Census 2027 will count caste – but the data it produces won’t be usable

Census 2027 will include caste as a category for the first time since 1931, but the data gathered using an open-column format is unlikely to serve the purposes of reservation policies or welfare programs, according to government officials. A pre-test for the second phase of the census, currently being conducted across 16 states and Union Territories, is evaluating an open column for self-reported caste. This approach differs from previous methods, allowing individuals to write down their caste rather than selecting from predefined options. The test runs until July 20, and its findings will influence how caste is recorded and managed for the coming decade. In the first phase of the census, information on housing conditions and amenities was collected. The second phase marks a significant shift, introducing caste as a new variable. Officials have confirmed that the form includes an open column for caste, enabling respondents to provide their own descriptions. However, this method has faced criticism due to its potential to generate an overwhelming number of unique caste names, making the data impractical for policy implementation. The 2011 Socio-Economic and Caste Census employed a similar open-column approach, which resulted in over 46 lakh distinct caste names. These included variations such as Gupta and Agarwal, which were essentially the same caste. Despite the effort, the data produced was deemed unusable for policy-making. Subsequent governments opted not to release the results, highlighting the challenges of relying on self-reported caste data. The current pre-test is perplexing given the government's prior stance on open self-reporting. In a 2021 affidavit submitted to the Supreme Court, the government acknowledged that self-description leads to fragmented caste classifications, rendering the data unreliable for allocating reservations in education, employment, or electoral constituencies. This admission underscores the limitations of the open-column method, despite its adoption as the preferred approach for the upcoming census. Another notable discrepancy in the first-phase forms concerns the categorization of gender and caste. While the form includes a specific transgender category for the head of household, a departure from the 2011 census, it omits Other Backward Classes and the General category. Instead, a single "Other" bucket is used for all individuals not classified as Scheduled Caste, Scheduled Tribe, or Other Backward Class. This omission raises questions about the comprehensiveness of the data collection process. Central government records indicate approximately 2,650 Other Backward Classes, along with 1,170 Scheduled Castes and 890 Scheduled Tribes. States also maintain their own lists of Other Backward Classes, demonstrating that caste names can be systematically categorized. However, the absence of standardized lists for non-Scheduled Caste and non-Other Backward Class groups, such as Bania and Brahmin, remains a critical gap. These groups, whose numerous variant spellings contributed to the inefficacy of the 2011 census, lack comparable structured listings. While listing caste names may never be fully comprehensive, the government is obligated to attempt systematic classification based on available evidence before resorting to open reporting. The closed method, which groups Other Backward Classes and dominant castes into a single category, contrasts sharply with the open method, which generates an unpredictable array of self-reported names. Both approaches result in the exclusion of the OBC and General categories, groups the census was intended to identify, thereby undermining the purpose of the exercise. The implications extend beyond mere methodology. Reservation policies, welfare initiatives, and electoral representation all rely on accurate and actionable caste data. A census producing either an unmanageable influx of names or a generic "Other" category without internal structure leaves policymakers with limited tools to address social inequalities effectively. As the pre-test concludes, the challenge remains whether the data collected will truly reflect the complexities of caste in modern Indian society.