Census data without the need for a separate Sikh ethnic tick box

The views and opinions expressed in this article are those of the author and do not necessarily reflect the official policy or position of the University of Birmingham

“On the basis of the evidence the decision not to include a separate Sikh ethnic tick box in the 2021 Census appears to be the correct one.”


Over the last couple of decades, certain Sikh groups have been arguing for a Sikh ethnic tick box to be included in the UK census on the grounds that the current religious tick box for Sikhs is inadequate to capture proper statistics for Sikhs and that public bodies are discriminating against them on this basis. Other Sikh groups have been vehemently against this move, arguing that it misconstrues Sikh religious identity and that data on Sikhs is already routinely collected and used by public institutions.

Weighing up the evidence, The Office for National Statistics (ONS) published its White Paper titled, ‘Help Shape Our Future: The 2021 Census of Population and Housing in England and Wales’ and did not recommend a separate Sikh ethnic tick box. Since then it has gone through the parliamentary stages in both Houses, and The Census (England and Wales) Order 2020 became legislation on 20 May 2020 without a separate Sikh ethnic tick box.1

The recent COVID-19 pandemic has really put a spotlight on the usefulness of the existing census data, which has been used in conjunction with other datasets to ascertain difference in rates of death between ethnic and religious groups, and therefore may allow us to answer the question of whether or not a separate ethnic tick box for Sikhs would have been beneficial or detrimental for the community.

One of the headline stories from the pandemic has been the higher rate of BAME deaths, including non-white Sikhs, versus the ‘white’ population. A number of Inquiries have been launched into understanding why this is. Between 2 March to 15 May 2020 deaths per 100,000 for Sikhs males was 128.6 and 69.4 for women compared to 92.6 for Christian males and 54.6 for women. Adjusting for geographic, socio-economic and demographic factors showed a lower risk of death for Sikhs; however, for Sikh males the risk of death was still significantly higher than for Christian males.

Religion and ethnicity data is not collected as part of the death registration, or on a Medical Certificate of Cause of Death. However, ONS was able to gather this information by linking death registrations during this pandemic to the 2011 Census, allowing information such as the religious group, ethnicity and other demographic factors of the deceased to be gathered. Therefore, for the Sikh community they were able to calculate the mortality rates due to COVID-19 by age and sex, as well as adjusting these for the other factors mentioned above.

Having data on ethnicity (Indian, White English etc.) from the census enabled the ONS to add this has an additional covariate in their regression analysis, which allowed them to adjust the death rates in religious groups by ethnicity as well, and thereby provides an additional tier of analysis. Further analysis adjusting for existing comorbidities within religious and ethnic groups will provide more insights, as will an analysis over a longer time horizon. Ultimately, there is likely to be a multitude of factors that will explain disparities between different religious and ethnic groups, including proportions of each community working on the front-line in hospitals. The analysis being conducted now will hopefully prepare us better to minimise deaths overall and the disparities in death rates across religious and ethnic groups should we get a second wave of Covid-19 or another future pandemic.

The prevailing argument for the Sikh ethnic tick box was that Sikhs were being ignored in analyses being conducted by government bodies because of their lack of an ‘ethnic’ status and that religion data in the census could not be used properly. This recent analysis by ONS on death by religious group has demonstrated that religion data in the census has, and can be used effectively to present findings on the Sikh community. It also highlights how adding a Sikh tick box to the current ethnicity classification on the census will not improve statistics for Sikhs. In fact, it may actually damage what is already there by reducing the amount of information, i.e. ethnicity, currently reported by Sikhs in the census. On the basis of the evidence the decision not to include a separate Sikh ethnic tick box in the 2021 Census appears to be the correct one.

With reference to public bodies, such as the NHS and the Civil Service it is clear that they collect the compulsory ethnicity data, however, religion while it may be collected is not compulsory. Under the Equality Act 2010 religion and belief is a protected characteristic like age and disability, so efforts in the future should be directed to engaging with public bodies to ensure that they fulfil their mandatory duty to collect religion data.

A judicial review challenge has been brought by Amrik Singh Gill, on behalf of the Sikh Federation UK, against the Cabinet Office in response to the draft Census (England and Wales) Order 2020 that was laid before Parliament on 2 March 2020.