Young Caregivers Aren't 'Resting,' Despite What Stats Say · Issue #203

Opening

The Federal Reserve Bank of St. Louis published a short piece last week. Young Americans aged 18 to 24 who are neither working nor in school — U.S. statistics call them “disconnected.” As of 2024, that’s 16%. So far, this is a familiar story.

But this time, instead of just counting people, the research team opened up their daily diaries — pooling 22 years (2003–2024) of the American Time Use Survey¹. The most striking finding wasn’t gaming or YouTube. It was caring for household members. About 41% of disconnected youth had provided care to a family member on the day surveyed, versus 16% of their peers.

Dear reader, this raises a question. We’ve been calling these people “disconnected” — so why did it take this long to actually look at how they spend their days?

Let me give you the conclusion up front. Today’s story isn’t really about American youth. It’s about how the label attached to a statistic decides the diagnosis before anyone looks at the data — and Korea’s “resting” category, known as swieosseum, the label Statistics Korea uses for economically inactive people who cite no specific reason, is, in this respect, engineered even more tightly than America’s. Those 41% of young people I just mentioned? Korean statistics exclude them by definition, from the start.

A Day in the Life of the Youth America Calls “Disconnected”

One clarification first: what the research team measured wasn’t “how many hours were spent” but “whether the activity happened at all.” That’s why the chart’s subtitle reads “Probability.” The 41% doesn’t mean 41% of the day — it means 41% of people did caregiving that day.

Seen through that lens, here’s the picture.

It’s not just caregiving. In household chores, too, disconnected youth scored 76% versus 64% for their peers — cooking, cleaning, running a household. None of this shows up as a single line in labor statistics.

Work and education, by contrast, were low — which makes sense, since being neither employed nor in school is the very definition of “disconnected.” But here’s what’s interesting: work-related activity isn’t 0%, it’s 10%. The research team pins down what that 10% is: job hunting, interviews, and informal income-generating activity like selling goods or providing small services. In other words, people statistically classified as “not employed” are out there actively trying to earn money. Education shows up at 22% too — even among people who aren’t enrolled anywhere.

The most decisive evidence is in the second chart. Socializing and leisure sit near 95% for both groups. Eating and drinking, also near 95%. Phone calls: 18% versus 17.5%. Shopping: 43% versus 42%. The gap is essentially zero. That means these young people aren’t disconnected because they’re goofing off. Their leisure time is identical to their peers’.

The one category that stood out as noticeably lower was travel: 82% versus 94%, a 12-point gap. They go outside less. Put that next to the caregiving and chores data, and the reason reads itself: they’re tied to home.

The research team’s conclusion, in their own words: it points to “constraints,” not “disengagement.”

The Bank of Korea Arrived at the Same Conclusion

Korea has a similar statistic. It’s called swieosseum — “resting.”

According to an issue note the Bank of Korea released in January 2026, among the economically inactive population² aged 20 to 34, the share classified as “resting” rose from 14.6% in 2019 to 22.3% in 2025 — a 7.7-point jump in 6 years.

And among what the Bank of Korea found, one point runs squarely against conventional wisdom.

Their expectations aren’t unrealistically high. This is the report’s strongest punch. The average reservation wage³ of “resting” youth was ₩31,000,000 — about the same as other categories of unemployed youth. Asked what kind of company they’d want to work for, 48.0% said small and medium-sized enterprises, the top answer, followed by public institutions at 19.9% and large corporations at 17.6%. Their expectations were, if anything, lower than other unemployed youth, who ranked large corporations and public institutions first.

Most have work experience. This is where most of the increase came from. “Resting” youth with prior job experience grew from 360,000 in 2019 to 477,000 — people who got in once and came back out.

Education levels do track conventional wisdom — though this one needs a precise reading. Among “resting” youth, those with an associate’s degree or less made up an average of 59.3% between 2019 and 2025 — 6 out of 10. Among youth with an associate’s degree or less, the “resting” rate was 8.6%, nearly double the 4.9% for those with a four-year degree or higher. Regression analysis also shows those with an associate’s degree or less running 6.3 points higher. That said, the Bank of Korea adds that “recently, the number of resting youth with a four-year college degree or higher has also been rising.”

How you read this third point is the fork in the road. Read it as “they can’t get in because their education level is low,” and you’re back to blaming the individual. But the Bank of Korea reads it the opposite way: if expectations are already low, and yet lower education still means fewer people get in, that’s not a question of willingness — it’s a question of barriers. Which is why the report’s policy recommendation is to “focus on drawing youth with an associate’s degree or less back into the labor market.”

The Bank of Korea points to two causes: AI-driven technological change, and companies’ preference for experienced hires. In other words, the number of entry-level positions itself is shrinking.

Two central banks, in the same year, using different data, arrived at the same sentence: the problem isn’t attitude, it’s structure.

But the Two Statistics Don’t Count the Same People

This is where today’s real point begins.

These two statistics are hard to place side by side. There are 2 layers to why.

First, the age ranges differ. The St. Louis Fed looks at ages 18 to 24. The Bank of Korea’s analysis covers 20 to 34 — 10 years wider, counting 10 years further up. Someone resting at 33 and someone resting at 19 are entirely different stories, yet one statistic includes both while the other includes only one.

Second — and this matters more — the two categories are built in opposite directions.

America’s “disconnected” has a simple definition: not employed plus not enrolled in school. That’s it. It’s additive. So people raising children, people caring for a sick parent, people who are ill themselves — all of them fall inside this category.

Korea’s “resting” is subtractive. Among the economically inactive population, it’s the residual category left over after removing everyone who cites a clear reason — childcare, housework, being enrolled in school or classes, physical or mental illness, job-hunt prep, prep for further schooling, waiting for military conscription. What’s left are people who couldn’t name any reason and simply answered, “I was just resting.”

Put that difference side by side, and it gets a little chilling.

The very people the St. Louis Fed uncovered by opening up the data — young people doing caregiving and housework — Korean statistics exclude by definition, from the outset. Do caregiving, and you’re sorted into “childcare.” Do housework, and you’re sorted into “chores.” There’s no way to end up in the “resting” box.

So the discourse inevitably drifts toward “why aren’t they doing anything?” Because the label already has “no reason” stamped into it. The statistical definition has, in effect, already rendered a moral verdict before anyone looks.

America’s label was loose, so it swept everyone in — and opening it up revealed something. Korea’s label is tight, filtering everyone out — designed so that opening it up reveals nothing at all.

One more thing worth flagging. There’s a figure the media often cites: “450,000 resting youth.” This is not the full “resting” youth population. It’s a subgroup within it — those who answered that they “don’t want a job at all.” That number grew 56.8% over 6 years, from 287,000 in 2019 to 450,000. For reference, per the 2025 employment trends data, the total “resting” population in their 20s and 30s is 717,000. A single number rides along in headlines, but what it’s actually a number of doesn’t travel with it.

We Count, But We Don’t Measure

It’s not that Korea has no tool for measuring time use. There’s the Time Use Survey run by the National Data Agency (formerly Statistics Korea). The 2024 survey results came out in July 2025, covering roughly 25,000 people aged 10 and up across 12,750 households.

The problem is frequency and resolution.

The Time Use Survey runs once every 5 years. It began in 1999, and 2024 marks its 6th round. America’s ATUS runs annually, and this analysis stacked 22 years of it to carve out a small subgroup. With a survey run once every 5 years on a sample of 25,000, it’s hard to isolate “resting youth” specifically and get statistically meaningful results. And 5 years is long enough for the youth labor market to turn into an entirely different market.

So this is where we stand: the headcount is refreshed every month, but we effectively have no idea how these people actually spend their days.

This matters because scale determines the budget, while content determines the design. Trillions of won in youth-support budgets get allocated every year — with the scale known and the design unknown.

Think it through, and the prescriptions diverge completely.

If barriers are the cause → the answer is loosening hiring practices and the preference for experienced hires.
If a career break is the cause → the answer is a re-entry ladder.
If learned helplessness is the cause → the answer is access to psychological and medical care.

These three are policies with different budget line items, different responsible ministries, different performance metrics. Yet right now, they’re all bundled into a single question — “what should we do for resting youth?” — because no one has measured how they actually spend their days.

Oswald’s Lens

I teach data management at Sejong University, and I always assign the same thing in the first class of the semester: before you interpret an indicator, pull apart its definition. It’s the part students find most boring — they want to jump straight to the numbers. But almost every misreading starts right there.

I’ve seen the same scene play out again and again while building go-to-market strategies. Faced with a dashboard showing “low conversion,” teams split. Team A concludes “users just aren’t interested” and raises the ad budget. Team B opens session replays to see exactly where users stall. More often than not, Team B was right. Aggregate numbers look like they’re telling you about attitude, but they’re actually hiding constraints. “Resting” and “disconnected” are exactly the same mistake, played out at national scale.

That said, let me flag 3 things honestly.

First, don’t over-read the American data. Because “disconnected” includes caregivers by definition, the high caregiving share is partly an artifact of the definition itself. Read it as “look, everyone’s caregiving,” and you’re making the same mistake in the opposite direction. The real finding in this data isn’t the caregiving share — it’s that leisure time matches their peers’ exactly. The conventional idea that they’re “just goofing off” simply didn’t show up in the data. That’s the point.

Second, constraints alone don’t explain everything either. The Bank of Korea’s analysis shows that each additional year of unemployment raises the probability of being “resting” by 4.0 points and lowers the probability of choosing to job-hunt by 3.1 points. Given enough time, constraints really do harden into learned helplessness. Miss this and simply chant “it’s all structural,” and you can’t design any policy at all. Sequence matters here: constraint is the cause, and helplessness is the effect. Keep that order straight, and you get a prescription. Reverse it, and you’re back to “it’s a matter of willingness.”

Third, I got something wrong once while writing this piece. In my first draft, I wrote that most “resting” youth are highly educated — because that image fit neatly with “refuting the expectations theory.” When I reopened the Bank of Korea’s original report, it was the opposite: those with an associate’s degree or less make up 59.3%. I’d picked the fact that fit my argument first and checked it later. The person who teaches students to pull apart definitions did exactly the opposite. Since that habit is precisely what this piece criticizes, I’m leaving the mistake in rather than erasing it.

Closing

To sum up:

When the Fed opened up 22 years of time diaries, the days of “disconnected” youth turned out to be filled with caregiving and chores — and their leisure time was identical to their peers’.
The Bank of Korea also concluded that the problem with “resting” youth is structural, not a matter of expectations. Their reservation wage was ₩31,000,000, and their top choice of employer was small and medium-sized enterprises. But the methods differed: America measured behavior; Korea asked about intent.
Korea’s “resting” category is a residual left over after filtering out childcare and housework by definition. Because the label already has “no reason” stamped into it, the discourse inevitably drifts toward “why aren’t they doing anything?”

So the one proposal I actually want to make here is this: when we discuss youth policy, let’s not just ask “how many” — let’s also ask “where does the day go?” The Fed didn’t need a new budget or a new survey. It just re-sliced data that already existed. All it changed was the angle.

And this isn’t only about policy. Whenever a team talks about “people with low engagement,” we face the same choice every time: count the numbers, or open up the day.

If you or someone around you has been through a period that would’ve been classified as “resting,” tell me in the comments where most of those days actually went. What I’m most curious about is which box the statistics failed to see. If enough responses come in, I’ll pull them together in a future issue.

💬 Tell me in the comments where those days actually went · 📨 Share this with someone who’s found the word “resting” uncomfortable

(Web version only) 📬 Subscribe · Every week, I write a newsletter that cuts across technology, economics, and the humanities.

References & Further Reading

Primary sources

William M. Rodgers III, Alice L. Kassens, “How Are ‘Disconnected’ Young Adults Spending Their Time?”, St. Louis Fed On the Economy, July 14, 2026. Link ··· This is where today’s piece starts. Just the second chart alone (the socializing/leisure figures) is enough to shake conventional wisdom. Every figure cited in this piece was read off that article’s two charts — the text itself contains no numbers.
Bank of Korea, “[Issue No. 2026-3] Characteristics and Assessment of ‘Resting’ Youth: A Comparative Analysis by Type of Unemployment,” BOK Issue Note, January 20, 2026. Link ··· The section covering reservation wage and preferred employer type is the key part — the cleanest rebuttal to the “expectations theory” out there.
National Data Agency (formerly Statistics Korea), “2024 Time Use Survey Results,” July 28, 2025. Link ··· Proof that Korea does have the tool for this. At the same time, it shows the limitation of a 5-year cycle.

Background

Korea Labor Institute, “An Analysis of the Recent Rise in the ‘Resting’ Youth Population,” Monthly Labor Review No. 218, May 2023. Link ··· Lays out, most cleanly, exactly which categories are subtracted to leave the “resting” residual. This is the basis for Chapter 4 of today’s piece.
St. Louis Fed, “Disconnected Young Adults: A Look at the Eighth Federal Reserve District,” October 2024. Link ··· The predecessor piece, covering the regional, racial, and income distribution of America’s “disconnected” youth. Read alongside the time-use analysis, the full picture comes together.
St. Louis Fed, “How Shifts in Labor Supply and Demand Shape Outcomes for Young Workers,” June 2026. Link ··· The second installment of the three-part series this piece belongs to. It covers why young workers get pushed out first.

📝 Glossary

The author, Kwangseob Ahn, is a professor of business administration at Sejong University and lead consultant at OBF (Oswarld Boutique Consulting Firm). He teaches statistics and data analysis, including business data management and business analytics, while leading GTM and AI strategy consulting in the field, designing the seam between technology and business. He has published academic research on a memory architecture for AI dialogue systems (HEMA) and runs Daily Arxiv, a daily curation of global AI papers. He holds a master's from Korea University's Graduate School of Technology Management and a KMBA. He is the author of Homo Brainless: The People Who Outsource Their Thinking.

Footnotes

American Time Use Survey (ATUS): A survey sponsored by the U.S. Bureau of Labor Statistics and conducted annually by the Census Bureau. It has respondents log, hour by hour, how they spent “yesterday.” Because it measures behavior rather than asking about intent, it’s relatively less prone to self-report distortion. ↩
Economically inactive population: People aged 15 and older who are neither employed nor unemployed. The unemployed are defined as people willing to work and actively job-hunting, so once someone stops job-hunting, they drop out of the unemployment rate and shift into this category. This is why the unemployment rate can look fine even as the situation for young people worsens. ↩
Reservation wage: The minimum wage someone would need to be offered before deciding to take a job. Below this threshold, they’d rather not work at all — making it the standard metric for putting a number on “expectations.” ↩