Skip to main content

Employers and social mobility: monitoring social diversity in the civil service

Posted by: , Posted on: - Categories: professions, Social Mobility

A number of people – including the Commission and many others – are concerned by the fact that those at the top of many of our professions are unrepresentative of the UK population.

One question we have been asked in response to our Elitist Britain report is what can employers do to tackle this issue?

Our Business Manifesto last year set out a number of steps that will help. One important one is collecting data about the social backgrounds of their current staff and those they recruit. This will enable employers to develop a far better understanding of how different aspects of their recruitment practices impact on social mobility and to ask the right questions and take the right action to address any barriers that are identified.

There has been an increasing focus on this, with organisations such as Professions for Good developing staff surveys that can be used by businesses to monitor the social background of staff and new recruits. An increasing number of organisations are now collecting this data and it is one of the key steps being taken by employers who are signed up to the Champions Tier of the Government’s Social Mobility Business Compact.

One of the leading organisations in this respect is the Civil Service which has, since 2011, collected detailed information about the social background of entrants to its graduate recruitment programme, the Fast Stream. It also announced plans in May 2012 to extend this monitoring across the entire senior civil service. While this latter commitment has – over two years later – not been followed through, the Civil Service Talent Action Plan confirmed that the Government still intend to collect and publish this data at some point in the future, and some data on new entrants to the senior civil service is now being published (though this is incomplete – barely half of new entrants responded to the survey and the report concludes that “we cannot consider the results to be representative”).

Our analysis from Elitist Britain showed that 55 per cent of Permanent Secretaries were educated privately, 87 per cent attended Russell Group universities and 57 per cent went to either Oxford or Cambridge University. The Fast Stream data tells us something important about the pool from which the Permanent Secretaries of the future will be drawn from.

Analysis by the Commission of data from the 2013 Fast Stream Recruitment round shows that:

  • Those accepted onto the Fast Stream are far less likely to come from working-class backgrounds1 than the general population (only 30 people from working-class backgrounds were successful in their application – 3.7% of those recommended for appointment compared to 30.2% of all those in employment)
  • Relatively few applications are made to the Fast Stream by people from working-class backgrounds – they only account for 7.6% of applications.
  • Applicants from working-class backgrounds are less than half as likely to be successful in their application as those from middle-class backgrounds2 (2.4% success rate for those from working-class backgrounds compared to a 5.3% success rate for those from middle-class backgrounds).
  • A quarter of successful applicants from the UK (26 per cent) were independently educated, with a success rate of 5.5%, compared to a 4.7% success rate for those from state schools.
  • A quarter of successful applicants had a degree from Oxford or Cambridge University (23.5% of successful UK applicants compared to 10.6% of UK applicants).
  • Three quarters of successful applicants had a degree from a Russell Group university (76.8% of successful UK applicants compared to 59% of UK applicants).

Looking at the Graduate Fast Stream – the main scheme recruiting the Permanent Secretaries of the  future - there were only three successful applicants from working-class backgrounds, only 0.8% of recruits. Middle-class applicants were seven times more likely to succeed in their application, with a success rate of 3.5% versus 0.5%. Almost a third (32.2%) of recruits had an Oxbridge degree and 83.5% had a degree from a Russell Group university. These are worrying statistics.

Of course this does not necessarily mean that the recruitment processes of the civil service are unfair – there is, for example, a strong correlation between school attainment and social class which we would expect to be reflected in the outcomes of highly competitive recruitment processes.

However, looking at one potential benchmark of performance which attempts to control for attainment suggests there may be a real issue here: we know from HESA data on the social background of young undergraduate students that 20% of students at the Russell Group and 11% of students at Oxford and Cambridge are from working-class backgrounds (though the HESA definition of working-class is slightly broader than the one used in the Fast Stream statistics3). This would suggest that the social background of entrants to the Fast Stream is more advantaged than the social background of those attending the most academically selective universities in the United Kingdom4.

It should be emphasised that this is only indicative of a potential problem - it is difficult to really get to grips with this issue and make conclusive statements without further interrogation of the raw (unpublished) data. The Commission hopes that there is someone in Whitehall carrying out the regression analysis that is necessary to understand the detailed reasons for these differential success rates (e.g. controlling for the impact of A-level results, degree class, degree subject) and develop a better understanding about where in the recruitment process disadvantaged students fall away. This analysis should then be used as the basis to develop a civil service-wide strategy to reform recruitment and selection processes to ensure they are designed to recruit the best people for the job regardless of social background.

This post illustrates why it is crucial that the professions collect and analyse information on social background of recruits: without this any potential issues of recruitment processes unfairly disadvantaging those from less well-off backgrounds will remain hidden with important questions about the fairness of recruitment processes never being asked. The likely result of this is that outcomes will remain unfair.


1. “Working-class backgrounds” has been taken to mean those with parents in the “Routine and Manual” group within the three class version of the National Statistics Socio-Economic Classification (NS-SEC)

2. “Middle-class backgrounds” has been taken to mean those with parents in the “Higher Managerial, Administrative and Professional Occupations” group within the three class version of the National Statistics Socio-Economic Classification (NS-SEC)

3. The HESA data looks at those in NS-SEC classes 4-7 within the eight class version of the National Statistics Socio-Economic Classification (NS-SEC). It includes “Small Employers and Own Account Workers” in addition to the occupations within the “Routine and Manual” group. Therefore comparisons are less favourable to Fast Stream recruitment than is warranted.

4. To check the definitional issues explained in note 3 are not driving this finding, we looked at the “Intermediate” and “Routine and Manual” groups combined. This is a broader group than in the HESA comparator as it includes “Intermediate Occupations” as well as “Small Employers and Own Account Workers”: it therefore has the opposite issue of making the comparison look more favourable to Fast Stream recruitment than is warranted. 16.1% of Fast Stream hires are in this group – showing it is still significantly more advantaged than the Russell Group (20.4% of entrants).

5. We have looked only at candidates with a known social background in analysing social class and looked only at UK candidates with a known school and university background in analysing educational background. Therefore our percentages differ slightly to those reported in the Fast Stream statistics.


Sharing and comments

Share this page