a year earlier. Over the same period, the youth population shrank by 141,000. Those are the youth figures in the October 2026 Employment Trends report, released on November 11 by the Ministry of Data and Statistics (MODS), Korea's national statistics agency. Put side by side, the two numbers read as if jobs were disappearing faster than people.

You can't simply subtract one from the other, though. The youth employment-to-population ratio, meaning the share of the population that has a job, fell from 44.6% a year earlier to 42.9%. That's roughly 43 employed out of every 100 young people. A shrinking population does pull employment down, but only by the population drop multiplied by the employment ratio.

So how much of the 196,000 decline comes from there simply being fewer young people, and which age groups account for the rest? PRISM split the figure using the five-year age-group table in the same release. About 100,000 is the share from population decline. The remaining roughly 90,000 is the share from lower employment ratios within each age group, and about 80,000 of that came from people aged 20–24 (PRISM calculation).

October's numbers alongside the months before

The table below carries only the youth figures from the release. All changes are year-on-year, against the same month a year earlier.

MonthChange in employed young peopleChange in youth populationYouth employment ratio (year earlier → that month)
Oct 2025−163,000−184,00044.6% that month
Jun 2026−197,000−147,00045.6% → 43.9%
Jul 2026−191,000−144,00045.8% → 44.2%
Aug 2026−143,000−144,00045.1% → 44.1%
Sep 2026−147,000−143,00045.1% → 44.0%
Oct 2026−196,000−141,00044.6% → 42.9%

Table sources: MODS October Employment Trends, "Employment-to-population ratio trends" table. Year-earlier employment ratios are from the "Employed persons and employment-to-population ratio by age group" table in the June, July, August and September Employment Trends. The June population change is the value in the September release's "Employment-to-population ratio trends" table.

Two things stand out. First, the gap between the fall in employment and the fall in population varies from month to month. In August and September the two were nearly identical; in June, July and October, employment fell by more. Second, in October last year the fall in employment (163,000) was smaller than the fall in population (184,000). This article doesn't break that October 2025 decline down by age group. Doing so would require the age-group table for October 2024.

In the same month, Korea's semiconductor exports rose 259% from a year earlier, to 3.6 times their level in October 2025 (Ministry of Trade, Industry and Resources, October Export and Import Trends; multiple calculated by PRISM). This article places the export figures and the youth employment figures side by side only as numbers from the same period.

Below, we split the 196,000 into the share from population decline and the share from age-group employment ratios. Then we look at how much manufacturing job figures can tell us about young people, and at whether the line "a shrinking population makes it easier to get hired" holds up for a 27-year-old job seeker.

Split by five-year age group: population share ~100,000, employment-ratio share ~90,000

A decomposition breaks a change in employment into two pieces. The first is the number of employed people who would have vanished as the population shrank, even if the employment ratio had stayed at last year's level. That's the "share from population decline." The second is the change that a shift in the employment ratio would have produced even if the population had held steady: the "share from the lower employment ratio." Add the two and you get the actual change.

You can run this calculation on 15–29-year-olds as one block or age group by age group. The MODS release publishes population, employment and employment ratios separately for ages 15–19, 20–24 and 25–29. The three groups look very different.

Over the past year, the population aged 15–19 grew by 47,000. The 20–24 group shrank by 91,000 and the 25–29 group by 96,000. October employment ratios were 5.8% for ages 15–19, 40.0% for 20–24 and 71.0% for 25–29. In other words, the youth population decline is concentrated among people in their twenties, where employment ratios are high. That makes the share from population decline larger, too.

Calculated age group by age group, October's 196,000 decline in youth employment breaks down like this (PRISM calculation):

  • Share from population decline: about 100,000. These are employed people who would have been lost even if each age group's employment ratio had stayed at last year's level.
  • Share from lower employment ratios within age groups: about 90,000.
  • That ~90,000 by age: about 80,000 for ages 20–24, and 14,000–15,000 for 25–29. Ages 15–19 contributed a positive 4,000.

The employment ratio for ages 20–24 fell 3.6 percentage points, from 43.6% to 40.0%. For ages 25–29 it slipped 0.5 points, from 71.5% to 71.0%. The 20–24 group had both the steepest drop and a population of 2.2 million. That's why most of the employment-ratio share comes from this age group.

One objection is that the answer depends on how you slice it. Treat 15–29-year-olds as a single block, and the share from population decline shrinks to about 60,000, while the employment-ratio share grows into the 130,000s (PRISM calculation). The single-block method assumes that the 47,000 additional 15–19-year-olds, along with the 91,000 fewer 20–24-year-olds and 96,000 fewer 25–29-year-olds, all work at the average youth employment ratio (44.6% a year earlier, or 42.9% this year). In reality, the ratio is around 6% for ages 15–19, in the 40% range for 20–24 and in the 70% range for 25–29. So to count "employed people who vanished as the population shrank," multiplying each age group separately is the better fit for the question. The single-block decomposition asks, "What if the overall youth employment ratio had stayed the same?" The five-year version asks, "What if each age group's employment ratio had stayed the same?" Five-year bands are the finest breakdown the release offers, but they can't filter out shifts in the age mix within each band.

Was October a one-month blip?

It's too early to draw conclusions from a single month. PRISM ran the same calculation on the same tables from the June–September releases.

2026Share from population decline (5-year bands)Share from lower employment ratio (5-year bands)Of which ages 20–24Reference: population share (15–29 as one block)
June−106,000−92,000−63,000−65,000
July−106,000−84,000−72,000−65,000
August−105,000−38,000−41,000−64,000
September−107,000−40,000−59,000−64,000
October−104,000−92,000−82,000−61,000

Table source: PRISM calculation from the original figures in the "Employed persons and employment-to-population ratio by age group" tables of the MODS June–October Employment Trends. Note: Every value in the table is the average of two versions, one that multiplies the population share by the year-earlier employment ratio and one that multiplies it by this year's (inputs are values rounded to the nearest thousand). October ranges: population share −102,000 to −106,000; employment-ratio share −90,000 to −94,000; ages 20–24 −80,000 to −84,000.

Three things emerge. First, the share from population decline was nearly constant across all five months, at 100,000–110,000. Second, what moved from month to month was the employment-ratio share. June, July and October were similar, at about 90,000, 80,000 and 90,000; August and September were smaller, at about 40,000. In those two months the 15–19 group's share was positive (about 18,000 and 21,000, respectively; PRISM calculation), offsetting part of the 20–24 share. Third, the employment-ratio share for ages 20–24 was negative in all five months, and in all five it was the largest of the three age groups on paper (sampling error not considered).

PRISM covered September's youth employment ratio of 44.0% (down 1.1 points from a year earlier) in Five Things, fourth week of October. Split by age group, that September's largest negative employment-ratio share also came from ages 20–24 (about 59,000).

PRISM

Advertise with Us

[email protected]

A caveat: this comparison covers just five months. Sampling error in monthly employment by age group isn't included in the calculation.

Ages 20–24: fewer unemployed, more people outside the labor force

So where are the 20–24-year-olds whose employment ratio fell? Labor statistics sort the population into three groups: the employed, who work; the unemployed, who are looking for work; and people outside the labor force, who are neither working nor looking. That last group includes students, people preparing for jobs, and those who "just rested" without any particular activity, a category of its own in Korea's survey. The labor force participation rate is the employed plus the unemployed as a share of the population.

Age (Oct 2026)Population changeEmployment ratio (year earlier → this year)Participation rate (change)Unemployment rate (change)Change in unemployedChange in people outside the labor force
15–19+47,0005.6% → 5.8%————
20–24−91,00043.6% → 40.0%41.9% (−4.2 pts)4.6% (−0.7 pts)−14,000+43,000 (PRISM calculation)
25–29−96,00071.5% → 71.0%75.2% (−0.3 pts)5.5% (+0.1 pts)−2,000—
15–29−141,00044.6% → 42.9%—5.2% (−0.1 pts)——

Table source: MODS October Employment Trends, tables on employed persons and employment-to-population ratio by age group, labor force participation rate by age group, and unemployment. People outside the labor force aged 20–24 are a PRISM calculation: population minus the employed and the unemployed (1,236,000 → 1,279,000). A dash marks cells this article doesn't use.

For ages 20–24, the participation rate fell 4.2 points, from 46.1% to 41.9%. The number of unemployed dropped by 14,000, and the unemployment rate fell 0.7 points to 4.6%. Subtract the employed and the unemployed from the population, and the number of people outside the labor force rose by about 43,000 over the year (PRISM calculation). Both the employed and the unemployed shrank, while those outside the labor force grew. In September, too, the 20–24 participation rate fell 2.6 points to 44.2%, and unemployed 20–24-year-olds decreased by 6,000 in August and 2,000 in September.

These numbers compare last year's 20–24-year-olds with this year's. Over that year, some people aged into the group and others aged out. So the figures can't tell us where any particular person moved from or to.

Figures on what people outside the labor force are actually doing exist only for broader age groups. Across all ages, the number in school or training rose by 131,000 and job preparers by 21,000, while those who "just rested" fell by 69,000. Among people aged 20–29, "just rested" stood at 353,000, down 49,000 (12.1%) from a year earlier, after declines of 58,000 in August and 51,000 in September.

This release can't answer why the 20–24 employment ratio fell. The activity-status table isn't broken down by age, and the "just rested" table uses a 20–29 band. So the decisive cells, the changes in school or training, job preparation and "just rested" for ages 20–24, are empty. Those values are what it would take to weigh delayed entry into work (supply) against reduced hiring (demand). The 14,000 drop in unemployed 20–24-year-olds fits either reading. People who stay in school without job-hunting and people who've put off their search both fall outside the unemployed count. Until figures for ages 20–24 are available, we set the two readings side by side and don't pick one.

  • Delayed entry (supply side). This reading holds that more people are staying in school longer or spending more time preparing for jobs. Widen the age band, and "just rested" fell among 20–29-year-olds while school or training rose across all ages. The limitation: those two figures are for ages 20–29 and for all ages.
  • Reduced hiring (demand side). This reading holds that people postponed job-hunting as openings shrank. One figure fits it: employment insurance enrollment for people aged 29 and under (wage workers only; a different dataset from the Economically Active Population Survey) fell by 54,000 year-on-year as of the end of October. It has limits, too. That figure covers everyone aged 29 and under, and the Ministry of Employment and Labor said the decline had "narrowed" (PRISM translation).

The unemployment rate and the broad unemployment measure belong in the picture together. In the same month that the unemployment rate for young people aged 15–29 fell 0.1 point to 5.2%, the broad unemployment measure (Supplementary Labor Underutilization Indicator 3) rose 0.4 points to 15.1%. The broad measure adds to the unemployed those working short hours who want more work, plus people who are willing to work but can't immediately look for or take a job. Both figures cover all 15–29-year-olds, so neither directly settles why the number of 20–24-year-olds outside the labor force rose.

Chip exports hit 3.6 times last October's level. What about factory jobs?

Two statistics count jobs. MODS's Economically Active Population Survey surveys households and counts the people who worked. The Ministry of Employment and Labor's employment insurance data counts wage workers enrolled in employment insurance as of month-end. Since they cover different people at different points in time, this article never adds or subtracts figures across the two.

Indicator (Oct 2026)LevelYear-on-year changeSourceCoverage / measure
Semiconductor exports—+259% (3.6 times, PRISM calculation)Ministry of Trade, Industry and Resources, export and import trendsCustoms basis, dollar value
Employment insurance enrollees, semiconductor industry113,000+7,900 (+7.5%)Ministry of Employment and Labor, employment administrative statisticsAll ages, month-end
Employment insurance enrollees, manufacturing3,858,000+14,000Ministry of Employment and Labor, employment administrative statisticsAll ages, month-end
Manufacturing enrollees aged 29 and under—−18,000Ministry of Employment and Labor, employment administrative statisticsAged 29 and under, month-end
All enrollees aged 29 and under2,203,000−54,000Ministry of Employment and Labor, employment administrative statisticsAged 29 and under, month-end
Employed in manufacturing4,357,000−7,000 (September: +16,000)MODS, Economically Active Population SurveyAll ages, household survey

Table sources: Ministry of Trade, Industry and Resources October Export and Import Trends; Ministry of Employment and Labor Labor Market Trends Based on October 2026 Employment Administrative Statistics; MODS October Employment Trends. Under the labor ministry's classification, semiconductors are a sub-industry of electronics and telecommunications manufacturing.

In the employment insurance data, semiconductor-industry enrollment reached 113,000 at the end of October, up 7,900 (7.5%) from a year earlier. The year-on-year gain widened every month, from 5,000 in June to 7,900 in October (all ages). In its October employment administrative statistics briefing, the Ministry of Employment and Labor said of electronics and telecommunications manufacturing that "as semiconductor exports increased sharply, enrollment growth expanded, centered on semiconductor manufacturing, but audio-visual equipment, which includes home appliances and displays, continued to decline" (PRISM translation). In the Q&A, a ministry official also said, "In manufacturing, as you mentioned, semiconductors seem to have the biggest effect" (PRISM translation). The link between exports and enrollment growth is the ministry's explanation. What PRISM can confirm with numbers from these releases is that semiconductor exports and semiconductor-industry enrollment rose together over the same period.

The same employment insurance data looks different for people aged 29 and under. Among manufacturing enrollees, that age group fell by 18,000. Total enrollees aged 29 and under stood at 2,203,000, down 54,000 over the year. The decline narrowed from 62,500 in June to 54,100 in October. The ministry said enrollment among those aged 29 and under "declined, centered on manufacturing, information and communications, health and welfare, and professional, scientific and technical services, amid the impact of a population decrease of 143,000" (PRISM translation).

The data doesn't show how many semiconductor-industry enrollees are 29 or younger. The semiconductor increase (7,900, all ages) and the decline among manufacturing enrollees aged 29 and under (18,000) come from the same employment insurance data, so they can sit side by side. Their age coverage differs, though, so you can't subtract one from the other to get a "youth share of semiconductors."

In the Economically Active Population Survey, manufacturing employment rose by 16,000 in September, then fell by 7,000 in October (all ages).

In short, the most we can say is that rising chip exports and a falling youth employment ratio showed up "at the same time." Exports are a customs-based dollar value and employment is a headcount, so even the units differ. These releases contain no figures that would let anyone judge how the chip boom has affected jobs for young people.

For a 27-year-old job seeker: does a shrinking population make hiring easier?

Say you're 27, you graduated in engineering from a university outside the Seoul area, and you've spent two years applying to major chipmakers and their suppliers. You've probably heard that "with fewer young people, there's less competition, so getting hired gets easier." Here's what the latest statistics confirm.

In October, the population aged 25–29 was 96,000 smaller than a year earlier. Yet its employment ratio failed to rise, slipping from 71.5% to 71.0%. The participation rate fell 0.3 points to 75.2%, and the unemployment rate rose 0.1 point to 5.5%. In August, September and October, the 25–29 unemployment rate (5.6%, 5.7% and 5.5%) was higher than a year earlier in all three months.

In the decomposition above, the employment-ratio share for ages 25–29 was negative in all five months from June to October (2,000 to 27,000; PRISM calculation). In September, the 25–29 participation rate rose 0.5 points to 76.0%, while the employment ratio held at 71.7%, the same as a year earlier. Put differently, the share of the population working or looking for work went up, and the share actually employed stayed flat.

These statistics show no sign that finding work has gotten "easier" for 25–29-year-olds. We won't use these numbers to predict whether it will.

If you're weighing applications to chip companies, the verifiable numbers are the two lines from the previous section. Semiconductor-industry employment insurance enrollment rose across all ages, and in the same month, manufacturing enrollees aged 29 and under declined. How enrollment among those aged 29 and under changed within the semiconductor industry is unknown, since the labor ministry's data doesn't include it. Nor do these statistics say anything about competition by industry or your odds of getting hired.

Summing up: what we know and what we don't

Split by five-year age group, about 100,000 of October's 196,000 decline in employed young people (ages 15–29) is the share from population decline. The remaining roughly 90,000 is the share from lower employment ratios within age groups, and about 80,000 of that is from ages 20–24.

The strength of the evidence varies from figure to figure. The inputs, meaning population, employment and employment ratios by age group, come straight from MODS's original release. The figures of about 100,000, 90,000 and 80,000 are PRISM calculations. Even within the same five-year decomposition, results shift by around 4,000 depending on whether last year's or this year's employment ratios are used. Treating 15–29-year-olds as a single block puts the share from population decline at about 60,000. The 20–24 share was the largest negative of the three age groups on paper in all five months from June to October (sampling error not considered). Still, that's a five-month comparison. If the 20–24 share doesn't come out as the largest of the three in the same table for November, we'll revisit the conclusion that most of the employment-ratio share came from ages 20–24.

Some things remain unknown. Pinning down why the 20–24 employment ratio fell would take that age group's changes in school attendance, job preparation and "just rested." The monthly release doesn't have those cells, and this check couldn't confirm whether or when they're published separately. The labor ministry's data also doesn't give the number of semiconductor-industry enrollees aged 29 and under.

What to watch next

  • MODS November Employment Trends (due in December). It adds a sixth month to the decomposition using the same five-year table. We'll check whether the 20–24 employment-ratio share is the largest negative of the three groups for a sixth straight month, and whether the employment-ratio share is closer to its June, July and October size or its August and September size.
  • DGBAS October manpower survey for Taiwan (due November 23). It lets us match October with October and recompare participation rates for ages 15–24 and 25–29.
  • The labor ministry's next monthly employment administrative statistics. We'll watch how the semiconductor-industry enrollment gain and the decline among enrollees aged 29 and under move.
  • Activity status of 20–24-year-olds outside the labor force. These cells aren't in the monthly release, and their publication schedule hasn't been confirmed yet. If it becomes clear which of school or training, job preparation or "just rested" grew, the weight given to each explanation of the cause could shift.

Figures are based on original releases from MODS, the Ministry of Employment and Labor, the Ministry of Trade, Industry and Resources, Taiwan's DGBAS and Taiwan's Ministry of Finance; calculations by PRISM. Cross-checked against the originals by PRISM on November 16, 2026.

This content is AI-generated based on source articles. While we strive for accuracy, errors may occur. We recommend verifying with the original source.

Thoughts

Related Articles

PRISM

Advertise with Us

[email protected]
PRISM

Advertise with Us

[email protected]