Open consultation

E: Technical Annex on the Adult Social Care (ASC) Relative Needs Formula (RNF) consultation

Updated 20 June 2025

Applies to England

1. Introduction

The ASC RNF is used to distribute central government ASC grants to local authorities. The ASC RNF was first developed in 2005 to 2006 and has not been updated since the 2013 to 2014 Local Government Finance Settlement (LGFS) - hereafter the “current ASC RNF”. A study published in 2018 proposed a revision of the ASC RNF (hereafter the “2018 RNF”) as part of the social care charging reforms, but these were not taken forward.

The Ministry of Housing, Communities and Local Government (MHCLG) has published a public consultation seeking views on proposed updates to the ASC RNF using improved data and approaches, hereafter the “proposed ASC RNF”. The proposed ASC RNF builds on advancements made in the 2018 RNF. This technical annex sets out the analytical details of the proposed ASC RNF, which could be subject to change depending on the outcome of the questions posed in the consultation.

Section 2 gives a broad overview of the methodological approach of this proposed ASC RNF and explains the key differences to the current ASC RNF. Section 3 provides more details on the low income adjustment (LIA). Section 4 provides a guide on how we calculate the final proposed ASC RNF allocation shares. Section 5 provides the data definitions and sources underlying the calculations we used for the indicative ASC RNF allocation shares published in the supporting Needs Calculator technical document.

2. Proposed ASC RNF methodology overview

The ASC RNF is designed to reflect the relative needs of local authorities providing ASC services by taking account of underlying factors that could explain the local variations in the cost of service delivery.

The proposed ASC RNF is built on a base formula which calculates relative needs amount per capita. A set of adjustment factors are applied to this base formula to reflect other local circumstances. As the ASC RNF only reflects the relative costs of local authority-funded adult social care services, the final ASC RNF outputs are expressed as a proportion of the total. These are referred to as “allocation shares” in this consultation. The proposed base ASC RNF has two components: younger adults (18 to 64 years) and older adults (65 and over). The older adults and younger adults component allocation shares are combined using weights to obtain the overall ASC RNF allocation shares.

The Department of Health and Social Care (DHSC) and MHCLG use the current ASC RNF allocation shares to distribute ASC funding as well as determine formula funding. Depending on the responses to the consultation on the proposed ASC RNF, MHCLG and DHSC will consider how to use this updated ASC RNF to distribute available resources in the future. These formulae do not influence the total amount of funding available.

2.1 Proposed base ASC RNF model

We commissioned academics from the Adult Social Care Research Unit (ASCRU) at the Personal Social Services Research Unit (PSSRU) at the University of Kent to update the base local authority-funded ASC relative needs models for this update. These models aim to estimate the relative gross expenditure per capita for local authority-funded ASC services. This section gives a summary of the models developed by ASCRU-PSSRU. Further details can be found at in the latest ASCRU-PSSRU report on Adult Social Care RNF.

Criteria for local authority-funded ASC services

Broadly, local authorities consider three criteria when assessing whether ASC should be local authority-funded^{[footnote 1]} or not. These are based on impairment needs, financial needs (in relation to the means test for local authority-funded ASC) and, once the first two needs are established, whether there are formal care needs (needs not already being met by a continuing, possibly unpaid, carer, for example).

Components of the proposed base ASC RNF model

ASCRU-PSSRU developed models by investigating underlying factors that drive the local authority differences for these three local authority-funded ASC assessment criteria. They used a utilisation-based approach that infers relative ASC needs from past local authority-funded ASC use patterns.

ASCRU-PSSRU produced separate models for older adults and younger adults given differences in their utilisation patterns and expected cost drivers (for example, younger adults are less likely to have accumulated as much wealth as older adults) for the local authority-funded ASC assessment criteria mentioned above.

Each of the younger adults and older adults-components are made up of two further sub-components for two care settings: community care and home care. Care settings were separated out given their likely differing relationships to cost drivers and the differences in their costs. The ASCRU-PSSRU models aim to estimate the ASC relative needs as expressed in cost terms. In particular, the community care setting model estimates the relative gross expenditure for local authority-funded ASC per capita. The home care setting model estimates the relative number of users per capita as proxies to the relative gross expenditure for local authority-funded ASC per capita.

Use of small area modelling for the proposed base ASC RNF model

To better account for differences within local authorities and to increase the number of model data points, ASCRU-PSSRU used small area modelling at lower layer super output area (LSOA) level^{[footnote 2]}. One challenge of small area modelling is the lack of routine data on local authority-funded ASC use at this level. Thus, ASCRU-PSSRU used survey data on LSOA level ASC use collected in 2012 to 2013 from 48 local authorities and adjusted the data to reflect 2022 to 2023 ASC use. ASCRU-PSSRU used the 2022 to 2023 financial year reported local authority supported clients on long term nursing and residential care as reported in the 2022 to 2023 Adult Social Care Activity and Finance Report (ASC-FR) to rescale the care home data. They used the 2022 to 2023 reported gross current expenditure (GCE) reported on local authority supported community-based care from the same source to rescale the home care data. As the ASC-FR data is reported only at local authority level not at LSOA level, the assumption underlying the rescaling is the changes in ASC use for LSOAs within the same local authority are the same.

Note that in 2023 to 2024, local authorities started reporting data to NHS England for the Adult Social Care Client Level Data (CLD). While this will be an important data source on small area ASC use in the future, it has not yet been made available to use at the time of this update. More information on data collected for the CLD can be found at Adult Social Care Client Level Data - NHS England Digital.

All the data feeding into the model needed to be available at LSOA level (for example, minimal use of data only available at local authority level). In addition, for ease of future updates, ASCRU-PSSRU also only considered data for the model that is frequently updated (for example, data from one-off surveys would not be used).

How non-needs factors that affect ASC use are accounted for in the proposed base ASC RNF model

One key assumption of the utilisation-based approach used for this model is that local authority-funded ASC relative needs patterns are reflected in the relative local authority-funded ASC use patterns. However, other non-needs factors could also drive use patterns and need to be accounted for in the model. A factor we do not consider to be a direct indicator of ASC needs, but that could affect the level at which adults use social care services is the supply or availability of these services, such as the number of care home beds in an area.

ASCRU-PSSRU minimised the confounding effects of two supply factors – the number of care beds in each LSOA^{[footnote 3]} and a proxy to ASC labour workforce in each middle layer super output area (MSOA)^{[footnote 4]} - by including these factors in the proposed base ASC RNF model. If we do not include these non-needs supply indicators in the model, we could incorrectly attribute supply effects on the ASC costs to the model needs indicators. For example, the model could potentially overestimate the effect of the proportion of people aged 65 or over who are 80 or older on the relative ASC costs if the models do not account for the observation that the number of care home beds per capita are higher in areas where there are a higher proportion of people aged 80 or over (in the 65 and over population). To account for potential interdependences between these supply factors and relative ASC local authority costs, ASCRU-PSSRU applied an instrumental variable approach.

Since the above supply factors are non-need factors and local authorities could have some control over them, ASCRU-PSSRU used the sample means of these two supply indicators for all local authorities, (and included the means in the constant terms of the final formulae). This means the supply indicators do not appear as direct relative needs indicators in the ASC RNF even if they are still controlled for in the proposed base RNF model. As a result, these two supply factors will not be directly used to calculate the allocation shares (and hence to allocate funding). We believe this to be the most appropriate way of ensuring the allocations are not distorted by indicators that do not directly drive ASC relative needs.

In addition, the proposed models controlled for local authority fixed effects to account for other non-need and non-supply drivers of relative ASC use patterns. For example, possible local authority policy differences in response to general funding pressures that are non-needs driven or other unobserved differences.

How we use the proposed base ASC RNF model to obtain the relative local authority level gross expenditure per capita estimates

We estimate the pseudo local authority-level ASC gross expenditure per capita by inputting local authority level data into the proposed base ASC RNF for the indicators in the formulae. Note these estimates are only suitable for use to estimate relative costs. While they might appear to estimate absolute costs, the assumptions for the models and indicators included in the models are not designed to estimate absolute ASC costs, only for relative costs. See section 4 for information on how we use these values to calculate the final ASC RNF allocation shares.

2.1.1 Changes from the current ASC RNF

The current ASC RNF and this proposal for the ASC RNF update use the same broad methodological approach. The underlying models to both are small area utilisation-based models and were developed considering factors that could drive the variations in the three local authority-funded ASC assessment criteria (impairment, financial and formal carer needs). ASCRU-PSSRU considered a selection of possible model indicators and selected the ones with strong associations with the relevant utilisation or cost distributions. Both sets of models aimed to remove the effects of selected supply factors on utilisation without directly including these indicators in the final RNF and hence they would not be used directly to calculate allocation shares. See linked reports for further details of the current RNF model and development (younger adults report, older adults report and  Methodology Guide for Adults’ Personal Social Services Relative Needs Formulae 2013 to 2014).

The exact data and factors considered were different for the ASC RNF update due to benefits system and policy changes over time (for example the introduction of Universal Credit to replace previous benefits and tax credits), and newer or improved data availability (for example the Census 2021 data, Department of Work and Pensions (DWP) benefits combinations data that allow public access to data on number of individuals claiming multiple benefits). The models underlying this proposed ASC RNF includes the following key improvements over the current RNF model:

• data at a smaller area; at LSOA level compared to ward-level used in the current RNF model developed in 2005 to 2006. Consequently, this proposed model was based on a higher number of observations, around 12,000 to 13,000 (depending on the component) compared to around 800 in the current ASC RNF
• proposed model developed using newer data, for example using Census 2021 data rather than 2001, small area ASC use survey data from 2013 to 2014 rather than 2003 to 2005, and newer 2022 to 2023 benefits data
• proposed model includes new indicators to better capture the distribution of wealth and impairment needs for older adults, such as interaction terms between the home ownership proportions and the proportion of dwellings in various Council Tax bands, and inclusion of older adults who are still able to claim Disability Living Allowance

The final base local authority-funded ASC RNF indicators for the current and this proposed ASC RNF are in tables 1 and 2. There are similar indicators included but they cannot be directly compared due to different data sources used and changes in data definitions over time.

Table 1 – Model indicators(1) in the younger adults component of the current and proposed ASC RNF

Indicators in the younger adults component of the current ASC RNF	Indicators in the younger adults component of the proposed ASC RNF
Proportion of households with no family	Proportion of household reference persons aged 16 to 64 living in one-family households
Proportion of people aged 18 to 64 who work in routine or semi routine occupations	Proportion of people aged 18 to 64 who are Universal Credit (No Work Requirements) or Employment Support Allowance or Personal Independence Payment, Disability Living Allowance or Attendance Allowance(2) claimants
Proportion of people aged 18 to 64 who are long term unemployed or have never worked	Proportion of people aged 16 to 64 who are aged 16 to 24
Proportion of people aged 18 to 64 who are in receipt of Disability Living Allowance

(1) Note that the indicators from the 2 models are not directly comparable due to definition changes over time and different data sources used.

(2) We note that it is not possible to claim Attendance Allowance under the State Pension age which includes people who are aged 18 to 64. However, this is the name of the indicator as included in the DWP Stat Xplore benefits combination dataset where this data is obtained. So, we have kept this name for consistency.

Table 2 – Model indicators (1) in the older adults component of the current and proposed ASC RNF

Indicator in the older adults component of the current ASC RNF	Indicator in the older adults component of the proposed ASC RNF
Proportion of people aged 65 or over who were in receipt of Attendance Allowance	Proportion of people aged 65 or over who are Personal Independence Payment, Disability Living Allowance, or Attendance Allowance claimants
Proportion of people aged 65 or over who are living alone	Proportion of household reference persons aged 65 or over living as a couple
Proportion of people aged 65 or over who are aged 90 or over	Proportion of people aged 65 or over who are aged 80 or over
Proportion of people aged 65 or over who were in receipt of Pension Credit (2)	Proportion of people aged 65 or over who are Pension Credit claimants aged 80 or over
Proportion of people aged 65 or over living in rented accommodation	Proportion of household reference persons aged 65 or over who own their home outright multiplied by the proportion of all dwellings in Council Tax bands A to E
	Proportion of household reference persons aged 65 or over who own their home outright multiplied by the proportion of all dwellings in Council Tax bands F to H

(1) Note that the indicators from the 2 models are not directly comparable due to definition changes over time and different data sources used.

(2) Pension Credit was the only benefit included during the 2005 to 2006 development of the final model for the older adults component of the current ASC RNF. However, in subsequent years, during the calculation of the estimated gross expenditure using local authority-level data, it appears the input data used also included information for other benefits such as income support, and so on.

An ethnicity indicator was considered in the model development of the current RNF. However, there was limited statistical evidence for its inclusion in the models underlying the current RNF. This contrasts with the updated ASC RNF where the model ethnicity indicator was statistically significant (it had important impacts on differences in local authority-funded ASC). Thus, the models underlying this proposed base ASC RNF include an ethnicity indicator: the proportion of the relevant population age group reporting being White using Census 2021 data.

ASCRU-PSSRU included ethnicity in their model because ethnicity can be related to both the model needs indicators and the relative ASC cost and can distort relationships between them if ethnicity is not accounted for. Therefore, we propose to include LSOA level data on ethnicity in the base ASC RNF model to minimise its distorting effects on other model indicators.

However, we do not consider ethnicity as a direct ASC relative needs indicator because it is usually not ethnicity per se, but other factors (such as ill health levels and cultural differences) associated with ethnicity that drive the differences in ASC relative needs. For example, existing evidence shows that people of non-White ethnicity are at higher risk of developing long-term health conditions from a younger age compared to people of white ethnicity (Hayanga et al., 2023). They are also more likely to have lower income and wealth (Byrne et al., 2020), and, at the same time, are less likely compared to ethnic White people with similar long-term health conditions to take-up disability benefits. (Salway et al., 2016). Not accounting for ethnicity in models predicting local authority-funded ASC support can, therefore, distort the true effects of other relative needs indicators, by causing the models to over or under-estimate their impact. This approach ensures that the effect of care need (as captured, for example, by benefit claims) on local authority funded support is not disproportionally impacted by any ethnic differences. This means, for example, that the ASC RNF do not allocate more funds to communities with a higher share of ethnic White people because they claim relatively more benefits per capita. In addition, the non-White group is a very heterogenous group and at a small area level it is often not appropriate or not possible to separate this group out further.

Therefore, as with supply, the direct effect of ethnicity was controlled for by taking the sample average proportion of the relevant aged population reporting as being White for each local authority and included it in the constant term for the proposed final base ASC RNF. As we are proposing to use the same proportion for all local authorities, ethnicity does not appear as a relative needs indicator in the final proposed base ASC RNF (although it was included in the underlying base ASC RNF models). This means the ethnicity mix in local authorities is not used directly to calculate allocation shares (and hence allocate funding). We believe this to be the most appropriate way of ensuring the allocations are not distorted by indicators that do not directly drive ASC relative needs.

In summary, we include ethnicity to make sure it doesn’t skew the relationships we are trying to measure, but we hold it at its average value for all local authorities as we do not consider it a direct ASC relative needs indicator.

Further details of the data definitions and sources underlying the current ASC RNF used in the 2013 to 2014 Local Government Finance Settlement can be found at Calculation of 2013 to 2014 Formula Funding and Definitions of Indicators for 2013 to 2014 Part 1, and in section 5 for this update.

2.2 Proposed scaling and further local variation adjustments made to the base ASC RNF

Scale to population size – the proposed base ASC RNF aims to estimate the relative gross expenditure per capita. Thus, these estimates need to be scaled up to the relevant population size of the local authority. For the younger adults component, we propose to use, when published, the 18 to 64 population projections rebased to the 2021 Census for each local authority.

For the older adults component, we propose to use the 65 and over population with a supported residents adjustment that aims to remove self-funders aged 65 or over from, when published, the population projections rebased to the 2021 Census for each local authority. We propose to use the following calculation:

i. Census usual residents in household population aged 65 and over, divided by
ii. Census usual residents population aged 65 and over, multiplied by
iii. projection of population aged 65 and over, plus
iv. reported number of local authority-supported clients aged 65 or over accessing long term nursing or residential support.

The indicative allocation shares use 2023 ONS population estimates for each LA, the most current population data available, to give respondents the most up-to-date data to inform their responses. Our view is that using more up-to-date population data would more accurately reflect the current population across England. Therefore, if the updated ASC RNF is implemented in 2026, we will adopt MHCLG’s proposal outlined in section 9.4 of the consultation document to use, when published, the local authority level population projections rebased to the 2021 Census, to adjust allocation shares during the multi-year settlement.

Area Cost Adjustment (ACA) - to account for the variation in the cost of delivering services for local authorities. MHCLG estimates the ACA for various services including for adult social care that DHSC proposes to use for the proposed ASC RNF. The ACA consists of 3 adjustment factors listed below:

• Labour Cost Adjustment (LCA) to account for local differences in the cost of labour
• Rates Cost Adjustment (RCA) to account for local differences in the cost of property rates/rents
• Accessibility adjustment to account for the impact of the differences in travel time to provide services on the cost of labour. It is measured using journey time data and combined with the LCA to account for additional labour cost.

The accessibility adjustment factor of the ACA is a new element not included in the 2013 to 2014 ACA used in the current ASC RNF. This means that the separate sparsity adjustment used in the older adults component of the current RNF is no longer needed. The sparsity adjustment was used in the current RNF to reflect that home care costs tend to be higher in more sparsely populated areas due to greater travel times between visits.

Further details on the proposed approach to the ACA can be found in Chapter 5 of the consultation document.

Low income adjustment – used for the older adults component of the proposed ASC RNF to account for the differences in income from local authorities supported user charges due to users’ differing income and wealth profiles. More details in section 3 below.

2.3 Proposed method to combine the older adults and younger adults components of the ASC RNF

We propose to use the relative England-level net current expenditure (NCE) on younger adults versus older adults as weights to combine the younger adults and older adults components of the proposed ASC RNF.

In this proposed ASC RNF update, we used the 2023 to 2024 ASC NCE as reported in the ASC-FR. We included the Planned Better Care Fund (BCF) expenditure on social care as reported in the 2023 to 2024 ASC-FR, and assumed the age distribution of the BCF expenditure is the same as the “Income from NHS” age distribution. The weights used for the proposed ASC RNF and the current ASC RNF are given in table 3.

Table 3 – weights used to combine the younger adults and older adults components of the ASC RNF

ASC RNF version	Younger Adults component	Older Adults component
Proposed ASC RNF weights	51.86%	48.14%
Current ASC RNF weights*	40.21%	59.79%

 *calculated from the control totals published in Calculation of 2013 to 2014 Formula Funding

3. Low Income Adjustment

DHSC calculates the Low Income Adjustment (LIA), an adjustment to the older adults component of the proposed base ASC RNF to account for differences in income from local authority-funded care users. The LIA could be considered as adjusting the base ASC RNF relative gross expenditure per capita towards a relative net expenditure, although local authorities also receive income from sources other than users of ASC services. We propose a weighted ordinary least squares regression with robust standard errors for the LIA model using local authority-level data. We propose to calculate a LIA value from the model regression coefficients for each local authority following the steps in section 3.3.

3.1 - The proposed LIA model

The proposed model indicators are as follows:

• dependent variable: proportion of ACA-deflated gross current expenditure (GCE) on LA funded short and long-term ASC for clients aged 65 and over that are client contributions
• explanatory variables:
  • balance of care variable: proportion of total GCE on local authority-funded long-term ASC for clients aged 65 and over spent on nursing care or residential care
  • low income variable: proportion of people aged 65 or over who are either Universal Credit (excluding the conditionality regimes “no work requirements” and “unknown”) or Guarantee Credit part of Pension Credit claimants
• the model is weighted by the value for the older adults component of the proposed base ASC RNF from this update and adjusted by the older adults population estimates with a supported residents adjustment.

For older adults, the split of ASC users between community and residential or nursing care is an important factor driving relative differences in client contributions as a proportion of the ACA-deflated GCE on ASC (see table 4), and hence its inclusion in the LIA model. However, the choice of care settings for local authority-funded users is largely within local authorities’ control and affects the amount of income raised from user charges. Thus, the balance of care variable is used only to investigate the relationship between the low income variable and the proportion of ACA-deflated ASC CGE from client contributions. The direct effect of the balance of care variable on the latter is assumed to be the same for all local authorities and so not directly used to calculate the final LIA values. Further information on the calculation is outlined in section 3.3.

We removed the Isles of Scilly and City of London during the proposed LIA model development as they are known outliers. This means they are excluded from the calculation underlying the regression estimates in Table 4. The LIA values for the Isles of Scilly and City of London are estimated separately using the final model regression coefficients.

We developed the proposed LIA model using four-year average data and adjusted expenditure data to the latest prices using the GDP deflator and took averages to smooth any yearly fluctuations in the ASC GCE or benefit claimant figures. The 2020 to 2021 data was excluded due to data quality concerns for some data underlying the ASC-FR as a result of the COVID-19 pandemic, especially comparability with previous years’ data and reporting between local authorities in 2020 to 2021^{[footnote 5]}.

Table 4: Proposed LIA model regression results

Model indicator	Indicator coefficients	Confidence Interval	P value
(intercept)	0.2017	0.1041 – 0.2993	<0.0001
Low income indicator	-0.3348	-0.5733 - -0.0964	0.0062
Balance of care	0.1576	0.0204 – 0.2947	0.0247

Observations: 151

R²/R² adjusted: 0.1873/0.1764

The regression results above could be subject to change depending on the outcome of the questions posed in the consultation.

The above regression results are broadly in line with expectation. For example, as client contributions are more important for nursing and residential care than other care settings, we would expect higher proportions of nursing and residential care GCE in overall GCE to lead to higher proportions of GCE from client contributions. This is what we observe in our model.

We found higher proportions of certain benefit claimants in the older adults population (as defined above) lead to lower proportions of ASC GCE from client contributions (with GCE ACA deflated). We would expect this because we believe the benefits data included in the proposed LIA model is more likely to capture people with lower income and fewer financial assets. These individuals are expected to be less able to contribute to local authority-funded ASC if they need care. However, we note that the statistical evidence for this relationship is relatively weak. This may be due to a lack of specific data that can distinguish between the poorest people, who are least able to contribute to local authority-funded ASC if they need care, and people with slightly more means who might need to contribute if they need care. In addition, we believe this lack of granular data is also likely to be the reason for the relative low model fit of the LIA model (17.6% of the variation in the proportion of ACA- deflated GCE from user charges is explained by the proposed LIA model indicators). However, we still included the LIA for the older adults component of the proposed ASC RNF since client contributions are not accounted for elsewhere in the model.

3.2 Changes from the current LIA model

The proposed LIA model remains broadly similar to the current LIA model first used in the 2011 to 2012 LGFS2. However, we updated the model using the latest data available at the time of modelling and improved the methodological approach. See section 5 for the data definitions and sources used for the LIA model.

The low income indicator has been updated to reflect the current benefits available. The current LIA model includes older adults on Income Support, Income-based Jobseeker’s Allowance, or guarantee element of Pension Credit^{[footnote 6]}. People on Income Support and Income-based Jobseeker’s Allowance have been removed as these benefits have now been replaced by Universal Credit. We have therefore now included the number of Universal Credit (excluding the “no work requirements” and “unknown” conditionality regimes) claimants aged 65 or over. The guarantee credit element of Pension Credit caseload remains. We note that the State Pension age has increased from 65 to 66 and is set to change again in the future, however, the inclusion of both Universal Credit and Pension Credit ensures the indicator is not sensitive to State Pension age changes. We considered other low income related variables, including the proportion of people aged 65 and over in the social rented sector (Census 2021). We chose the model using benefits data because the social rented data are less frequently updated and did not significantly improve the model. We also investigated including additional variables in the model such as house ownership data from the Census. However, this did not improve the model and since the model has a limited number of data points, adding more variables could make the model less reliable.

3.3 Calculating the proposed LIA value

The following steps are used to produce the final proposed LIA value for each local authority using the proposed LIA model outlined in section 3.1. The direct effect of the proportion of ACA-deflated ASC GCE on nursing care or residential care for clients aged 65 or over is removed by assuming the same value for all local authorities; it was calculated by multiplying its model coefficient by its national weighted average value for every local authority. The LIA value for people aged 65 and over is calculated as follows:

i. 0.2017 plus

ii. 0.0990 (0.1576 multiplied by the weighted mean of the proportion of total GCE on local authority-funded long-term ASC for clients aged 65 and over-spent on nursing care or residential care), plus

iii. -0.3348 multiplied by the proportion of people aged 65 or over who are either Universal Credit (excluding the conditionality regimes “no work requirements” and “unknown”) or Guarantee Credit part of Pension Credit claimants

The results of the above:

iv. divided by the ACA and subtracted from 1,

v. divided by the minimum value

4. Calculating the proposed ASC RNF allocation shares

We calculate the allocation shares for the younger adults and older adults components separately and then combine these using weights as outlined in section 2.3 to calculate the final proposed ASC RNF allocation shares. For the population data used to scale to population size, see section 2.2.

4.1 Proposed younger adults component

The proposed younger adults component of the ASC RNF value is calculated as follows using local authority-level data:

i. 6.15 plus
ii. 19.06 multiplied by the proportion of people aged 18 to 64 who are Universal Credit (No Work Requirements) or Employment Support Allowance or Personal Independence Payment, Disability Living Allowance or Attendance Allowance* claimants, plus
iii. -3.06 multiplied by the proportion of household reference persons aged 16 to 64 living in one-family households, plus
iv. -6.15 multiplied by the proportion of the people aged 16 to 64 who are aged 16 to 24

The result of the above multiplied by:

i. ACA for adult social care, multiplied by
ii. number of people aged 18 to 64 years

*We note that it is not possible to claim Attendance Allowance under the State Pension age which includes people who are aged 18 to 64. However, this is the name of the indicator as included in the DWP Stat Xplore benefits combination dataset where this data is obtained. We have kept this name for consistency.

To calculate the younger adults component allocation share for each local authority, take its younger adults component ASC RNF value and divide by the sum of the younger adults ASC RNF values for all 153 local authorities with social care responsibilities, and then multiply by 100 to calculate the local authority younger adults component allocation share expressed as a percentage.

4.2 Proposed older adults component

The proposed older adults component of the ASC RNF value is calculated as follows using local authority-level data:

i. 22.42 plus
ii. 14.88 multiplied by the proportion of people aged 65 or over who are Personal Independence Payment, Disability Living Allowance, or Attendance Allowance claimants, plus
iii. -12.73 multiplied by the proportion of household reference persons aged 65 or over living as a couple, plus
iv. 12.99 multiplied by the proportion of people aged 65 or over who are aged 80 or over, plus
v. 25.95 multiplied by the proportion of people aged 65 or over who are Pension Credit claimants aged 80 or over, plus
vi. -9.12 multiplied by the proportion of household reference persons aged 65 or over who own their home outright multiplied by the proportion of all dwellings in Council Tax bands A to E, plus
vii. -18.32 multiplied by the proportion of household reference persons aged 65 or over who own their home outright multiplied by the proportion of all dwellings in Council Tax bands F to H

The result of the above multiplied by:

i. ACA for adult social care, multiplied by
ii. number of people aged 65 and over with a supported residents adjustment, multiplied by
iii. low income adjustment

To calculate the older adults component allocation shares for each local authority, take its older adults component ASC RNF value and divide by the sum of the older adults ASC RNF values for all 153 local authorities with social care responsibilities, and multiply by 100 to calculate the local authority older adults component allocation share expressed as a percentage.

4.3 Proposed ASC RNF allocation shares

The proposed younger adults and older adults component allocation shares are combined using weights listed in table 3 in section 2.3 to obtain the final ASC RNF allocation shares. As these are relative shares, if the value for one local authority changes then the shares will change for all local authorities.

5. Data definitions

This section contains the definitions and sources used for the calculation of the allocation shares based on the adult social care relative needs formulae.

5.1 - Proposed ASC RNF

Data from the Office for National Statistics (ONS)

Population aged 18 to 64

Estimate of the number of people aged 18 to 64 in mid-2023. Note this is what has been used to produce the indicative allocation shares. See section 2.2 for the population data we propose to use during the multi-year settlement.

Source: ONS mid-2023 population estimates summed over age categories 18 to 64 (inclusive) from: Estimates of the population for England and Wales - Office for National Statistics (ons.gov.uk).

Population aged 16 to 24

Estimate of the number of people aged 16 to 24 in mid-2023. Note this is what has been used to produce the indicative allocation shares. See section 2.2 for the population data we propose to use during the multi-year settlement.

Source: ONS mid-2023 population estimates summed over age categories 16 to 24 (inclusive) from: Estimates of the population for England and Wales - Office for National Statistics (ons.gov.uk).

Population aged 16 to 64

Estimate of the number of people aged 16 to 64 in mid-2023. Note this is what has been used to produce the indicative allocation shares. See section 2.2 for the population data we propose to use during the multi-year settlement.

Source: ONS mid-2023 population estimates summed over age categories 16 to 64 (inclusive) from: Estimates of the population for England and Wales - Office for National Statistics (ons.gov.uk).

Census household reference persons aged 16 to 64 living in one-family households

The number of household reference persons aged 16 to 64 in a single family household on 21 March 2021.

Source: Census 2021 - Create a custom dataset, ‘Population type: all Household Reference Persons; Age: aged 16 to 64 years; Household composition: One-person household, Single family household: Couple family household, Single family household: Lone parent household’.

Census household reference persons aged 16 to 64

The number of household reference persons aged 16 to 64 on 21 March 2021.

Source: Census 2021 - Create a custom dataset, ‘Population type: all Household Reference Persons; Age: aged 16 to 64 years’.

Population aged 65 and over

Estimate of the number of people aged 65 or over in mid-2023. See section 2.2 for the population data we propose to use during the multi-year settlement.

Source: ONS mid-2023 population estimates summed over age categories 65 to 89 (inclusive) and age category 90+ from: Estimates of the population for England and Wales - Office for National Statistics (ons.gov.uk).

Population aged 80 and over

Estimate of the number of people aged 80 or over in mid-2023. See section 2.2 for the population data we propose to use during the multi-year settlement.

Source: ONS mid-2023 population estimates summed over age categories 80 to 89 (inclusive) and age category 90+ from: Estimates of the population for England and Wales - Office for National Statistics (ons.gov.uk).

Census household reference persons aged 65 or over living in a couple

The number of household reference persons aged 65 or over living in a couple on 21 March 2021.

Source: Census 2021, ‘Age: aged 65 years and over; Living Arrangements: Living in a couple’ from RM066 - Living arrangements by age - Household Reference Person – Nomis.

Census household reference persons aged 65 or over

The number of household reference persons aged 65 or over on 21 March 2021.

Source: Census 2021 - Create a custom dataset, sum the ‘Population type: all Household Reference Persons; Age: aged 65 years and over’

Census household reference persons aged 65 or over who own home outright

The number of household reference persons aged 65 or over who own home outright on 21 March 2021.

Source: Census 2021, select ‘Age: aged 65 years and over; Household Tenure: Owns outright’ from RM201 - Tenure by age - Household Reference Persons - Nomis.

Census residents in households aged 65 or over

The number of usual residents in households aged 65 or over on 21 March 2021.

Source: Census 2021 - Create a custom dataset, ‘Population type: all usual residents; Age: aged 65 years and over’.

Census all usual residents aged 65 or over

The number of usual residents aged 65 or over on 21 March 2021

Source: Census 2021, select ‘Age: aged 65 years and over; Sex: All persons’ from RM121 - Sex by age - Nomis.

Data from the Department for Work and Pensions (DWP)

Universal Credit (No Work Requirements) or Employment Support Allowance or Personal Independence Payment, Disability Living Allowance, or Attendance Allowance claimants aged 18 to 64

The number of Universal Credit (No Work Requirements) or Employment Support Allowance or Personal Independence Payment, Disability Living Allowance, or Attendance Allowance claimants aged 18 to 64, averaged across August 2023, November 2023, February 2024, and May 2024.

Source: Statistics at DWP: Stat-Xplore, under ‘Benefit Combinations – Data from May 2019 for England and Wales’. Filter age categories between ‘18 to 24’ and ‘60 to 64’ (inclusive), under ‘Benefit’ filter ‘ESA’, ‘PIP, DLA & AA’, and ‘UC’, for August 2023, November 2023, February 2024 and May 2024.

Pension Credit claimants aged 80 or over

The number of Pension Credit claimants aged 80 or over, averaged across August 2023, November 2023, February 2024 and May 2024.

Source: Statistics at DWP: Stat-Xplore, under ‘Benefit Combinations – Data from May 2019 for England and Wales’. Filter age categories between ’80 to 84’ and ’95 and over’ (inclusive), under ‘Benefit’ filter ‘PC’ for August 2023, November 2023, February 2024 and May 2024.

Personal Independence Payment, Disability Living Allowance, or Attendance Allowance claimants aged 65 or over

The number of Personal Independence Payment, Disability Living Allowance, or Attendance Allowance, claimants aged 65 or over, averaged across August 2023, November 2023, February 2024 and May 2024.

Source: Statistics at DWP: Stat-Xplore, under ‘Benefit Combinations – Data from May 2019 for England and Wales’. Filter age categories between ’60 to 65’ and ’95 and over’ (inclusive), under ‘Benefit’ filter ‘PIP, DLA & AA’ for August 2023, November 2023, February 2024 and May 2024.

Data from other sources

Dwellings in Council Tax bands A-E

The number of dwellings in Council Tax bands A, B, C, D or E in England as at 31 March 2024.

Source: Table CTSOP1.1: number of properties by Council Tax band, local authority and super output area as at 31 March 2024 from Council Tax: stock of properties, 2024. Sum across Council Tax bands A to E.

Dwellings in Council Tax bands A-H

The number of dwellings in Council Tax bands A, B, C, D, E, F, G or H in England as at 31 March 2024.

This is the same as ‘all dwellings’ in England.

Source: Table CTSOP1.1: number of properties by Council Tax band, local authority and super output area as at 31 March 2024 from Council Tax: stock of properties, 2024. Sum across Council Tax bands A to H.

Dwellings in Council Tax bands F-H

The number of dwellings in Council Tax bands F, G and H in England as at 31 March 2024.

Source: Table CTSOP1.1: number of properties by Council Tax band, local authority and super output area as at 31 March 2024 from Council Tax: stock of properties, 2024. Sum across Council Tax bands F to H.

ACA for Adult Social Care

MHCLG’s proposed Adult Social Care ACA value for local authorities with adult social care responsibilities.

Source: MHCLG.

Clients aged 65 or over accessing long term nursing or residential support

The number of clients aged 65 or over accessing long term nursing or residential support at the end of the 2023 to 2024 financial year.

We used the 2019 to 2020 reported figure for Hackney, the latest reported data from Hackney on this indicator. We assumed this figure was 4 for Isles of Scilly and City of London where the number of clients were less than 5 and hence supressed for reporting in the ASC-FR.

Source: Adult Social Care Activity and Finance Report - NHS England Digital, ‘Table 38: Number of clients accessing long term support at the end of the year, by age band and support setting’.

NCE on short term and long term social care for clients aged 18 to 64

The local authority NCE on short term social care for clients aged 18 to 64 plus the local authority NCE on long term social care for clients aged 18 to 64 in the 2023 to 2024 financial year.

Source: Adult Social Care Activity and Finance Report - NHS England Digital, NCE Data, ‘Net Current Expenditure on long and short term care, by care type and age band, year on year comparison, 2023-24’.

NCE on short term and long term social care for clients aged 65 or over

The local authority NCE on short term social care for clients aged 65 or over plus the local authority NCE on long term social care for clients aged 65 or over in the 2023 to 2024 financial year.

Source: Adult Social Care Activity and Finance Report - NHS England Digital, NCE Data, ‘Net Current Expenditure on long and short term care, by care type and age band, year on year comparison, 2023-24’.

Planned Better Care Fund expenditure on social care

The Planned Better Care Fund expenditure on social care in the 2023 to 2024 financial year.

Source: Appendix C – Expenditure on adult social care, 2009-10 to 2023-24 - NHS England Digital, ‘Table 5: Net current expenditure on adult social care services in cash and real terms: by source of funding, 2023-24’.

5.2 Low Income Adjustment Model

Client contributions received from clients aged 65 or over accessing short or long term local authority-funded ASC

Income from client contributions aged 65 or over for long term care plus income from client contributions aged 65 or over for short term care, averaged across the 2019 to 2020, 2021 to 2022, 2022 to 2023 and 2023 to 2024 financial years.

Source: Adult Social Care Activity and Finance Report - NHS England Digital, ‘Income, by finance description, care type and age band’ for each year from 2019 to 2020, to 2023 to 2024.

Nursing care and residential care expenditure for local authority-funded ASC clients aged 65 or over

GCE on nursing care for clients aged 65 or over plus GCE on residential care for clients aged 65 or over, averaged across the 2019 to 2020, 2021 to 2022, 2022 to 2023 and 2023 to 2024 financial years.

Source: Adult Social Care Activity and Finance Report - NHS England Digital, ‘Gross Current Expenditure on long term care for clients aged 65 and over, by support setting’ for each year from 2019 to 2020, to 2022 to 2023.

GCE on local authority-funded short term ASC for clients aged 65 and over

GCE on short term care for clients aged 65 or over, averaged across the 2019 to 2020, 2021 to 2022, 2022 to 2023 and 2023 to 2024 financial years.

Source: Adult Social Care Activity and Finance Report - NHS England Digital, ‘Gross Current Expenditure on short term care for clients aged 65 and over, by purpose and primary support reason’ for each year from 2019 to 2020, to 2023 to 2024.

GCE on local authority-funded long term ASC for clients aged 65 and over

GCE on long term care for clients aged 65 or over, averaged across the 2019 to 2020, 2021 to 2022, 2022 to 2023 and 2023 to 2024 financial years.

Source: Adult Social Care Activity and Finance Report - NHS England Digital, ‘Table 44: Gross current expenditure on long term care for clients aged 65 and over, by support setting’ for each year from 2019 to 2020, to 2023 to 2024.

People aged 65 or over claiming Universal Credit

The number of people aged 65 or over claiming Universal Credit (excluding the conditionality regimes ‘no work requirements’, ‘unknown or missing regime’ and ‘not available prior to April 2015’), averaged over May, August, November and February of each year between May 2019 and February 2024 (inclusive), excluding May 2020, August 2020, November 2020 and February 2021.

Source: Statistics at DWP: Stat-Xplore, under ‘People on Universal Credit’: Filter age categories ’65’, ‘Over 65’; under ‘Conditionality regimes’ filter ‘Searching for work’, ‘Working – with requirements’, ‘Working – no requirements’, ‘Planning for work’, ‘Preparing for work’ for May, August, November and February of each year from May 2019.

People aged 65 or over claiming Guarantee Credit only part of Pension Credit

The caseload of Guarantee Credit only part of Pension Credit for those aged 65 or over, averaged over May, August, November and February of each year between May 2019 and February 2024 (inclusive), excluding May 2020, August 2020, November 2020 and February 2021.

The number of caseloads and number of claimants are the same for Guarantee Credit.

Source: Statistics at DWP: Stat-Xplore, under ‘Pension Credit – Data from May 2018’. Filter age categories ’65-69’ to ’90 and over’ (inclusive), under ‘Type of Pension Credit’ filter ‘Guarantee Credit only’ for each quarter from May 2019.

Population aged 65 and over

Estimate of the number of people aged 65 or over on 30 June of the reference year, averaged across 2019, 2021, 2022 and 2023.

Source: Estimates of the population for the UK, England, Wales, Scotland, and Northern Ireland, 2011 to 2023 estimates.

Regression weights for the proposed LIA model

The allocation shares for each local authority from the older adults component from this proposed base ASC RNF with the adjustment for the older people adults population estimate with a supported residents adjustment.

Source: DHSC.

ACA for Adult Social Care

As above section “ACA for Adult Social Care”.

GDP deflator

The December 2024 Gross Domestic Product deflator to adjust expenditure data to 2023-24 prices.  

Source: GDP deflators at market prices, and money GDP - GOV.UK, GDP deflators at market prices, and money GDP December 2024 (Quarterly National Accounts).

1. Further information on the assessment criteria can be found in the Care and support statutory guidance ↩

2. LSOAs are the second lowest statistical geography with usual resident population of between 1,000 to 3,000 people Statistical geographies - Office for National Statistics (ons.gov.uk) ↩

3. The methodology accounts for the distance of care homes to each LSOA and estimates the number of beds for youngers adults and older adults using the number of care beds from the Sept 2023 CQC care directory. ↩

4. Number of residents employed in caring and personal service occupations (SOC2020-61) in human health and social care (SIC2007-Q) per capita (18 to 64 years or 65 and over) using Census 2021 data. ↩

5. See the summary in Adult Social Care Activity and Finance Report, England - 2020 to 2021 for more detail ↩

6. See the data definitions of indicators for 2013 to 2014 ↩