Social Workers
MSW graduates grew for a decade and then leveled off, but there are still more than HRSA assumes. HRSA projects the largest gaps in mental health and healthcare social work.
WatchOne source points to a shortage. The others can't tell or don't apply. See each source · How we decide
HRSA projects a shortage. Why isn't this labeled one?
HRSA projects that in 2038 the supply of social workers will be 95% of what's needed. On its own, that would be a shortage. But HRSA's math assumes 28,724 new social workers start each year, and we count 34,350. With our count in place of HRSA's, supply comes to 112% of need. That's above 100%, so with our count there's no shortage in 2038. The two don't agree, so this source can't tell either way.
We use the word shortage, and the Signal label, only when two separate sources point to one and none shows enough supply. Right now only one does: our comparison with BLS openings. The others can't tell either way, or don't apply to this job.
Three counts, and the one this page uses
All related degrees
54,436
The awards at the levels we reviewed (master's, bachelor's, post-master's certificates, doctorates) in the 2 program codes in the table below, 2023–24. Other levels in the same codes aren't counted here. Some of these lead to other jobs.
Degrees that lead to the license
34,196
Only the programs that lead to a social worker license, 2023–24.
This page's count of new social workers
New practitioners
25,369
New provider IDs (NPIs) that list this job, 2024. An NPI is a registration, not a license. Not everyone in the job gets one, and anyone can list a job code.
How many people finish social worker training each year?
34,196 in 2023–24 (provisional), up 39% from 24,566 in 2013–14.
Show as a table
| Academic year | Graduates |
|---|---|
| 2013–14 | 24,566 |
| 2014–15 | 26,164 |
| 2015–16 | 27,191 |
| 2016–17 | 28,150 |
| 2017–18 | 29,447 |
| 2018–19 | 30,969 |
| 2019–20 | 32,203 |
| 2020–21 | 32,012 |
| 2021–22 | 33,255 |
| 2022–23 | 34,350 |
| 2023–24 | 34,196 |
Is that enough to meet projected need?
HRSA projects social worker supply at 95% of need in 2038. Using the new social workers we count instead of HRSA's assumption, our estimate is 112%.
Projected supply as a share of need
100% means supply meets need.
Newly trained each year vs. projected openings
Openings include people retiring or leaving the field. Some BSW graduates go on to earn an MSW, so the two graduate counts overlap over time.
Demand
Four kinds of demand: students who want to train, employers who want workers, people who need the service, and changes that could raise need. Where HRSA's projected need and BLS's projected openings fit this job, they're in Supply and need and count toward the label. Pay counts for some jobs, and its card says so when it does. Nothing else here does. Each card links to the full numbers.
Students who want to train
Counts from the groups that run each profession's programs or applications. They count students and applications, not graduates.
CSWE, 2023–2024 Social Work Education in the United States Observed · CSWE, national totals entered by hand
| Count | 2023–24 |
|---|---|
| Students enrolled, all years | 55,935 |
- Enrollment counts every year of the program, not only new students.
- What each count covers: MSW students at the 225 programs that answered.
Employers
Projected openings from BLS are in Supply and need and count toward the label.
People who need the service
HRSA's projected need, in Supply and need, is built from population and rates of care. It counts toward the label. The figures below give context. They aren't HRSA's inputs, and they don't change the label.
How many people may need this service, and how that's changing. Context only: these counts don't say how many workers are needed, and they don't change the label.
Census Bureau population estimates Observed · Census population estimates
| Count | 2019 | 2020 | 2021 | 2022 | 2023 | 2024 |
|---|---|---|---|---|---|---|
| People 65 and older | 54.0 million | 54.5 million | 55.9 million | 57.5 million | 59.3 million | 61.2 million |
| Share who are 65 and older | 16.5% | 16.4% | 16.8% | 17.2% | 17.6% | 18.0% |
- Everyone living in the U.S. on July 1 of each year. Estimates, not a count: Census updates past years with each new release.
Census Bureau 2023 national projections, middle series Projection · Census 2023 national projections
| Count | 2025 | 2030 | 2035 | 2038 | 2040 |
|---|---|---|---|---|---|
| People 65 and older | 63.3 million | 71.2 million | 75.8 million | 77.7 million | 78.3 million |
| Share who are 65 and older | 18.7% | 20.6% | 21.6% | 22.0% | 22.0% |
- The projections start from 2022 and assume a path for births, deaths, and immigration. Newer estimates are higher, so the projected total for 2025 is below the 2024 estimate. Compare years within this table, not with the estimates above.
CMS Medicare Monthly Enrollment Observed · CMS
| Count | 2020 | 2021 | 2022 | 2023 | 2024 | 2025 |
|---|---|---|---|---|---|---|
| Enrolled in Medicare | 62.8 million | 63.9 million | 65.1 million | 66.5 million | 68.0 million | 69.5 million |
| Of them, enrolled because of age (65 and older) | 54.5 million | 55.9 million | 57.4 million | 59.0 million | 60.8 million | 62.5 million |
- CMS's yearly figure for each year. The rest of Medicare enrollees qualify because of a disability or kidney failure.
CMS Medicaid and CHIP enrollment (Performance Indicator data) Observed · CMS
| Count | 2020 | 2021 | 2022 | 2023 | 2024 | 2025 |
|---|---|---|---|---|---|---|
| Enrolled in Medicaid or CHIP, December | 80.2 million | 86.7 million | 92.8 million | 85.8 million | 79.2 million | 76.2 million |
- The sum of the 50 states and DC, from each state's final December report. Enrollment rose during the pandemic, when states kept people enrolled, and fell after that ended in 2023.
SAMHSA National Survey on Drug Use and Health Observed · SAMHSA NSDUH (survey estimate)
| Count | U.S. |
|---|---|
| Adults with any mental illness in the past year | 23.1% (22.7% to 23.6%), 2023–2024 |
| People 12 and older with a substance use disorder in the past year | 16.9% (16.6% to 17.3%), 2023–2024 |
| Adults who got mental health treatment in the past year | 22.9% (22.3% to 23.5%), 2024 |
- Survey estimates, with 95% intervals. A rate of need isn't a count of people who will seek care, and it doesn't say how many workers are needed.
What's changing
- Federal grants to hire school mental health staff. Status: Stalled. We watch for: A final ruling, or an appeals court decision, on the June 2026 terminations, by June 30, 2027.
Research records from Emerging demand. They don't change the label.
Other evidence
Sources we checked
Every source on this page, what it says for this job, and whether the label uses it. The rules are on How we measure.
| Source | What it says for this job | Used in the label? |
|---|---|---|
| HRSA projection | HRSA's own 2038 figure: 95% of need, so HRSA projects a shortage. HRSA with our count: 112%. Our reading: HRSA's own figure and HRSA with our count fall on different sides of 100%, so this test can't tell either way. | Yes. Can't tell either way |
| BLS openings | Our comparison: 69,100 projected openings a year against graduates together (52,534). Openings are 32% above our highest count. That is more than 4%, the largest yearly change in our count over the last 5 years, so the gap is bigger than the count's normal movement. BLS revises its projections each year. In its last 5 releases (2021–31 to 2025–35), openings ranged from 63,800 to 74,700 a year, and the gap holds in each one. | Yes. Shows a shortage |
| Pay (BLS OEWS, after inflation) | Median pay after inflation, May 2020 to May 2025: −5% for this job, −3% for all jobs. | No. BLS's pay group is broader than this job, so it isn't used. |
| Who works in the job (Census ACS) | 21% age 55 or older, 11% under 35 hours, 1.9% unemployed; all workers: 23%, 20%, 5.2%. | No. Context only |
| Job postings (Indeed) | Community & Social Service sector at 105 in October 2026 (February 2020 = 100), 112 a year earlier; all postings 104. | No. Context only |
| Where graduates work (Census PSEO) | 80% of graduates with master's degrees in public administration and social services worked in their school's state a year later (2016 to 2020). The whole field, not only this job. | No. Context only |
| CSWE, 2023–2024 Social Work Education in the United States | Students enrolled, all years: 55,935 (2023–24). | No. Shown only |
| Census Bureau population estimates | People 65 and older: 61.2 million (2024); Share who are 65 and older: 18.0% (2024). | No. Shown only |
| Census Bureau 2023 national projections, middle series | People 65 and older: 78.3 million (2040); Share who are 65 and older: 22.0% (2040). | No. Shown only |
| CMS Medicare Monthly Enrollment | Enrolled in Medicare: 69.5 million (2025); Of them, enrolled because of age (65 and older): 62.5 million (2025). | No. Shown only |
| CMS Medicaid and CHIP enrollment (Performance Indicator data) | Enrolled in Medicaid or CHIP, December: 76.2 million (2025). | No. Shown only |
| SAMHSA National Survey on Drug Use and Health | Adults with any mental illness in the past year: 23.1% (22.7% to 23.6%), 2023–2024; People 12 and older with a substance use disorder in the past year: 16.9% (16.6% to 17.3%), 2023–2024; Adults who got mental health treatment in the past year: 22.9% (22.3% to 23.5%), 2024. | No. Shown only |
| Emerging demand: Federal grants to hire school mental health staff | Status: Stalled. | No. Research record |
What we don't know yet
- Which MSW programs are CSWE-accredited, school by school. That needs CSWE's permission to use its list. Needs permission
- How many MSW graduates pass the licensing exam, by school. ASWB publishes national pass rates, and we use those, with credit. Its pass rates by program go to schools only. Nobody publishes a national count of new licenses. Fourteen states publish license files or yearly counts we can use. Those counts are under By state. Working on it
- Which jobs graduates take. HRSA splits them among child and family, healthcare, and mental health social work by current employment shares. No public source
- Many social workers never bill insurance, so new NPIs undercount new social workers. No public source
More about this job
Questions and answers
The short version of this page.
Is there a social worker shortage?
One source points to a shortage. The others can't tell or don't apply. MSW graduates grew for a decade and then leveled off, but there are still more than HRSA assumes. HRSA projects the largest gaps in mental health and healthcare social work.
What do we know about the supply of social workers?
- 34,196 people finished training programs in 2023–24 (provisional: schools can still correct it). That is up 39% from 24,566 in 2013–14.
- HRSA projects that supply will be 95% of need in 2038, if people keep using care at today's rates.
- HRSA with our count: HRSA's model assumes 28,724 new social workers start each year. We count 34,350. Part of the gap with HRSA's assumption is HRSA's own choice: its count of new graduates leaves out training positions funded by the federal Behavioral Health Workforce Education and Training (BHWET) program. We add the difference for each year from 2024 to 2038, assume 85% of those people are still working in 2038, and hold HRSA's other assumptions fixed. That puts 2038 supply at 112% of need, compared with HRSA's 95%. That is on the other side of 100% from HRSA's own figure, so it doesn't show a shortage or enough supply.
- BLS projects 69,100 openings a year from 2025 to 2035. That is more than graduates at all levels together (52,534). Openings are jobs to fill as people leave or retire, not jobs sitting empty today.
- We also label each state from its own data. Of the 50 states and DC, 33 are Watch, 8 show no shortage, and 10 have not enough evidence. Watch: Arizona, California, Colorado, Connecticut, Delaware, and 28 more.
Cite as: Scarcity, “Social Workers: supply pipeline,” release 2026-11-17, data as of November 17, 2026. https://scarcity.co/professions/social-workers/
For a reference manager: BibTeX · RIS
The data files for this release stay at /releases/2026-11-17/, so a citation can be checked later.
Corrections and data sources
Corrections: Found an error? Email corrections@scarcity.co. We fix errors and log every change, with the old and new numbers, in the page history. NCES scheduled 2024–25 completions for July 2026. They had not been posted as of November 2, 2026. We check for new IPEDS, HRSA, and BLS files on the 1st of each month and update NPI counts by hand.
Data: U.S. Department of Education, NCES, IPEDS; U.S. Census Bureau, Vintage 2024 population estimates; CSWE 2023–2024 Annual Survey of Social Work Programs; ASWB exam pass rates, 2025; California Board of Behavioral Sciences, Exam Results by School reports 2018–2024; HRSA NCHWA workforce projections (FY2025) and technical documentation (behavioral health, Exhibit VIII-1); BLS Employment Projections 2025–35; CMS NPPES NPI file, August 2026, via the NBER mirror; HRSA mental health HPSA designations, November 2026. Pulled November 2026.
Download and cite
The evidence, point by point
2038 by HRSA group
In schools today
What we know
The facts behind the label, from the data on this page. Each one is tagged by kind.
For schools reviewing or planning a program
How many people finish these programs each year, and where. This is training output: it shows the programs already running, not student demand or whether a new program would fill. Each number links to the details.
This career at a glance
The main facts for someone deciding whether to train for this job. Each one links to the details.
Compare this job with others you could train for, by pay in your state and years of school
If you're considering this career
What the evidence shows
In your state
Getting licensed
What this data can't tell you
Pay, cost, earnings, and training What each figure at the top of the page means, and where it comes from
Pay
Cost
Earnings and debt after graduating
How long training takes
By state
Every state
New licenses from state boards
New licenses in the states that publish them
Associate and pre-license credentials
New licenses in states that publish yearly counts
Notes on each state
Texas: people licensed today, by the year they got their license
Mental health shortage areas by county
Counties HRSA designates as short of mental health care. These designations are based mostly on the number of psychiatrists per person, not on this profession, so they show where mental health care is hard to reach. Hover a county for its name.
Loading county map…
Largest programs
Compare schools
Every school on this page, with size, how it offers the program, its tuition and fees, and what graduates earned and borrowed. Select a column heading to sort.
Download the school list (CSV) · Every download for this job
The table is wider than the screen. Scroll it sideways to see every column.
Programs over time
The usual path. Exact requirements vary by state, so check with the licensing board in the state where you plan to work.
Steps to a license
The job market
How we know this
What we don't know yet, the data behind the page, and where every number comes from.
What we don't know yet
No public source means nobody publishes it, so we can't fix it. Needs permission means another organization has it and we've asked or are asking. Working on it means we can build it.
What would change this picture
How not to use these numbers
Terms used on this page
IPEDS is the federal survey every college that takes student aid fills out. A new NPI is the ID a provider gets to bill insurance, so it marks someone starting to practice. HRSA is the federal health workforce agency that projects supply and need. BLS is the Bureau of Labor Statistics, which projects job openings. A CIP code is the federal code a school uses to report each program; the codes we count are listed under How we know this. Every term is in the glossary.
Data details
Download the data
License
License: Scarcity's own data on this page (our counts, matches, and estimates) is free to reuse under CC BY 4.0. Credit "Scarcity" and link to this page. Federal data (IPEDS, Census, BLS, HRSA, CMS) is public domain. Figures from other organizations are cited to them and stay under their terms.
Codes we counted and left out
Coverage check
Pay, from BLS
U.S.
By state
U.S. median by year
Change after inflation
Change after inflation, by state
Where graduates work, from Census
By the school's state
Who works in this job now, from Census
By state, where people live
Job postings, from Indeed
See the numbers as a table
Published tuition, from IPEDS
Earnings and debt after graduating, from College Scorecard
All programs on this page
Largest programs
Training length, from accreditors and certifying bodies
Each requirement comes from the group that accredits programs or certifies people in this field, or from an association of state licensing boards. We download each source and check that the quote and every number are in it. States set their own rules, and they often differ. State detail is coming with our licensure guides.
From graduate to practice
Enrollment in CACREP-accredited programs, every report
Checks we ran on each source
Supply and need numbers
Top 15 schools as a table
Page history
Method and limits
Other professions
Each page uses the same method.