Invocation of Muslim population growth has been a recurring feature of Indian electoral politics. In September 2002, ahead of that year’s Gujarat assembly elections, Narendra Modi, then the state’s caretaker chief minister, used the phrase “hum paanch, hamare pachees” (we are five, ours are twenty-five) in reference to Muslim family size, a remark widely reported at the time and later. Since then, similar themes have surfaced periodically.In 2015, a BJP MP implored Hindu women to bear more children while in 2024, a Union minister’s campaign attributed the rising share of the Muslim population to “appeasement.” Most recently, in the run-up to the April 2026 West Bengal assembly elections, the BJP attributed the growth of the state’s Muslim population to sustained cross-border infiltration from Bangladesh (the so-called ghuspaithiyas). It won the elections by a landslide.What is the factual basis underlying this narrative? It undoubtedly works as politics, but the question is whether it works as demography. Can Muslim population growth, whether from natural increase or from illegal immigration, invert the religious composition of a country in which Hindus form the overwhelming majority?The decennial census, conducted in its post-independence series since 1951, with the most recent religion data from 2011, allows the question to be answered based on data and mathematics rather than assertion.What the census countsA threshold point matters for everything that follows: The Indian census does not enumerate citizens; it enumerates people. Conducted under the Census Act, 1948, it counts every “usual resident” of India, defined as a person who has lived in a locality for six months or more or intends to, irrespective of nationality or legal status. Thus, a Bangladeshi national who crossed the border without documents and settled in Murshidabad is, for census purposes, as countable as anyone else.This cuts against the instinct to dismiss census data as blind to illegal immigration. On the contrary, precisely because the census counts everyone present, any mass settlement of foreign migrants would leave a footprint in the total population, its religious composition, and above all, in the relative growth rates of the regions where migrants settle. The census cannot tell us who is a citizen, but it can tell us whether the people are there.Also read: Why Nirmala Sitharaman Is Wrong to Say ‘Growing’ Muslim Population Means All Is WellNote that India’s census methodology is not a domestic idiosyncrasy. The United Nations Statistics Division has, since 1958, issued and periodically revised international Principles and Recommendations for Population and Housing Censuses. These guidelines were designed to ensure that national censuses are internationally comparable so that country-level counts can be aggregated consistently into regional and world totals.India’s post-1951 census series has been conducted broadly in line with this framework, including its “usual residence” concept of enumeration, and Indian census officials have participated directly in the UN expert groups that periodically revise the standard. Hence, the population data in the three tables below are produced by the government of India using the same internationally vetted methodology used to compile population statistics worldwide.Table 1: Population of India, religion-wise demographics (1951-2011)Source: Decennial Population Censuses, 1951, 1961, 1971, 1981, 1991, 2001, 2011; Government of India.Table 2: West Bengal Population Trend (1951-2011)The RRI returns the same rate of growth as provided by the LOGEST function based on the log-linear trend line fitted to all observations. Source: Census of India, 1951 to 2011.Table 3: Total Fertility Rate in India (1992-93 to 2019-21)Source: Total Fertility Rate (TFR) based on National Family Health Surveys (NFHS-1 to NFHS-5); Government of India.Basic population dynamicsAt independence in 1947, India’s population stood at roughly 340 million. Improvements in public health, nutrition and medicine drove rapid growth thereafter: from 361 million at the 1951 census to 1.21 billion at the 2011 census, an annual trend rate of growth of about 2% by either of the two methods used throughout this article – the endpoint compound annual growth rate (CAGR) and a log-linear trend (LOGEST) fitted to all seven censuses (see tables 1, 2 and 3).Between 1950 and 2023, sustained investment in famine prevention, childhood vaccination and rural health services cut infant mortality from about 176 to roughly 25 deaths per 1,000 live births, and life expectancy rose to about 72 years for men and 74 for women. By mid-2026, the UN Population Division put India’s population at approximately 1.48 billion – the world’s largest, having overtaken China.The 2011 census remains the most recent source of religion-wise population data. The long-delayed next census, originally scheduled for 2021, is now underway. The house-listing phase began on April 1, 2026, and the population enumeration, including collection of data on religion, is scheduled for early 2027. While the following analysis excludes speculation about what that census will show, the results will provide a direct test of the propositions advanced here.That belt in the dark is not a snakeBetween 1951 and 2011, the Hindu population grew from 303.7 million to 966.3 million, a compound annual rate of about 1.9%, while the Muslim population grew from 35.4 million to 172.3 million, about 2.7% per annum. As a result, the Hindu share of the population declined from 84.1% in 1951 to 79.8% in 2011, while the Muslim share rose from 9.8% to 14.2%.In West Bengal, the shift was somewhat larger: Muslims rose from 19.5% of the state’s population in 1951 to 27.0% in 2011, while the Hindu share fell from 78.9% to 70.5%. However, like the proverbial belt in the dark mistaken for a snake, the fears these figures evoke are misplaced for several reasons.First, the arithmetic of the base effect forecloses demographic inversion. Because Hindus have always vastly outnumbered Muslims, even six decades of higher rates of Muslim population growth have widened rather than narrowed the absolute gap between the two communities.Consider that while Hindus outnumbered Muslims by roughly 268 million in 1951, the gap expanded to nearly three times as much (794 million) by 2011. A minority growing faster from a base one-eighth the size of the majority does not overtake it, on present trends, within any horizon relevant to policy or politics.Also read: How Does India See Its Muslim Population?Second, and central to the infiltration narrative: if illegal immigration from Bangladesh had materially inflated West Bengal’s Muslim population, the census would show West Bengal’s Muslim population growing substantially faster than the Muslim population in the rest of India. It does not. Over 1951–2011, the annual trend rate of growth of Muslims inside West Bengal was essentially identical to, in fact marginally below, that of Muslims in the rest of the country. This result holds whether we use the endpoint compound annual growth rate or a log-linear trend fitted to all seven census counts.Both measures are reassuringly close throughout – rarely more than a few hundredths of a percentage point apart (see table 2).More importantly, they agree on the finding that matters: Muslim population growth outside West Bengal ran fractionally ahead of, not behind, growth inside the state. In short, we find no evidence in the growth-rate data of the sustained mass illegal immigration into West Bengal that would be necessary to tilt the demographic scales of that state, let alone India.Consistent with this, West Bengal’s share of India’s total Muslim population was almost perfectly flat across six decades: 14.4% in 1951 and 14.3% in 2011. The border state that is supposedly the principal receptacle of a decades-long deluge holds exactly the share of India’s Muslims it held sixty years earlier. Thus, as far as the census data are concerned, the “ghuspaithiyas” are behaving like Godot.Of course, one might object that Bangladeshi migrants dispersed beyond West Bengal in spite of the fact that cultural, linguistic and dietary affinities make West Bengal the natural destination. Even so, for argument’s sake, let us suppose a significant number of migrants did move to other parts of India. In that case, the Muslim population outside West Bengal should have grown conspicuously faster than the national trend. However, it did not.The near-identity of growth rates inside and outside the state is consistent with border crossings that were sporadic, plausibly numbering in the thousands over the decades. They are inconsistent with a sustained inflow of millions necessary to alter the religious composition of a state of 91 million, let alone a country now approaching 1.5 billion. That said, nothing in these observations denies the reality of sporadic infiltration across India’s long and porous borders and hence the need to strengthen border security.Two caveats on the data deserve note. The 1981 census could not be conducted in Assam and the 1991 census excluded Jammu and Kashmir; decade-to-decade comparisons spanning those years should be read accordingly, although the 1951 and 2011 endpoints used for the CAGR are unaffected. Further, the 1951 Muslim share itself reflects the enormous Partition-era population transfers, so part of the subsequent rise represents recovery from an artificially depressed base.The narrowing fertility gapWhat, then, explains the rise in the Muslim population share, in West Bengal and nationally? The answer is unglamorous: a fertility differential, which is now closing fast. The National Family Health Surveys (NFHS) show that the Muslim total fertility rate fell from about 4.4 children per woman in 1992–93 (NFHS-1) to 2.36 in 2019–21 (NFHS-5), the steepest decline recorded for any religious community, while the Hindu rate fell from about 3.3 to 1.94 over the same period.Also read: RSS Claims About Rapid Growth of the Muslim Population are Simply FalseThe Hindu–Muslim fertility gap has thus narrowed from roughly 1.1 children per woman to about 0.4 – a convergence visible in the census record itself. In fact, the last decade (2001–2011) produced the slowest decadal growth for Muslims recorded in the six decades shown – a sharper deceleration for Muslims than for Hindus.Higher Muslim fertility in earlier decades, compounding on an established domestic population base, fully accounts for the observed shift in shares. We do not need to introduce illegal immigration to account for that shift. And because the differential that drove the shift is disappearing, the shift itself is decelerating. The religious composition of India’s population is heading toward stability, not inversion.This is worth pausing on. The census counts and the NFHS fertility surveys are two independent data collection systems: different instruments, different sampling frames, fielded at different times of year, run by different survey mechanisms, decades apart in their respective schedules.That the deceleration in decadal population growth visible in the census data lines up with the deceleration in fertility rates recorded separately by the NFHS is not a coincidence one would expect from two unrelated or unreliable data sources. It is the kind of cross-validation that gives demographers confidence in a country’s underlying statistical system. Two independently sourced datasets that both point to a narrowing Hindu-Muslim demographic differential make it considerably harder to dismiss either as an anomaly or a data artifact (see chart below).Optical illusionWhy, then, does Hindu anxiety about demographic inversion persist? This is because street-level perception is a poor instrument for measuring macro-demographics. The illusion is not unique to India.An observer judging American demographics from a walk through Edison, New Jersey or parts of Queens, New York, might conclude that Indian immigrants were rapidly displacing the native population. In fact, Indian Americans comprise about 1.4% of the US population. What the observer sees is not national replacement but localised spatial concentration driven by economic clustering. The visible minority presence in Kolkata’s informal economy is the same phenomenon: urban density mistaken for demographic transformation.The illusion is also compounded by a classification error. Just as the casual American observer collapses Pakistani, Bangladeshi, Nepali and Sri Lankan immigrants into a single (“Indian”) mental category, the political discourse in West Bengal treats any visible Muslim or Bangladeshi-accented, Bengali-speaking presence as evidence of cross-border infiltration. This erases internal demographic scenarios entirely: Muslim laborers moving from Bihar, Jharkhand or Uttar Pradesh to Kolkata who, as Indian citizens, are exercising a constitutional right.To the politically primed onlooker, that Muslim is indistinguishable from a foreign infiltrator, yet the fact is that in the Indian subcontinent, national origin cannot be verified on sight. Political drumbeating sharpens the illusion it feeds on. History reinforces the data here: the creation of Bangladesh in 1971, won through a war of liberation fought for linguistic and cultural sovereignty, gave its citizens every incentive to build their homeland rather than abandon it permanently.Fear not the teeming masses but the looming declineThe irony of the infiltration panic is that India’s actual demographic challenge runs in the opposite direction. India launched the world’s first national family-planning program in 1952, and from the late 1970s onwards, rights-based family planning, expanding maternal and child health services and rising female education and incomes accelerated the fertility transition.NFHS data show the total fertility rate falling from about 3.4 in 1992–93 to 2.0 in 2019–21, below the replacement level of 2.1. The latest Sample Registration System report (2024) puts the national TFR at 1.9, with West Bengal – the state at the center of the population panic – at an ultra-low 1.3, among the lowest in the country.Population growth will therefore markedly slow and eventually peak and India’s age structure is already turning. As fertility declined, the working-age share (15–64 year-olds) rose and the dependency ratio fell. The UN projects that India is in the middle of its demographic-dividend window, with the dependency ratio bottoming out near 44% during the 2030s before rising again as the population ages. The window is expected to close around mid-century.Also: Lynchings Are Increasing Because of Muslim Population Growth: BJP MPThe UNFPA’s India Ageing Report 2023 projects that the share of persons aged 60 and over will double to more than 20% by 2050, that the elderly will outnumber children around 2046, and that the 80-plus population will grow by 279% between 2022 and 2050, with a predominance of widowed, highly dependent women. Over 40% of elderly Indians are in the poorest wealth quintile and nearly a fifth report no income.Reaping the dividend before it closes requires that the working-age population be productively employed: good-quality education and skills; job creation through labour-intensive manufacturing, infrastructure, and digital connectivity; continued reductions in maternal and child mortality and universal access to healthcare; and a decisive lift in female labour-force participation.Simultaneously, India must prepare for rapid ageing by broadening pension and social-security coverage, improving geriatric healthcare, extending more support to family caregivers, and providing age-accessible housing, transport, and public spaces, with particular attention to rural areas where state support is thinnest.ConclusionThe census data admit of one reading. India’s religious composition has shifted modestly over six decades, driven by a fertility differential that is itself vanishing. The majority is still nearly six times the size of the minority – down from over eight times in 1951, even after six decades of a higher Muslim population growth rate.That arithmetic forecloses inversion under any realistic scenario. And the signature that mass illegal immigration would necessarily have left in the data – divergent Muslim growth in West Bengal, a rising West Bengal share of India’s Muslims – is simply absent.A census that counts every resident regardless of citizenship found West Bengal holding the same one-seventh of India’s Muslims in 2011 that it held in 1951. Over time, thousands likely crossed the border illegally, not millions.The infiltration narrative is, in short, a political instrument, not a demographic finding. Its electoral potency – amply demonstrated in the run-up to the West Bengal elections in April 2026 – rests on an optical illusion that mistakes urban clustering and internal migration for foreign influx, amplified by rhetoric designed to do exactly that. The forthcoming 2027 census will provide a fresh test. But nothing in six decades of data suggests it will vindicate the panic.Meanwhile, the demographic development that will actually shape India’s future – fertility below replacement, a dividend window closing by mid-century and a rapidly ageing population concentrated among the poor, presents a policy challenge far more difficult to meet than winning elections based on a misleading narrative. Chasing phantom infiltrators can win an election. It cannot reopen a demographic dividend window once it has closed.Dev Kar is the chief economist emeritus at Global Financial Integrity and a Fellow at Yale University. Prior to this, he was a senior economist at the International Monetary Fund (IMF) where, in a career spanning nearly 32 years, he worked in various departments and visited many countries on missions.