Inherited, Chosen, and Abandoned Faith:
What Global Religious Demographics Reveal About Christianity, Islam, and Religious Disaffiliation
Introduction
Most human beings do not begin life by comparing the world's religious traditions and selecting the one they find most persuasive. Religion is generally encountered first through family, geography, language, culture, and community. A Christian child is ordinarily introduced to Christianity by Christians, just as a Muslim child is ordinarily introduced to Islam by Muslims. Only later, if ever, does the individual independently evaluate the religious identity received in childhood.
That observation creates an important problem for the study of religion. When approximately 2.3 billion people identify as Christian and approximately 2.0 billion identify as Muslim, what exactly do those numbers measure? Do they primarily measure theological conviction? Conversion? Childhood socialization? Fertility? Geography? Culture? Community? Political environment? Or some combination of all of them?
Recent demographic research provides unusually strong data with which to examine these questions. Pew Research Center's global religious-change research found substantial differences among religious traditions not only in population size and growth but also in rates at which adults retain, leave, or enter the religions in which they were raised. Islam and Hinduism exhibit exceptionally high retention, while Christianity and Buddhism experience considerably larger losses through religious switching. At the same time, religiously unaffiliated populations have made substantial gains through switching (Hackett et al. 2025a).
These differences invite obvious questions, but they also demand methodological restraint. A retention rate cannot tell us by itself why someone remains religious. A person may remain Muslim because she profoundly believes Islam is true. Another may remain because Islam is inseparable from family and identity. Another may fear social consequences. Another may live under laws restricting conversion. The same complexity applies to Christians, Hindus, Buddhists, agnostics, atheists, and adherents of other traditions.
Demographics can tell us what is happening. Determining why requires another layer of evidence.
Before the Numbers: How Old Are the World's Religions?
Comparing religions requires first acknowledging that the word religion encompasses traditions that emerged in radically different historical circumstances.
Some have identifiable founders. Others do not. Some possess scriptures closely associated with a particular historical period and language. Others developed from much older oral, ritual, philosophical, and textual traditions. Assigning each religion a neat “founding date” therefore creates a false impression of historical certainty.
Indigenous and animistic religious traditions are the oldest category in the broadest sense, because religious practices existed long before written history and developed independently throughout Africa, Asia, Australia, the Americas, Europe, and Oceania. There is no single founder, founding date, geographical origin, or original language that can responsibly be assigned to them.
The traditions now collectively called Hinduism likewise have no single historical founder. Their antecedents developed across the Indian subcontinent over a long period, with the Vedic tradition becoming particularly important during the second millennium BCE. Vedic Sanskrit is the language of the Rigveda and other early Vedic literature, but Hinduism itself developed through numerous textual, philosophical, devotional, and regional traditions over subsequent centuries.
Judaism also resists a simple founding date. Religious tradition traces its ancestry to figures including Abraham and Moses, while historical scholarship traces the development of Judaism through ancient Israelite religion in the Levant, the kingdoms of Israel and Judah, the Babylonian exile, Second Temple Judaism, and later rabbinic Judaism. Biblical Hebrew is the principal language of the Hebrew Bible, with portions written in Aramaic. A precise historical date at which ancient Israelite religion became “Judaism” cannot be established.
Zoroastrianism emerged in the ancient Iranian world and is associated with Zarathustra, or Zoroaster. Dating Zarathustra remains contested, with scholarly proposals spanning many centuries. The oldest portions of the Avesta are written in Old Avestan, making Avestan foundational to the tradition, even though assigning Zoroastrianism one universally accepted founding year would be misleading.
Jainism emerged in ancient India from religious currents that predate its best-known historical teacher, Mahavira. Mahavira is traditionally regarded as the twenty-fourth tirthankara, rather than the creator of an entirely new religion. Early Jain canonical traditions are associated particularly with Prakrit languages, including Ardhamagadhi.
Buddhism emerged in South Asia around the fifth century BCE through the teachings associated with Siddhartha Gautama, the Buddha. Even here, “original language” requires qualification. The Buddha probably taught using vernacular Middle Indo-Aryan speech rather than the later literary language of Sanskrit. Buddhist teachings subsequently survived in several textual traditions, including Pali, Buddhist Hybrid Sanskrit, Sanskrit, and later Chinese, Tibetan, and other translations.
The Confucian tradition developed in ancient China around the teachings associated with Confucius during the sixth and fifth centuries BCE, although it incorporated older Chinese intellectual and ritual traditions. Its foundational texts were transmitted in Classical Chinese.
Daoism, or Taoism, likewise emerged from ancient Chinese philosophical and religious currents. Texts associated with Laozi and Zhuangzi became foundational, but Daoism developed gradually rather than appearing as a fully formed religion on one historical date. Its foundational literary language is Classical Chinese.
Shinto developed from indigenous Japanese ritual and religious traditions rather than from a single founder. Its practices predate Japan's earliest surviving written records. Those early records were produced using Chinese characters and literary conventions to represent Japanese material, making any claim about a single “original language” difficult. Ancient Japanese religious culture preceded the written sources through which historians reconstruct it.
Christianity emerged during the first century CE from the Jewish environment of Roman Judea and the eastern Mediterranean. Jesus and his earliest followers were Jewish. Jesus probably spoke primarily Aramaic, while also existing in a multilingual environment in which Hebrew and Greek were present. The earliest surviving New Testament writings, however, are written in Koine Greek. Christianity therefore presents an important distinction between the probable principal spoken language of Jesus and the language in which the earliest surviving Christian canonical documents were composed.
Islam emerged in seventh-century western Arabia, beginning with Muhammad's (PBUH) prophetic career, traditionally dated from the first Qur'anic revelations around 610 CE. Mecca and Medina became its foundational geographical centers. The Qur'an is an Arabic text, and Arabic remains uniquely important to Islamic scripture and ritual even though most Muslims worldwide are not native Arabic speakers.
Sikhism emerged in the Punjab during the late fifteenth and early sixteenth centuries through Guru Nanak and the succeeding Sikh Gurus. Punjabi, particularly as represented through the Gurmukhi script, is central to Sikh textual history, although the Guru Granth Sahib contains a richer mixture of linguistic forms than describing it simply as a Punjabi-language book would suggest.
The Bahá'í Faith emerged in nineteenth-century Persia through the Báb and Bahá'u'lláh. Its foundational writings were composed principally in Persian and Arabic.
Finally, religious unaffiliation and agnosticism should not be placed on this timeline as though they constitute another religion. Philosophical skepticism about religious claims is ancient, while Thomas Henry Huxley coined the term agnostic in the nineteenth century. Today's demographic category of “religiously unaffiliated” includes atheists, agnostics, and people who simply report having no religion. It has no founder, scripture, sacred language, or historical founding date.
This chronology exposes an important methodological problem. When modern demographic studies compare “Christian,” “Muslim,” “Hindu,” “Buddhist,” and “unaffiliated,” they are comparing contemporary identity categories, not necessarily equivalent historical phenomena.
Language, Translation, and Religious Inheritance
The linguistic history of religion adds another dimension to the question of inherited belief.
Billions of people practice religions whose foundational texts they encounter primarily through translation. A modern English-speaking Christian ordinarily reads an English translation of biblical manuscripts written principally in Hebrew, Aramaic, and Greek. The canonical Gospels preserve the teachings of an Aramaic-speaking Jewish teacher principally through Greek texts, which are subsequently translated again into modern languages.
Islam presents a different textual relationship. The Qur'an remains inseparably associated with Arabic in Islamic religious tradition, and Muslims throughout the world recite Qur'anic passages in Arabic even when they do not speak Arabic conversationally. Translation nevertheless becomes important for the majority of Muslims who require another language to understand the text's meaning.
Buddhism provides still another model. Buddhist teachings moved across languages and civilizations for more than two millennia, producing major Pali, Sanskrit, Chinese, Tibetan, Japanese, Korean, and other textual traditions.
These linguistic differences do not establish that one religion is more or less true than another. They do establish something relevant to the study of religious belief: the religion a person inherits is frequently mediated through centuries of preservation, interpretation, translation, institutional development, and cultural transmission before it reaches the individual believer.
That reality makes historical and linguistic literacy relevant to religious inquiry.
The Global Religious Landscape
Pew Research Center's most recent comprehensive estimates cover 201 countries and territories and draw upon more than 2,700 censuses, surveys, population registers, and other sources. They estimate that in 2020 Christianity remained the world's largest religious category, with approximately 2.3 billion adherents, or 28.8 percent of humanity. Islam followed with approximately 2 billion, or 25.6 percent. The religiously unaffiliated numbered approximately 1.9 billion, or 24.2 percent. Hindus numbered approximately 1.2 billion, or 14.9 percent, while Buddhists numbered approximately 324 million, or 4.1 percent (Hackett et al. 2025a).
The unaffiliated category requires particular caution. It includes respondents reporting no religion as well as people identifying as atheist or agnostic. Pew explicitly measures religious identity, not philosophical belief about whether God exists. A person categorized as unaffiliated can therefore retain substantial spiritual or even theistic beliefs (Hackett et al. 2025a).
Consequently, 1.9 billion religiously unaffiliated people does not mean 1.9 billion atheists or agnostics.
Growth Does Not Mean Conversion
Between 2010 and 2020, the global Muslim population increased by approximately 347 million people, reaching 2 billion. Christianity added approximately 122 million people, reaching 2.3 billion. Buddhism was the only major category in Pew's analysis to decline numerically, losing approximately 19 million people and reaching approximately 324 million (Hackett 2025).
Islam's growth was driven primarily by demographics rather than large-scale conversion. For 2015–2020, Pew estimated an average lifetime fertility of approximately 2.9 children per Muslim woman compared with 2.2 per non-Muslim woman. Muslims also had a global median age approximately nine years younger than non-Muslims, 24 compared with 33 (Hackett 2025).
Christianity illustrates the opposite complexity. Its population increased numerically by approximately 6 percent from 2010 to 2020, yet its share of humanity declined from roughly 31 percent to 29 percent because non-Christian populations grew faster and substantial numbers of Christians disaffiliated (Hackett et al. 2025a).
The distinction is essential:
Population growth is not synonymous with conversion or persuasive success.
Births, deaths, age structure, migration, and religious switching are separate demographic variables.
Who Stays, Who Leaves, and Who Joins?
Pew defines religious switching as a difference between the religion in which an adult reports being raised and that person's current religious identity. It deliberately uses switching rather than conversion because many people leave religion without joining another one (Hackett et al. 2025a).
Among adults ages eighteen to fifty-four, for every 100 people raised Christian, approximately 17.1 leave Christianity while 5.5 enter, producing a net switching loss of 11.6 per 100 raised Christian. Among Buddhists, approximately 22.1 leave and 12.3 enter, producing a net loss of 9.8 (Hackett et al. 2025a).
Islam and Hinduism exhibit strikingly different patterns. Approximately one person out of every 100 adults raised Muslim leaves Islam, while a roughly similar number enters. Hinduism exhibits similarly low switching in both directions. The religiously unaffiliated show the reverse pattern: approximately 7.5 people leave for every 100 raised unaffiliated, while 24.2 enter, producing a net gain of 16.7 (Hackett et al. 2025a).
This produces approximate retention rates of 99 percent for Islam and Hinduism, 93 percent for the religiously unaffiliated, 83 percent for Christianity, and 78 percent for Buddhism.
Retention, however, is not necessarily conviction.
A 99 percent Muslim retention rate establishes that nearly all surveyed adults raised Muslim continue to identify as Muslim. It does not establish that 99 percent accept every Islamic theological claim, observe Islamic practice, never doubt, or would necessarily report the same identity under different familial, cultural, political, or legal circumstances.
Likewise, leaving Christianity does not necessarily mean rejecting God. A former Christian may become Muslim, Buddhist, agnostic, atheist, spiritual but unaffiliated, or simply decline to identify with an organized religion.
Identity, belief, practice, and institutional affiliation overlap. They are not interchangeable.
Christianity in the United States
The United States offers considerably richer data concerning religious switching.
Pew's research documents substantial movement away from Christianity toward religious unaffiliation. Globally, Christianity has the largest net loss through switching, and the United States is one of the countries in which that pattern is particularly pronounced (Hackett et al. 2025a).
The decline should not be confused with the disappearance of Christianity. Christianity remains America's largest religious tradition, and Pew has reported evidence that the decline in Christian identification may recently have slowed or temporarily leveled. Nevertheless, the longer-term switching imbalance is substantial.
Why Are People Leaving Christianity?
There is no single explanation.
Pew's 2025 survey of 8,937 U.S. adults found that among Americans who left their childhood religion, 46 percent said they stopped believing its teachings, 38 percent said religion was no longer important in their lives, and 38 percent said they gradually drifted away. Other reasons included religious teachings involving social and political issues and scandals involving clergy or religious leaders (Alper et al. 2025).
This evidence complicates claims that people are abandoning Christianity for one simple reason. Loss of belief, gradual disengagement, institutional distrust, moral disagreement, scandals, unmet spiritual needs, and changing social environments can all contribute.
Why Do People Become Religiously Unaffiliated?
Globally, the religiously unaffiliated population increased from approximately 1.6 billion in 2010 to 1.9 billion in 2020, an increase of about 17 percent. The category grew from approximately 23 percent to 24 percent of humanity. North America's unaffiliated population nearly doubled, increasing approximately 92 percent to 114 million people (Hackett et al. 2025a).
Religious disaffiliation is particularly prominent in Europe, North America, Australia, and New Zealand. In 2020, religiously unaffiliated people constituted majorities in ten countries and territories (Hackett et al. 2025a).
But becoming unaffiliated should not automatically be described as becoming atheist. It may represent rejection of institutional religion, agnosticism, atheism, spiritual independence, uncertainty, or simply unwillingness to claim a religious label.
Islam's 99 Percent Retention Rate
Islam presents perhaps the most intriguing contrast in the data.
Globally, only approximately one in every 100 adults ages eighteen to fifty-four raised Muslim reports leaving Islam, while approximately the same proportion enters it (Hackett et al. 2025a).
One possible explanation is straightforward: many Muslims remain Muslim because they sincerely believe Islam is true and find their religious tradition meaningful.
But a 99 percent retention rate cannot, by itself, tell us why people remain.
Muslim identity may intersect with family, ethnicity, nationality, marriage, culture, and community. In some countries, leaving Islam may also carry legal or social consequences. Consequently, the possible relationship between freedom of religious exit and measured retention deserves examination.
Apostasy, Hadith, and Islamic Law
Classical Islamic jurisprudence developed legal doctrines concerning ridda, or apostasy, substantially drawing upon hadith literature. The dominant premodern juridical tradition prescribed capital punishment for certain forms of apostasy, although jurists disagreed about conditions, procedures, repentance, gender, and related questions (Saeed 2012).
The Qur'an itself discusses apostasy and unbelief but does not straightforwardly prescribe an earthly death penalty merely for ceasing to believe. This distinction has become important in contemporary Muslim scholarship.
Some contemporary scholars defend significant elements of the classical jurisprudential tradition. Others argue that the relevant early Islamic precedents concerned apostasy associated with political rebellion, treason, warfare, or abandonment of the political community, rather than private change of belief. Saeed's analysis demonstrates why the historical development of apostasy law must be understood within both theological and political contexts (Saeed 2012).
It would therefore be historically inaccurate to claim either that Islam simply commands Muslims to kill everyone who leaves Islam or that Islam possesses no historical jurisprudential tradition prescribing capital punishment for apostasy.
The historical reality is more complicated.
Does Apostasy Doctrine Help Explain the 1 Percent Departure Rate?
This is a legitimate research question.
It is not yet a demonstrated causal conclusion.
The global Pew dataset measures religious identity. It does not isolate whether a respondent remained Muslim because of religious conviction, family attachment, cultural identity, community pressure, fear of ostracism, legal restrictions, or some combination of these influences.
Where abandoning a religion entails significant social or legal costs, however, it is reasonable to investigate whether those costs influence both actual departure and willingness to report departure.
The strongest study would compare Muslim populations living under substantially different legal and social environments while controlling for age, education, migration, minority status, family structure, religiosity, and other variables.
Until such evidence is available, the academically defensible conclusion is straightforward:
Islam's approximately 99 percent measured retention is well supported by the available demographic data. The proposition that apostasy law explains that retention rate is not.
The latter remains a hypothesis requiring evidence.
Churches and Mosques in the United States
The institutional landscape provides another comparison, although seemingly simple counts quickly encounter methodological problems.
The 2020 U.S. Religion Census reported 356,739 congregations associated with 372 religious bodies and approximately 161.4 million adherents. That number encompasses multiple religious traditions and therefore should not be described simply as 356,739 Christian churches.
A national study of American mosques identified approximately 2,769 mosques in 2020, compared with 2,106 in 2010, an increase of approximately 31 percent (Bagby 2021).
Obtaining an equally clean national count of active Christian churches is more difficult because American Christianity includes Catholic parishes, Orthodox churches, hundreds of Protestant denominations, nondenominational churches, independent congregations, and house churches. Any comparison therefore requires consistent definitions of church, mosque, congregation, and active.
What About Churches in Muslim-Majority Countries?
There is no defensible single number for active Christian churches across Muslim-majority countries.
Different governments count denominations, licensed buildings, congregations, religious organizations, church compounds, and recognized houses of worship differently.
Unregistered or underground congregations may not appear in government statistics at all. Multiple congregations may also use a single building.
Even the phrase Islamic country requires clarification. A Muslim-majority country is not necessarily an Islamic state. A state with Islam as its official religion is not necessarily governed primarily by Islamic jurisprudence. Membership in the Organization of Islamic Cooperation constitutes yet another category.
A more useful comparison would therefore examine Muslim-majority countries individually, recording the estimated Christian population, recognized churches or congregations, legal ability to construct churches, conversion laws, restrictions on proselytism, apostasy provisions, and documented treatment of converts.
That analysis could eventually become a separate Round Table investigation.
What Can These Numbers Actually Tell Us?
Several conclusions can be drawn with reasonable confidence.
Religious inheritance remains enormously important. Most people encounter religion through family and community before they possess the intellectual capacity to conduct an independent investigation of competing theological claims.
Religious traditions also differ dramatically in measured retention. Islam and Hinduism show approximately 99 percent retention among the population included in Pew's switching analysis, compared with approximately 83 percent for Christianity and 78 percent for Buddhism. Religious unaffiliation, meanwhile, has gained substantially through switching (Hackett et al. 2025a).
Christian disaffiliation is particularly consequential because Christianity remains the world's largest religious tradition. Islam's rapid numerical growth, meanwhile, is driven primarily by age structure and fertility rather than mass conversion (Hackett 2025).
None of these observations establishes theological truth.
A statistic can tell us that someone identifies as Christian, Muslim, Hindu, Buddhist, Jewish, Sikh, atheist, agnostic, or unaffiliated. It cannot tell us whether that person examined competing claims and became convinced; inherited an identity and never seriously questioned it; remains because the religion provides meaning; fears losing family or community; privately doubts while publicly conforming; or simply finds the inherited identity comfortable enough to retain.
That distinction becomes even more significant when the historical timeline is placed beside the demographic data.
Human beings have been developing, transmitting, interpreting, translating, reforming, defending, questioning, and sometimes abandoning religious traditions for thousands of years. Some traditions emerged gradually. Others developed around identifiable historical figures. Their scriptures were composed in different languages, historical periods, political environments, and cultures. Most modern believers encounter those traditions long after their foundational events and usually through communities into which they were born.
If most human beings inherit their initial religious identity, if retention varies dramatically across cultural and legal environments, if religious texts reach billions of people through translation and interpretation, and if demographic growth is substantially influenced by birth rather than conversion, then the numerical success of a religion cannot establish the truth of its theological claims.
Nor does widespread departure establish their falsity.
Population statistics measure human affiliation. They do not measure God.
The more consequential questions are therefore not simply how many people believe, but why they believe, what they believe, how freely they believe it, what evidence they have examined, what it would cost them to change their minds, and whether they evaluate another tradition's claims by the same standards they apply to their own.
For anyone seriously pursuing religious truth, those questions may matter considerably more than the size of any religion.
References
Alper, Becka A., Patricia Tevington, Asta Kallo, and Jeff Diamant. “Why Do Some Americans Leave Their Religion While Others Stay?” Pew Research Center, December 15, 2025. Pew Research Center
Bagby, Ihsan. The American Mosque 2020: Growing and Evolving. Washington, DC: Institute for Social Policy and Understanding, 2021.
Hackett, Conrad. “Islam Was the World's Fastest-Growing Religion from 2010 to 2020.” Pew Research Center, June 10, 2025. Pew Research Center
Hackett, Conrad, Marcin Stonawski, Yunping Tong, Stephanie Kramer, Anne Fengyan Shi, and Dalia Fahmy. “How the Global Religious Landscape Changed from 2010 to 2020.” Pew Research Center, June 9, 2025. https://doi.org/10.58094/fj71-ny11. Pew Research Center
Hackett, Conrad, Marcin Stonawski, Yunping Tong, Stephanie Kramer, Anne Fengyan Shi, and Nick Zanetti. “Religious Composition by Country, 2010–2020.” Pew Research Center, June 9, 2025. https://doi.org/10.58094/5shf-2d69. Pew Research Center
Saeed, Abdullah. “Pre-Modern Islamic Legal Restrictions on Freedom of Religion, with Particular Reference to Apostasy and Its Punishment.” In Islamic Law and International Human Rights Law, edited by Anver M. Emon, Mark S. Ellis, and Benjamin Glahn, 226–246. Oxford: Oxford University Press, 2012.
United States Commission on International Religious Freedom. Anti-Conversion Laws Compendium. Washington, DC: USCIRF, 2023.
U.S. Religion Census. 2020 U.S. Religion Census. Association of Statisticians of American Religious Bodies, 2023.
Comments