Summary It should be noted that the names used to convey ethnicity were more successful for some ethnic groups than others(the 'black Caribbean'names were not as well attributed to that group as the 'Indian'names for instance).However,on this evidence,it does not appear that differences in labour outcomes between minority ethnic groups are the result of differences in the level of discrimination in the application phase of the recruitment process. Gender and area Although the number of tests in the sample is relatively small,we can look at the level of discrimination for various subgroups relating to the type of applicant,the nature of the role,the nature of the employer and the process for applying for the work. There was a high level of racial discrimination for applications of both genders. The level of discrimination was somewhat higher among male applicants (32 compared to 26 per cent),but this was not statistically significant. Similarly,there was little to suggest that racial discrimination was a problem confined to particular cities in Great Britain.The numbers of useable sets of applications in each of the seven cities included in the test were too small for differences between them to be statistically significant,but the results suggest high levels of discrimination across the board. Nature of employer and occupation The range of occupations included in the study aimed to provide a mix of higher and lower skilled jobs.The occupations in our test can be grouped together simply into 'higher level'(IT Technician,Accountant,Human Resources Manager, Teaching Assistant)and 'lower level'(IT Support,Accounts Clerk,Office Assistant, Care Assistant).There was some suggestion that racial discrimination was lower for higher level vacancies(23 per cent compared to 33 per cent),although this was not a statistically significant difference. The Civil Service and other public sector organisations are required by law to promote equality and make efforts to reduce racial and other discrimination.Public sector employers in our test were considerably less likely to have discriminated on the grounds of race than those in the private sector(four per cent compared to 35 per cent).The difference between them was statistically significant. It might be supposed that organisations with larger workforces would be more likely to have a dedicated human resources function and have documented procedures for recruitment.The number of employees at the site of the organisation could be obtained for around half of the cases in our sample.The results are only suggestive due to the small base sizes,but larger employers were found to discriminate less in our test
4 It should be noted that the names used to convey ethnicity were more successful for some ethnic groups than others (the ‘black Caribbean’ names were not as well attributed to that group as the ‘Indian’ names for instance). However, on this evidence, it does not appear that differences in labour outcomes between minority ethnic groups are the result of differences in the level of discrimination in the application phase of the recruitment process. Gender and area Although the number of tests in the sample is relatively small, we can look at the level of discrimination for various subgroups relating to the type of applicant, the nature of the role, the nature of the employer and the process for applying for the work. There was a high level of racial discrimination for applications of both genders. The level of discrimination was somewhat higher among male applicants (32 compared to 26 per cent), but this was not statistically significant. Similarly, there was little to suggest that racial discrimination was a problem confined to particular cities in Great Britain. The numbers of useable sets of applications in each of the seven cities included in the test were too small for differences between them to be statistically significant, but the results suggest high levels of discrimination across the board. Nature of employer and occupation The range of occupations included in the study aimed to provide a mix of higher and lower skilled jobs. The occupations in our test can be grouped together simply into ‘higher level’ (IT Technician, Accountant, Human Resources Manager, Teaching Assistant) and ‘lower level’ (IT Support, Accounts Clerk, Office Assistant, Care Assistant). There was some suggestion that racial discrimination was lower for higher level vacancies (23 per cent compared to 33 per cent), although this was not a statistically significant difference. The Civil Service and other public sector organisations are required by law to promote equality and make efforts to reduce racial and other discrimination. Public sector employers in our test were considerably less likely to have discriminated on the grounds of race than those in the private sector (four per cent compared to 35 per cent). The difference between them was statistically significant. It might be supposed that organisations with larger workforces would be more likely to have a dedicated human resources function and have documented procedures for recruitment. The number of employees at the site of the organisation could be obtained for around half of the cases in our sample. The results are only suggestive due to the small base sizes, but larger employers were found to discriminate less in our test. Summary
Summary 5 Application process There was virtually no net discrimination (one per cent)for sets of applications where the employer's own form had been used.This compared to 38 per cent where a CV had been sent.This difference was statistically significant.This result may relate to employer forms often being designed so that the section containing personal details(including name)can be detached before the sifting process.This is in addition to the standardisation of applications in favour of characteristics pertinent to the job.Further,these measures may be associated with organisations with dedicated human resources functions and well-developed procedures Part of the explanation for the absence of net discrimination among public sector employers may be the widespread use of standard application forms.Forms were used in 79 per cent of public sector applications compared to six per cent of those to private sector employers. Conclusions The random assignment of names to convey ethnicity in applications in this correspondence test mean there are no plausible explanations for the difference in treatment found between white and ethnic minority names other than racial discrimination. High levels of name-based net discrimination were found in favour of white applicants.This is consistent with the high levels of discrimination found in studies in other countries in recent years.This relates only to the early stage of the recruitment process,and there are limitations with the approach in terms of the occupations and vacancies that it was possible to cover.Nevertheless,candidates were denied access to a range of jobs in a range of sectors across British cities as a result of having a name associated with an ethnic minority background. The level of discrimination was consistent across the ethnic minority groups included in the study,suggesting that it accounts for a proportion of ethnic penalties for all ethnic groups.However,it does not appear to account for the difference between minority groups. The findings point to the potential effectiveness of a practical lever for tackling the problem.No discrimination at the first stage of recruitment was found where employers were using their own forms for the process(as opposed to CVs)
5 Application process There was virtually no net discrimination (one per cent) for sets of applications where the employer’s own form had been used. This compared to 38 per cent where a CV had been sent. This difference was statistically significant. This result may relate to employer forms often being designed so that the section containing personal details (including name) can be detached before the sifting process. This is in addition to the standardisation of applications in favour of characteristics pertinent to the job. Further, these measures may be associated with organisations with dedicated human resources functions and well-developed procedures. Part of the explanation for the absence of net discrimination among public sector employers may be the widespread use of standard application forms. Forms were used in 79 per cent of public sector applications compared to six per cent of those to private sector employers. Conclusions The random assignment of names to convey ethnicity in applications in this correspondence test mean there are no plausible explanations for the difference in treatment found between white and ethnic minority names other than racial discrimination. High levels of name-based net discrimination were found in favour of white applicants. This is consistent with the high levels of discrimination found in studies in other countries in recent years. This relates only to the early stage of the recruitment process, and there are limitations with the approach in terms of the occupations and vacancies that it was possible to cover. Nevertheless, candidates were denied access to a range of jobs in a range of sectors across British cities as a result of having a name associated with an ethnic minority background. The level of discrimination was consistent across the ethnic minority groups included in the study, suggesting that it accounts for a proportion of ethnic penalties for all ethnic groups. However, it does not appear to account for the difference between minority groups. The findings point to the potential effectiveness of a practical lever for tackling the problem. No discrimination at the first stage of recruitment was found where employers were using their own forms for the process (as opposed to CVs). Summary
Background 1 Background 1.1 Introduction This report shows the findings from a field experiment to estimate the extent of racial discrimination in key areas of the British labour market.The task of this study was to collect factual evidence to test the assertion that discrimination is a significant factor affecting labour market outcomes for members of ethnic minorities. The Department for Work and Pensions(DWP)commissioned the National Centre for Social Research(NatCen)to carry out this study in a context of continued policy focus on how best to promote equal opportunities.The research was commissioned after the employer-led Business Commission on Race Equality recommended 'matched CV'testing to measure progress towards eliminating the ethnic minority employment gap.The Chancellor of the Exchequer accepted the recommendation on 'matched CV'testing and asked the Ethnic Minority Employment Task Force to oversee delivery and report back to him in writing by December 2009.The Equality Bill was before Parliament at the time of writing, and we hope that the findings in this report represent a useful addition to the evidence to inform debate. 1.2 Ethnic minority penalties in the labour market That there are ethnic penalties in employment in Britain is a well-established fact. A recent study by Professor Anthony Heath and Professor Yaojun Li looked at differences in labour market attainment for males belonging to different ethnic groups during the period 1972 to 2005 (Li and Heath,2008).The study assessed ethnic penalties after controlling for measures of human capital,predominantly education.Heath and Li found that black and Pakistani/Bangladeshi men became more likely to be unemployed during the period.While Indian men's labour market status improved over the period,the situation of black Caribbean,black African and Pakistani/Bangladeshi men worsened (Li and Heath,2008)
Background 7 1 Background 1.1 Introduction This report shows the findings from a field experiment to estimate the extent of racial discrimination in key areas of the British labour market. The task of this study was to collect factual evidence to test the assertion that discrimination is a significant factor affecting labour market outcomes for members of ethnic minorities. The Department for Work and Pensions (DWP) commissioned the National Centre for Social Research (NatCen) to carry out this study in a context of continued policy focus on how best to promote equal opportunities. The research was commissioned after the employer-led Business Commission on Race Equality recommended ‘matched CV’ testing to measure progress towards eliminating the ethnic minority employment gap. The Chancellor of the Exchequer accepted the recommendation on ‘matched CV’ testing and asked the Ethnic Minority Employment Task Force to oversee delivery and report back to him in writing by December 2009. The Equality Bill was before Parliament at the time of writing, and we hope that the findings in this report represent a useful addition to the evidence to inform debate. 1.2 Ethnic minority penalties in the labour market That there are ethnic penalties in employment in Britain is a well-established fact. A recent study by Professor Anthony Heath and Professor Yaojun Li looked at differences in labour market attainment for males belonging to different ethnic groups during the period 1972 to 2005 (Li and Heath, 2008). The study assessed ethnic penalties after controlling for measures of human capital, predominantly education. Heath and Li found that black and Pakistani/Bangladeshi men became more likely to be unemployed during the period. While Indian men’s labour market status improved over the period, the situation of black Caribbean, black African and Pakistani/Bangladeshi men worsened (Li and Heath, 2008)
8 Background Another authoritative study is the report for DWP by Professor Anthony Heath and Dr Sin Yi Cheung of Oxford University (2006,DWP Research Report No, 341).They make an important distinction between the overall or 'gross'level of inequality and the 'net patterns'that remain after controlling for differences in characteristics of the various ethnic minority groups,such as the age profiles and levels of education of members.The way in which these differences are controlled is through multivariate statistical analysis.While the net differences are smaller than the overall gap,there remains evidence of a'net ethnic penalty'in employment. The size of the penalty was shown to vary across different ethnic groups as well as for men and women.Those ethnic minority members born and educated in Britain,so-called 'second generation migrants',experience ethnic penalties in a similar way to the first generation. Heath and Cheung are careful to emphasise that the ethnic penalty may arise for a number of reasons and that their evidence does not identify the causal processes. They identify the following factors as plausible contributors to the gap: lack of information about job opportunities; lack of contacts with potential employers; difficulties of transport to areas where job vacancies are located. They make it clear that their research method is suggestive of the role of discrimination,but is not able to demonstrate the extent to which it contributes to the ethnic penalties. Carmichael and Woods came to a similar conclusion in 2000 when they argued that the ethnic penalties experienced by minority workers were not fully explained by differences in human capital endowments and personal characteristics.They attributed at least some of this difference to discriminatory selection practices among employers(Carmichael and Woods,2000). According to this study: 'Discrimination in selection practices is [therefore]consistent with lower occupational status as well as higher unemployment and lower average earnings for ethnic minorities.' (Carmichael and Woods,2000 pg 73) However,the evidence of ethnic minority penalties has been challenged.For example,the eminent US economist,Professor James Heckman,summarised considerable literature on the US situation in these terms(1998 Heckman): 'Most of the disparity in earnings between blacks and whites in the labor market of the 1990s is due to the differences in skills they bring to the market,and not to discrimination within the labor market. (1998:101)
8 Another authoritative study is the report for DWP by Professor Anthony Heath and Dr Sin Yi Cheung of Oxford University (2006, DWP Research Report No, 341). They make an important distinction between the overall or ‘gross’ level of inequality and the ‘net patterns’ that remain after controlling for differences in characteristics of the various ethnic minority groups, such as the age profiles and levels of education of members. The way in which these differences are controlled is through multivariate statistical analysis. While the net differences are smaller than the overall gap, there remains evidence of a ‘net ethnic penalty’ in employment. The size of the penalty was shown to vary across different ethnic groups as well as for men and women. Those ethnic minority members born and educated in Britain, so-called ‘second generation migrants’, experience ethnic penalties in a similar way to the first generation. Heath and Cheung are careful to emphasise that the ethnic penalty may arise for a number of reasons and that their evidence does not identify the causal processes. They identify the following factors as plausible contributors to the gap: • lack of information about job opportunities; • lack of contacts with potential employers; • difficulties of transport to areas where job vacancies are located. They make it clear that their research method is suggestive of the role of discrimination, but is not able to demonstrate the extent to which it contributes to the ethnic penalties. Carmichael and Woods came to a similar conclusion in 2000 when they argued that the ethnic penalties experienced by minority workers were not fully explained by differences in human capital endowments and personal characteristics. They attributed at least some of this difference to discriminatory selection practices among employers (Carmichael and Woods, 2000). According to this study: ‘Discrimination in selection practices is [therefore] consistent with lower occupational status as well as higher unemployment and lower average earnings for ethnic minorities.’ (Carmichael and Woods, 2000 pg 73) However, the evidence of ethnic minority penalties has been challenged. For example, the eminent US economist, Professor James Heckman, summarised considerable literature on the US situation in these terms (1998 Heckman): ‘Most of the disparity in earnings between blacks and whites in the labor market of the 1990s is due to the differences in skills they bring to the market, and not to discrimination within the labor market.’ (1998: 101) Background
Background 9 Heckman went on to describe discrimination as 'the problem of a previous era'. Employers are interested in the productivity or employability of their workforce These can be seen as synonyms for the term'skills'used by Heckman.It is by no means clear that the statistical approach used by Heath and Cheung,being limited to the characteristics recorded on major surveys,is able to approximate to these characteristics of employees or potential employees. An alternative view of some significant causes of ethnic penalties is that they may arise through a process known as 'statistical discrimination'.As described by Pager and Karafin(2009)much of the discussion on the causes of discrimination focuses on the rationality of employer decision making.Information is scarce when employers are making recruitment decisions.Knowing about the age and education of an applicant may be insufficient,given that the employer is also interested in other factors that influence productivity,such as motivation and social skills.In this situation,employers may rely on observable characteristics that they believe are correlated with the unobserved characteristics.The implication of this view is that employers are behaving in a manner that may appear rational(albeit what they are doing contravenes legislation on equal opportunities),rather than being motivated by 'preference-based'discrimination,when they make decisions that result in ethnic penalties. Discriminatory outcomes may also result from processes that are not consciously intended to be discriminatory but which nevertheless have that effect.Roberts and Campbell(2006)find no evidence of overt discrimination in their analysis of interviews for low-skilled jobs,but identify penalties for first generation migrants who are not familiar with the conventions of British job interview question style and organisational culture. Another aspect of this more indirect discrimination would be where recruiters do not apply strictly job-related criteria in their recruitment process,and instead base choices on how personable they themselves find a candidate to be.In some cases this may have the effect that they recruit people from their own ethnic and socio- demographic group. 1.3 Policy measures to reduce inequality Heath and Cheung(2006)provide a useful discussion of potential policy measures that may address the ethnic penalties in employment.Among the policy measures they discuss are: addressing educational inequalities; active labour market policies(not necessarily targeted at ethnic minorities,but towards residents in deprived areas); improved careers services at further education colleges and universities;
9 Heckman went on to describe discrimination as ‘the problem of a previous era’. Employers are interested in the productivity or employability of their workforce. These can be seen as synonyms for the term ‘skills’ used by Heckman. It is by no means clear that the statistical approach used by Heath and Cheung, being limited to the characteristics recorded on major surveys, is able to approximate to these characteristics of employees or potential employees. An alternative view of some significant causes of ethnic penalties is that they may arise through a process known as ‘statistical discrimination’. As described by Pager and Karafin (2009) much of the discussion on the causes of discrimination focuses on the rationality of employer decision making. Information is scarce when employers are making recruitment decisions. Knowing about the age and education of an applicant may be insufficient, given that the employer is also interested in other factors that influence productivity, such as motivation and social skills. In this situation, employers may rely on observable characteristics that they believe are correlated with the unobserved characteristics. The implication of this view is that employers are behaving in a manner that may appear rational (albeit what they are doing contravenes legislation on equal opportunities), rather than being motivated by ‘preference-based’ discrimination, when they make decisions that result in ethnic penalties. Discriminatory outcomes may also result from processes that are not consciously intended to be discriminatory but which nevertheless have that effect. Roberts and Campbell (2006) find no evidence of overt discrimination in their analysis of interviews for low-skilled jobs, but identify penalties for first generation migrants who are not familiar with the conventions of British job interview question style and organisational culture. Another aspect of this more indirect discrimination would be where recruiters do not apply strictly job-related criteria in their recruitment process, and instead base choices on how personable they themselves find a candidate to be. In some cases this may have the effect that they recruit people from their own ethnic and sociodemographic group. 1.3 Policy measures to reduce inequality Heath and Cheung (2006) provide a useful discussion of potential policy measures that may address the ethnic penalties in employment. Among the policy measures they discuss are: • addressing educational inequalities; • active labour market policies (not necessarily targeted at ethnic minorities, but towards residents in deprived areas); • improved careers services at further education colleges and universities; Background