Physical therapist licence

The number the pay figure leaves out

Of the 322,992 unique physical therapists holding an active US licence counted in 2025, federal employment records account for 267,330 working in the occupations it is for — 82.8%. That leaves 55,662 holders it does not account for.

That gap is an upper bound, not an unemployment rate. Holding a credential is not the same as looking for work with it. The gap contains everyone who moved into management, teaching, sales, inspection or a related job that federal data files under a different occupation; everyone who retired or left without surrendering the credential; everyone working part time, agency or self-employed in a way the employment survey does not capture; and everyone holding it as a second string to another job. So the gap is the CEILING on how many trained people are not working in this field — the real figure is lower, and it is not a count of anyone who is unemployed. What the pair does establish is the size of the pool the published pay figure is silent about, which is the thing no pay site shows you.

Both sides are standing counts. Holders on a date against employment in a period — the same kind of quantity, which is the only reason the two may be divided at all.

What that count includes: De-duplicated individuals, which is unusual and is why this pairing is unusually clean: FSBPT matches licence records across jurisdictions, so a therapist licensed in four states is counted once rather than four times. Physical therapist assistants (134,457) are a separate licence and are not in this number. It counts everyone whose licence is active, including people not currently practising.

Denominator: unique physical therapists holding an active US licence = 322,992 (Standing count of people holding the credential, as at 31 December 2025) · FSBPT, Census of Licensed PTs and PTAs in the USA, 2025 · view

Four years after completing — a figure for the programme code, not for this job alone

Of 11,740 people who completed Rehabilitation and therapy, doctoral degree and were no longer studying four years later, 119 had no earnings on record at all — 1.0%. The ones who were working earned a median of $82,893, with the middle half between $68,836 and $95,078.

“No earnings on record” is not “unemployed”. The measure is the absence of any wage record and any positive self-employment income for the whole year. Nobody was asked whether they wanted work, so the count holds people who left the workforce by choice — raising children, caring, travelling, living on someone else's income, moving abroad — alongside everyone who wanted a job and had none. Nobody publishes the split and we will not guess it. What it does establish is how many people finished this training and were earning nothing from it or from anything else four years on, which is exactly what the pay figure above cannot tell you.

And this figure is not specific to this job. The federal government publishes programme outcomes at four-digit granularity and no finer, and this credential's programmes share their code with others. So the figure describes everyone who completed anything under that code at this level, not only the people who went on to do this work.

What that programme code covers: CIP 51.23 covers physical therapy, occupational therapy, therapeutic recreation, art and music therapy and rehabilitation counselling together. The doctoral row is mostly the Doctor of Physical Therapy alongside occupational therapy doctorates; the master's row is mostly occupational therapy and counselling. Physical therapy alone cannot be extracted from any of them, so none of these is the physical therapy figure.

The doctoral level of this code is the Doctor of Physical Therapy together with occupational therapy doctorates. It is the same blending that stops this page carrying a cost figure, so the outcome rate is a rehabilitation-professions figure rather than a physical-therapy one.

Who is in the cohort: completers who received federal student aid — roughly half of US undergraduates, and the other half are not measured here. People who went on to further study are in neither column. The Department suppresses institutions with cohorts too small to publish and adds statistical noise to the rest, so this is a sum over the institutions large enough to report: 150 of them, recovering 89.0% of the Department's own national count for this programme. Dropping the smallest reporting programmes moves the rate by 0.0 percentage points.

Source: US Department of Education, College Scorecard, Field of Study data, 2026 — release of 10 June 2026; measured four years after completion · view · the same figure for every programme we can publish