Counting the Third Gender: Lessons from 2011 for an Inclusive 2021 Indian Census
The 2011 Census was the first major census accounting of third gender persons in India. It showed us that over 490,000 third gender individuals live in the country. According to the Times of India, “Over 66% of the population identified as third gender lived in rural areas, very close to the 69% of the overall population that lives in villages. The census data also revealed the low literacy level in the community, just 46%, compared to 74% literacy in the general population”. However, all mainstream articles written about this census present a very limited picture of the status quo for third gender persons. To better understand the community, I began the project of mapping the noteworthy 2011 Primary Census Abstract for Others table (hereafter called the 2011 Third Gender table). Here is my attempt to contextualize some of the findings, discuss the limitations of the 2011 Third Gender table, and provide some suggestions for restructuring the 2021 census counting of third gender persons.
Findings
Let’s start with a simple metric: population. The 2011 census tells us that the ratio of third gender population to total population is higher in Andhra Pradesh, Orissa, Madhya Pradesh, Jharkhand, Uttar Pradesh, Uttarakhand, and Manipur. Gujarat and Kerala seem to have the lowest ratios. This may be simply an artifact of population — i.e. since states like Uttar Pradesh and Andhra Pradesh are already more populated than others, the number of third gender individuals may seem higher by comparison. Alternatively, these states may house communities which allow for open self-identification as third gender, and better livelihoods for third gender persons through work opportunities or less stigma.
The 2011 census also shows that children comprise a significant percentage of India’s third gender population. Uttar Pradesh, Bihar, Meghalaya, Mizoram, Manipur, and Nagaland have the highest percentage of third gender children. This information is not just useful for counting purposes — it may help redefine healthcare, food access, and early childhood education in these regions to make these systems more inclusive of third gender children.
Next, we examine literacy rates of third gender and total populations. The map on the left shows that third gender individuals in Rajasthan, Madhya Pradesh, Jammu and Kashmir, Bihar, Jharkhand, and Meghalaya are, on average, significantly less literate than the rest of the states. Of these states, Madhya Pradesh and Jharkhand also have a higher ratio of third gender individuals. Combined with lower literacy rates, third gender individuals in these states may face further difficulties related to occupations and livelihoods. Third gender persons in Kerala, Karnataka, Goa, Maharashtra, and Mizoram have literacy rates at par with, or exceeding, their state average. It would be interesting to examine which local and state policies lead to higher literacy rates for third gender persons in Western coastal states.
Third gender persons in 28 of 36 Indian states face unemployment rates higher than the total population
Unemployment trends for third gender individuals are also very interesting. Comparing the percentage of non-working persons in both third gender and total populations, we see that third gender individuals in 28 of 36 states and Union Territories face unemployment rates higher than the total population. Of these states, third gender unemployment is highest in the states of Kerala, Orissa, and Chattisgarh. Many states in the northeast have very low unemployment ratios between the two communities— which is largely driven by high unemployment rates in these states overall. Combined with low education and high unemployment, third gender persons in states like Orissa and Chattisgarh may face further challenges.
Third gender persons in urban areas of Madhya Pradesh, Chattisgarh, Jharkhand, West Bengal, Assam, and Nagaland are 23–25% more literate than those in rural areas.
Rural/Urban Gaps
The rural/urban gap is more stark for third gender persons than for the total population. As seen from the map of rural/urban literacy gap, there is a range of disparity in literacy across rural and urban areas of each state. Third gender persons in Kerala, Sikkim, Arunachal Pradesh, and Manipur are at least as literate as the rest of the population. However, rural and urban literacy is very disparate in Madhya Pradesh, Chattisgarh, Jharkhand, West Bengal, Assam, and Nagaland. Urban areas in these states are 23–25% more literate than rural areas, indicating a severe gap between people and the educational institutions that serve them.
Evaluating employment in rural and urban areas also shows a severe gap. Third gender persons consistently report higher employment rates in rural areas than in urban areas in most states. In Andhra Pradesh, Chattisgarh, Jharkhand, Meghalaya, and Sikkim, rural areas trump urban areas for jobs by over 16%. In northern states like Jammu and Kashmir, Punjab, and Haryana, the alternate is true. Since these states have highly populated urban areas, it is easier to find a job in the city, and the gap stands at about 3–7%. In the Union Territories of Delhi and Chandigarh, this rural/urban gap jumps to 26% and 38% respectively, predictably in favor of urban areas. Despite this concentration, most of the third gender population of India reports higher employment in rural areas, as shown by the prevalence of red-orange states. Comparing both maps, notice that Chattisgarh and Jharkhand report large gaps in both literacy and employment between rural and urban areas.
Challenges
Publicly available data lacks the required spatial and demographic resolution
State-level findings can only tell us so much. Active inclusion of third gender individuals in the census requires data on these communities at the district or taluk level. Available information in the 2011 Third Gender table is not nearly enough to fully understand the varying spatial distribution of access and opportunity. However, I realize that data at the district or taluk level comes with privacy concerns. In many areas, third gender individuals already face stigma, harassment, and violence due to their gender identity. Being able to isolate sensitive data to the district level may not be advisable in many cases, and may even compromise the identity of third gender individuals. To address this issue, I would like to see a thoroughly-defined data protocol and terms of use that allow only registered individuals to access potentially sensitive census data. This ensures that crucial data is not being withheld, and allows many NGOs and public sector organizations to access data and mobilize resources.
41% of all employed third gender persons in India report their occupation as “Other” in the 2011 census
The 2011 Third Gender table is also demographically truncated. While traditional Indian census tables display population age information in 10-year increments from 0–80+, the 2011 Third Gender table only contains age-specific counts for third gender children from 0–6 years old. While this population is of particular interest for childcare and health purposes, not sharing age information limits decision-making for third gender persons of other ages. Similarly, traditional census occupations cover a range of activities, including agriculture, manufacturing, utilities, trade, public administration, education, etc. The 2011 Third Gender table examines only a limited set of occupations: Cultivators, Agricultural Laborers, Household Industry, and Other. This further limits the analysis of labor for working-age third gender persons, which is important to ensure inclusion in spheres of productive labor. While cultivators and agricultural laborers dominate in almost all states, 41% of all employed third gender persons in India report their occupation as “Other” in the 2011 census. Therefore, a larger variety of occupations must be studied to better understand the types of livelihoods accessible to third gender individuals.
Indian Census data release is very slow, and not transparent
Data release for the Indian census has historically been very slow. The 2001 Indian Census was fully compiled and uploaded to the International Household Survey Network in February 2014, 13 years after the data was collected. The 2011 data tables are still in the process of being compiled and released, one by one. Important tables on worker occupations and migration patterns have not been made publicly available as of January 2016. The public has no way of knowing the timeline of census data table release, and there is no unified system to curate, regulate, and distribute this data. I am not sure if district-level third gender population statistics will be shared, or if so, when. But if data is not quickly and efficiently compiled and shared, we lose the window of opportunity to learn from, and make decisions based on, this data. Many individuals, especially marginalized communities, would greatly benefit from timely and transparent disclosure of census data.
The third gender is not yet treated as a “true” gender — it is currently treated as a separate category
The creation of a separate table for an “Other” gender, instead of incorporation with traditional census sex categories of Male and Female, indicates that this population is not yet fully integrated in the census. Incorporation of the third gender into the traditional census would resolve many of the issues of spatial and demographic resolution discussed above. Of course, this is easier said than done, and no country has figured out how to count its third gender population accurately yet. Mona Chalabi of FiveThirtyEight describes various studies that attempted to count the number of transgender individuals in the United States, with varying levels of success. But the consensus seems to be that all counts are gross underestimates — there is no magic formula for census counting of third gender persons anywhere yet.
Framing Solutions
Census enumerators have a monumental task of gaining access to individuals, and recording private information about them. The Associated Press reports in the video below that “The enumerators say they often have to face hostility when it comes to personal questions about income and contribution of women in the household…. Many people don’t let them in — they just make them stand outside and just ask them to tick the answers. It is [also] very difficult to convince them that the information collected is confidential and will remain so.”
Given these challenges, the 2011 Third Gender table is proof that we are capable of collecting this crucial information. As India formulates the methodology for its 2021 census, I hope the Registrar General and Census Commissioner continue to expand on the foundations of the 2011 Census for third gender individuals. We know that state-level information is a great start, but further spatial and demographic resolution is necessary for planning, resource allocation, and decision-making purposes.
This exercise led me to several recommendations for improving the quality of data collected from the census. Here are my recommendations towards a truly inclusive Census 2021, with varying degrees of feasibility:
- Better educate informants about their choices for each census question. The accompanying introduction to the 2011 Third Gender table indicates: “In Census 2011, three Codes were provided for recording gender i.e. Male-1, Female-2 and Other-3. In case the respondent wished to record neither ‘1’ nor ‘2’ the enumerator was instructed to record sex as ‘Other’ and give code ‘3’. Thus, category of ‘Other’ would not only include transgender but also any person who desired to record sex under the category of ‘Other’. It is also possible that transgenders might have returned themselves either male or female depending on their choice”. By default, this enumeration protocol selects to report only third gender persons who are vocal and open about their gender identity, and excludes third gender persons who may not know that reporting an “Other” gender identity is an option. If confidentiality was better enforced, and informants were better educated about all their response options and rights, enumerators may be able to record a more complete picture of third gender persons and their lives.
- Openly publish census data release timelines. The Registrar General and Census Commissioner should inform the public about when key datasets will be published, and encourage data sharing, analysis, and visualizations using this data.
- Create and maintain a better system for curating and disseminating census data. The Registrar General and Census Commissioner should require users to register to protect sensitive census data, and create user interfaces that allow general public to navigate variables, spatial resolutions, and population subsets. These information delivery systems should make sure that the metadata for all census variables is easily accessible, well-defined, and included with all downloaded census products.
- Collaborate with other countries trying to achieve similar goals. In 2011, Nepal also included the third gender in its national census for the first time. However, “due to confusion and discrimination among census enumerators and a data entry system that only allowed for two genders, those who identified as third gender were not accurately measured”. This process and its findings are discussed in detail in Surveying Nepal’s Third Gender: Development, Implementation, and Analysis by Kyle G. Knight, Andrew R. Flores, and Sheila J. Nezhad. Collaborating with the Nepali Central Bureau of Statistics on gender inclusion and census counting methods would be a great opportunity for both India and Nepal to learn from each other and innovate. Similar collaboration with other countries’ census bureaus may also be beneficial for all organizations and populations.
Many countries, including the United States, are currently struggling with the inclusion of transgender individuals in their national censuses. The New York Times recently reported the challenges faced by the United States Census Bureau in recording transgender individuals. Given the international interest in gender inclusion, the 2021 Census offers India an unprecedented opportunity to develop an innovative census protocol that can be followed by many for years to come. I hope the Registrar General and Census Commissioner of India learn from the pilot 2011 Third Gender table, and continue their efforts towards a truly inclusive Indian census through education, transparency, data sharing, and collaboration.
Technical note: All maps in this article were created using Quantum GIS, and all mapped categories are based on quantile distributions of data from the 2011 Indian Census.
