Sunday, 7 June 2015

Following the 2011 riots which did so much damage to Croydon much effort was put into discovering possible causes. Attention was given to the strained relationship between the police and young people from a Black and Minority Ethnic (‘BAME’) background, with police use of stop-and-search a particular source of friction. As a result, significant changes have now been made to how stop and search is carried out.

A second, less-publicised area of concern are claims that BAME individuals are disproportionately likely to die in police custody. A recent report, “Dying for Justice” by the Institute for Race Relations (IRR) asserted this to be so, quoting data from a charity, Inquest, which provides advice to people bereaved by a death in custody. If true, this would be a serious matter requiring urgent attention.

The evidence on stop and search was, as the Home Secretary stated, overwhelming. In contrast, I could find less detailed analysis of deaths in custody. I decided to see what I could do to help plug that gap.

Data comes from the Independent Police Complaints Commission (IPCC) and from Inquest. Comparisons between these datasets aren’t straightforward due to differing definitions of what constitutes ‘in custody’. To complicate things further, the IPCC reports on a fiscal year basis whereas Inquest reports by calendar year. I therefore analysed the figures separately, and compared results.

Mercifully, the number of deaths is very small. But this does mean that on an annual basis, the question of statistical significance arises. To mitigate this, it helps to take a longer time period, say five years or more, bearing in mind that the ethnic make-up of the population also changes over time, albeit slowly.



Taking the last five years for which all figures are available, according to the IPCC in England and Wales there were 79 deaths in custody of which 6 were from the BAME group: 7.6%. Over ten years, the figure is 12.6%. Inquest data shows 256 deaths, of which 32 were from the BAME group: 12.5%. Over ten years their figure is 12.0%.

These figures compare with 14.1% of the population being of BAME ethnicity at the 2011 Census. The two datasets are consistent in showing no significant bias.

The above graph shows the Inquest information in brown compared to the general population in grey. It is clear that BAME deaths are systematically less than one might expect so it would appear that this group is actually less at risk of dying in custody, although the difference is small and might be explained by other factors. For example, the in-custody population ethnic mix may not be the same as the general population

The IPCC has a second category – “other deaths following police contact”. This includes, for example, deaths that occur after the police are called to attend a domestic incident that results in a fatality. This accounts for the significantly higher number of deaths in the Inquest data compared to the IPCC figures.

In the last five years there have been 203 deaths in this category. These data are plotted in blue. Here we can see that BAME individuals are indeed systematically over-represented. Another characteristic of this category is the much higher proportion of females, which over the last three years has averaged 50%, around four times higher than the 13% of female deaths in custody. The victims are also younger by around ten years compared to the victims of in custody death.

That BAME individuals are more likely to die in police custody is not supported by this evidence. Both are important, but the current narrative means that resources may be disproportionately allocated to addressing the deaths of BAME individuals in custody (around two per year with no evidence of ethnic bias) rather than younger females, of all ethnicities, who have interactions with the police outside of custody (around twenty per year with some evidence of ethnic and age bias). Sensitivities can also be inflamed and unnecessary damage done.

Before drawing any firm conclusions further investigation is needed. The published information does not allow me to do more, but that additional information will be available to the IPCC and Inquest. I hope they are making use of it.