44 | The Evolution of the Census after Independence: From 1947 to the 2027
Updated: 8 hours ago
By Gaurav Kalyani and Shivakumar Jolad
Published on: 6 November 2026
“Few people outside realise that census is an administrative operation of great dimensions and, in addition, it is a scientific process. The Indian census, in particular, covers the largest population in the world and it is also one of the most economical administrative operations. Census as an institution goes back to the remote past, but it is no longer a mere counting of heads; it involves extraction of information which plays a vital role in the determination of many of our administrative policies. The facts elicited during the course of this operation yield valuable scientific data of sociological importance. In many matters it provides a useful guide for the effectiveness or otherwise of our economic policies. The theory of population is in itself an interesting part of economics. The census helps us to test and adapt that theory”
- Sardar Vallabhbhai Patel, as quoted in (Gopalaswami, 1953a)
Introduction
After India gained its independence in 1947, the census undertaking went under significant changes, not only in the territory it covered, what it counted and how it did so; while preserving the basic skeleton of the precedence set by the previous colonial censuses. This essay traces that change from the wartime census of 1941, until the now ongoing digital census of 2027.
Statutory Census in Independent India
The Indian census began as a patchwork and became a permanent institution only after independence. The 1881 census was the first synchronous census but it excluded the mountainous and hilly tracts and some isolated areas.
The whole of India was counted on modern lines for the first time only in 1971, but that completeness has been broken since by the censuses that could not be held in Assam (1981) and Jammu and Kashmir (1991) (Natarajan, 1972a; ORGI, 2023).
In the colonial period, each census was built, run and dismantled within about three years. Yeatts, who directed the 1941 census, called this the Phoenix System . He complained that every decade's organisation started from nothing, with no standing house lists, no permanent house numbering and weak registration of births and deaths (Gopalaswami, 1953a; Natarajan, 1972a). The legal footing was just as temporary as the 1941 count rested on the Census Act of 1939.
The passage of the Census Act of 1948 established a permanent legal and administrative framework for census operations in post-independence India. Under Section 3 of this act, the Central Government officially initiated a census through a formal notification. Section 4 defined the administrative hierarchy, providing for a Census Commissioner at the national level and Directors of Census Operations (designated as "Superintendents" until 1974) in each state. To ensure compliance, Section 8 made answering census questions legally mandatory, with a specific historical exception regarding the names of female household members. Furthermore, Section 15 safeguards respondent privacy by declaring all records strictly confidential and inadmissible as evidence in civil or criminal court proceedings, except in prosecutions under the Act itself (Census Act, 1948).
Central Government officially initiated a census through a formal notification. Section 4 defined the administrative hierarchy, providing for a Census Commissioner at the national level and Directors of Census Operations (designated as "Superintendents" until 1974) in each state. To ensure compliance, Section 8 made answering census questions legally mandatory, with a specific historical exception regarding the names of female household members. Furthermore, Section 15 safeguards respondent privacy by declaring all records strictly confidential and inadmissible as evidence in civil or criminal court proceedings, except in prosecutions under the Act itself (Census Act, 1948).
  "Ever since I assumed charge of the Home Ministry, I have been taking some interest in census matters and I was really very happy to have been instrumental in placing on the Statute Bock a permanent enactment dealing with the whole operation of Census…” “Hitherto, the Census used to be looked upon as a decennial operation for which haphazard temporary arrangements used to be made. I have already stated that there is now a permanent Census Act on the Statute Book and Government have already a permanent office of the Registrar General and Census Commissioner. It is our intention through this unified organisation to effect continuous improvement over the whole field of population data including the Census and vital statistics and to conduct experiments in sampling which would reduce not only the elaboration of these operations but also the cost." -Sardar VALLABHBHAI PATEL, then Home Minister at the inauguration of the first Census Conference (1950). |
Building on this statutory foundation, the government created a permanent administrative core in 1949 within the Ministry of Home Affairs under the Registrar General, who serves ex-officio as Census Commissioner. The vital statistics were integrated under the same authority (Natarajan, 1972a).
As census administrator Ashok Mitra noted, the central census apparatus survived in part because West Bengal maintained its state-level census office through 1958, an effort that ultimately helped anchor and formally affirm the permanent central office that year (Mitra, 1994).
Over time, practical administrative experience was incorporated directly into the Census Act of 1948 through successive statutory amendments. Geographically, initial restrictions limiting the Act's jurisdiction to "Provinces and Acceding States" were removed in 1950, followed by the removal of the exception for Jammu and Kashmir in 1959 (Census Act, 1948). Major statutory reforms occurred in 1994 to resolve operational challenges encountered during the 1991 Census, where house-listing lacked clear legal authorization and prosecutions stalled due to a lack of sanction (ORGI, 2023).
The 1994 amendments significantly expanded the legal scope and administrative authority of census operations. Section 17A formally extended statutory backing to pre-tests, house-listing, and post-enumeration surveys. Section 4A legally obligated local authorities to provide staff for census operations. Sections 7A to 7H authorized administrators to requisition premises and vehicles for field logistics. Penal provisions were also established of up to three years' imprisonment for offenses committed by census personnel or those required to assist in census duties (Census Act, 1948).
From 1941 to 1951
Despite the narrative emphasizing a sharp contrast between the colonial and post-independence administration, the 1951 Census inherited significant operational methods from the wartime 1941 Census, even as it fundamentally transformed the census's legal foundation, purpose, and nationwide scope. Conducted during World War II under Census Commissioner M. W. M. Yeatts, the 1941 Census introduced major methodological shifts driven by severe staffing shortages and heightened communal tensions. Yeatts abandoned the traditional single-night headcount in favor of a 10- to 15-day enumeration window anchored to a March 1 reference time. He replaced large household schedules with individual paper slips, and cut the field workforce nearly in half. Although, the wartime constraints ultimately restricted data processing to only a fraction of British India's slips (Natarajan, 1972a, 1972b).
By contrast, as the first census of the independent Indian republic, the 1951 Census operated under the statutory authority of the Census Act of 1948. Yeatts, who was appointed Census Commissioner in 1948, died before the count in August 1950, and R. A. Gopalaswami took charge. An army of roughly 700,000 field workers enumerated 64.4 million occupied houses and completed 356.9 million individual slips between February 9 and 28, 1951, followed by a three-day revisional round (Gopalaswami, 1953a).
The 1951 enumeration introduced some key structural innovations. It established a uniform nationwide definition of a census household (sharing a main entrance and common kitchen) (Gopalaswami, 1953a), added specialized schedules to track people displaced by Partition, and conducted India's first post-enumeration field check, revealing a net undercount of 1.1 percent (Natarajan, 1972a, Appendix 10).
When compared directly, the primary distinction between the 1941 and 1951 censuses lies not in their basic counting methods, but in the administrative, legal, and financial infrastructure supporting them. Both operations abandoned the traditional single-night headcount in favor of extended multi-day enumerations anchored to a fixed sunrise reference time. It was a framework officially classified as an extended de facto count (ORGI, 2023). The 1951 census established a permanent statutory foundation under the Census Act of 1948, created a standing central office, enforced uniform nationwide definitions, introduced post-enumeration accuracy checks, and fully funded a complete tabulation program (Gopalaswami, 1953a; Natarajan, 1972a).
This structural expansion required a dramatic increase in state expenditure. Central outlays rose from a net Rs 17.5 lakh in 1941 to Rs 149 lakh in 1951, more than an eightfold increase. Although the 1941 figure excluded unpaid volunteer labor and funded significantly less data tabulation (Gopalaswami, 1953a).
Alongside these administrative changes, post-Partition territorial realignments redefined population baselines. Recalculated within India's new borders, the 1941 population was adjusted to 318.7 million, while the 1951 Census recorded 361.1 million inhabitants (including an interpolated figure for Jammu and Kashmir), representing a 13.3 percent population growth over the decade (Gopalaswami, 1953a).
National Register of Citizens in 1951 Census During the 21-day enumeration phase of the 1951 Census, enumerators recorded individual particulars of about 35.69 crore citizens across 356.9 million individual census slips, which served as personal dossiers. Either during enumeration or shortly thereafter, the most important items from these dossiers were transcribed directly into a separate compilation known as the National Register of Citizens (NRC). One dedicated part of the register was allotted to every village and every ward of every town in the country. For every enumerated individual, the NRC recorded basic demographic, economic, and socio-cultural attributes, including name, relationship to head of household, birthplace, sex, age, marital status, nationality, religion, mother tongue, literacy level, economic and employment status, principal and subsidiary means of livelihood, membership in special groups (such as Scheduled Castes or Scheduled Tribes), and displacement particulars for refugees. The NRC was designed as a permanent baseline register to facilitate specialized socio-economic studies and detailed cross-tabulations without requiring fresh field surveys (Gopalaswami, 1953a, 1953b). It is important to note that this NRC was a byproduct of the 1951 census under Census Act of 1948, and was not linked with any citizenship statute at the time. The later NRC in the 21st century was, however, explicitly mandated under Section 14A of the Citizenship Act, 1955. |
Omission of Caste in Census
One significant change from previous colonial censuses was the dropping of caste census. It shows how a 1941 retreat due to wartime crunch, became a policy 1951 onwards. Yeatts dropped all-India caste tables to save money, noted that many Hindus in Bengal and elsewhere declined to name a caste, and suggested that caste returns might eventually be replaced by the plain community question. In 1951 the elaborate caste classification was discarded and caste-wise data were kept only for Scheduled Castes, Scheduled Tribes, Anglo-Indians and backward classes (Natarajan, 1972a, Appendix 6; Sen, n.d.).
Dwaipayan Sen (n.d., working paper) demonstrates that the decision was the culmination of a much wider ideological consensus shared across early postcolonial leadership that included Prime Minister Jawaharlal Nehru, Home Minister C. Rajagopalachari, parliamentarians like R.R. Divakar, and senior census administrators such as Census Commissioner R. A. Gopalaswami and Ashok Mitra. Although standard public and scholarly narrative heavily attributes the removal of caste categories solely to Home Minister Vallabhbhai Patel (Sen, n.d.). Addressing the 1950 Conference of Census Commissioners,
Sardar Patel argued that earlier elaborate caste tables satisfied colonial theories of a "caste-ridden country," whereas the new nation needed to focus its administrative energy on economic means of livelihood.
As administrator Ashok Mitra noted in his memoirs (The New India), the 1951 Census was envisioned as a fundamental break from the colonial era (Mitra, as cited in Sen, n.d.).
Anticolonial nationalists, drawing on sociological works like G.S. Ghurye's Caste and Race in India, argued that British policy in the army, legislature, and census had actively fostered "communal consciousness" and manufactured caste divisions (Sen, n.d.). Consequently, removing caste from government registers and census forms was seen as a patriotic necessity to undo colonial policy. However, writing in his 1955 book Thoughts on Linguistic States, Dr. B. R. Ambedkar expressed deep disappointment over the government's decision to drop caste tables from the first census of independent India (Sahu, 2026).
Pinning responsibility on Home Minister C. Rajagopalachari, Ambedkar sharply criticized the official rationale stating, “The Home Minister of the Government of India who is responsible for this omission was of the opinion that if a word does not exist in a dictionary it can be proved that the fact for which the word stands does not exist. One can only pity the petty intelligence of the author” (Ambedkar, 1955, as cited in Sahu, 2026). Ambedkar argued that carving out states on a purely linguistic basis without accounting for local caste structures risks transferring power to a dominant "communal majority". Rejecting generic counts of religious groups (Hindus, Muslims, Christians, etc.) as virtually meaningless for understanding power dynamics, he insisted, “What we want to know is the distribution of castes in different linguistic areas” (Ambedkar, 1955, as cited in Sahu, 2026).
Language controversy in 1951 Census Another controversy that occurred during the 1951 census was the communal tensions and field level falsifications in North-West India. While language enumeration proceeded smoothly across most of India, the language question triggered widespread agitation and public controversy across Punjab, PEPSU, Delhi, Himachal Pradesh, and Bilaspur. A section of Hindus launched a campaign urging people to declare Hindi as their mother tongue instead of Punjabi. This campaign met with intense resentment from Sikhs, triggering aggressive statements and counter-statements in newspapers and public speeches (Gopalaswami, 1953b). The agitation escalated into violence on February 10, 1951, when an altercation between Hindu and Sikh groups in a village in the Jullundur district resulted in a clash and the murder of one person. As a direct result of this public conflict, respondents extensively falsified their answers, rendering the recorded returns for mother tongue (whether Hindi or Punjabi) highly inaccurate across a large portion of the Punjab and PEPSU areas (Gopalaswami, 1953b). The census administrators decided not to separate or publish individual figures for these languages in the affected region. During sorting and tabulation, all returns were combined under a single unified heading: "Hindi-Punjabi-Urdu-Pahari" in both mother-tongue and bilingualism tables. Registrar General R. A. Gopalaswami noted the "regrettable result" that because of this grouping, no accurate All-India total could be specified for Hindi, nor could distinct figures be provided for Punjabi or Urdu speakers across North-West India in 1951 (Gopalaswami, 1953b). |
The 1951 report authored by Gopalaswami, devotes extensive analysis to land area, agricultural land use, livelihood structure, and the historical trend of cultivation relative to population growth. It recorded 81 crore acres of land for a population of 36 crore people, yielding a basic land area per capita of 2.25 acres (225 cents) (Gopalaswami, 1953a).
The report identifies 1921 as the "Great Divide" in Indian demographic and agricultural history. Prior to 1921, population growth was repeatedly checked by severe famines and epidemics. After 1921, the population grew rapidly without natural checks while land extension slowed. Across the country, the net cultivated area sown per capita averaged 79 cents. Although this ratio varied widely by region. A specialized study brought together land-use data across 5 natural regions, 15 sub-regions, and 52 natural divisions to analyze the relationship between terrain, soil quality, rainfall belts, and population carrying capacity (Gopalaswami, 1955 ).
The report emphasizes that land area and cultivated acreage statistics were among the most accurate national statistics in India. To feed a projected population of 52 crores by 1981, the report estimated that agricultural productivity would need to increase by 54% above 1951 levels. The Registrar General concluded that land expansion and irrigation alone could only increase productivity. The report also urged that agricultural planning must be paired with voluntary birth control to stabilize the population (Gopalaswami, 1953a).
From 1951 to 1961
Between 1951 and 1961 the census gained territory far more than it lost. Jammu and Kashmir and the north-eastern frontier were drawn in, former French and Portuguese enclaves were also brought under the Indian count, and the methods were rebuilt around housing, maps and work.
The 1951 Census contained three major coverage gaps arising from legal, political, and logistical constraints. The state of Jammu and Kashmir was completely omitted from field enumeration due to post-Partition security, political, and administrative conditions (Gopalaswami, 1953a). This exclusion was also statutory, as the original Census Act of 1948 excluded Jammu and Kashmir until Parliament amended the law in 1959 (Census Act, 1948).
Similarly, direct enumeration of North-Eastern Frontier Tribal Areas (Arunachal Pradesh), also known as Part B tribal areas of Assam, leaving an estimated 6 lakh individuals uncounted. Operational coverage in the Tuensang area occurred for the first time in 1951, but remained restricted to 7,025 residents across 129.5 square kilometers. Lastly, Areas beyond formal Indian administration or contact were also excluded. Population figures for Pondicherry relied on a 1948 French-era count, while indigenous groups in the Andaman Islands, such as the Jarawa and Sentinelese, remained entirely uncontacted (Natarajan, 1972a).
New territorial inclusions
The 1961 Census marked a major administrative expansion of demographic coverage across previously omitted, frontier, and newly integrated territories in India. Jammu and Kashmir was formally enumerated for the first time post-independence. while the North-East Frontier Agency (NEFA) was censused for the first time alongside expanded field operations covering 5,356.1 square kilometers in Tuensang (Natarajan, 1972a).
In former colonial enclaves, Pondicherry was directly enumerated, while Goa, Daman, and Diu were counted as of December 15, 1960, and Dadra and Nagar Haveli were enumerated as of March 1, 1962. Conversely, Sikkim remained outside the scope of the Indian census. Although official records group Sikkim under the 1961 Census, statutory census authority was extended there only in 1976, with full enumeration taking place in 1981. While subsequent official tables retroactively back-filled population figures for 1951 and 1961, these estimates fall outside the formal boundaries of the 1961 Indian census. The uncontacted groups in Andaman Islands remained uncounted (Roy, 1992).
Language, linguistic boundaries, and 1961 Census Inside the old frontiers the map had also changed significantly. The 1956 linguistic reorganization had redrawn states largely along mother-tongue lines, so each 1961 state report opened with an appendix tracing how the state's area had changed since the last census. During the state reorganisation process, the States Reorganisation Commission (SRC) utilized figures and raw data from the 1951 Census as a critical empirical tool to evaluate territorial claims, determine language majorities, and delineate state boundaries. However, the Commission treated census data as one of several factors rather than an absolute rule.
It used the 1951 language data to test whether areas met the 70% threshold recommended by the Dhar Commission for a unilingual classification (SRC, 1955). In several highly contentious regions, political and communal groups disputed the accuracy of the 1951 Census, prompting the Commission to look beyond census figures. The SRC also drew on 1951 Census demographic data (such as Census Paper No. 5 of 1954) to assess regional literacy rates and administrative efficiency when evaluating proposals to split large states. For example, 1951 Census figures highlighting Uttar Pradesh’s low literacy rate (10.8%) were analyzed alongside state social service expenditure to examine arguments for partitioning Western U.P (SRC, 1955). |
The 1961 Census introduced major infrastructural, economic, and methodological advancements that modernized Indian demographic data collection. Operationally, administrators prepared notional maps for roughly one million villages and urban blocks to define enumerator beats, establishing a standardized geographical foundation for field operations. Urban areas were classified under a uniform nationwide definition, while "town groups" were created to capture continuous urban spread. Economically, enumeration returned to a work-based framework rather than an income metric, introducing distinct categories for household industries and family workers, as well as a specialized Household Schedule for landholdings and cottage industries (Mitra, 1994).
The 1961 Census marked a major methodological shift by introducing a standardized, uniform Houselist schedule across all States for the first time. Driven by the data needs of the Planning Commission, the National Buildings Organisation (NBO), and the Ministry of Commerce and Industry, Census Commissioner Ashok Mitra expanded the routine house-numbering drive into a comprehensive, nation-wide Housing Census (Mitra, 1994).
Advanced logistical planning, beginning in 1958 under Ashok Mitra, allowed post-enumeration accuracy checks to be built directly into the schedule, reducing net under-enumeration to 0.7 percent (down from 1.1 percent in 1951) (Natarajan, 1972a, Appendix 10). Concurrently, publication output expanded dramatically from 17 volumes and 307 district handbooks in 1951 to 26 project series encompassing 1,476 volumes in 1961 (Mtra, 1994).
From 1961 to 1971
The 1971 Census of India marked the centenary of modern census-taking in the country (1872–1971) and achieved universal national coverage across all 18 States (excluding Sikkim, which was not yet part of India) and 10 Union Territories, enumerating a population of 54.81 crores. The census reference date was postponed from March 1 to sunrise on April 1, 1971, due to the mid-term general elections, with field enumeration carried out between March 10 and 31, followed by a three-day revisional round (Natarajan, 1972a; ORGI, 2023).
This operation introduced a comprehensive Establishment Schedule covering all manufacturing, commercial, and service units, while replacing the earlier "Town Group" framework with Urban Agglomeration and Standard Urban Area (SUA) to evaluate continuous urban growth. Additionally, the 1971 Census initiated a major technological transition toward Electronic Data Processing (EDP) by tabulating data on electronic computers, while replacing the dedicated Household Schedule with a Census Population Record for each household (Natarajan, 1972a, 1972b).
The 1971 questionnaire incorporated significant enhancements to meet the analytical requirements of government ministries, the Planning Commission, and demographic researchers. For the first time, the census collected data on current fertility, recording age at marriage and births within the last year for currently married women. It also expanded internal migration metrics to record the place of last residence and duration of stay. Economic questions were substantially restructured on a "work" basis into a two-part frame capturing Principal Activity and Secondary Activity. Furthermore, the administration distributed an expanded Degree Holders and Technical Personnel Card across all academic disciplines and launched ancillary socio-economic initiatives, including intensive urban and village re-studies (Natarajan, 1972a).
Indian Census Operations After 1971
1981
The 1981 Census of India operated across two main phases - Houselisting in 1980 and Population Enumeration from February 9 to 28, 1981, with a revisional round from March 1 to 5. Field data collection was carried out by a massive workforce of 12.5 to 17 lakh enumerators and supervisors, primarily local government schoolteachers and revenue officials. However, due to severe political agitation and disturbed law-and-order conditions, population enumeration could not be conducted in the state of Assam. Methodologically and technologically, the 1981 Census introduced several pioneering shifts to refine economic measurements and improve processing turnaround times (ORGI, 2023).
For the first time, the Houselisting operation was integrated with the Central Statistical Organization’s Economic Census to canvass an Enterprise List. The nationwide data on physically handicapped individuals (covering the blind, crippled, and dumb) was collected for the first time since 1931. To capture secondary and female labor force participation more accurately, economic questions were restructured into Main Work (183 days or more), Marginal Work (less than 183 days), and Secondary Work, alongside explicit queries for unemployed seeking work. This census adopted a two-tier schedule system, canvassing a Universal Slip for the total population and a 20 percent Sample Slip for detailed migration and fertility data. It also modernized data processing by replacing manual card punching with direct key-to-tape entry and 100 percent computerization of Household Schedules (ORGI, 2023).
1991
Conducted under a two-phase framework with the standard reference moment on 1 March, the 1991 census operation mobilized a field workforce of approximately 1.6 million (16 lakh) enumerators and supervisors. While the state of Jammu and Kashmir was excluded from primary enumeration due to disturbed law-and-order conditions, Assam was successfully censused following its 1981 omission. Methodologically, the 1991 Census expanded the Houselist Schedule from 14 to 24 columns to capture detailed living conditions—such as predominant floor material, electricity, sanitation, drinking water availability, and, for the first time in census history, cooking fuel and female-headed households. Additionally, an intensive collaboration with UNIFEM and academic research centers introduced probing questions and pictorial training to capture women's economic contributions more accurately, raising the recorded national female work participation rate (ORGI, 2023).
Technologically and logistically, the 1991 Census marked a pivotal transition point by introducing large-scale computerization for data entry and processing, while leveraging the National Informatics Centre Satellite Link (NICNET) to transmit provisional population totals rapidly from district headquarters. Field training was modernized using standardized audio-visual overhead projector (OHP) modules, and census outputs were distributed for the first time on electronic media such as magnetic tapes, floppy disks, and CD-ROMs alongside traditional printed volumes. However, the operational experience exposed critical statutory deficiencies in the Census Act of 1948, notably the lack of explicit legal authorization for houselisting and severe prosecution bottlenecks. It directly led to catalyze major legislative reforms in the 1994 Census Act amendments (ORGI, 2023; Census Act, 1948).
2001
The Census of India 2001 (termed the "Millennium Census") was the 14th decennial population count of India and the first of the 21st century. It covered 35 States and Union Territories, 593 districts, and nearly 600,000 villages and towns. It followed a similar structural pattern of the previous census to execute the census operations. the 2001 enumeration successfully navigated major localized disruptions, including the catastrophic January 2001 Gujarat earthquake in Kachchh, flash floods in Himachal Pradesh's Kinnaur district, and non-synchronous winter coverage in Jammu & Kashmir (ORGI, 2023).
Methodologically and technologically, the 2001 Census introduced several structural breakthroughs. It abolished separate individual slips in favor of a single, unified Household Schedule containing 23 questions. It mandated respondent signatures or thumb impressions for authentication, and introduced the Permanent Location Code Numbering (PCLN) system, which assigned a continuous 8-digit unique code to every village to ensure seamless tracking across future boundary realignments (ORGI, 2023).
Furthermore, the census revived disability enumeration across five specific categories (seeing, hearing, speech, movement, and mental) after its omission in 1991, while expanding queries on commuting modes and domestic agricultural production. Crucially, it marked a landmark technological transition by adopting Intelligent Character Recognition (ICR) high-speed image scanning across 15 regional centers to achieve 100 percent data digitization, slashing publication turnaround times from 5–10 years down to 2–3 years (ORGI, 2023).
2011
The census of 2011 was the 15th decennial population count since 1872 and the last until it was postponed in 2021 due to Covid-19 pandemic. It operated under the national motto “Our Census, Our Future”. It introduced for the first time the option for gender self-identification ('Others'). Other structural innovations such as capture of exact dates of birth alongside completed age, the expansion of disability categories from 5 to 8, and updated asset tracking for mobile phones and internet-connected computers (ORGI, 2023).
To manage logistics, the census established a strategic partnership with India Post to transport over 21,000 metric tonnes of scannable materials, utilizing high-speed Intelligent Character Recognition (ICR) image scanning across 17 regional data centers to achieve 100 percent data digitization. Despite nationwide completion, operations encountered localized disruptions including field halts across 103 villages in Chhattisgarh and 260 villages in West Bengal, data withholding pending verification in three sub-divisions of Manipur's Senapati district, and the transit loss of 22 schedule boxes in Jammu and Kashmir (ORGI, 2023).
Caste question in 2011
In late 2008 and early 2010, a coordinated campaign by anti-caste activists, public intellectuals, and parliamentarians across party lines pressured the government for a full caste count. Following an extensive May 2010 Lok Sabha debate, the ruling United Progressive Alliance (UPA) government, led by Prime Minister Manmohan Singh and Finance Minister Pranab Mukherjee, publicly conceded to conduct a nationwide caste-wise enumeration for the first time since 1931. Despite the political concession, senior administrators within the Office of the Registrar General and Census Commissioner, India (ORGI) and the Ministry of Home Affairs strongly opposed adding caste questions to Census 2011, contending that it would jeopardize the "integrity" and operational accuracy of the headcount. Due to this ‘bureaucratic deflection’, 2011 census operations were concluded without caste questions (Vithayathil, 2025).
However, in May 2011, the central government rerouted the caste count out of the statutory decennial census into the Below Poverty Line (BPL) poverty survey, renaming the non-statutory exercise the Socio-Economic and Caste Census (SECC) or SEC Survey 2011. Ground-level SECC operations began in July 2011 using handheld tablet PCs managed by private contractors. It had several internal irregularities and flaws and the logistical and software bottlenecks delayed field enumeration across states until 2015. In July 2015, the central government released the provisional rural poverty and SC/ST data but withheld the caste-wise dataset. In July 2015, the government appointed an expert committee chaired by NITI Aayog however, the committee never published a report and reportedly never formally met. The central government ultimately declared the SECC caste data "unusable" and "unreliable" due to technical flaws, leaving the nationwide caste data unpublished and highly contended (Vithayathil, 2025).
Check out The Appendix, which covers an extended and comprehensive summary of 1981-2011 censuses.
The evolution of data work
The post-independence evolution of census data in India reflects a deliberate transition from colonial-era demographic profiling toward a dynamic, policy-oriented framework for national planning and socio-economic welfare. Over subsequent decennial cycles, data collection continuously expanded in scope and analytical depth, along with profound technological shifts in data processing. It went from progressing from manual punch-card sorting and early electronic mainframe computers to 100% Intelligent Character Recognition (ICR) image scanning. It was first utilized in 2001 and then in 2011, achieving 100% data digitization and slashed processing time to 2–8 years. For the 2021 cycle, a complete shift toward digital enumeration via Mobile Applications under a Bring-Your-Own-Device (BYOD) mechanism was developed and pre-tested in 2019 (ORGI, 2023).
Item / Census Year | 1961 | 1971 | 1981 | 1991 | 2001 | 2011 | 2021* |
Population (Million) | 439 | 548 | 683 | 846 | 1028 | 1210 | NA |
% Field Data Collection | 100% | 100% | 100% | 100% | 100% | 100% | NA |
% Data Digitized | 5% | 15% | 25% | 45% | 100% | 100% | NA |
Data Capture Method | HandPunch | KeyPunch | OfflineData Entry | PC BasedData Entry | Scanning /ICR | Scanning /ICR | CAPI /ICR |
Digitization Time | 5–10 Yrs | 5–10 Yrs | 5–10 Yrs | 5–10 Yrs | 3–7 Yrs | 2–8 Yrs | NA |
* Note: Mobile Applications were developed for data collection in Census 2021 along with an option for paper mode. Source: ORGI, 2023
Roy's 1992 paper identifies the problem that remained. No mechanism tied a block in one census to the corresponding block in the next, so changes in boundaries and place names were invisible to users. His example is the Kochi urban agglomeration, which had 20 components in 1991 whose constituents and spellings had shifted since 1981. He proposed geocoding enumeration areas for future censuses (Roy, 1992).
The 2027 Census
The 2027 census keeps the Census Act and the two-stage design that 1961 made standard, but it changes the medium, the questions and the political stakes. After being postponed in 2021, the intention to hold it was gazetted on 16 June 2025, and in December 2025 the Cabinet approved ₹11,718.24 crore and about 30 lakh personnel (India Tribune, 2025). Fieldwork for this census will be digital for the first time. Enumerators will use their own phones, which send data to a central server, with paper as a fall-back that must be keyed in within seven days. For the first time digital self-enumeration is introduced, where a Households could complete an optional online self-enumeration before the official visit by the census enumerator (PIB, 2026a). More than 1.44 crore had done so by mid-May 2026. Houselisting blocks are drawn on satellite-based web maps, and a monitoring portal lets supervisors follow progress.
Administrative boundaries were frozen on 31 December 2025 and may not change until 31 March 2027 (PIB, 2025). This follows the freezes of 1990 to 1991 and 1999. Houselisting ran in 30-day state windows between 1 April and 30 September 2026, with 33 questions against 35 in 2011. Ladakh and the snow-bound parts of Jammu and Kashmir, Himachal Pradesh and Uttarakhand were enumerated in September 2026 against a 1 October reference date. The rest of the country follows in February 2027, with midnight of 1 March 2027 as the reference moment. Goa, Uttar Pradesh, Punjab and the rest of Uttarakhand are scheduled from 1 December 2026 to 4 January 2027 because of assembly elections, an echo of 1971, when elections moved the reference date from 1 March to 1 April (Natarajan, 1972a).
The population schedule notified on 14 August 2026 has 40 questions against 29 in 2011 (PTI, 2026b). Question 10 reads "Scheduled Caste (SC)/Scheduled Tribe (ST)/Caste", and the enumerator records the caste the respondent declares. It is the first caste enumeration across all communities since 1931 and the first since independence (DTE Staff, 2026).
No territory is excluded outright, but Manipur's houselisting, due to begin on 1 September, was deferred "till further declaration" by a notification of 5 September that amended the January one (News On Air, 2026).
On 17 April 2026 the Lok Sabha rejected the 131st Amendment Bill, which would have enabled delimitation on 2011 data (Shrivastava, 2026). The existing Delimitation Act, 2002 therefore continues, and the 2027 count is expected to underpin the next delimitation. In late October the Registrar General told state units not to release or use any figures meanwhile, citing the Act's confidentiality.
Concluding Reflection
The largest break in method came in 1941, not 1951. Wartime shortage forced the extended count, the individual slip and sampling, and 1951 made them permanent by adding what 1941 lacked - a statute, a standing office, uniform definitions and a tabulation plan fixed in advance. The change that independence itself brought was institutional, together with a change of purpose of the census.
The 2027 Census exposes the unresolved tensions that have run through the record since 1951. The caste question, the deferred houselisting in Manipur, the confidentiality instruction to state units and the delimitation shadow all show that the census remains a site where statistical accuracy and political demand pull against each other. This long view underscores that census is not merely a neutral record of the nation, it remains an active negotiation with what the state is willing to see. The 2027 count, with digital and self-enumeration and inclusion of caste, will be the test of whether that negotiation has become more inclusive or merely more efficient. The stakes are also higher than in most past rounds, because the count underpins the next delimitation, and the schedule now asks for identity numbers while the Act’s promise of confidentiality stays the same.
(Authors: Gaurav Kalyani works as Research Associate at the Center for Legislative Education and Research, FLAME University, Pune;
Dr. Shivakumar Jolad works as Associate Professor (Public Policy), and is the Chair of Center for Legislative Education and Research and Director India State Stories, FLAME University, Pune
Gaurav did primary research and writing; Shivakumar contriuted to conceptualization, research, writing and editing)
Bibliography
Press Information Bureau. (2026a). No more pen and paper: Digital Census 2027. New India Samachar, 16. https://newindiasamachar.pib.gov.in/WriteReadData/flipbook/2026/May/2nd/English/files/basic-html/page16.html
Press Trust of India. (2025, June 29). Boundaries of administrative units to be frozen on Dec 31, 2025 for Census. Arunachal Times.
Press Trust of India. (2026b). Centre notifies 40 questions for Census 2027, caste details included. Telangana Today. https://telanganatoday.com/centre-notifies-40-questions-for-census-2027-caste-details-included
Census Act, No. 37 of 1948 (India) (as amended by Act No. 11 of 1994).
DTE Staff. (2026). Census 2027 to include question related to caste, a first in independent India. Down To Earth. https://www.downtoearth.org.in/governance/census-2027-to-include-question-related-to-caste-a-first-in-independent-india
India Tribune. (2025, December 12). Census 2027 gets Cabinet nod; digital count to deploy 30 lakh personnel and include caste data. https://indiatribune.com/public/census-2027-gets-cabinet-nod-digital-count-to-deploy-30-lakh-personnel-and-include-caste-data
Gopalaswami, R. A. (1953a). Census of India, 1951: Vol. 1, India. Part I-A — Report. Manager of Publications.
Gopalaswami, R. A. (1953b). Languages — 1951 census (Census of India Paper No. 6). Office of the Registrar General, India.
Gopalaswami, R. A. (1955). Census of India, 1951: Vol. 1, India. Part I-B — Appendices to the census report. Manager of Publications.
Mitra, A. (1994). Census 1961: New pathways. Economic and Political Weekly, 29(51/52), 3207–3221. https://www.jstor.org/stable/4402155
Natarajan, D. (1972a). Indian census through a hundred years: Part I (Chapters I & II) (Census of India 1971, Centenary Monograph No. 2). Office of the Registrar General, India, Ministry of Home Affairs.
Natarajan, D. (1972b). Indian census through a hundred years: Part II (Chapters III–V) (Census of India 1971, Centenary Monograph No. 2). Office of the Registrar General, India, Ministry of Home Affairs.
News On AIR. (2026, September 5). Centre defers house-listing operations for Census 2027 in Manipur until further declaration. https://newsonair.gov.in/centre-defers-house-listing-operations-for-census-2027-in-manipur-until-further-declaration/
Office of the Registrar General & Census Commissioner (ORGI), India. (2023). A treatise on Indian censuses since 1981. Ministry of Home Affairs, Government of India.
Roy, B. K. (1992). The census mapping: Retrospectives and some futuristic views by 2001 AD. Demography India, 21(1), 73–83.
Sahu, S. N. (2026). Why Ambedkar’s 1951 regret still echoes in the 2026 caste enumeration. The Leaflet. https://theleaflet.in/opinion/why-ambedkars-1951-regret-still-echoes-in-the-2026-caste-enumeration
Sen, D. (n.d.). The exclusion of caste enumeration from the 1951 census of India [Unpublished working paper].
States Reorganisation Commission. (1955). Report of the States Reorganisation Commission. Government of India Press.
Vithayathil, T. (2025). Counting caste: Census politics, bureaucratic deflection, and Brahmanical power in India. Cambridge University Press.
Comments