The Demographic Siege
BY KOENRAAD ELST
1. Visions of a demographic doomsday
1.1. Hindus, the dying race
Demography can change the cultural orientation of a country or region by making an originally small group numerically dominant. Thus, demographic success was a decisive factor in the Christian takeover of the Roman Empire: in the first three centuries, Christian numbers grew by 40% per year, due to conversion and to the rejection of birth control (whether primitive contraception, abortion or infanticide). Demographic shifts can lead to violence: in Lebanon, the tilt of the demographic balance in favour of the Muslims, leading to demands of a redistribution of political power shared with the Christians, was one of the causes of the civil war. Hindus ought to know, for they themselves have also done it once: "After years of immigration by ethnic Nepalis, [Sikkim] lost its sovereignty when a Nepali majority chose to be part of India." Finally, a certain demography can lead to selfannihilation: many members of India's Parsi community have resigned to the perspective that their community will soon disappear due to a suicidal demographic selfforgetfulness.
In today's India, demography is a hot item, not just because of the economic and ecological burden of overpopulation, but even more because of the differential between Hindus and Muslims with its real or perceived political implications. One of the classic statements of this concern is Hindu Sangathan, Saviour of the Dying Race (Delhi 1926), in which Swami Shraddhananda briefly sketches the problem of demographic decline threatening Hindu survival: "while Muhammadans multiply like anything, the numbers of the Hindus are dwindling periodically".
1.2. Ominous census figures
Swami Shraddhananda quotes from the 1911 Census Report (para 172 ff.) to show the reasons why the Muslim population is growing faster than the Hindu population, whose percentage of the total population is steadily declining. The Census Director had written: "The number of Muhammadans has risen during the decade  by 6.7 per cent as compared with only 5 p.c. in the case of Hindus. There is a small but continuous accession of converts from Hinduism and other religions, but the main reason for the relatively more rapid growth of the followers of the Prophet is that they are more prolific." Follow a number of social customs which encourage the Muslim birth rate, e.g. fewer marriage restrictions and common remarriage of widows, and the Muslim insistence that the children of mixed marriages be brought up as Muslims.
In 1909, on the basis of demographic trends visible in the census results (1881, 1891, 1901), Colonel U.N. Mukherji had projected the rate of Hindu decline into the future in a strictly linear fashion, and calculated logically (if somewhat simplistically) that it would take less than 420 years for the Hindu race to disappear completely from the face of India. This was a "correction" on 1891 Census Commissioner O'Donnell's prediction that the Hindus would die out in 620 years. The series of articles in the Bengalee of 1909 in which Mukherji proposed his analysis was titled: "Hindus, a Dying Race".
This pessimistic prognosis of the numerical HinduMuslim proportion defines the problem to which Swami Shraddhananda (who knew Mukherji personally) tried to offer a solution. One of Mukherji's concluding sentences, "They count their gains, we calculate our losses", became the title of a Hindu Mahasabha pamphlet as late as 1979. If anything, the fever of Hindu demographic pessimism is only rising.
1.3. "Hindus talking nonsense"
The Hindu suspicion that Islam is using demography to increase its strength and to wrest territories from Hinduism is a constant theme in Hindu Revivalist writing from at least 1909 till today. The rhetoric is often shrill and exaggerated and the case is wrapped in the wrong arguments, most notably the claim that "Muslims have lots of children because they have four wives". A typical example, referring to the official birthcontrol slogan, "we are two, our [children] are two", is the following: "For the Hindu the slogan is: We are two, and we have two. The slogan for a Moslem is: We are five and we have twentyfive." Sometimes, outside authorities (the BBC, the WHO) are falsely claimed as confirming the Hindus' worst fears: "The United Nations census projections have indicated that the uncontrolled birthrate of the Moslems of India coupled with huge infiltrations will turn India into a Moslem majority country before the year 2000 AD."
It is, therefore, no surprise that The Economist ridicules these demographic doomsday scenarios: "Hindu militants are talking nonsense by predicting that chunks of the country will gain Muslim majorities and then secede". There is no doubt that some of the rhetoric generated by this Hindu unrest about the future is plain nonsense, but it doesn't follow that the proportional decline of the Hindus is mere fantasy. The Economist itself acknowledges the numerical gains of the Indian Muslim community, and explains that Muslims are less willing to use birth control, and that the infant mortality rate is lower among Muslims because they are more concentrated in the cities where medical care is better.
For those who dismiss U.N. Mukherjee's reasoning as an obvious and ridiculous case of paranoia, it may be useful to verify this prediction for the subsequent 80 years. Official census data show that the Hindu percentage has declined, and the Muslim percentage increased, in every single successive census in British India, free India, Pakistan and Bangladesh. As we are about to demonstrate in some detail, the demographic trends confirm Mukherji's general prediction of a steady decline, though a quantitative prognosis is more complex than he envisaged.
1.4. Demography: the facts
Considering the pivotal role of the Hindu Revivalist perception that Hinduism is besieged and that demography is one of the weapons used by Islam against Hinduism, I believe that a meaningful discussion of Hindutva ideology is only possible after verification of this fundamental perception. Indeed, on this verification depends whether we ought to consider the Hindutva movement as a bunch of dangerous lunatics spreading lies and paranoia (a fairly common assumption among Indiawatchers) or merely a group of realistic people who try to face up to real challenges.
The following table shows the percentage of Hindus and Muslims in British India in every successive census since 1881:
yr. 1881 1891 1901 1911 1921 1931 1941
H. 75.09 74.24 72.87 71.68 70.73 70.67 69.46
M. 19.97 20.41 21.88 22.39 23.23 23.49 24.28
And these are the figures for the Indian Republic:
year 1951 1961 1971 1981 1991 est.
H. 84.98 83.51 82.72 82.29 81.8
M. 9.91 10.70 11.21 11.73 12.2
These official figures are not altogether accurate for 1981 and 1991, for Assam was not counted in 1981 and neither was the state of Jammu & Kashmir in 1991. Adjusted on the basis of an estimate for these states, the figures for 1991 become: Hindus 81.54%, Muslims 12.60%. The figure of 12.60% for Muslims in 1991 is confirmed by independent secularist observers on the basis of official data and standard procedures for extrapolation. I will use that figure in preference to the Government figure. This brings the Hindu percentage down to ca. 81.5%.
In truncated India, the Muslim population has grown 2.69% in forty years (from 9.91% to 12.6% in 195191), but Muslim leaders like Imam Bukhari routinely claim that the true figure of the Muslim population in the Indian Republic is about 3% higher. There are indeed some problems with the official figures for the Indian Republic, e.g. there is a suspicion that many illegal Bangladeshi immigrants are lying low and avoiding the census personnel because they are used to a regime which is not so lenient with unsolicited immigrants (Bangladesh pushed back the Muslim Rohingya refugees from Myanmar in 199293). But for the present discussion, it is probably best to keep these alleged unregistered millions outside our considerations and stick to verified figures. Even without this unknown "dark figure" of unregistered Muslim inhabitants, it is only very slightly exaggerated to say that in the Indian Republic, ever since 1951, "the proportion of Muslims has been gradually but steadily increasing every decade by roughly one percentage point".
All kinds of local and regional data confirm the faster muslim growth rate. The two provinces with the highest relative population growth between 1981 and 1991 are Kashmir (28%) and Lakshadweep (27%), both with a Muslim minority though in very divergent economic and political conditions; followed by Madhya Pradesh (26%) and Uttar Pradesh (25%). In Uttar Pradesh, between 1981 and 1991, the Muslim percentage rose from 15.53% to 17.33 %, in Bihar from 14.13% to 14.81%, in West Bengal from 21.51 to 23.61%. While in Bihar birth control seemed to have a slight effect in a decrease of the decadal increase from 30.03% in 197181 to 29.5% (still more than 5% higher than the Hindu figure) in 198191, there was a much larger increase in the decadal increase in U.P. from 29.11% to 36.54%, and in West Bengal from 29.55% to 36.89%.
How does all this work out for U.N. Mukherji's demographic doomsday prognosis? Mani Shankar Aiyar, the Congress Party's leading antiHindutva polemicist, predicted in 1993 that until at least the mid21st century, the proportion of Muslims will remain stable "bar a decimal point up or down from time to time, at 11.2 per cent." But even the conservative estimate for 1991 (not yet published at the time of his writing) already refuted his prediction by putting the Muslim population more than 1% higher. In the contest between Mukherji's prediction of a continuous Muslim growth and Aiyar's prediction of a stable percentage for the next sixty years, Mukherji has obviously won.
Other secularist observers admit that "it is true that the growth rate amongst Muslims is higher than amongst Hindus", and have calculated, on the basis of the 1971 and 1981 census figures, that "if both the communities continue to grow at the same rate, Muslims at the turn of the century will account for only 13.55% of the country's population." That constitutes a refutation of the stillrecent rumour that Muslims will outnumber Hindus by the year 2,000, but also an admission that the Muslim percentage shows a continuous substantial increase. For the long term, their estimate is: "It will take 316 years for Muslims to outnumber Hindus". Here, the difference with Mukherji is merely one over the exact quantity of time needed to do the job.
The projection just quoted is a strictly linear extrapolation of the HinduMuslim differential in the decade 19711981. But this approach fails to take into account at least two important factors which we will now consider more closely: the Muslim increase is not linear, but is itself increasing; and there is a large immigration of Muslims from Pakistan and Bangladesh, which can only increase.
1.6. The longterm trend
Ever since regular census operations were started, the percentage of Muslims has grown every decade in British India, independent India, Pakistan and Bangladesh. The only seeming exception is Pakistan between 1971 and 1981, due to the official declaration of Ahmadiyas (ca. 3%) as non‑Muslims in 1974. We will include the Ahmadiyas in the Muslim category, as they themselves also do.
In the subcontinent, Muslims were 19.97% in 1881 and 24.28% in 1941 when the last prePartition census was held. According to the last census (1991), the Muslim percentage in the subcontinent was as follows:
Bangla Desh: 86.8% of 108,760,000, or ca. 94,403,608.
Pakistan: 97.0% of 126,406,000, or ca. 122,613,820.
India: 12.6 % of 846,349,050, or ca. 106,639,980.
Total: 29.92% of 1,081,515,050, or ca. 323,657,480.
The Muslim percentage has not only increased, but the rate of increase itself has increased. This is very clear when we take a longterm perspective: in the fifty years between 1941 and 1991, their percentage has risen 5.64% (from 24.28% to 29.92%), substantially more than the 4.31% gain in the sixty years between 1881 and 1941. At this rate, the Muslims in the Subcontinent must have passed the 30% mark in the mid1990s and will pass the milestone of becoming more than half the number of Hindus (ca. 32% to ca. 64%) before the census of 2011.
Similarly calculating from the available figures for the three countries, the Hindu percentage had come down to 65.15% in 1991. To evaluate the trend of the Hindu percentage, we must take into account that the preIndependence census always had a tentative category "tribal" or "animist", variously defined and therefore making odd quantitative jumps (but always between 2.26% and 3.26%), from 2.57% in 1881 to 2.26% in 1941. After Independence, this category was included in the Hindu category. So, putting everything on the postIndependence denominator, we include the "animists" in the Hindu percentage to get a total Hindu percentage of 77.35% for 1881, 71.72% for 1941, and 65.15% for 1991. Here again, we see a longterm acceleration of the observed trend: a decrease of 5.63% in the sixty years between 1881 and 1941, and a larger decrease of 6.57% in the shorter period of fifty years between 1941 and 1991.
So, all the predictions quoted above are far too conservative, for they are based on a linear projection. In reality, the observed trends are accelerating, so Muslims will need far less than 316 years to outnumber the Hindus. According to Mukherji, the Hindu percentage of exBritish India (including India, Pakistan, Bangladesh and Burma) should now have declined by about 13%, down to 54% of the total. This was too pessimistic, he overestimated the rate of Hindu decline, but in the future the increasing rate of Hindu decline will catch up with Mukherji's estimate.
1.6. Muslim testimony
A very unexpected prediction which more than confirms this trend is implicitly given by one of those authors who ridicule Hindu fears about the demographic evolution: Rafiq Zakaria claims that to outnumber Hindus, "Muslims will need no less than 365 years", but in another discussion in the same book he complains that according to an allIndia governmental survey in 1984, Muslim children were underrepresented among primary school pupils: "the enrolment of Muslim children at the primary school level in the relevant period was 12.39 per cent as against the child population of 16.81 per cent." So, Rafiq Zakaria claims that in 1984, Muslim children in the primaryschool age group constituted 16.81% of the total.
This means that approximately by the time this group reaches the median position on the age pyramid of their community (i.e. when the number of people younger than them will equal the number older than them), certainly before 2010 (when they will be in their midthirties), the native Muslim population of India, not counting the millions of post1984 immigrants, will be 16.81% of the total. That is even faster than all the above (admittedly cautious) calculations suggest, as if the religious differential in the use of birth control since the 1960s is having a bigger impact than hitherto assumed.
It gets even more dramatic when you look at it this way: in 1984, a generation of Muslims which was about 12% of the population had produced a generation of children, certainly not more than 30 years younger on average, which constituted more than 16%. This would mean an unprecedented growth rate of more than 4% in less than 30 years, or rather, a growth with over a third of the original percentage (4 to 12). For a little thought experiment: if this differential growth rate is kept constant, we get 16.81% of Muslims in ca. 2014, over 22% in 2044, nearly 30% in 2074, 40% in 2104, crossing 50% in ca. 2125 etc., all without counting the effect of Muslim immigration.
Of course, demographic prediction is a difficult task, which has to factor in many different types of data and influences, so we should not take any amateur predictions too seriously, nor those of specialists paid by political institutions with an interest in popularizing this or that impression about demographic trends. Let us not pin ourselves down on precise predictions but let us not ignore the generally visible trend either. The one general prediction to which the data certainly compel us, is that the Muslim percentage will be increasing at an accelerating rate for at least another generation; and also beyond that, unless the present generation of young adult Muslims brings it procreation rate down to the average Indian level.
1.7. The Subcontinental context
The subcontinental total shows a faster growth of the Muslim percentage than the official figures for the Indian republic, and should be considered the more important indicator for the long‑term evolution, given the increasing spill‑over of population surplus from Pakistan and especially Bangladesh. Apart from the millions of Bangladeshis staying illegally in India, there is also an increasing number of Pakistanis overstaying their visas or otherwise settling in India.
In Pakistan and Bangladesh, the Muslim percentage has continually increased, partly by pestering the nonMuslims out, partly by conversions under pressure (pressurizing people to marry their daughters off to Muslims, allocating jobs on conditon of conversion, etc.), and partly by higher birthrates. Bangladeshi Muslim expansion has already destroyed the Chakmas and other nonMuslim populations in the Chittagong Hill Tracts, with the ethnically cleansed minorities fleeing to India's NorthEast, there to create friction with the host population. But the most worrying from the Indian viewpoint is not the rise in percentage but the rise in absolute figures: in parts of Pakistan and in the whole of Bangladesh, sheer living space is becoming extremely scarce, and these countries may pursue a policy of pushing their surplus population into India.
Incidentally, in Nepal the breakthrough of Islam (nearly nonexistent in the 1970s) is simply spectacular, and is again due in large measure to immigration from Bangladesh. In Sri Lanka, the Muslim percentage is slowly rising by demographics alone.
So, every decade the Muslim percentage in the Subcontinent increases by more than 1%, with the rate of increase itself increasing. In India, the rate of increase in the Muslim percentage is considerable, though lower than the subcontinental total, but is rising faster due to the differential in the use of birth control and the increasing Muslim immigration. In Hindutva circles, this remarkable demographic differential is interpreted as the result of Muslim "demographic aggression".
 Estimate given by Rodney Stark: The Rise of Christianity (1996); this book has received jubilant reviews in Christian circles because it downplays the factor of suppression of Paganism by Christian emperors and highlights positive factors such as the effectiveness of Christian charity in attracting converts, see e.g. the review by Marcel van Nieuwenborgh: "Toen Nieuwe Testament nog nieuw was", De Standaard (Brussels), 2761996.
 Among the Parsis, 36% is over 55, and 25% of the adults are unmarried. Their numbers have plummeted to ca. 50,000, half of what it was in mid20th century. See e.g. Nergis Dalal: "The Parsis are allowing themselves to die out", Times of India, 2031990; Bachi J. Karkaria: "Dying embers", Illustrated Weekly of India, 2811990; and Arnavaz Mama: "Survival strategies", Sunday Observer, 711990.
 See also P.K. Datta: "'Dying Hindus' Production of Hindu communal common sense in early 20th century Bengal", Economic and Political Weekly, 1961993, p.1307; and C. Jaffrelot: Hindu Nationalist Movement (Viking/Penguin 1996), p.24. Typically, both exclusively discuss the presumed sociological determinants and other externals of Mukherji's analysis, not its degree of accuracy.
 S.K. Bhattacharyya: Genocide in East Pakistan/Bangladesh (A. Ghosh, Houston 1987), p.159. In the Hindi original: Ham do hamâre do Ham pânch hamâre pachîs. The saying is sometimes accompanied by a cartoon showing the Government poster (father, mother, boy, girl) plus its Muslim variant: a man with goatbeard and four veiled wives surrounded by a sea of children.
 Syed Shahabuddin: "Census 1991, Muslim Indians and Sangh Parivar", Muslim India, September 1995, p.386. The estimate for Jammu & Kashmir is based on the assumption of a constant ratio between the religions, which understates the Muslim percentage by ignoring the higher Muslim birth rate and the Hindu emigration from the state.
 Most explicitly in the title of retired police officer Baljit Rai's book Demographic Aggression against India: Muslim Avalanche from Bangladesh (B.S. Publ., Chandigarh 1993), about the spillover of Bangladesh's demographic surplus into India.