Should we be paying public servants more? Some have expressed disquiet over the long-term sustainability of the 1% cap on pay settlements first introduced in 2010 and due to continue until 2019/20. Independent experts who advise government on setting pay for the 2.5 million public servants covered by Pay Review Bodies (PRBs) have cited pay restraint as a reason for the difficulties recruiting and retaining high quality staff to deliver health services, education and other public services.
This week the Office of Manpower Economics, the body supporting PRBs, published a report commissioned from us, into trends in public sector pay. The report tracks average real earnings within each of 394 occupations over the period 2005-2015, including 32 occupations covered by the five major PRBs (NHS staff; School teachers; Doctors and Dentists; Police Officers; and Prison Service Staff). We focus primarily on the 10 largest PRB occupations. Chart 1 from the report reveals four important findings:
- Median real gross hourly occupational earnings have fallen by 5.8% since 2005
- The decline was steeper among non-PRB occupations (6.1%) than it was among PRB occupations (3.1%)
- Changes in real median earnings varied considerably between PRB occupations, even among those whose pay was set by the same PRB
- Eight of the ten largest PRB occupations experienced a drop in real median earnings.
Of course, average earnings within a particular occupation may vary from year to year with changes in the composition of the workforce. But even after we account for those changes, we still find substantial differences in the fortunes of PRB occupations. The findings challenge the popular misconception among commentators and politicians that a 1% cap on pay settlements has a uniform impact on the earnings of those covered by the settlement. In fact, changes in actual earnings rarely resemble changes in pay settlements because they reflect not only changes in base pay, but also changes in overtime, shift and other premia, the share of premium paid hours, promotion rates, regrading and other matters. This leads to gaps between pay settlements, on the one hand, and actual earnings on the other – what economists refer to as “wage drift” (Phelps Brown, 1962). That is why it is important to look at what has happened to actual earnings in an occupation over time.
How can we evaluate whether these changes in PRB median real earnings are big or small? Comparisons are often made with the private sector (NHSPRB, 2016; Jenkins, 2014; Cribb, 2017). Others compare to groups of occupations. For instance, the School Teachers’ Review Body (2016) draws a comparison between the pay of schoolteachers and that of all other Professionals. All these approaches suffer from the same drawback: they are not comparing PRB earnings growth with earnings growth in comparable occupations.
To overcome this problem, we adopt a very different approach, one which has not been tried before, to compare earnings growth in PRB occupations directly with that in similar non-PRB occupations. PRB occupations are “matched” to non-PRB occupations based on their characteristics in 2005, including earnings levels, together with trends in earnings over the period 2002-2005. For each of the 10 PRB occupations we find a “nearest neighbour” non-PRB occupation against which to compare changes in median real earnings over time. We make comparisons both before and after adjusting for compositional change in the workforce for each occupation.
Two findings emerge:
- Relative to their nearest non-PRB comparators, earnings growth was higher for the PRB group in five cases and lower in five cases. However, differences were only statistically significant in three instances, with PRB Nurses and PRB Nursing Auxiliaries experiencing higher earnings growth than their non-PRB comparators, while PRB Radiographers experienced significantly lower growth than their non-PRB comparator occupation (Table 1).
- After accounting for compositional differences in the workers entering and leaving different occupations between 2005 and 2015, relative to their nearest non-PRB comparators, earnings growth was higher for the PRB group in seven cases and lower in three cases (Table 2).
Concern over public servants’ pay is compounded by the low pay many of them receive. Perceptions of wage differentials underpin attitudes to wage inequality (Kuhn, 2015). Of our 10 PRB occupations, 5 were outside the top 100 occupations in terms of median hourly earnings over the whole period. In 2015, Nursing Auxiliaries were the lowest paid of our 10 PRB occupations with median earnings of £10 per hour (Table 3), putting them in 276th position in the occupational earnings rankings among the 394 occupations in our study. Five of the 10 PRB occupations had fallen in the occupational earnings rankings since 2005 with Radiographers and Physios dropping furthest (30 and 20 places respectively).
Taken together, a picture emerges of earnings stagnation or decline for most occupations, PRB and non-PRB, since 2005. The big difference between PRB employees and those in non-PRB occupations in the private sector is that PRB employees are public servants. As such, the government can determine their annual pay settlement. When deciding what to do government will weigh public concern over pay equity for groups like nurses, and experts’ concerns regarding public employers’ ability to recruit, retain and motivate staff, against the potential costs of lifting the 1% cap for the public purse. That cost could be quite considerable, but history reminds us that periods of public sector pay restraint are rarely successful and are often followed by a period of wage inflation. Whether this is likely given current macro-conditions remains to be seen, but it appears unlikely.
|PRB occupation||% growth||Nearest non-PRB comparator||% growth||Percentage point difference||Significant?|
|Occupational Therapists||-8.0||Sports Coach/instructor||-0.6||-7.4||No|
|Nursing Auxiliaries||7.5||Telephone Sales People||-5.8||+13.3||Yes|
|Police Officers||-7.5||Fire Service Officers||-6.8||-0.8||No|
|Prison Officers||-9.1||Fire Service Officers||-6.8||-2.3||No|
Notes: (1) Source: Authors’ calculations based on Annual Survey of Hours and Earnings. (2) Earnings are deflated using the CPI. (3) Rounding means percentage point differences in column 5 may not be exactly the difference between column 2 and column 4
Table 2: Growth in Median Real Gross Hourly Earnings Having Netted Out Changes in Workforce Composition, 2005-2015
|PRB occupation||% growth||Nearest non-PRB comparator||% growth||Percentage point difference|
|Occupational Therapists||1.1||Sports Coach/instructor||-0.9||+2.0|
|Nursing Auxiliaries||13.4||Telephone Sales People||-1.7||+15.1|
|Police Officers||-6.3||Fire Service Officers||-13.4||+7.1|
|Prison Officers||-4.9||Fire Service Officers||-13.4||+8.5|
Notes: (1) Source: Authors’ calculations based on Annual Survey of Hours and Earnings. (2) Earnings are deflated using the CPI.
|£ per hour||Average annual growth (%)|
Notes: (1) Source: Authors’ calculations based on Annual Survey of Hours and Earnings. (2) Earnings are in constant (2015) prices (deflated using the CPI) and rounded to the nearest pound to prevent disclosure of individual earnings values.
Alex Bryson is Professor of Quantitative Social Science at UCL’s Department of Social Science (firstname.lastname@example.org). John Forth is a Fellow at the National Institute of Economic and Social Research (email@example.com).