Say you are 24, choosing between a certificate in medical coding and a bookkeeping course, and you open a career profile that says “faster than average.” What does that phrase rest on? The answer is a ten-year estimate from the Bureau of Labor Statistics (BLS), and the newest edition landed on August 27, 2026.
The headline is modest. BLS projects total employment will rise from about 170.3 million jobs in 2025 to about 176.2 million in 2035. That is 5.9 million more jobs, or 3.5 percent, compared with 10.9 percent growth over 2015 to 2025. Below, you’ll see how the numbers are built, what the columns mean and how to use them without over-trusting them. Nothing here is a promise about any particular job or paycheck.
Key takeaways
- The 2025-35 projections show total jobs growing 3.5 percent, slower than the previous decade.
- Healthcare drives much of the growth, while office and administrative support shrinks the most.
- Growth and openings are different numbers. Check both.
- BLS calls the projections a set of “what if” scenarios, not a forecast. Compare occupations with them, then verify with local postings.
What the program actually produces
The Employment Projections program estimates employment ten years out for the whole economy, for industries and for hundreds of detailed occupations. The Occupational Outlook Handbook (OOH), the BLS career guide, draws on the occupational part and covers roughly 600 detailed occupations.
When you see a profile, the “job outlook” and “employment projections” figures come from this program. The same tables also feed state agencies, school counselors, workforce boards and many private career sites. That is why a small change in assumptions can ripple through a lot of advice.
How the numbers are built, step by step
The methodology page describes six linked steps. Each one feeds the next, and BLS can loop back through the sequence to check that the pieces fit together. In plain English:
- Start with people. Census Bureau population projections are converted into the working-age population, then participation rates by age, sex, race and ethnicity are applied to estimate the labor force. BLS calls this the most important outside input to its economic model.
- Project the whole economy. A commercial macroeconomic model (BLS names the S&P Global US Macro Model) estimates total output and the main categories of demand, with BLS supplying the labor force numbers and key assumptions such as energy prices and fiscal and monetary policy.
- Split demand into goods and services. Analysts estimate how much consumers, businesses, government and foreign buyers will spend on different kinds of products, with industry experts refining the details.
- Translate demand into industry output. Input-output tables track what each industry buys from others, which converts final demand into how much each industry must produce.
- Estimate industry employment. A regression links output, relative wages and time to hours worked, then converts hours into jobs. A productivity check catches results that don’t make sense.
- Divide jobs among occupations. A national “staffing pattern” table shows which occupations each industry employs. Economists adjust those patterns using historical data, expert interviews and news reports.
Separately, BLS estimates occupational separations: workers who leave an occupation, either for another one or to exit the labor force. That comes into play when we get to openings.
Why this matters to you
Occupations are projected through industries. An occupation concentrated in a shrinking industry usually gets a weaker outlook, even if the work itself is still valuable. Ask which industries employ the job you want.
The big picture for 2025-35
Three assumptions shape the headline. BLS assumes the economy reaches full employment in the target year, meaning employment roughly matches the labor force minus ordinary job-switching unemployment. It assumes productivity and technological progress follow historical patterns. And it relies on the Census population outlook, where an aging population holds back labor force growth.
Indeed Hiring Lab’s Cory Stahle, quoted by SHRM, pointed to that aging population and shrinking labor supply as the main reason growth slows. That is an analyst’s reading, not a BLS statement, but it fits the method: the labor force is the first input.
By sector, private healthcare and social assistance is projected to grow 9.5 percent and add more than 2.2 million jobs, about 37 percent of all new jobs. Utilities grows fastest in percentage terms at 9.8 percent, but that adds only 58,800 jobs because the sector starts small. Professional, scientific and technical services adds 926,700 jobs.
Occupational groups: winners and decliners
The table below uses BLS figures for major occupational groups. Percent change is across 2025 to 2035.
| Group | Projected change | Jobs added or lost |
|---|---|---|
| Healthcare support | +13.3% | about +1.12 million |
| Healthcare practitioners and technical | +8.0% | about +820,000 |
| Life, physical and social science | +7.5% | about +123,000 |
| Community and social service | +7.4% | about +243,000 |
| Computer and mathematical | +7.3% | about +403,000 |
| Management | +6.2% | about +844,000 |
| Sales and related | -1.4% | about -207,000 |
| Office and administrative support | -4.0% | about -752,000 |
Office and administrative support has the steepest decline of any group. BLS’s Economics Daily write-up attributes it to automation, including AI. Still, this group began with about 18.8 million jobs, so even after a 4 percent drop it remains one of the largest in the economy.
Pay differs a lot too. For all occupations, the 2025 median annual wage was $50,980. Management had the highest group median at $126,520, followed by computer and mathematical at $109,280. Food preparation and serving was lowest at $35,050. These wage figures cover wage and salary workers only, not the self-employed.
Percent growth versus jobs added
The “fastest growing” list and the “most new jobs” list tell different stories, and many readers only see one.
| Measure | Top examples (2025-35) |
|---|---|
| Fastest percent growth | Nurse practitioners (41.0%), solar photovoltaic installers (36.5%), data scientists (34.6%), wind turbine service technicians (29.5%), medical and health services managers (24.2%) |
| Most new jobs | Home health and personal care aides (+847,300), stockers and order fillers (+250,700), fast food and counter workers (+223,300), registered nurses (+194,700), general and operations managers (+181,300) |
Look at the data scientists row. They are projected to grow about 34.6 percent, from roughly 275,600 jobs to 371,000. That is about 95,400 jobs. Home health and personal care aides grow a slower 18.1 percent, yet add nearly nine times as many jobs because the starting base is 4.7 million.
Neither number is “better.” Percentages show momentum in a niche. Counts show how many positions exist. If you want the broadest set of openings, look at counts. If you want a field that is expanding quickly, look at percentages, and then check how big it really is.
Growth is not the same as openings
This is where most readers go wrong. Net growth counts only the change in the number of jobs between 2025 and 2035. But employers hire constantly to replace people who leave.
The OOH registered nurse profile shows the difference. Employment is projected to rise by about 194,700 jobs, from roughly 3.47 million to 3.66 million, which BLS calls faster than average. Yet the profile lists about 180,800 openings every year. Most of those openings come from nurses who transfer to other occupations or leave the labor force, such as to retire.
That logic applies in reverse as well. An occupation can be shrinking and still hire thousands of people each year, because retirements and career changes keep opening seats. Declining does not mean closed.
Don’t read “decline” as “no jobs”
A negative projection means fewer total positions in ten years. It does not mean employers stop hiring. Check annual openings, and consider whether the field is still worth entering given your skills and local demand.
A quick example
Imagine two occupations. Occupation A is small and projected to grow 30 percent, adding 20,000 jobs. Occupation B is large and projected to grow 1 percent, adding 5,000 jobs, but 150,000 workers leave each year through retirement or career change. For a first-year job seeker, B might offer more entry points, even though A looks more exciting in a headline. This example is illustrative, not real data. Use the actual openings column in the OOH to run your own comparison.
What the AI exposure data does and doesn’t say
The 2026 release adds something new: an AI exposure rating for occupations, sorted into Low, Moderate, High and Very high. BLS built it from five outside data sources, three that estimate which tasks AI could assist or complete and two that track observed use of AI tools. Occupations are ranked relative to each other, then grouped.
BLS is blunt about the limits. The categories are not a forecast of growth, decline, wages or adoption. They don’t separate automation from helping a worker. They reflect AI capabilities as of mid-2023 for the theoretical measures, and mostly cover language models.
The employment projections themselves assume technology follows historical patterns. BLS says that if AI raised productivity evenly across all industries, projected employment would not change, and it lacks data to model uneven effects. So the projections are not an AI forecast.
SHRM’s coverage of the release reported that across 831 occupations there was no relationship between AI exposure and projected growth or decline. Appcast’s Julius Probst warned that AI could still make it harder for junior workers to enter fields like software development. Adecco’s Amy Glaser said employers are seeing more job transformation than elimination. Those are industry views, and they conflict in emphasis, so treat them as perspectives rather than facts.
What can you do with this? Look at the tasks in a job, not only the label. Ask what the work involves day to day, and which parts a tool could speed up. Then build skills that complement the tools.
How to read a profile in the Occupational Outlook Handbook
Open any OOH profile and you’ll see a few key fields. Here is what to do with each.
| Field | How to use it |
|---|---|
| Summary outlook percent | Compare it with the all-occupations average (3 percent for 2025-35 in the handbook), but also look at the job count |
| Employment change | Shows how many jobs are added or lost, not just the percent |
| Annual openings | Best clue to how many people get hired each year, including replacements |
| Median pay | The midpoint wage in 2025, so half earn more and half less. Entry pay is usually lower |
| Typical entry education and experience | Tells you what employers usually ask for before you apply |
| Work environment and similar occupations | Helps you spot adjacent paths if the first choice is crowded |
The registered nurse profile, for example, shows a 2025 median pay of $97,550 and lists a bachelor’s degree as the typical entry education. Median pay is a snapshot of all nurses, including veterans, not a starting offer.
Myth versus fact
Myth: A shrinking occupation is a dead end. Fact: Office and administrative support is projected to lose about 752,000 jobs, yet it employed roughly 18.8 million people in 2025. Turnover alone creates plenty of hiring. The risk is different: the entry-level tasks may change, so the skills you bring matter more.
Myth: The fastest-growing list is where the jobs are. Fact: Nurse practitioners top the percent list at 41 percent, but that is about 138,000 added jobs. Home health and personal care aides add roughly 847,000. If your goal is simply to find work soon, the second list says more about how many seats exist.
Myth: BLS knows what will happen. Fact: The agency says the projections are not intended to be a forecast of what the future will be. They show what follows if assumptions hold. Recessions, policy shifts, new technology or a change in immigration or birth rates would move the results.
Myth: Healthcare means every healthcare job is safe. Fact: Healthcare support and practitioner groups grow fast, and together account for almost a third of new jobs. But individual roles, regions and employers vary, and many require licenses or years of training.
A worked example using real figures
Suppose you’re weighing data work against healthcare support. Here is how the pieces from the BLS tables fit together.
- Data scientists: about 275,600 jobs in 2025, projected to reach about 371,000. That is about 34.6 percent growth, a small occupation expanding quickly. Entry usually calls for advanced skills, and competition for junior roles can be stiff.
- Home health and personal care aides: about 4.68 million jobs, projected to add about 847,000. Entry barriers are generally lower, though pay is too. Check the OOH profile for the actual median pay.
- Registered nurses: a bigger commitment in education, but about 180,800 openings per year and a $97,550 median.
None of these is a recommendation. The point is the comparison: three occupations with very different sizes, entry routes and pay, all of which look good under a broad label like “healthcare” or “tech.” Projections help you ask the right questions, and the rest of the work is yours.
Why revisions happen
Each year BLS reruns the model with new data: updated Census population estimates, new survey results on employment by occupation and industry, and revised assumptions. That is why an older table will not match this year’s. If you saw a number on a blog or a video, check the base year before you cite it.
Common mistakes
- Treating the label as a job offer. “Much faster than average” says nothing about openings in your town or your odds of landing one.
- Chasing percentages from tiny bases. Solar installers and wind technicians are among the fastest growing, yet the electric power sector adds few jobs overall.
- Ignoring openings. A shrinking occupation can still offer steady hiring, and a fast-growing one can be hard to enter.
- Skipping entry requirements. Licenses, degrees and clinical hours don’t disappear because demand is high.
- Using old tables. The program updates yearly. Check that the profile shows 2025-35 data.
- Reading AI exposure as a danger score. BLS says it is not a prediction.
- Choosing only by the numbers. If you would dislike the work, a bright outlook won’t help.
A five-step method for using projections
- List three occupations you could see yourself doing. Include one you know little about.
- Record five fields for each from the OOH: percent change, employment change, annual openings, median pay and typical entry education.
- Read the work description and ask whether the daily tasks appeal to you.
- Check your state. State labor market agencies publish their own projections, and CareerOneStop links to them. National figures can hide big regional gaps.
- Test with live postings. Search for jobs near you, note the requirements, pay ranges and how many are open. Our guides to the JOLTS openings report and ghost jobs can help you judge what you find.
Then talk to a person who works in the field. Two informational chats will tell you more than any table. If asking feels awkward, see our tips on asking for a referral without feeling awkward.
A short message to a worker in the field
Hi Dana, I’m exploring careers in health information and noticed you work in the field. Government data suggests demand is growing, but I’d like to hear what it’s like in practice. Could I ask you three quick questions over a 15-minute call? I’m happy to work around your schedule. Thank you for considering it.
Projections versus current data
The projections look ahead. Other reports show what is happening now. The monthly jobs report tells you how many jobs were added last month, and what the unemployment rate measures explains the headline number most people quote. Use them together.
A good rule: projections tell you which direction the tide is moving over a decade, while current reports and local postings tell you what the water looks like this month. You need both when you decide on training that takes a year or two.
The bottom line
BLS projections are a structured way to compare occupations under clear assumptions. They are not a promise, and BLS itself advises focusing on the direction and relative size of changes, not precise values.
Your next steps:
- Pick three occupations and pull their OOH profiles this week.
- Compare openings and entry requirements, not only the growth label.
- Look up your state’s projections and search real postings nearby.
- Speak to one working professional before you commit to training.
- Revisit the numbers each year, since BLS updates them.
