The Shrinking Workday

The American Time Use Survey averages across all Americans age 15 and older — working and not working, weekdays and weekends. When the BLS reports that the average person worked 3.43 hours in 2024, that includes the retired schoolteacher who worked zero hours, the part-time barista who worked four, and the surgeon who logged twelve. It includes Saturdays and Sundays, holidays and sick days. It is a population-level measure of how much of America’s collective time goes to work.

By that measure, 2024 marked a new low. The previous bottom was 2010, when the lingering effects of the Great Recession pushed work down to 3.50 hours. The economy recovered, and work time climbed back to 3.61 hours by 2019. Then came COVID. The 2021 reading was 3.50 again, matching the recession floor. By 2022 and 2023, work edged back up to 3.50 and 3.56. But in 2024 it fell to 3.43 hours — below both the recession and pandemic troughs. This happened during a year with unemployment near historic lows, which makes it all the more puzzling. More Americans had jobs in 2024 than in 2010, yet the average person spent less time working.

The paradox resolves when you separate two things: how many people work on a given day, and how much the workers work. In 2003, 46.2% of Americans performed some work on an average day. By 2024, that had fallen to 42.6%. That’s 3.6 percentage points fewer people showing up to work on any given day — driven by an aging population (more retirees), more young people in school longer, and the rise of non-traditional work schedules. Among those who do work, full-time employees put in 8.67 hours on the days they worked in 2024, down from 8.81 in 2003 — a modest 8-minute decline over 21 years. The real story of the shrinking workday is not that workers are slacking off. It’s that a smaller share of the population works on any given day.

No demographic split in the time-use data is as dramatic as gender. In 2003, the average American man spent 4 hours and 34 minutes per day on work and work-related activities. The average woman spent 2 hours and 53 minutes — a gap of 100 minutes, or an hour and forty minutes. Men worked nearly 60% more than women in terms of daily time.

By 2024, men had dropped to 3 hours and 55 minutes and women had inched up to 2 hours and 57 minutes. The gap shrank from 100 minutes to 58 — a 42% reduction. This is one of the largest social transformations visible in the data, and it happened without a single piece of legislation targeting it directly. Men lost 38 minutes of daily work time. Women gained 4 minutes. The convergence was driven almost entirely by men working less.

Why are men working less? Several forces are at play. The manufacturing sector — historically male-dominated and built on long shifts — shed millions of jobs over this period, from 14.3 million workers in 2003 to about 12.8 million in 2024. Construction employment recovered from the housing bust but never returned to the overtime hours of the mid-2000s boom. Meanwhile, the fastest-growing sectors (healthcare, education, professional services) have more balanced gender ratios and more standardized hours. The gig economy and part-time work have also grown, and men are disproportionately represented among discouraged workers who have left the labor force entirely.

Women’s near-flat trajectory is itself remarkable. Female labor force participation peaked at 60.0% in 2000 and has been slowly declining since, to about 57% in 2024. Yet women’s daily work time held steady at roughly 2.9 hours. This suggests that the women who are working are working more intensively — longer hours per worker offsetting the slight decline in participation. The ATUS data confirms this: among employed women, work hours barely changed.

Work time follows a predictable arc: low in youth, peaking in the 45–54 age bracket, and collapsing after 65. But the shape of that arc has changed in telling ways.

Young Americans (15–24) worked 2.68 hours per day in 2003 and just 2.38 in 2024 — an 18-minute decline. This reflects more time in education, later labor-force entry, and a generation less likely to hold summer jobs. A 2019 Pew Research Center study found that only 35% of teens worked during the summer, down from 55% in the late 1970s. The flip side of young Americans sleeping 9 hours and 40 minutes (as we saw in Episode 2) is that many of them aren’t working.

The prime-age brackets tell divergent stories. Workers aged 25–34 shed just 5 minutes (4.82 to 4.73), while the 35–44 group actually gained 16 minutes (4.79 to 5.05). This is the bracket most affected by dual-income households with children, where both parents are working and the logistical burden of school schedules and childcare makes every working hour more compressed and intentional. The 45–54 group, which had the most work time in 2003 at 4.92 hours, was essentially flat at 4.91 in 2024.

The 55–64 group tells the story of delayed retirement and the knowledge economy. These workers logged 3.69 hours in 2003 and 3.74 in 2024 — they’re working slightly more, not less. Many are in professional or managerial roles that don’t require the physical stamina that sent earlier generations to the golf course at 55. And retirees 65 and older held steady at 0.75 to 0.80 hours, reflecting a persistent minority who continue working part-time well past traditional retirement age.

Americans are working less on both weekdays and weekends, but the weekday decline is sharper. On non-holiday weekdays, average work time fell from 4.68 hours in 2003 to 4.37 in 2024 — a drop of 19 minutes. On weekends and holidays, it fell from 1.34 to 1.22 hours — just 7 minutes.

The weekday decline maps almost perfectly onto the rise of remote work and flexible schedules. When your commute evaporates, the boundaries between “work time” and “personal time” blur. A remote worker might take 30 minutes at 2 p.m. to walk the dog, making up the work at 8 p.m. The ATUS diary captures this — it records what you did at each moment, not what you intended to do. The result is that the measured “work” block looks shorter even if total productive output hasn’t changed, because personal activities are now interspersed throughout the day.

Weekend work, by contrast, was always low and has stayed that way. The 7-minute decline from 1.34 to 1.22 hours is within the range of year-to-year noise. What’s notable is that 1.22 hours — about 73 minutes — is still the weekend average, meaning a substantial number of Americans work a few hours on Saturday or Sunday. This includes shift workers, healthcare professionals, retail employees, and the growing population of gig workers for whom the weekend is just another workday.

The ATUS offers three lenses on work time, and they tell very different stories. Understanding which one applies matters for interpreting claims about whether Americans are “working less.”

Lens 1: All persons. This is the headline number — 3.43 hours in 2024, down from 3.69 in 2003. It includes retirees, students, stay-at-home parents, and the unemployed. It fell 16 minutes. But it’s a population-level average that mixes working and non-working populations.

Lens 2: Employed persons. If you filter to only people who have jobs, the picture changes. Employed Americans worked 5.37 hours per day in 2024, down from 5.62 in 2003 — a 15-minute decline. Employed full-time workers averaged 6.05 hours, down from 6.30 — also 15 minutes. This is a meaningful decline but much smaller than the headline figure suggests. It includes weekends and days off, which pull the average down.

Lens 3: Full-time workers on work days. This is the most apples-to-apples comparison. How much does a full-time worker actually work on a day they go to work? In 2003, the answer was 8 hours and 49 minutes. In 2024, it was 8 hours and 40 minutes — a decline of just 8 minutes over 21 years. The workday for people who actually show up to work has barely budged. What changed is everything around it: how many people work, how often they work, and how the non-work days dilute the average.