September 21, 2021

Religious Composition of India

2. Causes of religious change

The religious composition of the global population is always in flux. For example, the share of Buddhists worldwide is shrinking because of their low average fertility rate, and Christians are declining in the U.S. and Western Europe as more people leave organized religion. Demographers attribute changes in the size of religious populations to three main causes:

  • Fertility rates
  • Religious switching (or conversion)
  • Migration

Causes of change: Fertility

Fertility rates have been the main driver of population change in India. In the decades since Partition, the number of children born to an average woman over her lifetime has been decreasing, and most gaps between religious groups have narrowed. However, some fertility differences between religious groups persist. Overall, the average woman in India is expected to have 2.2 children in her lifetime. Hindu women are expected to have 2.1 children, on average, while Muslims have 2.6 and Christians have 2.0.

The role of age structures and life expectancy in religious change

Related to fertility rates, the age distributions of India’s largest religious groups differ significantly. As of 2020, Pew Research Center estimates that Hindus have a median age of 29, compared with 24 for Muslims and 31 for Christians. India’s other religious groups have a combined median age of 34. Groups with younger populations have more women entering their early reproductive years and are expected to grow faster than older populations. This can create what demographers call “population momentum”: High fertility rates produce a relatively young age distribution, which in turn accelerates population growth. By the same token, declining fertility rates act as brakes on population growth.

In theory, mortality rates also could be a significant factor in the changing size of religious groups if some groups face much lower life expectancies than others. But religious differences in life expectancy in India are estimated to be modest: Christians have a life expectancy at birth of 68 years, followed by Muslims (66) and Hindus (65).

A decline in infant mortality, meanwhile, means that parents can now attain their desired family sizes with fewer births, which in turn has helped slow India’s fertility rates and further reduce the growth differences between religious groups.

Fertility and its connections to education, economics and location

Within religious groups, adherents tend to share certain demographic characteristics, such as levels of education and household wealth – and some of these factors impact fertility. Often, these shared characteristics are tied to a group’s regional concentration. Sharing a location often means sharing a distinct set of local cultural norms, and decisions about family size are influenced by what seems common and desirable among neighbors, nearby friends and family. Economic conditions that may impact access to education and health care (including family planning resources) also vary from place to place.

Globally, a woman’s level of education is the best predictor of the number of children she will have in her lifetime. Higher education usually coincides with childbearing years, and as a result, highly educated women often marry later and have their first child later than women with less education. Education also frequently gives women more economic opportunities – including jobs that might lead to further delays in childbearing – and better access to family planning resources. In India, there are sizable religious differences in the average years of education women receive, with Christians receiving the most and Muslims the fewest.

A note on large numbers

India uses a number system that differs from the international number system. This report presents numbers in the international system and, in parentheses, the Indian system. The Indian number system uses units such as lakhs and crores and places commas at different intervals than the international system. Some examples of equivalents:

International number system vs. Indian number system

One hundred thousand (100,000) = 1 lakh (1,00,000)
One million (1,000,000) = 10 lakh (10,00,000)
Ten million (10,000,000) = 1 crore (1,00,00,000)

Wealth is another measure that is tied to fertility around the world. Poorer women tend to have more children, not only because wealth is connected with access to education and health care, but also because children can contribute labor and earnings to a household with limited means. For similar reasons, women living in rural areas typically have more children than those in urban areas. Patterns by religion that influence fertility do not always pull in the same direction. For example, Indian Muslims on average live in poorer households – a characteristic that is tied to higher birth rates. But Muslims are also more concentrated in urban areas than are Indians overall, a characteristic that is tied to lower birth rates.

A preference for sons and aversion to daughters also may play a role in overall fertility. There is evidence that some parents in India resort to sex-selective abortions to reduce their number of daughters – causing an estimated deficit of 20.7 million (2.07 crore) girls compared with what would naturally be expected between 1970 and 2017 – and that this practice is more common among Indian Hindus than among Muslims and Christians.7

Women in central India tend to have more children

These interconnected patterns raise questions: Can fertility gaps between religious groups be fully explained by factors other than religion? Are religious patterns simply an accident of correlation, and not a meaningful differentiator? Perhaps Muslim women tend to have more children because they spend fewer years in school and are less wealthy, on average, than other women in India, or because many Muslims happen to live in areas where women of all faiths tend to have large families.

Indeed, many factors contribute to family sizes, making it impossible to pinpoint exactly how much religious affiliation alone impacts fertility. Unlike levels of education and wealth, some of these factors are cultural or historical and not easily measured in surveys.

However, a closer statistical examination of the available data can shed light on the ways that fertility, education, wealth and place of residence are related in India – and the extent to which religious gaps in fertility can be explained by these factors.

Controlling for group differences among Indian women in their 40s

To try to disentangle the interrelated influences on fertility – and determine if there would still be differences in the number of children that Hindus, Muslims and Christians tend to have if they lived in the same place and had equal levels of wealth and education – researchers used a statistical technique called a multilevel mixed-effects model.

This analysis focuses on a cohort of women who were in their 40s when they took the National Family Heath Survey in 2015. Women in this age group typically have completed both schooling and childbearing. This group of women also began their childbearing years at a time when larger families were the norm, and consequently they have had more children than younger Indian women today are expected to have in their lifetimes. In fact, the number of children that women in their 40s have had is in line with the fertility rates of the 1990s: The average Indian woman in this cohort has had 3.2 children, and there are relatively large differences by religion. Christian women have the smallest families, with an average of 2.6 children. Muslims have the most children, with an average of 4.2. Hindu women have had an average of 3.1 children.

Women in this group are geographically distributed among states and territories in a way that mirrors India’s overall population. Muslim women in their 40s are more likely to live in urban areas than Hindus or Christians and, after accounting for this difference, Muslims have lower levels of household wealth (on a measure that accounts for a wide range of factors including whether homes have running water or flooring). About half of women in this cohort have received no formal schooling at all, including 50% of Hindus, 57% of Muslims and 28% of Christians. The average number of years of education for women in this group is 4.2 for Hindus, 3.2 for Muslims and 7.0 for Christians.

Fertility gaps by religion remain after controlling for other differencesThe results of the analysis – controlling for years of education, household wealth, the state where women reside, whether they live in an urban or rural area and their age – show that if women in their 40s were equal in all these other ways, a Hindu woman would be predicted to have had 3.2 children (up slightly from the 3.1 they actually had), on average, compared with 4.1 for a Muslim woman (a little below the actual 4.2).8 In other words, if the women in this group all had an average amount of wealth and education, were the same age and lived in the same places, Hindu women still would be predicted to have 0.9 fewer children than their Muslim counterparts, on average.

An average Christian in her 40s, on the other hand, would be expected to have 3.5 children – nearly one full child more than Christian women in this age group actually have – if she were otherwise similar to all other Indian women. This large shift between actual children and predicted children among Christians is largely due to differences in education levels. If Christian women did not have more years of schooling, they would be predicted to be much more similar to Hindus in the number of children they have.

While these factors – years of education, household wealth, place of residence and age – all are statistically associated with the number of children women have in India, they do not fully explain the fertility gaps among religious groups.

As is generally the case elsewhere, education has a major impact on fertility, as each additional year of education correlates with a significant drop in fertility. Wealthier women in this cohort have fewer children than poorer women, and women in urban areas have fewer children than their rural counterparts.

But these results indicate that if all women in this cohort were exactly the same in their levels of education and wealth, lived in the same place, and were the same age, there would still be a difference in the number of children ever born to Hindu, Muslim and Christian women in their 40s. (See Methodology for more details on modeling and results.)

Of course, Hindu, Muslim and Christian women in India differ in many ways that are not, and often cannot be, included in this analysis of survey data. These results do not quantify how much, if any, of the gaps in fertility are due to religious affiliation alone. They show only that differences remain after accounting for several other factors known to impact fertility and to vary between religious groups.

Causes of change: Migration

Millions (tens of lakhs) of people have left India or moved there in recent years, but because their numbers are small relative to the overall population, migration does not have a substantial impact on the country’s religious composition.

Emigrants leaving India far outnumber newcomers, by a ratio of more than three-to-one. India has been one of the top countries of origin for immigrants to other countries since the United Nations started tracking such statistics three decades ago.9 As of 2019, about 17.5 million (1.75 crore) people who were born in India are now living in other countries. This represents a tiny fraction of India’s population: Only about 1% of people born in India live elsewhere. This means that Indians, who account for more than 17% of the global population, make up about 6% of all the world’s people living outside of their country of birth.

The largest number of migrants who were born in India and now live abroad are in the United Arab Emirates (3.4 million, or 34 lakh), followed by the United States (2.7 million, or 27 lakh), Saudi Arabia (2.4 million, or 24 lakh), Pakistan (1.6 million, or 16 lakh) and Oman (1.3 million, or 13 lakh). All of these countries, except for the U.S., have large Muslim majorities. Movement between India and Persian Gulf countries is often circular, with migrants periodically leaving India to be temporary workers in fast-growing Arab countries with booming economies.

Census and survey data summarized in a 2012 Pew Research Center report on the religious affiliation of migrants indicates that religious minorities are overrepresented among emigrants from India. India is a top origin country of Muslim migrants around the world, with more than 3 million (30 lakh) Indian-born Muslims residing elsewhere – many more Muslims than are estimated to have been born elsewhere and now reside in India.

The UN estimates there were roughly 5.2 million (52 lakh) foreign-born people living in India in 2019 – including refugees, asylum-seekers and other irregular migrants – or about 0.4% of India’s population that year. By comparison, there were more than 50.7 million (5.07 crore) immigrants in the United States, accounting for 15% of the U.S. population.

Most of India’s immigrants (3.1 million, or 31 lakh) were born in Bangladesh, which was part of India until Partition and then part of Pakistan until 1971. The next most common origin countries are Pakistan (1.1 million, or 11 lakh), Nepal (530,000, or 5,30,000 in India’s number system), Sri Lanka (150,000, or 1,50,000) and China (110,000, or 1,10,000). Bangladesh and Pakistan are both countries with large Muslim majorities; Nepal is primarily Hindu, and Sri Lanka has a large Buddhist majority. China’s population is mostly religiously unaffiliated, but the regions that neighbor India – Tibet and Xinjiang – are unusual in having a significant presence of Buddhists and Muslims, respectively.

There is, however, some evidence to suggest that immigrants to India do not necessarily match the religious composition of the countries they come from. For example, more than 40% of all emigrants out of Bangladesh are Hindu, even though the country is about 90% Muslim, according to Pew Research Center’s 2012 estimates, which found that about two-thirds of India’s immigrants overall are Hindu. On the whole, among people whose families migrated to India in recent generations, 87% were Hindu, 6% were Muslim, another 6% were Sikh and about 0.5% were Christian, according to the most recent India Human Development Survey.10

Unauthorized immigration is a controversial subject in India and practically impossible to measure accurately over time, especially as laws about legal or protected status have shifted through the years. While India has been willing to host refugees, they typically have not been granted legal status and are expected to return to their home countries as soon as conditions allow. According to some reports, up to tens of millions (crores) of people from Muslim-majority countries are living without legal status or documentation in India, but a lack of evidence for corresponding outmigration and other indicators have led to doubts about the plausibility of such high estimates. (This report relies on UN migration estimates that are intended to include all migrants, regardless of their legal status.)

Sidebar: Controversies over migration and citizenship in India

Tensions over citizenship and migration have intensified in India during the past decade. In 2019, Parliament passed the Citizenship Amendment Bill granting expedited citizenship to many Hindu, Christian, Jain, Sikh, Parsi and Buddhist – but not Muslim – immigrants who have fled persecution in neighboring countries.

People who come to India either as refugees or as undocumented immigrants often are from nearby countries, and in recent years speculation has circulated that up to tens of millions (crores) of Muslims have moved from Bangladesh and other neighboring countries to live illegally in India.

The sources and methodologies behind such high estimates are unclear, and reliable estimates of undocumented people are difficult to come by. But if tens of millions (crores) of Muslims from nearby countries had indeed migrated to India, demographers would expect to see evidence of such mass emigration in data from their countries of origin, and this magnitude of outmigration is not apparent. On the contrary, the United Nations Population Division estimates that as of 2019, there were fewer than 8 million (80 lakh) people who were born in Bangladesh and now reside (either legally or illegally) in all other countries combined. About 3.1 million (31 lakh) migrants born in Bangladesh are thought to reside in India – a large share of all immigrants to India, but far fewer than some figures cited widely, often without supporting evidence.

The available data indicates that Muslims are more likely than Hindus to leave India, and immigrants into India from Muslim-majority countries are disproportionately Hindu. Migration of Hindus from Bangladesh to India, driven by sectarian conflict, is largely responsible for the steady decrease in the Hindu share of Bangladesh’s population.11 (See “Causes of change: Migration” in this chapter for more on the demographics of migrants into and out of India.)

The recent mass exodus of the primarily Muslim Rohingya people from Myanmar also has been cited as a source of illegal migration to India. While it is plausible that many Rohingyas are in India, only about 1 million (10 lakh) Rohingyas lived in Myanmar prior to the mass exodus in 2017-2018. Afterward, the UN estimated that there were approximately 50,000 people living in India who originated from Myanmar, including about 18,000 Rohingya refugees registered with the United Nations High Commissioner for Refugees (UNHCR). Indian officials estimate that there are 40,000 Rohingyas dispersed throughout the country.

Causes of change: Religious switching

The share of Indians who switch religions is modest and does not appear to be a major factor in demographic change. While India’s constitution guarantees citizens the freedom to “practice, profess and propagate” their religion, nine Indian states have laws that restrict proselytizing and conversion to Islam and Christianity. Such laws were introduced during the final decades of British occupation in the 1930s and 1940s and proliferated in the 2000s.

In a 2020 Pew Research Center survey of nearly 30,000 adults across India, very few indicated they had switched religions since childhood. Among adults who say they were raised as Hindus, 99% still identify as Hindu. Fully 97% of those raised as Muslims are still Muslims in adulthood. And among Indians who were raised as Christians, 94% are still Christians. Moreover, those who do switch religions tend to cancel each other out; among all Indian adults, for example, 0.7% were raised Hindu but do not currently identify as such, while 0.8% were raised outside of the religion but are now Hindu.

Interfaith marriage also is rare, and widely frowned upon. In the same survey, 99% of married Hindus, 98% of married Muslims and 95% of married Christians say they have a spouse of the same faith. Similar shares of Hindus and Muslims, as well as 92% of Christians, say their spouse was also raised in their current religion. Furthermore, 82% of Indians say that it is at least somewhat important to prevent women in their community from marrying someone of another religion, and 81% said the same for men, including about two-thirds who say each is very important.

  1. See Chao, Fengqing, Patrick Gerland, Alex R. Cook and Leontine Alkema. 2019. “Systematic Assessment of the Sex Ratio at Birth for All Countries and Estimation of National Imbalances and Regional Reference Levels.” Proceedings of the National Academy of Sciences. Note that an earlier version of this paper had a lower missing girls estimate due to a calculation error. See also Bhalotra, Sonia, Abhishek Chakravarty, and Selim Gulesci. 2020. “The Price of Gold: Dowry and Death in India.” Journal of Development Economics.
  2. While caste is a measured characteristic that could also impact fertility in this context, it is too closely associated with these other factors to include in the statistical model.
  3. UN estimates include recently arrived refugees and other migrants who might not have been counted in the 2011 census.
  4. The India Human Development Survey is a high-quality, nationally representative panel survey of over 40,000 households.
  5. According to census data, the share of Bangladeshis who are Hindu declined from 22.1% in 1951 (when Bangladesh was East Pakistan) to 8.5% in 2011.