The Union government has announced that caste enumeration would be a part of the next nationwide census. Population enumeration by castes has not been published after 1931. The reason being that caste was a colonial invention that tried to replace varna and jati, and later amended with ethnographic surveys, occupation and social hierarchies.
This attempt made caste a fluid identity, making it inconvenient to fit into the diverse social systems in India. However, caste debates based on the censuses from 1881 to 1931 continue till today. A list of castes can be enforced by law, while social identities continue to be more varied—the consequent problems should be tackled.
Another real problem may crop up about estimating the socio-economic characteristics of each caste to design reservation policies and deliver welfare schemes. This can be captured only when a full-fledged census that includes a large number of socio-economic variables and caste information is conducted under the purview of the Census Act.
Even after enumeration, a better validation of the census data is always required, as it is also prone to errors. It is to be noted that the net omission rate, measured as the ratio of the number of omitted persons per 1,000 persons enumerated in the census after adjusting for duplication, was nearly 20 percent on average in the last four decennial exercises.
A general practice of a census operation is that it is followed by a Post-Enumeration Survey (PES), which involves carrying out an independent survey for a sample of the population to ascertain the accuracy of the information. The PES helps in identifying under-enumeration or over-enumeration across coverage areas. India has conducted a PES after every decennial census, but has not published the data gathered or the error corrections. This only increases our apprehension about the quality of enumeration and estimation from the census.
The census is only a short-term solution to the enumeration of population by castes and the estimation of the relative socio-economic status of each caste. The long-term solution lies in continuously tracking changes in the population by caste and their socio-economic characteristics.
The reservation systems and welfare measures are bound to have a different impact on different castes and groups within a caste. The only way to address this issue is to create a comprehensive population registry that is continuously updated to capture the population and caste dynamics.
A decennial census for any country provides a static snapshot of its population composition at a point in time, while population is a dynamic variable. Most countries are moving towards maintaining a continuous population registry (CPR) to record all types of events from birth to death of their citizens.
CPR is a century-old idea that started in Denmark in the 1920s. Thereafter, the system spread to other Nordic countries in the 1940s and strengthened in the 1960s. Today, all countries have some form of CPR with varying data standards, coverage and use.
Among countries with large populations, China has a robust system of CPR called hukou. The CPR starts with a personal identity number assigned at birth or to a migrant on entering a country. Thereafter, the sequence of events from healthcare, education, employment, movement across regions, voting registration, tax payments, and many others are seeded with the personal identity number. The CPR, if created in India, should have caste as an additional information.
Although this CPR is ideal for a country to capture the changing nature of its demography for a better and efficient delivery of welfare schemes, it has drawbacks, such as not recording all the events in the life-cycle due to non-reporting or mis-reporting. The extent of the digital divide in a country like India and a lack of a strong legal framework to protect data privacy are issues to be addressed while creating a CPR.
Until a reliable and authentic CPR is created, the second-best alternative is integrating the administrative records held by the Union and state governments of India and using them to validate the census data. The existing databases distributed across government departments can be used to create a population registry. After masking for anonymity, these databases can be provided to experts to develop a population registry and estimate the caste and socio-economic characteristics.
State governments should be involved and empowered in this process, as the Union and the states possess different caste lists and sub-categories. Also, castes listed in different categories in different states add to the complexity. As the states are closely connected and have a better understanding of the social and anthropological aspects of castes, they should be allowed to conduct surveys for region-specific socio-economic and demographic information. They should also be permitted to access the administrative records held by the Union government to validate these surveys.
The next census is expected to be technology-driven and backed by a strong data protection and privacy policy. A robust validation framework is also expected to be designed to authenticate the numbers. Integrating the administrative records held by states and the Union can be to authenticate the data. But a better, long-term solution would lie in creating a full-bodied, continuous population registry.
R Srinivasan
Member, Tamil Nadu State Planning Commission
S Raja Sethu Durai
Professor, BITS-Pilani, Dubai
(Views are personal)