Care businesses do not all survive at the same rate, and the gap between the two halves of the sector is large enough to change how a new operator should think about structure, lending and exit. ONS Business Demography data, which tracks every UK business birth cohort by year and measures what share are still trading one, two, three, four and five years later, shows residential care companies surviving at a materially lower rate than non-residential social work and domiciliary-type providers. This page works through the full curve, the sub-sector breakdown, and what the pattern means in practice.1
The headline five-year numbers
Of the 825 residential care businesses (ONS SIC group 87) born in 2019, 45.5% were still trading five years later. Of the 2,135 non-residential social work businesses (SIC group 88, which includes domiciliary care and supported-living-type provision) born the same year, 63.9% survived to the same point. That is an 18.4 percentage point gap between the two halves of the care sector, using the only birth cohort (2019) with a complete five-year curve published so far.2
Combined, 58.8% of the 2,960 residential and non-residential care businesses born in 2019 survived to year five. The combined figure sums raw births and survivor counts across both groups before dividing, so it is a true weighted figure rather than a simple average of the two percentages.
Where the attrition actually happens
Both groups start close together and pull apart as the cohort ages. First-year survival is high for both (96.3% combined), which means the risk that shows up in this data is not a start-up cliff-edge in year one; it is a widening gap that opens from year two onward and keeps widening every year to five.
| Years since birth | Residential care (SIC 87) | Non-residential social work (SIC 88) | Combined |
|---|---|---|---|
| 1 | 94.5% | 97.0% | 96.3% |
| 2 | 77.6% | 85.7% | 83.4% |
| 3 | 63.0% | 76.1% | 72.5% |
| 4 | 50.9% | 69.6% | 64.4% |
| 5 | 45.5% | 63.9% | 58.8% |
The gap between the two lines is 2.5 points at year one, 8.1 points at year two, 13.1 points at year three, 18.7 points at year four and 18.4 points at year five. Nearly all of the separation between residential and non-residential care survival builds up in years two through four, the period after initial trading is established but before the business has had time to become a mature, stable operation.
Residential care is not one thing: the sub-sector spread
Averaging all residential care into a single 45.5% figure hides a wide internal spread. Within the 2019 birth cohort, five-year survival by residential sub-type ranges from 36.8% to 54.9%, an 18.1 point spread, almost as wide as the gap between residential and non-residential care as a whole.
| Residential care sub-type (SIC code) | 2019 births | 5-year survival |
|---|---|---|
| Nursing care (871) | 190 | 36.8% |
| Learning disability, mental health and substance misuse residential care (872) | 180 | 44.4% |
| Residential care for the elderly and disabled (873) | 200 | 42.5% |
| Other residential care activities (879) | 255 | 54.9% |
Nursing care carries the lowest five-year survival of any sub-type in the dataset, residential or non-residential, at 36.8%. Other residential care (879), a broader catch-all category, has the highest survival within residential care at 54.9%. The dataset records the outcome, not the cause: it does not attribute the nursing-care figure to any specific driver such as staffing intensity, clinical registration complexity or capital cost, and this page does not speculate beyond what the data shows.
Non-residential care is far more uniform
Non-residential social work shows a much tighter internal spread than residential care. The two sub-types with enough 2019-cohort data to compare sit within half a point of each other at five years.
| Non-residential social work sub-type (SIC code) | 2019 births | 5-year survival |
|---|---|---|
| Social work without accommodation for the elderly and disabled (881) | 385 | 63.6% |
| Other social work without accommodation, including supported living (889) | 1,750 | 64.0% |
The 0.4 point spread between these two sub-types is small next to the 18.1 point spread inside residential care. On this data, which specific corner of non-residential social work a business sits in matters far less to its survival odds than which corner of residential care it sits in.
Is the gap new, or consistent across years?
The 2019 cohort is the only one with a full five-year curve, but ONS has published at least a one-year figure for every cohort from 2019 through 2023. The residential-versus-non-residential gap shows up in every single one of those five years, at the one-year mark, not just in the single cohort that has aged all the way to five years.
| Birth year | Residential care (SIC 87) 1-year survival | Non-residential social work (SIC 88) 1-year survival |
|---|---|---|
| 2019 | 94.5% | 97.0% |
| 2020 | 95.3% | 97.1% |
| 2021 | 94.8% | 97.1% |
| 2022 | 95.1% | 97.4% |
| 2023 | 93.1% | 96.2% |
Residential care's one-year survival has never matched non-residential social work's in any of the five cohorts shown, and both groups post their weakest one-year figure in the most recent (2023) cohort, 93.1% and 96.2% respectively. Neither series shows a clear multi-year trend in either direction; the gap between the two groups is the consistent feature, not a drift over time.
Reading this data correctly
Two points from the methodology behind this index matter before drawing conclusions. First, ONS measures survival at enterprise level: a single care company with multiple registered locations counts once, so this is a different unit of measurement from location-level datasets such as CQC's own registered-locations figures. Second, survival here means the enterprise's VAT or PAYE registration stayed active at each anniversary; it does not mean the same care service, ownership or CQC registration continued unchanged underneath that registration. A business can be a five-year "survivor" on this measure while it has changed registered provider or registered manager in the interim.
SIC is also self-reported at the point of incorporation, and some domiciliary care agencies register under a residential SIC 87 code rather than 88100, so the split between the two headline groups is indicative rather than exact. Counts are control-rounded to base 5 by ONS disclosure rules, so small movements between adjacent cohorts can reflect rounding as much as real change. None of this undermines the scale of the gap between residential and non-residential care, which is far wider than any plausible rounding or misclassification effect, but it is worth knowing before quoting a single decimal point as precise.
What the gap means for structuring, lending and exit
An 18.4 percentage point gap in five-year survival between residential and non-residential care is large enough to be relevant to decisions made well before a business is five years old.
Structuring. A residential care business carries materially different survival characteristics from a domiciliary or supported-living business, even though both sit under the same "care" umbrella in general conversation. Where an operator runs both a residential home and a domiciliary or supported-living arm, keeping the two in separate entities (a common propco/opco or multi-entity group structure in the sector) means each business's financial position is not obscured by pooling a lower-survival and a higher-survival activity together. This also has capital-allowances and business-rates implications covered in our capital allowances for care-home fit-out guide and business rates for care homes guide, both of which already treat residential and non-residential care differently for tax and cost purposes.
Lending. A lender assessing a new residential care home is underwriting against a sector where fewer than half of comparable businesses were still trading five years after opening in the most recent complete cohort. A lender assessing a new domiciliary or supported-living business is underwriting against a sector where close to two-thirds survived the same period, with far less variation by sub-type. This is not a reason to avoid residential care lending, but it is a reason the two should not be modelled on the same risk assumptions, and it strengthens the case for the kind of financial-viability evidence covered in our CQC Financial Viability Statement walkthrough.
Exit planning. A buyer evaluating an established residential care home that has already passed the year-three or year-four mark (the point where most of the residential-versus-non-residential gap has already opened up) is looking at a business that has already cleared the period where the bulk of sector attrition occurs. That survivorship itself is a data point worth weighing in valuation conversations, separate from the underlying trading figures.
Related research
This survival data sits alongside two other proprietary indexes built from public sector data: the UK Care Home Density and Quality Index, which maps CQC-registered locations and ratings, and the Care Provider Business Index, which tracks quarterly formation and dissolution counts by care sub-segment. The full birth-cohort dataset behind this article, including every sub-segment and cohort year referenced above, is published at the UK Care Business Survival Index.
Sources
- UK Care Business Survival Index, Care Home Tax, data compiled from ONS Business Demography reference tables under the Open Government Licence v3.0, pulled 2026-07-23.
- Business Demography 2024 reference tables, Office for National Statistics, released 2025-11-20, Open Government Licence v3.0.