Scarcity

Social Workers: data and method

WatchOne source points to a shortage. The others can't tell or don't apply. See each source

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.

About half of MSW growth came from schools with fully online MSW programs. Their graduates went from 4,663 to 9,706, 52% of the growth since 2013–14. All other schools grew 23%. The total was flat in 2023–24: 34,350 the year before, 34,196 this year.

Observed · IPEDS34,196graduates in 2023–24, up 39% since 2013–14
Our estimate112%of projected need in 2038, using the graduates we count
HRSA projection95%of projected need in 2038, HRSA's own figure

Latest graduates 2023–24, provisionalTrend from 2013–14Last updated Nov 17, 2026Release 2026-11-17Sources IPEDS, HRSA, BLS, CMS NPI fileDownload the dataCite this pagePage historyMethod

Download and cite

Download the data 16 spreadsheets
  • All 24 professions, one ZIPEvery file below for every profession, plus combined yearly and state files. Also by release under data files by release.
  • Data dictionaryWhat every column in every file here means, its unit, and where it comes from. Start here.
  • Yearly totalsGraduates by year, with the split shown in the small charts.
  • StatesEvery state: graduates, per 100,000 residents, share from online programs, new NPIs, and HRSA 2038 where published.
  • All schoolsEvery school with at least one graduate in any year, by year.
  • Graduates by school, code, and award levelEvery count on this page, one row per school, program code (CIP), award level, and year, first majors only, with the IPEDS file, its release (revised or provisional), and the date we pulled it. The rows add up to the totals on this page.
  • Mental health shortage areas by countyEvery county with an HRSA mental health shortage designation, and what kind (whole county, part of county, population group, or proposed for withdrawal only).
  • Codes counted and left outEvery CIP code and award level we counted or left out, with official titles, counts, and why.
  • New licenses from state license filesLicenses by year first issued, every year, from the Washington, Illinois, Colorado, Connecticut, and Delaware license files. Includes associate credentials and Colorado telehealth-only registrations, marked by kind.
  • New licenses from state board reportsEvery count we read from board reports, every year and level, with the report, page, the exact text, and the result of our check against the downloaded report.
  • Texas licensees by yearPeople holding a Texas license on the day we pulled the list, by license level, status, and the year they got their current level. Every year in the list.
  • Pay by yearEmployment and yearly pay (25th percentile, median, 75th percentile) for this page's BLS job codes and for all jobs, U.S. and each state, every May release since 2003, as BLS printed them, with the file each row came from.
  • Pay change after inflationOur calculation: the change in the median after inflation over 5 and 10 years, for this job and for all jobs, U.S. and each state, and the difference in percentage points. Each median is divided by the May CPI-U, U.S. city average (BLS series CUUR0000SA0).
  • Where graduates work (Census PSEO)Graduates of taking-part schools with master's degrees in public administration and social services (CIP 44), by the school's state and graduating years: how many had a job 1, 5, or 10 years later, and how many of them worked in the same state as their school. Census statewide rows, counts flagged OK only. This is the whole field, not only this job.
  • Who works in this job now (Census ACS)People working in this job and all workers, U.S. and every state, from the Census American Community Survey microdata: how many, and the shares age 55 or older, usually under 35 hours a week, and living in another state a year earlier, each with a standard error from the replicate weights. States with small samples are in the file but not on the page.
  • Job postings (Indeed)Indeed Hiring Lab Job Postings Index for its Community & Social Service sector and for all postings, U.S., monthly average of the daily index, February 2020 = 100. CC BY 4.0, Indeed Hiring Lab. Postings are not openings, and the sector is broader than this job.
  • Programs over timeSchools IPEDS lists as offering a program at this page's codes and award levels, each year, with new and dropped listings. New means the school listed none of them the year before; a school that moves a program to a code this page counts looks new too. Public domain, from the IPEDS Completions programs-offered file.

Graduates each year

34,196 in 2023–24 (provisional), up 39% from 24,566 in 2013–14.

Degrees and certificates awarded, 2013–14 to 2023–24. Academic years run July to June. Every year's count is in the table under the chart.

Observed · IPEDS Completions
10,00020,00030,00040,000’13–1424,566’14–15’15–16’16–17’17–18’18–19’19–20’20–21’21–22’22–23’23–2434,196
Show as a table
Social workers: graduates each year
Academic yearGraduates
2013–1424,566
2014–1526,164
2015–1627,191
2016–1728,150
2017–1829,447
2018–1930,969
2019–2032,203
2020–2132,012
2021–2233,255
2022–2334,350
2023–2434,196

Schools with fully online MSW programs vs. everyone else

Schools that offered a fully online MSW in 2023–24 (all their MSW graduates) compared with all other schools. Same scale on both panels.

Schools with fully online programs grew every year. All other schools peaked at 25,308 and were at 24,490 in 2023–24.

Schools with fully online MSW programs

4,663 → 9,706
20,00040,0002013–142023–24
Show as a table
Schools with fully online MSW programs
Academic yearGraduates
2013–144,663
2014–154,792
2015–165,405
2016–175,745
2017–186,086
2018–196,690
2019–207,353
2020–217,647
2021–228,573
2022–239,042
2023–249,706

All other schools

19,903 → 24,490
20,00040,0002013–142023–24
Show as a table
All other schools
Academic yearGraduates
2013–1419,903
2014–1521,372
2015–1621,786
2016–1722,405
2017–1823,361
2018–1924,279
2019–2024,850
2020–2124,365
2021–2224,682
2022–2325,308
2023–2424,490

Supply and 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%.

Bachelor's degrees in social work are falling. They went from 22,366 in 2017–18 to 18,338 in 2023–24, down 18%. BSW graduates are one source of MSW students, often through one-year advanced standing programs, and many take social work jobs that don't need an MSW.

The evidence, point by point
  • HRSA projects social workers as three groups.
  • Added together, supply is 101% of need in 2025 and 95% in 2038.
  • The groups differ: by 2038, child, family, and school social workers are at 103%, healthcare social workers 91%, and mental health and substance use social workers 85%.
  • HRSA assumes 28,724 new social workers a year; we see 34,350 MSW degrees in 2022–23.
  • Using that count, 2038 supply would be about 112% of need, or 105% if only half of the difference is still working in 2038.
  • HRSA can't say which group a graduate joins, so the scenario range is for all three together.
  • BLS projects about 69,100 openings a year across all social work jobs, more than the 34,196 MSW and 18,338 BSW degrees in 2023–24 combined.

Status: Watch

Only one source meets our shortage rules (BLS openings). The others can't tell or don't apply. The label follows published criteria, applied the same way to every profession. The criteria and the reasons for them are on the method page.

SourceWhat it showsOur reading
HRSA projection, checked with our countHRSA'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.Can't tell
BLS openings, compared with our count of new social workersOur 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.Shows a shortage
Pay after inflation, compared with all jobs (BLS OEWS)BLS's pay group is broader than this job, so it isn't used.Not used

Projected supply as a share of need

100% means supply meets need.

Bars: projection · HRSABand: our estimate, 2038
What to show for 2038

HRSA with our count: 112% of need in 2038. HRSA's 2038 total, adjusted for the number of new social workers we count, with 85% of the difference still working full time in 2038 and HRSA's other assumptions held fixed. This is our estimate, and it is what the label uses with HRSA's own figure.

0%50%100%125%2025101%2030101%203895%

Shaded band: HRSA with our count, 2038, 112%. It adjusts HRSA's 2038 total for the number of new social workers we count instead of HRSA's assumption, holding HRSA's other assumptions fixed. It is a sensitivity range, not a confidence interval. If only half of the difference is still working in 2038: 105%. HRSA's need means demand at today's use rates.

2038 by HRSA group

HRSA projects each group separately. HRSA with our count applies only to the total, because IPEDS has one code for all MSWs and HRSA splits new graduates among the groups by current job shares.

Projection · HRSA
Child, family, and school103%Healthcare91%Mental health and substance use85%100%

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.

Openings: projection · BLSGraduates, new provider IDs: observed
Projected openings per year, all social workers(BLS, 2025–35)69,100MSW graduates, 2023–2434,196BSW graduates, 2023–2418,338First-time Clinical exam takers, 2025 (ASWB;75.7% passed)26,439

What we know

The facts behind the label, from the data on this page. Each one is tagged by kind.

  • Observed · IPEDS34,196 people finished training programs in 2023–24 (provisional: schools can still correct it). That is up 39% from 24,566 in 2013–14.
  • Projection · HRSAHRSA projects that supply will be 95% of need in 2038, if people keep using care at today's rates.
  • Our estimate · Scarcity, from HRSAHRSA 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.
  • Projection · BLSBLS 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.
  • Our estimate · Scarcity, state rulesWe 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, Florida, Georgia, Indiana, Iowa, Maryland, Michigan, Minnesota, Mississippi, Missouri, Nebraska, Nevada, New Hampshire, North Carolina, Ohio, Oklahoma, Oregon, Pennsylvania, Rhode Island, South Carolina, Tennessee, Texas, Vermont, Virginia, Washington, District of Columbia, West Virginia, Wisconsin, Wyoming

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.

  • Needs permissionWhich MSW programs are CSWE-accredited, school by school. That needs CSWE's permission to use its list.
  • Working on itHow 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.
  • No public sourceWhich jobs graduates take. HRSA splits them among child and family, healthcare, and mental health social work by current employment shares.
  • No public sourceMany social workers never bill insurance, so new NPIs undercount new social workers.

What would change this picture

  • Whether MSW degrees keep growing after flat 2023–24 results, as BSW degrees keep falling.
  • ASWB's 2026 exam counts and pass rates.
  • Final 2023–24 data and the first 2024–25 data from IPEDS.

How not to use these numbers

  • Graduate counts are degrees and certificates awarded, not new workers. Some graduates never get licensed or take the job.
  • New provider IDs are registrations, not new workers. Some people register and never bill, and providers pick their own specialty code.
  • BLS projected openings are not unfilled jobs. They include openings from people who leave the job, not only new jobs. State projections come from each state's labor agency and don't add up to the BLS figure.
  • HRSA's need is demand at today's use of care. It isn't a count of unfilled jobs.
  • Graduates count in the state where the school is based. Students in online programs may live anywhere.
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

Data pulled November 17, 2026Years 2013–14 to 2023–24 (latest year provisional)Coverage Complete for MSWsCodes counted CIP 44.0701Main source IPEDS, the federal data on degrees every college reports
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.

Observed: a count from a published sourceProjection: another agency's modelOur estimate: our calculation from observed counts

Where a table has no number, it says why. Not collected: The source doesn't collect or publish this figure. Withheld: The source has the figure but holds it back, usually because the count is small. Not applicable: The question doesn't apply here. Not released yet: The source hasn't published this figure yet. Not yet pulled: The source has it, but Scarcity hasn't added it yet. The gap is ours. Not used: The source has it, but we chose not to use it for this job. The page says why. A number, zero included, is a reported figure. More on the method page.

Codes we counted and left out 1 code counted

Schools choose which code to report each program under. These are the codes and award levels this page counts, and the nearby ones it leaves out, with 2023–24 counts. To check a count, look up the code in IPEDS (first majors only). Code titles are from the CIP 2020 list.

CodeOfficial titleAwards2023–24Why
Counted
44.0701Social Workmaster's34,196Counted on this page
Coverage check Complete for MSWs

IPEDS counts every MSW degree from colleges that take federal aid. CSWE's own survey counts fewer because not every program answers it. ASWB's exam counts fit what you'd expect if most MSW graduates go on to take a licensing exam.

SourceWhat it countsPer year
IPEDS, 2022–23MSW degrees (44.0701, master's)34,350
IPEDS, 2023–24MSW degrees34,196
CSWE annual survey, 2022–23MSW graduates at the 225 programs that responded21,092
ASWB, 2025First-time Masters exams (pass rate)20,636 (73.9%)
ASWB, 2025First-time Clinical exams (pass rate)26,439 (75.7%)
CMS NPI file, 2024New social worker NPIs (clinical and general codes)25,369

Codes and levels not counted here (2023–24)

  • Bachelor's in social work (BSW), 44.0701: 18,338. A BSW leads to a bachelor's-level license in some states and to many social work jobs, but not to clinical licensure.
  • Social work doctorates (DSW and PhD), 44.0701: 1,004. Most doctoral students already hold an MSW.
  • Human services, 44.0000, master's: 842. These don't lead to social work licensure.
  • Post-master's certificates, 44.0701: 56.

The Clinical exam is taken after about two years of supervised work, so 2025 Clinical test-takers mostly finished their MSW around 2022 or 2023. Not every state requires the Masters exam. ASWB counts exams in all its member jurisdictions, which may include Canadian provinces. Many social workers never bill insurance and never get an NPI.

Pay, from BLS
Observed · BLS Occupational Employment and Wage Statistics, May 2025

BLS social worker groups include jobs open to people with a bachelor's degree. This page counts master's graduates. Pay is yearly wages for everyone in the job, at every career stage. A quarter of workers earn less than the 25th percentile and a quarter earn more than the 75th. Withheld means BLS didn't release the figure, usually because too few people in the state hold the job to report it reliably.

U.S., May 2025

BLS groupSOC codeWorkers25th percentileMedian75th percentile
All social workers21-1020775,930$49,230$61,780$79,040
Child, family, and school social workers21-1021392,550$48,270$59,550$76,070
Healthcare social workers21-1022187,630$56,710$67,880$82,240
Mental health and substance abuse social workers21-1023132,810$47,070$60,280$79,730
Social workers, all other21-102962,930$51,900$71,900$97,040

By state, May 2025

StateMedian, child, family, and school social workersMedian, healthcare social workersMedian, mental health and substance abuse social workersMedian, social workers, all other
Alabama$46,660$60,000$41,500$94,270
Alaska$62,630$62,000$56,710$61,220
Arizona$50,150$64,030$46,380$62,150
Arkansas$43,330$55,640$58,590$47,790
California$65,200$89,660$79,000$63,010
Colorado$63,720$63,480$65,620$61,610
Connecticut$78,970$79,900$78,640$69,080
Delaware$48,520$64,620$60,060$93,250
District of Columbia$73,550$94,100$73,310$96,380
Florida$56,550$64,820$51,590$58,550
Georgia$47,510$62,420$52,030$78,260
Hawaii$68,270$85,420$54,390$116,130
Idaho$59,840$74,500$52,520$88,610
Illinois$63,220$64,270$49,960$64,030
Indiana$50,470$64,580$48,300$84,570
Iowa$52,250$63,050$48,530$91,170
Kansas$54,300$64,440$56,330Withheld
Kentucky$59,100$64,080$42,570$85,160
Louisiana$56,800$61,080$47,100$69,230
Maine$63,290$71,610$57,050$60,280
Maryland$76,390$76,440$73,580$75,600
Massachusetts$75,570$76,440$62,500$101,280
Michigan$59,270$67,600$60,830$62,140
Minnesota$67,450$73,970$75,020$79,790
Mississippi$46,530$62,250$48,970$91,560
Missouri$48,520$59,870$49,560$98,740
Montana$56,780$57,090$45,160$51,650
Nebraska$47,770$62,420$43,240$76,520
Nevada$61,080$78,470$62,900$121,830
New Hampshire$71,010$78,260$63,900$103,710
New Jersey$76,280$82,690$66,590$76,640
New Mexico$56,160$65,930$65,640$90,820
New York$65,100$73,070$79,380$84,010
North Carolina$59,610$63,280$58,210$63,490
North Dakota$64,560$62,860$60,980$81,330
Ohio$51,520$63,610$46,080$59,780
Oklahoma$43,950$56,120$43,120$101,860
Oregon$64,130$84,350$68,210$67,820
Pennsylvania$50,820$61,900$48,290$74,430
Puerto Rico$43,620$36,050$26,500$47,580
Rhode Island$69,870$75,390$60,990$111,590
South Carolina$51,800$60,840$39,930$96,350
South Dakota$51,100$57,510$48,920$93,760
Tennessee$56,310$61,240$46,500$51,460
Texas$55,680$69,990$46,630$84,240
Utah$52,000$78,150$50,770$66,440
Vermont$67,340$81,140$73,210$73,070
Virginia$61,330$66,550$64,240$92,930
Washington$74,480$78,570$70,460$102,520
West Virginia$52,220$60,140$50,070$62,590
Wisconsin$59,280$66,910$61,860$62,830
Wyoming$56,420$62,260$59,310$69,020

U.S. median by year

May ofAll social workers
2025$61,780
2024$61,330
2023$58,380
2022$55,350
2021$50,390
2020$51,760
2019$50,470
2018$49,470
2017$47,980
2016$46,890
2015$45,900
2014$45,500
2013$45,050
2012$44,200

Pay in this table is not adjusted for inflation. BLS started this group in 2012, when it changed its job codes, so the years start there. Source: BLS Occupational Employment and Wage Statistics. We keep every file BLS has posted, back to 1997.

Change after inflation

Our calculation · BLS OEWS and CPI-U

The change in the median from May of each starting year to May 2025, in dollars of the same value. We divide each year's median by the May CPI-U, U.S. city average. The difference is this job's change minus the change for all jobs, in percentage points. Pay isn't used for this job's status label. Many low-wage jobs were lost early in the pandemic, which raised the all-jobs median for May 2020, so changes from 2020 start from a high point.

Group5 years, 2020 to 202510 years, 2015 to 2025
All social workers (21-1020)−5%0%
All jobs (00-0000)−3%+4%
Difference−1.7 points−4.6 points

Change after inflation, by state

Change after inflation, 5 years, by state
StateThis jobAll jobsDifference
AlabamaWithheld
AlaskaWithheld
ArizonaWithheld
ArkansasWithheld
CaliforniaWithheld
ColoradoWithheld
ConnecticutWithheld
DelawareWithheld
District of ColumbiaWithheld
FloridaWithheld
GeorgiaWithheld
HawaiiWithheld
IdahoWithheld
IllinoisWithheld
IndianaWithheld
IowaWithheld
KansasWithheld
KentuckyWithheld
LouisianaWithheld
MaineWithheld
MarylandWithheld
MassachusettsWithheld
MichiganWithheld
MinnesotaWithheld
MississippiWithheld
MissouriWithheld
MontanaWithheld
NebraskaWithheld
NevadaWithheld
New HampshireWithheld
New JerseyWithheld
New MexicoWithheld
New YorkWithheld
North CarolinaWithheld
North DakotaWithheld
OhioWithheld
OklahomaWithheld
OregonWithheld
PennsylvaniaWithheld
Puerto RicoWithheld
Rhode IslandWithheld
South CarolinaWithheld
South DakotaWithheld
TennesseeWithheld
TexasWithheld
UtahWithheld
VermontWithheld
VirginiaWithheld
WashingtonWithheld
West VirginiaWithheld
WisconsinWithheld
WyomingWithheld

Withheld means BLS didn't publish the median for that year or state, or put it at its top wage. BLS says its pay survey isn't designed for comparing years, because its methods and job codes change. Small jobs have wide error ranges in state data. Every state uses the national CPI-U, so state figures aren't adjusted for living costs or for prices in the state. Sources: BLS Occupational Employment and Wage Statistics and the BLS Consumer Price Index (CPI-U, May 2015, 2016, 2017, 2018, 2019, 2020, 2021, 2022, 2023, 2024, 2025).

Where graduates work, from Census
Observed · Census Post-Secondary Employment Outcomes (PSEO), release 2026Q2 (V4.14.1)

Census links graduates of the schools that take part to the jobs they hold later. Of graduates with master's degrees in public administration and social services from 2016 to 2020 who had a job a year after graduating, 80% worked in the same state as their school (taking-part schools in 34 states). Five years after graduating, it was 72% (graduates of 2011 to 2015, 34 states).

This is for all master's degrees in public administration and social services, not only this job. Census groups programs by broad field only. 100% of this page's graduates in 2023–24 earned master's degrees in public administration and social services. Only some schools in some states take part, so this doesn't cover all graduates. Graduates Census didn't find working aren't counted. The share is shown only. It doesn't change the status label.

By the school's state

School's stateWorked in the state, 1 year after, of those with a job (2016 to 2020)5 years after (2011 to 2015)
Alabama69% of 1,08971% of 989
Arizona70% of 2,57077% of 1,686
Colorado85% of 1,32078% of 978
Connecticut84% of 1,62178% of 1,190
Georgia78% of 2,95473% of 2,425
Hawaii81% of 49585% of 360
Illinois82% of 7,18271% of 6,076
Indiana68% of 3,19961% of 2,589
Iowa68% of 72078% of 544
Louisiana65% of 1,66281% of 1,356
Maine95% of 34878% of 456
Maryland77% of 3,64065% of 1,140
Massachusetts85% of 1,77479% of 1,325
Michigan63% of 1,90252% of 1,647
Minnesota88% of 3,62582% of 3,258
Missouri70% of 1,75173% of 1,587
Montana94% of 20171% of 214
New Jersey81% of 5,03179% of 3,765
New Mexico55% of 86176% of 326
New York92% of 8,53584% of 7,297
North Carolina88% of 4,05577% of 3,445
Ohio86% of 5,17183% of 4,199
Oklahoma87% of 1,17480% of 883
Oregon80% of 1,40078% of 1,329
Pennsylvania70% of 1,67665% of 1,533
South Carolina64% of 1,43352% of 1,423
South Dakota42% of 29544% of 399
Tennessee74% of 2,02873% of 1,466
Texas88% of 8,91780% of 6,875
Utah88% of 1,58986% of 1,117
Virginia63% of 3,98039% of 7,647
West Virginia77% of 81165% of 570
Wisconsin86% of 2,05879% of 1,631
Wyoming90% of 12858% of 161
All taking-part states80% of 85,19572% of 71,886

States with fewer than 30 graduates with a job, or where Census held back a count, are left out or marked withheld. States that take part, with Census's estimate of the share of the state's graduates covered (2015): Alabama (21%), Arizona (24%), Colorado (73%), Connecticut (69%), District of Columbia (3%), Georgia (43%), Hawaii (72%), Iowa (69%), Idaho (24%), Illinois (69%), Indiana (87%), Louisiana (74%), Massachusetts (32%), Maryland (91%), Maine (54%), Michigan (10%), Minnesota (74%), Missouri (50%), Montana (83%), North Carolina (38%), New Jersey (80%), New Mexico (87%), New York (44%), Ohio (67%), Oklahoma (66%), Oregon (70%), Pennsylvania (16%), Rhode Island (10%), South Carolina (86%), South Dakota (79%), Tennessee (52%), Texas (77%), Utah (74%), Virginia (87%), Vermont (5%), Wisconsin (45%), West Virginia (44%), Wyoming (86%). Online-only schools that take part: Western Governors University. Source: Census Post-Secondary Employment Outcomes.

Who works in this job now, from Census
Observed · Census American Community Survey, 2024 5-Year microdata (survey estimate)

About 818,000 people worked in this job in 2020 to 2024, on average, by the Census survey (a broader group: see below). Here's how they compare with everyone who works.

ShareThis jobAll workers
Age 55 or older20.6% ±0.523.2% ±<0.1
Usually work under 35 hours a week11.5% ±0.420.4% ±<0.1
Lived in another state a year earlier2.3% ±0.22.6% ±<0.1
Unemployed, of people whose last job was this one1.9% ±0.25.2% ±<0.1
  • All social workers, including jobs open to people with a bachelor's degree.
  • Age 55 or older: how many may retire in the next 10 years or so.
  • Under 35 hours: the hours people usually work. It shows part-time work, not people leaving.
  • Lived in another state a year earlier: people who moved between states in the past year, for any reason.
  • Unemployed: people looking for work whose last job, in the past 5 years, was this one, out of them plus everyone working in it. For all workers it also counts only people with a job in the past 5 years, so it's a little lower than the official rate. The 5-year file includes 2020 and 2021, when unemployment was high, so compare the job with all workers rather than with today's rate. Low unemployment fits a shortage, but it isn't proof of one.
  • The numbers after each share are 90% margins of error. These are context only. They don't change the status label.

By state, where people live

Age 55 or older, by state
StateThis jobAll workersDifference
Alabama18.4% ±4.323.0%−4.6 points
Alaska11.2% ±5.420.5%−9.3 points
Arizona19.6% ±3.122.1%−2.5 points
Arkansas19.6% ±5.422.1%−2.5 points
California17.9% ±1.222.0%−4.1 points
Colorado14.3% ±3.020.8%−6.5 points
Connecticut23.7% ±3.827.2%−3.6 points
Delaware18.6% ±5.626.1%−7.5 points
District of Columbia32.2% ±11.016.7%+15.4 points
Florida22.8% ±2.425.3%−2.5 points
Georgia18.4% ±2.521.7%−3.3 points
Hawaii26.0% ±6.725.8%+0.2 points
Idaho18.9% ±7.821.3%−2.4 points
Illinois19.7% ±2.123.3%−3.6 points
Indiana18.7% ±3.423.1%−4.3 points
Iowa17.7% ±4.524.0%−6.3 points
Kansas21.1% ±4.523.5%−2.4 points
Kentucky15.3% ±3.222.2%−6.8 points
Louisiana26.3% ±5.323.0%+3.4 points
Maine32.1% ±7.728.5%+3.6 points
Maryland24.2% ±2.825.5%−1.2 points
Massachusetts25.7% ±2.725.6%+0.1 points
Michigan22.6% ±2.223.7%−1.1 points
Minnesota19.2% ±3.323.5%−4.2 points
Mississippi21.1% ±5.122.7%−1.6 points
Missouri24.3% ±3.623.3%+1.0 points
Montana18.3% ±7.724.4%−6.1 points
Nebraska18.2% ±6.123.8%−5.6 points
Nevada20.7% ±4.821.8%−1.1 points
New Hampshire22.4% ±5.428.6%−6.2 points
New Jersey21.9% ±2.326.3%−4.4 points
New Mexico26.0% ±6.023.4%+2.6 points
New York24.0% ±1.524.8%−0.8 points
North Carolina18.7% ±2.122.7%−4.0 points
North Dakota18.1% ±8.221.9%−3.9 points
Ohio22.7% ±2.224.2%−1.5 points
Oklahoma20.2% ±4.022.3%−2.1 points
Oregon18.0% ±4.221.6%−3.6 points
Pennsylvania20.9% ±2.225.5%−4.6 points
Rhode Island21.8% ±5.826.0%−4.2 points
South Carolina21.5% ±4.223.5%−2.1 points
South Dakota28.5% ±12.725.4%+3.1 points
Tennessee15.0% ±3.022.8%−7.8 points
Texas17.4% ±1.620.3%−2.9 points
Utah15.5% ±3.417.1%−1.6 points
Vermont18.7% ±7.128.8%−10.1 points
Virginia22.5% ±3.023.6%−1.1 points
Washington20.8% ±2.521.1%−0.4 points
West Virginia16.5% ±4.624.1%−7.6 points
Wisconsin18.4% ±3.324.3%−5.9 points
Wyoming20.7% ±8.823.6%−2.8 points

States with fewer than 50 people in this job in the survey sample are left out. State is where people live, not where they work. ACS job group: 211021;211022;211023;211029 (Census SOC codes). Source: Census American Community Survey Public Use Microdata Sample, 2024 5-Year.

Job postings, from Indeed
Observed · Indeed Hiring Lab Job Postings Index (CC BY 4.0)

Indeed groups this job into its Community & Social Service sector, which also covers other jobs (examples Indeed gives: case managers and social workers). In October 2026, postings in that sector were at 105.4, where February 2020 = 100, compared with 111.5 a year earlier. All postings on Indeed were at 103.8, compared with 101.0 a year earlier.

50100150200202120222023202420252026
See the numbers as a table
MonthCommunity & Social ServiceAll postings
October 2026105.4103.8
September 2026104.7103.2
August 2026104.6101.9
July 2026103.1101.6
June 2026101.5100.8
May 2026102.0101.9
April 2026104.4102.7
March 2026109.2103.9
February 2026111.6104.6
January 2026110.0103.3
December 2025111.1102.9
November 2025111.1101.3
October 2025111.5101.0
September 2025112.1102.8
August 2025112.8104.8
July 2025114.5104.3
June 2025114.9105.3
May 2025116.8105.9
April 2025120.3106.8
March 2025121.3107.8
February 2025124.1109.3
January 2025125.4110.2
December 2024126.4110.3
November 2024125.5109.5
October 2024125.8110.1
September 2024129.4112.4
August 2024131.1113.4
July 2024132.7113.7
June 2024132.3114.2
May 2024134.6115.9
April 2024137.6118.5
March 2024138.9118.6
February 2024138.3118.2
January 2024138.9119.4
December 2023140.6121.3
November 2023142.8123.3
October 2023145.6124.9
September 2023148.1127.0
August 2023147.4128.5
July 2023148.3129.4
June 2023153.6131.8
May 2023156.5134.7
April 2023158.7136.0
March 2023158.4135.0
February 2023158.3136.6
January 2023159.2140.6
December 2022161.3144.4
November 2022163.2147.7
October 2022162.2148.5
September 2022160.9150.6
August 2022162.9153.4
July 2022163.8155.7
June 2022164.4158.6
May 2022165.1160.4
April 2022163.6159.8
March 2022163.8160.4
February 2022162.3159.2
January 2022158.7156.5
December 2021158.5157.2
November 2021156.0152.9
October 2021152.7147.0
September 2021148.1143.1
August 2021143.5139.1
July 2021141.9137.4
June 2021135.1133.4
May 2021128.3128.7
April 2021119.2122.0
March 2021108.2110.6
February 2021101.4103.2
January 202197.197.7
December 202095.694.8
November 202094.893.5
October 202091.090.5
September 202084.986.3
August 202081.083.0
July 202077.278.3
June 202069.470.1
May 202064.562.9
April 202072.968.2
March 202096.394.5
February 2020100.2100.0
  • One posting isn't one opening. Some openings are never posted, and some postings cover several openings or stay up after the job is filled.
  • Indeed's sectors are built from job titles and are broader than this job, so the trend is for the sector, not the job.
  • Each month is the average of Indeed's daily index, seasonally adjusted. This is context only. It doesn't change the status label.
  • Source: Indeed Hiring Lab, Job Postings Index, CC BY 4.0. Indeed's data covers postings on Indeed only.
Published graduate tuition, from IPEDS
Observed · IPEDS Institutional Characteristics, 2023–24

The price each school published for a full-time graduate student for the year: average tuition plus required fees, and the price per credit hour for part-time students. It covers graduate students across the school, not this program alone, so a program can cost more or less. It is the list price, not what students pay after grants and aid. Program length differs, so the total cost of a degree depends on how many credits it takes. Source: IPEDS, Institutional Characteristics.

SchoolStateControlIn-state tuition and fees, per yearOut-of-state, per yearPer credit hour, in-state
Walden UniversityMNPrivate for-profit$10,657$10,657$750
University of KentuckyKYPublic$14,330$35,459$722
Fordham UniversityNYPrivate nonprofit$25,954$25,954$1,045
The University of Texas at ArlingtonTXPublic$10,853$25,345$509
Rutgers University-New BrunswickNJPublic$22,348$36,244$826
CUNY Hunter CollegeNYPublic$11,398$20,828$470
University of DenverCOPrivate nonprofit$58,347$58,347$1,612
Columbia University in the City of New YorkNYPrivate nonprofit$56,279$56,279$2,276
Aurora UniversityILPrivate nonprofit$15,363$15,363$618
University of Southern CaliforniaCAPrivate nonprofit$68,128$68,128$2,244
Simmons UniversityMAPrivate nonprofit$28,662$28,662$1,230
New York UniversityNYPrivate nonprofit$41,124$41,124$2,157
University of Michigan-Ann ArborMIPublic$27,524$55,052$1,467
Indiana University-IndianapolisINPublic$10,856$27,992$402
Yeshiva UniversityNYPrivate nonprofit$35,150$35,150$1,250
Earnings and debt after graduating, from College Scorecard
Observed · College Scorecard, each programOur estimate · middle program

Each program's figures are College Scorecard's medians for that school, field, and degree level. Earnings are for graduates who got federal aid and were working and not in school, measured in 2022 and 2023 and shown in 2024 dollars. Debt is the median federal loan debt (Stafford and Grad PLUS) borrowed at that school, for graduates who borrowed. Scorecard holds back figures for small programs. "Middle program" is our calculation: half of the programs with a published figure are above it and half below. It is not a national median of graduates. Release: October 2026. Source: College Scorecard data and its documentation.

All programs on this page

Field and levelProgramsMiddle program, 1 year afterMiddle program, 4 years afterMiddle program, debt
44.07 Social work, Master's339$56,086 (291 published)$62,619 (269)$38,013 (297)

Largest programs

SchoolLevelEarnings, 1 year afterEarnings, 4 years afterDebt
Walden UniversityMaster's$58,916$63,559$68,334
University of KentuckyMaster's$52,636$60,970$30,750
Fordham UniversityMaster's$62,142$70,342$61,000
The University of Texas at ArlingtonMaster's$55,607$62,890$34,166
Rutgers University-New BrunswickMaster's$59,432$66,438$45,168
CUNY Hunter CollegeMaster's$67,802$76,071$41,000
University of DenverMaster's$56,800$65,203$93,505
Columbia University in the City of New YorkMaster's$64,626$75,980$66,998
Aurora UniversityMaster's$55,865$61,710$35,633
University of Southern CaliforniaMaster's$68,115$73,829$114,877
Simmons UniversityMaster's$61,660$69,292$71,750
New York UniversityMaster's$63,502$72,690$97,509
University of Michigan-Ann ArborMaster's$55,399$64,788$41,000
Indiana University-IndianapolisMaster's$53,963$59,903$42,556
Yeshiva UniversityMaster's$55,713$70,744$52,500
Training length, from accreditors and certifying bodies
Observed · Accreditor and certifying body rules

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.

StageRequirementSourceWhat the source says
SchoolAbout 2 years, or about 1 year with a bachelor's in social work.CSWE, Prepare for Your Education"MSW programs last 2 years and include a minimum of 900 hours of supervised field instruction"
During the degreeAt least 900 hours of field education.CSWE, 2022 Educational Policy and Accreditation Standards"a minimum of 900 hours of field education for master's programs"
After the degreeFor a clinical license, 1,500 to 6,400 hours of supervised work after the MSW, depending on the state. 3,000 hours over about 2 years is the most common.ASWB, Clinical social work supervision: Comparison of requirements"Post MSW degree supervised experience is required in all 56 jurisdictions"
From graduate to practice
StageCountSource
MSW degrees, 2022–2334,350IPEDS
First-time Masters exams, 2025 (not required in every state)20,636, 73.9% passedASWB
First-time Clinical exams after supervised work, 202526,439, 75.7% passedASWB
New clinical licenses (LCSW)Not published nationallyEach state board keeps its own count
New social worker NPIs, 202327,000CMS NPI file; many social workers never bill
Checks we ran on each source
SourceCheckResultUsed as evidence
HRSA projectionsAssumed new social workers per year vs. observedAssumes 28,724; we see 34,350 MSW degrees in 2022–23. HRSA with our count, 2038 supply: 112% of need.Partly. The national total depends on how many graduates enter practice, and HRSA's split across the three groups is an assumption.
BLS openings (SOC 21-1020)Occupation matchPartly. BLS counts all social work jobs, including many open to BSW graduates.Yes, compared with BSW and MSW graduates together.
  • Projections. Before a projection counts as evidence, we compare its assumed new entrants with what we observe. If they differ a lot, we show a scenario with our count instead, and the projection counts only if its conclusion holds across that range.
  • Occupation match. Job-opening projections count as evidence only when the occupation matches the profession on the page. A broader group is shown as context, not evidence.
Supply and need numbers
MeasureValue
HRSA: projected supply as a share of demand, 2025 / 2030 / 2038101% / 101% / 95%
HRSA: projected full-time-equivalent supply vs. demand, 2038382,810 vs. 404,590
HRSA: child, family, and school social workers, supply as a share of need, 2038103%
HRSA: healthcare social workers, supply as a share of need, 203891%
HRSA: mental health and substance abuse social workers, supply as a share of need, 203885%
HRSA's assumed new social workers per year vs. MSW degrees we count, 2022–2328,724 vs. 34,350
Rough estimate: HRSA 2038 supply as a share of need, all three groups, using MSW degrees112%
Lower case: only 50% of the difference still working in 2038105%
BLS: average openings per year, 2025–35 (SOC 21-1020, all social workers)69,100
California licensing board (BBS): first-time LCSW Clinical Exam takers, 2024 (pass rate)2,743 (74%)
BSW and MSW degrees, 2023–2418,338 and 34,196

HRSA counts full-time equivalents and splits new graduates among its three social worker groups by current employment shares. BLS counts jobs; its openings include people retiring, leaving, or changing occupations. We count MSW degrees only. Post-master's certificates in social work (about 55 a year) go to people who already have an MSW, so they don't add new social workers. HRSA's need is demand at today's use rates: the number of workers it would take to serve people the way they are served now, as the population grows and ages. It isn't the number it would take to meet all unmet need. The state map and state table show HRSA's state figures as published; state labels also use our state scenario. HRSA's assumed new entrants are graduates. HRSA counts every one as a new worker and doesn't model the licensing exam. Where we know what share pass, the low end of our range counts only those, which is the same as cutting HRSA's count and ours by that share. The high end counts every graduate. The page also shows a lower case. Where an end of the range uses new NPIs, it is an upper bound: NPIs include people who aren't fully licensed yet, and they jump when billing rules change. The scenario range is a rough Scarcity estimate, not HRSA's. It is a sensitivity calculation around HRSA's projection, not a correction and not a rerun of HRSA's model. HRSA simulates each year's entry, work, hours, and exits; we don't. We adjust HRSA's 2038 total and hold HRSA's other assumptions fixed. The range comes from different assumptions; it is not a confidence interval. It adds the difference between the new graduates we count and the number HRSA assumes, for each of the 15 years from 2024 to 2038, and assumes 85% of them are still working full time in 2038. The lower case assumes only 50%. HRSA says its counts come from "2023 IPEDS" without saying whether that means 2022–23 or 2023–24; we use 2022–23. Scenarios on different pages can't be added together, because some people are counted on more than one page. HRSA's behavioral health model leaves out graduates whose training positions are funded by the federal Behavioral Health Workforce Education and Training (BHWET) program, to show what happens if that funding ends. So HRSA's count of newly trained people is lower on purpose, and part of the gap between its count and ours comes from that choice.

Top 15 schools as a table
School2018–192023–24
Walden University
MN · Private for-profit · Fully online
4261,197
University of Kentucky
KY · Public · Partly online
108861
Fordham University
NY · Private nonprofit · Partly online
663718
The University of Texas at Arlington
TX · Public · In person
725672
Rutgers University-New Brunswick
NJ · Public · Partly online
674671
CUNY Hunter College
NY · Public · In person
528587
University of Denver
CO · Private nonprofit · Partly online
294582
Columbia University in the City of New York
NY · Private nonprofit · Partly online
561546
Aurora University
IL · Private nonprofit · Partly online
488518
University of Southern California
CA · Private nonprofit · Partly online
1,367507
Simmons University
MA · Private nonprofit · Partly online
481431
New York University
NY · Private nonprofit · Fully online
533419
University of Michigan-Ann Arbor
MI · Public · Fully online
343417
Indiana University-Indianapolis
IN · Public · Partly online
369396
Yeshiva University
NY · Private nonprofit · Fully online
100377
Page history Last changed Nov 17, 2026
  • Added counts of students who want to train to the Demand section, from CSWE: national totals entered by hand from each report. Other numbers unchanged.
  • Pay is now one of the sources for the label. Here it isn't used for this job. BLS's pay group is broader than this job, so it isn't used. Label unchanged: Watch. No numbers changed.
  • Our method. This page is now five pages: an overview, Career, Schools, States, and Data and method. A bar at the top links them. The overview answers three questions: is there a shortage, how many people finish training, and how that compares with need. No numbers or labels changed.
  • Our method. The page now opens with one plain statement that comes from the label, such as "The evidence points to a shortage." No numbers or labels changed.
  • Our method. Reworded new NPI counts as new provider IDs. Registering for an NPI doesn't prove a license or that someone is working, so the page no longer calls them providers billing insurance. No numbers changed.
  • Our method. Label rules. HRSA now shows a shortage only when HRSA's own 2038 figure and our scenario are both below 100% of need, so a mistake in our count can't create a shortage on its own. We call our range a scenario, not a correction. Signal no longer says the sources are independent, because the HRSA scenario and the BLS comparison use the same counts. The label Noise is now called No shortage shown. States get labels for the 50 states and DC only. No national label changed; 10 state labels changed across professions; see What changed.
  • Our method. Added a note that HRSA's behavioral health model leaves out graduates in training positions funded by the federal BHWET program, so HRSA's count of new people is lower on purpose. No counts or labels changed.
  • The Minnesota label now also uses the Minnesota DEED Job Vacancy Survey. The Minnesota label didn't change. National label unchanged.
  • Added earnings and debt after graduating, from College Scorecard (October 2026 release), for the programs on this page. Other numbers unchanged.
  • Added training length: the degree, school, and supervised hours, each from the accreditor or certifying body, with a link. Other numbers unchanged.
  • Added published graduate tuition and fees for the largest programs, from IPEDS 2023–24. Other numbers unchanged.
  • Added pay from BLS OEWS, May 2025: median, 25th and 75th percentile, for the U.S. and each state (21-1020: median $61,780). Other numbers unchanged.
  • Method notes now say why our corrected 2038 estimate is one number (112% of need), not a range: we have one count of new social workers, MSW degrees. No numbers changed.
  • Added Texas: people holding a Texas license today, by the year they got it, from the Texas Behavioral Health Executive Council's licensee list (LCSW: 2,348 in 2025). Other numbers unchanged.
  • Added a status label: Watch. Only one source meets our shortage rules (BLS openings). The others can't tell or don't apply. The rules are on the method page.
  • State comparisons of graduates now count only in-person and hybrid programs. Graduates of fully online programs are shown on their own line, because those programs enroll students from many states and IPEDS counts them where the school is based. The school view now shows where each online school's fully online graduate students live (IPEDS, fall 2023, whole school). National counts unchanged.
  • Each section now opens with a question and a short answer. Added a state view (pick a state, compare it with one other state and with the U.S.) and a school view (pick a school, compare it with one other school, the national trend, or its share of the U.S. total). No counts changed.
  • The state map and table now compare new NPIs with HRSA's projected need for workers in each state in 2024, instead of residents. HRSA's need reflects each state's population, such as its ages and insurance. No counts changed.
  • Added a table of every code we count and leave out, with official titles, award levels, 2023–24 counts, and why. Also in the downloads as codes.csv.
  • Corrections now use one rule on every page: we compare HRSA's assumed new graduates with the graduates we count, with no 70% cut on our side only, and show a lower case where only half of the difference is still working in 2038. The estimate is now one number, 34,350 MSW degrees in 2022–23 (was 24,045 to 34,350). Corrected 2038 supply changed from 80%–112% to 112% of need; lower case 105%.
  • Added HRSA's 2038 projection for each of its three social worker groups to the supply vs. need section. Counts unchanged.
  • Added a map view: share of each state's residents in an HRSA mental health shortage area (active designations only), and a county map of the same designations. Counts unchanged.
  • Page published: 34,196 MSW degrees in 2023–24, up 39% from 2013–14.
Method and limits
What is counted
Master's degrees (award level 7) in CIP 44.0701 (Social Work), first majors only, at all Title IV institutions.
Left out
Bachelor's degrees (BSW), which lead to bachelor's-level licenses and many social work jobs but not to clinical licensure; doctorates; post-master's certificates; and human services degrees.
Online
A school is in the online panel if it reported a fully online MSW program in 2023–24. All of that school's MSW graduates are counted there, in every year.
New NPIs
Individual providers in the CMS NPI file (August 2026, via the NBER mirror) with the clinical social worker (1041C0700X) or general social worker (104100000X) code, by the year the NPI was issued and the state they practice in.
Benchmarks
CSWE 2023–2024 Annual Survey of Social Work Programs (2022–23 graduates); ASWB 2025 exam pass rates for first-time test-takers.
Shortage areas
HRSA's mental health Health Professional Shortage Area (HPSA) file, downloaded November 2026. We count people in active geographic and population designations (status "Designated") and divide by July 2024 state population. Designations proposed for withdrawal are shown separately because many overlap newer ones; HRSA's own totals include them. Facility designations (clinics, prisons, hospitals) have no population and are left out. Mental health HPSAs are scored mainly on psychiatrists per person, sometimes with other core mental health providers. The county map shows the kind of designation in each county: the whole county, part of it (some census tracts or towns), or a population group such as low-income residents. When a county has more than one kind, it shows the broadest. County boundaries are Census Bureau cartographic boundaries (us-atlas 3.0.1).
Supply and need
HRSA National Center for Health Workforce Analysis projections (FY2025 release): supply and need in full-time equivalents, 2023–2038, nationally (HRSA doesn't publish state numbers for this profession). HRSA's need is demand at today's use rates, not all unmet need. The state map shows HRSA's state figures as published. State labels also use our state scenario. BLS Employment Projections 2025–35: average openings per year, including openings from people retiring or leaving the field.
State projections
Each state's labor market information agency projects employment and average annual openings ten years out (2022–32), with the same method BLS uses nationally. Collected by Projections Central, a U.S. Department of Labor program, in its combined long-term file (last modified August 2026). Openings cover all social work jobs (SOC 21-1021 to 21-1029), including many open to BSW graduates, while graduates here are MSWs. So graduates per opening is low on this page. 51 of 51 states publish this occupation. Openings include people retiring or leaving the field. A projected opening is not an unfilled job. We don't use state projections for the status label.
Online programs
From the IPEDS programs-offered file. Fully online means the school reports the program as offered entirely by distance education. Partly online means some required parts are in person. IPEDS began separating the two in 2019–20.
Per 100,000
Divided by July 2024 state population from the Census Bureau's Vintage 2024 estimates.
Released
2023–24 is the latest year IPEDS has published. It is provisional: schools can still correct it, and NCES will post a final version later. Older years use IPEDS's final, revised data.
  • Completions count awards, not people. One person can earn more than one award in different years.
  • MSW programs take one year for students with a BSW (advanced standing) and two years for others, so graduates in a given year started in different years.
  • Our 2038 scenario is one number, not a range, because we have one count of new social workers: MSW degrees. Pages with two counts (for example graduates and new NPIs) show a range. The lower case, where only half of the difference from HRSA's assumption is still working in 2038, is shown next to it. Until Nov 3, 2026 the low end assumed only 70% of graduates get licensed; the method review dropped that cut on every page, because no source measures it.
  • Completions lag enrollment by two to three years. They show who finished, not who is in the pipeline now.