The 24-Hour Budget

Since 2003, the Bureau of Labor Statistics has conducted the American Time Use Survey — a continuous study that asks roughly 26,000 Americans each year to account for every minute of the previous day. Not what they think they do. Not what they wish they did. What they actually did, in 15-minute increments, from 4:00 AM to 4:00 AM. The survey captures 127 distinct activities, from sleeping to commuting to watching television to “relaxing and thinking” — a category that sounds like meditation but more likely describes staring at a wall.

The result is 22 years of data (the survey was suspended in 2020 due to COVID) covering how Americans ages 15 and older allocate their most fundamental resource. The ATUS doesn’t ask how people feel about their time. It doesn’t measure quality or satisfaction. It measures quantity: hours and minutes, activity by activity, day by day. It is the most precise accounting of daily American life that exists.

What makes the ATUS uniquely valuable is its granularity. It doesn’t just tell you that Americans work “about 8 hours a day.” It tells you that the average American — including weekends, including the unemployed, including retirees — works 3.43 hours per day. That number shocks people because it contradicts the lived experience of anyone with a full-time job. But it’s a population-level average. On a given day, only about 52% of Americans work at all. Those who do work average roughly 7.5 hours. Averaged across the entire population — including the 48% who work zero hours on any given day — you get 3.43. The number is correct. The intuition is wrong.

This 10-part series will explore every dimension of the ATUS data: how the American day has changed since 2003, how men and women spend time differently, how age transforms our daily lives, what the pandemic did to our routines, and what the data says about the activities we’re choosing to do less of. This first episode establishes the baseline: how, exactly, does the average American spend their 24 hours in 2024?

Here is how the average American (ages 15+) spent their 24 hours in 2024, according to the ATUS:

Read that table slowly. It contains surprises.

Sleep dominates. At 9.04 hours — 9 hours and 2 minutes — sleeping consumes more than a third of the American day. This is significantly higher than the 7–8 hours that sleep researchers typically recommend and that most people report getting. The discrepancy exists because the ATUS measures time in bed, not necessarily time asleep. It also includes naps, dozing, and the increasingly common practice of lying in bed scrolling a phone before rising. Still, the trend is unmistakable: Americans spent 8.57 hours sleeping in 2003 and 9.04 in 2024. We are, collectively, spending 27 more minutes in bed every day than we did two decades ago.

Work is smaller than you think. The 3.43-hour average captures a population-level truth that individual experience obscures. A 35-year-old employed full-time works about 8 hours on a workday and perhaps an hour on a weekend day, averaging roughly 5.0 hours per day across the week. But the ATUS population includes 15-year-olds, retirees (the 65+ population works 0.80 hrs/day), the unemployed, stay-at-home parents, and people on vacation. When everyone is counted, work shrinks from the centerpiece of daily life to its second-largest component — and a distant second to sleep.

Television is the third-largest activity in America. At 2.60 hours per day — 2 hours and 36 minutes — watching TV consumes more time than household chores, more than eating, more than socializing, more than exercise, more than education, more than childcare. On any given day, 72.8% of Americans watch at least some television. That participation rate exceeds cooking (63.1%), shopping (35.9%), reading (16.1%), and exercise (which the ATUS doesn’t report at this level but other surveys place around 25%). Television is not America’s pastime. It is America’s default state.

Socializing has shrunk to 35 minutes. The average American spends 35 minutes per day in direct social interaction — talking with friends, hosting visitors, attending parties. In 2003, that number was 47 minutes. The 24% decline over two decades tracks with every measure of social isolation that researchers have produced: declining club membership, fewer dinner parties, smaller friend groups, and the rise of solitary digital entertainment. We now spend more time playing video games (34 minutes) than socializing face-to-face (35 minutes). The margin is a single minute.

The “average day” is, of course, a statistical fiction. No one lives an average day. Americans live weekdays and weekends, and the two are dramatically different.

On a non-holiday weekday in 2024, the average American sleeps 8.80 hours, works 4.37 hours, enjoys 4.56 hours of leisure, and watches 2.36 hours of television. On a weekend or holiday, they sleep 9.59 hours, work 1.22 hours, enjoy 6.24 hours of leisure, and watch 3.16 hours of television. The weekend adds 47 minutes of sleep, subtracts 3.15 hours of work, and adds 1.68 hours of leisure. Television absorbs 48 extra minutes on weekends — about half the leisure time freed by not working.

The weekday/weekend gap has narrowed over 21 years, and the reason is work. In 2003, weekday work averaged 4.68 hours; by 2024, it was 4.37 — a 19-minute decline. Weekend work barely changed (1.34 to 1.22). The implication: the American weekday is becoming slightly more weekend-like. Remote work, flexible schedules, and the structural decline in full-time manufacturing and retail employment have all contributed to a Monday-through-Friday that looks less rigidly defined by labor than it once did.

This convergence shows up in sleep. Weekday sleep rose from 8.26 hours (2003) to 8.80 (2024) — a gain of 32 minutes. Weekend sleep rose from 9.29 to 9.59 — a gain of 18 minutes. The weekday caught up because the weekday became less demanding. When you don’t commute, you can sleep longer. When your first meeting is at 9:30 instead of 8:00, you can stay in bed an extra 30 minutes. The remote work revolution didn’t just change where Americans work. It changed when they wake up.

The table below tracks how the American day shifted between 2003 and 2024. Some changes are gradual trends; others are step functions that occurred during the pandemic and never reversed.

The trends fall into three categories.

Growing: sleep, household work, games/computer, eating. Americans sleep 27 more minutes per day than in 2003 — equivalent to 164 extra hours per year, or nearly seven additional full days of sleep annually. Household work rose 11 minutes, driven almost entirely by a post-pandemic jump (from 1.78 in 2019 to 2.01 in 2024) as remote workers spent more time cooking, cleaning, and maintaining homes they now occupied all day. Games and computer leisure nearly doubled, from 17 minutes to 34 minutes per day. Eating gained 2 minutes — small in percentage but notable because it reversed a pre-pandemic decline.

Shrinking: socializing, shopping, reading, work. Socializing lost 11 minutes per day — a 24% decline that represents roughly 67 fewer hours of human contact per year. Shopping lost 8 minutes, reflecting the shift from physical stores to online purchasing that simply takes less time. Reading lost 5 minutes — a 22% decline from a small base. Work lost 16 minutes, consistent with longer-run trends toward later career entry, earlier retirement, and the rise of part-time and gig employment.

Flat then volatile: television. TV watching rose steadily from 2003 (2.58 hrs) through 2019 (2.81 hrs), then spiked during the pandemic (2.86 in 2021) before declining sharply to 2.60 in 2024. This is the first sustained decline in television viewing in the survey’s history. The decline coincides with the rise of short-form video (TikTok, YouTube Shorts, Instagram Reels), which the ATUS may classify differently depending on device and context. Television isn’t disappearing. But the definition of “watching TV” is blurring as screens fragment.

The ATUS reports two different measures for most activities: average hours per day for all persons (including those who didn’t do the activity at all) and average hours per day for participants only (those who actually engaged). The gap between these two numbers reveals how concentrated an activity is.

Sleeping: 99.9% of Americans sleep on any given day (the 0.1% presumably includes overnight workers surveyed mid-shift). Participation is universal. Eating and drinking: 95.9% — nearly universal, with the small remainder likely reflecting survey timing or fasting. Television: 72.8% — nearly three-quarters of Americans watch TV on any given day. Among those who do, they watch an average of 3.57 hours (much higher than the population average of 2.60, which dilutes by the 27% who watch nothing).

Food preparation: 63.1% cook or clean up on a given day. Shopping: 35.9% buy something on a given day. Reading: only 16.1% read for personal interest on a given day — meaning five out of six Americans do not pick up a book, magazine, or long-form article in any 24-hour period. Playing games: 15.3% play games on a given day — nearly as many as read. Religious activities: 8.7%. Volunteering: 4.7% — fewer than one in twenty Americans volunteers on any given day.

The participation rates matter because they tell you whether an activity is widespread but brief, or concentrated among a dedicated minority. Reading, at 16.1% participation, averages only 17 minutes per day across the population — but among the 16% who actually read, they average roughly 1 hour 45 minutes. Volunteering at 4.7% averages just 3 minutes per day population-wide, but volunteers spend about an hour when they do it. The “average American day” is actually a statistical composite of many different days lived by different types of people. The median American doesn’t read. The median American doesn’t volunteer. But 72.8% of them watch television.

If you think of the 24-hour day as a budget, the American allocation in 2024 looks like this: 37.7% sleep, 14.3% work, 10.8% television, 8.4% household maintenance, 5.2% eating, and the remaining 23.6% split among socializing, games, education, shopping, childcare, exercise, reading, religion, grooming, and everything else. Sleep and work together claim 52% of the day. Television alone claims more than one-tenth.

Compare this to 2003. Sleep was 35.7%, work was 15.4%, television was 10.8%, household was 7.6%. The big shift: sleep gained 2 percentage points of the day (from 35.7% to 37.7%) and work lost 1.1 points (from 15.4% to 14.3%). Those 3.1 percentage points — about 45 minutes — are the structural reallocation that defines the last two decades. Americans traded work minutes for sleep minutes. Not because they became lazier, but because the economy, technology, and demographics conspired to make work take less of the day: fewer people work full-time, more people are retired, remote work eliminated commutes, and the labor force participation rate declined from 66.2% to 62.6%.

The time economy also reveals what economists call “revealed preferences.” When Americans gain free time — whether from shorter commutes, earlier retirement, or reduced work hours — where do they allocate it? The data is unambiguous: sleep and screens. The 27 minutes gained in sleep and 17 minutes gained in games/computer account for 44 of the 45 minutes freed. Socializing, reading, exercise, and volunteering gained nothing. When given more time, Americans do not socialize more, read more, or volunteer more. They sleep more and play more video games. The preferences are revealed.