Adult social care labour supply: pay increase impact assessment

4. Results

Our headline results are based on the sample of ASC-WDS observations located in England’s North and Midlands regions (i.e., North East, North West, Yorkshire and The Humber, East Midlands, and West Midlands), excluding home care staff. The exclusion of observations from home care staff aims to avoid the measurement issues related to wage reporting for home care (see Section 3.1), while the restriction to observations from Northern and Midlands regions aims to account for the differences in ASC labour markets between Scotland and England (see Section 3.1.1). Our preferred specification also uses post-sampling weights to address the difference in relative employment share of the public sector in Scotland versus our sample, as discussed in Section 3.1.1.

The job separation estimations, on which our predicted employment effects are based, are presented in the Technical Annex, Table A6; Columns 1 and 2 report unweighted estimates, while Columns 3 and 4 report the preferred weighted estimates. The wage elasticities of separation and labour supply at each wage decile, as well as corresponding predicted incremental and total effects on employment of our preferred model (with weights) are presented in the Technical Annex, Table A10. As expected, wage elasticities of labour supply from the weighted estimations are slightly smaller than those from the unweighted estimations (Technical Annex, Table A9), as job separation outcomes of staff employed in the public and voluntary sectors, and likely having better pay and employment conditions, are given higher weights.

A summary of our preferred model predictions is given in Table 1. The predicted total employment effects were obtained (as described in Section 3.3) by adding up incremental employment effects from increases of the wage floor between deciles of the wage distribution. These were calculated using wage elasticities of labour supply to the sector, obtained from wage elasticities of separation (overall and to jobs inside ASC). Because only wage decile information is available for Scottish ASC wages, the analysis has had to make assumptions about the distribution of wages between deciles. This means we are only able to report upper- and lower-bound estimates of the employment effect of raising the ASC wage floor. These bounds result from contrasting assumptions and should be read together and interpreted as implying that true employment effects lie somewhere between the two estimates.

As the Scottish wage distribution data are for 2023, we deflated the new 2024 wage floor to its equivalent 2023 value, assuming a 3.75% inflation rate, based on Bank of England Q2 2024 estimates. For example, raising the wage floor from £10.90 to £12.00 in 2024 (a 10.1% nominal increase) represents a 6.1% increase in real terms to £11.57. The predicted policy effects are then estimated based on the 2023 equivalent.

Our preferred model predicts that the increase in the ASC wage floor to £12.00 in 2024/25 could lead to a predicted increase in ASC employment of between 2.7% and 4.0%. Table 1 also reports predicted employment increases under alternative policy scenarios where the ASC wage floor is increased to £12.50 or £13.00. All else constant, raising the wage floor to £12.50 (i.e., 14.7% nominal or 10.5% real wage increase) could lead to an ASC labour supply increase of between 5.4% and 7.5%, while an increase of the wage floor to £13.00 (i.e., 19.3% nominal or a 15.0% real wage increase) could lead to an ASC labour supply increase of between 8.4% and 11.2%.

As mentioned above, the expected employment effects should be interpreted as being somewhere between the upper and lower estimated values, as the two assumptions represent extremes. Moreover, the estimates assume that any other factors influencing labour supply are being kept constant, including wages in other sectors. Therefore, as discussed in more detail in the next section, to the extent that non-social care employers respond to the higher ASC wage floor by raising pay, the predicted effects are likely to overstate the net impact of the policy on social care employment.