Black, qualified and unemployed: how the figures were found
Today the TUC released analysis showing that for each level of education, BAME adults are significantly more likely to be unemployed than their white counterparts. In this post, I hope to provide a technical discussion about how these figures were calculated.
For an important discussion of the underlying problems, as well as what the TUC thinks needs to be done, I would strongly recommend Natasha’s blog.
Earlier analysis revealed a staggering pay gap, which did not decrease with the level of educational qualification. Following this analysis, and the discussion which ensued, we were interested in whether BAME people would suffer a similar penalty in terms of employment. To examine this question we performed original analysis of the UK Labour Force Survey, using the UK Data Service’s statistical package ‘Nesstar’.
In brief we:
- Aggregated different ethnic groups to from a ‘white’ and ‘BAME’ category. Although this is arguably problematic, we wanted to highlight the discrimination that all BAME people face.
- Looked at the highest level of qualification people achieved. This is a long list (with around 80 categories), ranging from Entry Level Awards to Higher Degrees.
- Calculated the numbers of people who were employed/unemployed (using the ILO definition), disaggregated by their ethnicity and highest level of qualification.
- Averaged the four quarters of 2015, to create an overall ‘2015’ figure.
- Calculated the unemployment rates for each group, using the formula:
- Due to limitations posed by survey size, we then filtered out qualifications which too few people had taken.
The overall figure was shocking: in 2015 BAME adults were twice as likely to be unemployed as white adults.
As a general rule of thumb, those with higher qualifications suffer lower levels of unemployment. The individual qualifications where BAME people suffered the highest levels of unemployment were: BTECs (29.1 per cent), trade apprenticeships (28.6 per cent – although note that the Labour Force Survey is bad at properly recording apprenticeships), low GCSE/CSE grades (21.5 per cent), AS level (21.0 per cent) and GCSE/O-levels (20.4 per cent).
Whilst these figures are disgraceful, we also noticed that the unemployment rates for white people with these qualifications were relatively high. As a simple measure of inequality, we divided BAME unemployment by white unemployment. We found that the highest ratio of BAME : white unemployment was for those with GNVQs (5.5), trade apprenticeships (5.2), NVQ level 4s (3.7), diplomas of higher education (3.5) and A-levels (3.2).
To give this analysis a historical perspective we compared these figures with those from 2012. This was as far back as we could take it due to changes in how the Labour Force Survey measures ethnicity.
The positive story is that between 2012 and 2015 unemployment fell for both white and BAME adults. For BAME people, the unemployment rate fell from between 13.6 per cent to 9.5 per cent, which is certainly very welcome.
However, the worrying thing is that whilst the unemployment rate has fallen, the gap has not closed. Overall inequality has very slightly increased, and although the picture is not uniform if we look at individual qualifications there is no evidence of convergence between BAME and white unemployment rates.
One final comment is that for the sake of the press release we did not want to include the entire plethora of qualifications considered above. As such we simplified the qualifications into more manageable groups comprised of qualifications of a similar level, based on the National Qualifications Framework. This is how the following table was produced:
Of course this analysis has limitations – we were unable to control for the quality of institutions people studied at, the subjects they studied for, the grades they achieved, or other potentially relevant factors.
However, we hope that this shows that BAME people do suffer from higher rates of unemployment, and this cannot be simply explained by their levels of education. Perhaps even more so than our pay gap analysis, these statistics would suggest discriminatory behavior by employers which the government must urgently act to tackle.