Sikhism, ethnicity and the 2021 Census: R (Gill)

In R (Gill) v UK Statistics Authority [2019] EWHC 3407 (Admin), Mr Amrik Singh Gill, as Chair of the Sikh Federation UK, sought judicial review of “the contemplated exercise of Her Majesty’s discretion to direct a census, based on [an] Order in Council made on the basis of a draft which does not include a Sikh ethnic tick box on the basis of the reasoning detailed in the White Paper” [1: emphasis added] and a declaration that “it would be unlawful for Her Majesty to make an Order in Council which follows the reasoning of the White Paper so as not to include the Sikh ethnic group tick box in the ‘particulars to be stated in returns’ section of the Order” [2].

In short, he claimed that the Office for National Statistics had adopted an unlawful approach in assessing whether or not the 2021 census questions about membership of ethnic groups should include a specific tick-box response option for Sikhs, in addition to the option of entering “Sikh” in the space marked “other” [5].


The 2011 Census questionnaire had asked “what is your religion?” and there was a list of tick-box response options, including one for “Sikh”. 423,158 respondents ticked the Sikh response option [15]. As to the topic of ethnic group, there had been a write-in option for “Any other Asian background” [16] and 83,362 respondents wrote in “Sikh (76,500 of whom had also identified as Sikh under the religion question) [17]. However:

“According to Mr Iain Bell, Deputy National Statistician and Director General of Population and Public Policy at the ONS, prior to the 2011 census, there was some demand for a Sikh tick box response option, mainly from Sikh community organisations, including the Sikh Federation. Other views were expressed by organisations and individuals. The ONS also considered that a different picture emerged from the cognitive testing by Ipsos MORI, commissioned for the Scottish census, in which most of the seven Sikh respondents ticked the Indian response option, and were uncomfortable at having to choose between ‘Indian’ and ‘Sikh’. Some felt that Sikhism was a religion and were confused about its inclusion as an ethnic group [18]”.

In March 2009 the ONS published an Information Paper, Deciding which tick-boxes to add to the ethnic group question in the 2011 England and Wales census, [19] and in January 2010, the Sikh Federation published a 28-page critique of the ONS’s Information Paper, commenting on the prioritisation tool and its application to Sikhs and concluding that “[w]hile the principles of the tool appear to be well reasoned and coherent the evidence base for Sikh scores and the scoring process applied are inconsistent, contradictory and not transparent” [21]. In December 2017, the ONS’s “Census topic research” report discussed the ethnicity topic and explained that the eight tick boxes had been reduced down to a final four for consideration: Jewish, Roma, Sikh and Somali [41] and it commissioned independent research from Kantar Public, a social research agency, to conduct a qualitative study in respect of the final four ethnic group categories [42]. Kantar concluded that older, male participants, who viewed Sikhism as an important part of their background, favoured a Sikh tick-box, while second-generation participants identified as British and were less likely to want Sikh included under ethnicity – and some raised concerns that Sikhism was not an ethnic identity [44].

The claim and counterclaim

Mr Gill and the Sikh Federation submitted that it would be unlawful for the Queen to exercise her discretionary power to make an Order in Council under the Census Act 1920 if the particulars to be stated in the census return (and scheduled to the Order) did not include a Sikh ethnic group tick-box response, as proposed in the because that proposal was based upon “unlawful reasoning” [61]:

  • Ground 1: Unlawful failure by the ONS to apply the evaluation criteria it had published and said it would apply (and until the final announcement had applied), including its published “public acceptability” test, when recommending not to include a Sikh tick box in the 2021 Census [62].
  • Ground 2: Unlawful failure by the ONS to apply its “public acceptability” criterion consistently across the questions and response options considered for inclusion under various topics/sub-topics in the 2021 Census [65]: it had treated the existence of negative responses as fatal to satisfying the “public acceptability” criterion in respect of the Sikh ethnic group question while taking a different approach to negative responses about the gender identity and sexual orientation questions [66].
  • Ground 3: Unlawful reliance by ONS, when recommending not to include the option of a Sikh tick-box, on a report by Kantar, a research agency commissioned to conduct focus groups to assess the inclusion of additional ethnic group tick-boxes [67]: it had been unlawful for Kantar to have regard to the view expressed in the focus groups that Sikhism was a religion rather than an ethnic group, because Kantar was bound to proceed on the basis that Sikhs were an ethnic group within the meaning of the Race Relations Act 1976, following the decision of the House of Lords in Mandla v Dowell Lee [1983] 2 AC 548 [69].

The UK Statistics Authority argued that the Claimant’s grounds were based on fundamental misconceptions:

  • As to Ground 1, the ONS had never promised to apply the topic criteria to evaluate the proposed additional tick-box responses for ethnic groups and had expressly stated in its published documents that it would use similar methodology to that used in preparation for the 2011 Census, subject to review and updating – and that had been made clear to stakeholders and all those organisations and individuals who engaged with the ONS on this issue, including the Sikh Federation [72]. Further, there was no legal obligation on the ONS to publish in advance “the relatively minor adjustments made to the methodology used for the 2011 census” [73].
  • As to Ground 2, sexual orientation was a topic, not a response option and, in any event, there was no valid comparison between the sexual orientation question and the Sikh tick box response [74].
  • As to Ground 3, the complaints about Kantar’s research had been based on a misunderstanding of the criteria and methodology used: the research was qualitative, not quantitative, and Kantar and the ONS were entitled to have regard to the views expressed in the focus group as to whether Sikhs were an ethnic group [75 & 76].

Further, the Defendant raised a preliminary objection to the Court determining the claim, on the ground that it was premature, because no ministerial decision or Order in Council had yet been made, and because it would be in breach of Parliamentary privilege: a declaration in the terms sought by Mr Gill would not respect the separation of powers between the legislature and the judiciary [78].


Lang J rejected the application. It was not an exceptional case that justified departing from the general rule that the Court would respect the separation of powers and not interfere with Parliamentary proceedings. Further:

“Under this legislative scheme, no justiciable decision has been made. The Minister has not yet made a draft Order in Council […] The claim is plainly premature […] As this claim concerns secondary, not primary legislation, the Claimant will be able to bring a challenge to the Order in Council once made, if the Sikh tick-box response is not included, and if he has valid grounds on which to issue a claim” [108].

Claim dismissed, “on the ground that it is premature, and in breach of parliamentary privilege and the constitutional convention of the separation of powers” [109].

Cite this article as: Frank Cranmer, "Sikhism, ethnicity and the 2021 Census: R (Gill)" in Law & Religion UK, 19 December 2019,

Leave a Reply

Your email address will not be published. Required fields are marked *