The Question of Minorities - part 3: Partition era migration of Hindus from East Pakistan

Pro-Indian agents, who never accepted the partition of India, claim that some 5.3 million Hindus had left East Pakistan between 1947 and 1971. Not only do they fail to share any credible source for such claims but they are also delinquently silent about the exodus figures of Muslims from India in the same period to Pakistan.[i] As noted by credible scholars, Indian statistics on such contentious matters were "not merely guesses”, 'but frequently demonstrably absurd guesses'.[ii] My concern in this article is with statistics relating to migration of people during the partition. As we shall see below unreliable statistics, let alone politicized claims with ulterior motives on such important issues, gravely impairs our judgment and only breeds mistrust and animosity between communities.

So, how many people did migrate between Pakistan and India? As I have noted elsewhere, in terms of mere size and rapidity, the Partition of British India to Pakistan and India in August of 1947 constitutes perhaps the largest example of voluntary and involuntary population movement in modern history. In the pre-partition Bengal, there were 5 divisions which were subdivided into a total of 28 districts, and that almost the entire population in Bengal was either Hindu or Muslim, apart from about 3 percent “Tribal”: 55 percent Muslim and 42 percent Hindu.[iii] [Note: The Hindu % in Bengal in 1931, 1941 and 1951 was 42.3, 41.2 and 42.4, while the Muslim % in the same period was 55.8, 55.7 and 56.4, respectively. In the same period, the tribal % was recorded as 1.8, 3.1 and 1.2, respectively. It goes without saying that the unusually high growth rate (e.g., more than doubling from 896,000 to 1,849,000) of the tribal population in a 10-year period between 1931 and 1941 simply cannot be explained by any natural growth phenomenon, and points to mis- or cross-categorization, and/or massive relocation into Bengal from other regions, including parts of Burma. The average annual growth rates amongst Hindus and Muslims in Bengal in 1931-1941 were 1.6% and 1.85%, respectively.]

Year =>	1931		1941		1951
Total in millions =>	48.811	%	58.838	%	61.653	%
Hindu	20.670	42.3%	24.244	41.2%	26.128	42.4%
Muslim	27.245	55.8%	32.745	55.7%	34.789	56.4%
Tribals	0.896	1.8%	1.849	3.1%	0.736	1.2%

Estimates of migrants to Pakistan and India between 1947 and 1951 vary between 10 and 17 million, while estimates of deaths associated with Partition range from 200,000 to over a million.[iv] Of these migrants, only about 3 million crossed the border in the eastern sector of what was once British India.

A further barrier to isolating the effects of Partition is the Bengal famine of 1943 when according to Professor Amartya Sen some 3 million people might have died. However, as we see from the table below the actual figure is probably 3 times as large (anywhere from 6.8 million to 9.3 million). [Note: the lower famine casualty figure omits ‘tribal’ data with abnormally high annual growth rate.]

	Population in 1000’s	1931	% avg. annual growth rate (‘31-‘41)	1941	% avg. annual growth rate (’41-’51)	1951	Expected population in 1951	Potential 1943 famine casualty
Bengal Male	Total	25338	2	30954	0.52	32583	37815	5232
	Hindu	10813	1.82	12970	0.72	13941	15557	1616
	Muslim	14060	1.92	17039	0.69	18263	20649	2386
	Tribals	465	7.35	945	-0.0875	379	1609	1230
Bengal Female	Total	23473	1.72	27884	0.43	29070	33124	4054
	Hindu	9857	1.34	11274	0.78	12187	12894	707
	Muslim	13185	1.75	15706	0.51	16526	18709	2183
	Tribals	431	7.70	904	-0.089	357	1521	1164

 

While the population of Bengal as a whole for both males and females grew, it grew more than one percent more slowly between 1941 and 1951 than between 1931 and 1941. This reduction in growth rate had much to do with the 1943 Famine. Both religious communities experienced reduced growth rates, but the reductions were larger for Muslims than for Hindus, and were least for Hindu females. [See http://drhabibsiddiqui.blogspot.com/ for detail analysis, and data.]

Then there is also the question about the reliability of the census data further complicating the issue. In 1941 Britain was at war against Japan that had captured Burma, and the Burmese Buddhist population collaborated with the Fascist Japanese Imperial forces and carried out an extermination and eviction campaign against non-Buddhist population that were racially Indian and considered an ally of the British. Many Indians working in Burma were killed and many others fled as a result of the war, which lasted for six years (1939-45). Could the above average growth rate in 1931-41 amongst Muslims in districts like Chittagong (e.g., 2.4% among males) and Noakhali (e.g., 3% overall), the two districts that are closer to Arakan in Burma, be a direct result of the World War II?

East Bengal Demography in thousands
District	Sex	Religion	Population (1931)	% Annual Growth Rate (1931-41)	Population (1941)
Noakhali	Male	Muslim	671	3.3	931
Noakhali	Male	Hindu	187	1.2	212
Noakhali	Female	Muslim	668	2.7	873
Noakhali	Female	Hindu	179	1.1	201
Chittagong	Male	Muslim	641	2.4	817
Chittagong	Male	Hindu	195	1.8	234
Chittagong	Female	Muslim	686	1.4	788
Chittagong	Female	Hindu	197	1.3	224
			West Bengal Demography in thousands
District	Sex	Religion	Population (1931)	% Annual Growth Rate (1931-41)	Population (1941)
Howrah	Male	Muslim	137	1.6	162
Howrah	Male	Hindu	469	3	632
Howrah	Female	Muslim	107	1.5	134
Howrah	Female	Hindu	391	2.8	520

Could the highly unusual 3% growth rate amongst Hindus in Howrah in 1931-41 be similarly attributed to the same reason, i.e., World War II? Or, was it due to undercounting in 1931? Nor should we ignore the importance of major cities to attract people for jobs. However, for Howrah, the Muslim population between 1931 and 1941 grew by only 1.5%, which is half the growth rate amongst Hindus. As such, this unusual growth rate amongst Hindus could not have been due to jobs but for some other factor(s). Dhaka’s (Dacca) population, on the other hand, grew in the same period for both Hindus and Muslims by 1.9% and 2.2%, respectively, which seems quite reasonable for a major city in East Bengal towards attracting all for jobs there.

District	Sex	Religion	(1931)  Population (thousands)	Growth Rate	(1941)  Population (thousands)
Dacca	Male	Muslim	1168	2.2	1453
Dacca	Male	Hindu	569	2.1	699
Dacca	Female	Muslim	1125	2.1	1388
Dacca	Female	Hindu	556	1.7	661

 

An Andrew W. Mellon Foundation and Harvard University-funded study (hereinafter referred to as the Mellon Report), entitled “The Demographic Impact of Partition: Bengal in 1947”  that was jointly collaborated between MIT, Harvard, Johns Hopkins and Fordham Universities in 2003 has shown that in Punjab, where the bulk of Partition related migration was over by the end of 1947, migration of Bengali Hindus to India and of Bengali and Urdu speaking Muslims to East Pakistan continued through 1951 and beyond further complicating the problem around reliable figures.[v] It says, “Bengal as a whole was little affected by Partition-related moves: the movement from East Pakistan into Assam was probably largely from the district of Sylhet, that had not been part of Bengal prior to Partition. West (Indian) Bengal gained about 1.4 million migrants, whereas East Pakistan lost a similar number.”

Most of the cross-border migration in 1947 happened in the western front. In 1941, in only four Punjab districts was the population more than two-thirds Hindu; in 12 districts the population was more than two-thirds Muslim; no district was majority Sikh; in only three districts was the population less than 28 per cent Muslim. By 1951, eight districts were more than two-thirds Hindu; all 15 of the districts that became part of Pakistan were more than 90 per cent Muslim; three districts were majority Sikh; and only one district that remained in India had more than 2.5 per cent Muslim population.

As to the case of Bengal, the Mellon Report continues, “Nothing remotely similar happened in Bengal, where the biggest change at the district level was the outflow of Hindus from Dacca district, reducing the proportion Hindu from 32 percent in 1941 to 21 percent in 1951.” This observation is understandable given the fact that Dhaka became the capital of East Pakistan in 1947, and many Muslims, including the Mohajirs (refugees) from India, gravitated to the capital city for a plethora of reasons including working in government sectors. On the other hand, the entire Hindu community was opposed to the partition of India, and felt that it would lose the privileged status that it had hitherto enjoyed in British India by continuing to live in what became East Pakistan.

It is worth noting here the strong opposition that came from famous Bengali Hindu luminaries like Rabindranath Tagore who some 40 years before Pakistan became a reality had strongly opposed the division of Bengal (1905), let alone the establishment of Dacca University (once known as the “Oxford of the East”) in East Bengal in 1921. Unlike the Hindus, who all supported undivided India, Muslims in British India were split on the Pakistan issue, and many Muslims, especially the prosperous ones, living in mainland India decided against migration and stayed in India. Even then, there was a net decrease in the number of Muslims living in 1951 (compared to 1941) in some of the districts of West Bengal, esp. in Howrah; the trend was just the opposite in most of East Pakistan, esp. in Dhaka, where there was a net decrease in Hindu population.

The overall Muslim proportion in West Bengal declined in 1951 compared to 1941. For instance, in West Bengal’s Burdwan, Birbhum, Bankura, Midnapore, Hooghly, Howrah and Murshidabad districts which had Muslim:Hindu proportion of 19:81, 29:71, 5:95, 8:92, 16:84, 20:80 and 58:42 in 1941 became 16:84, 27:73, 4.6:95.4, 7:93, 13:87, 16:84 and 55:45, respectively, in 1951. Correspondingly, in East Bengal’s (later East Pakistan’s) Rangpur, Bogra, Pabna, Dacca (Dhaka), Mymensingh, Bakerganj, Tippera (Comilla), Noakhali and Chittagong districts which had Muslim:Hindu proportion of 53:47, 85:15, 77:23, 68:32, 78:22, 73:27, 77:23, 81:19. 78:22 in 1941 became 63:37, 87:13, 84:16, 79:21, 84:16, 80:20, 81:19, 84:16, and 80:20 in 1951.

So, it is not difficult to see the impact of the partition – the proportion of Hindus in East Pakistan shrank considerably.  The Mellon Report says, “In the period 1941 to 1951, the Hindu populations of East Bengal districts all contract, though by very variable amounts, whereas the Hindu populations of West Bengal tend to grow somewhat faster than 1931 to 1941, as would be expected given the migration of over 2 million people, presumably predominantly Hindu, from East Bengal into West Bengal. The Moslem [sic] populations of East Bengal tend to grow more slowly between 1941 and 1951 than in the earlier period, as did the Moslem[sic]  populations of West Bengal... These patterns of growth suggest a more substantial relocation of the Hindu population about the time of Partition than of the Moslem [sic] population.” (p. 18)

The Mellon Report analyzed the demographic data for the aforementioned 7 districts of West Bengal and 9 districts of East Bengal. In those districts, the total number of Hindus and Muslims are tabulated below, showing that the net Hindu and Muslim gains were, respectively, 191,000 and 1,443,000 in 1941-1951 in what was once used to be known as joint Bengal. Noting that the pre-partition Bengal had a Muslim majority (55%, compared to 42% Hindus and 3% tribals), this disproportionate net gain among the Muslims could not have happened without a massive migration of non-Bengali Muslims to East Pakistan.

Population in thousands		1931	1941	1951	Net Gain (1941-1951)
West Bengal	Hindu	7753	8727	10349	1622
	Muslim	1995	2358	2343	-15
East Bengal (East Pakistan)	Hindu	6124	7024	5593	-1431
	Muslim	16462	20058	21516	1458

 

A Chi-square test to check for association between religious communities with census year showed that while the loss in religious minority numbers in both East Pakistan and West Bengal was statistically significant in 1951, the situation in Punjab was worse. As can also been seen the total number of cross-border migration of Muslims and Hindus to and from East Pakistan (1947-1951) possibly could not have been more than 3 million. 

The Report concludes “Bengal remained quite heterogeneous in terms of religion in 1951. The second feature is the very low growth of the population of Bengal in the 1940s relative to its growth in the 1930s: had 1931-41 growth rates continued between 1941 and 1951, the population of Bengal as a whole would have been nearly 9 million larger than it actually was in 1951. This excess loss cannot be explained by net out-migration, since our analysis suggests that Bengal as a whole was little affected by net migration in 1947 and thereafter. Given the relatively small scale of migration, it is unlikely that this population loss resulted from violence at the time of Partition; it is a grim reminder of the magnitude of the disastrous Bengal famine of 1943.”

My statistical analysis of the census data agrees with the findings of the Andrew Mellon Foundation and Harvard University-funded study and shows that there is no truth to the exaggerated claims made by certain quarters about millions of Hindus migrating out of the then East Pakistan as a result of partition of British India.

 ==========
Appendix: Tables on population demography

1, West Bengal Demography in '000s
District	Sex	Religion	Population 1931	%Growth Rate	Population 1941	%Growth Rate	Population 1951
Burdwan	Male	Muslim	152	1.5	177	0.4	183
Burdwan	Male	Hindu	639	1.4	737	2.7	969
Burdwan	Female	Muslim	140	1.3	160	-0.1	159
Burdwan	Female	Hindu	600	0.9	657	2.8	866
Birbhum	Male	Muslim	126	1.3	143	0	144
Birbhum	Male	Hindu	317	0.8	344	1.4	394
Birbhum	Female	Muslim	127	1.3	144	-0.1	143
Birbhum	Female	Hindu	319	0.7	342	1.1	381
Bankura	Male	Muslim	26	0.9	29	0.8	31
Bankura	Male	Hindu	506	0.8	546	1	605
Bankura	Female	Muslim	25	0.9	27	0	27
Bankura	Female	Hindu	505	0.5	533	1.1	597
Midnapore	Male	Muslim	108	1.4	125	0.1	126
Midnapore	Male	Hindu	1263	0.9	1376	1.4	1575
Midnapore	Female	Muslim	105	1.5	122	-0.6	114
Midnapore	Female	Hindu	1230	0.6	1306	1.4	1508
Hooghly	Male	Muslim	97	1.5	113	-0.7	106
Hooghly	Male	Hindu	489	1.8	587	2	716
Hooghly	Female	Muslim	83	1.2	94	0.7	101
Hooghly	Female	Hindu	435	0.6	462	3.1	628
Howrah	Male	Muslim	137	1.6	162	-1.3	143
Howrah	Male	Hindu	469	3	632	1.6	744
Howrah	Female	Muslim	107	1.5	134	-1.3	118
Howrah	Female	Hindu	391	2.8	520	1.5	601
Murshidabad	Male	Muslim	379	1.2	465	0.3	477
Murshidabad	Male	Hindu	295	1.6	346	2	391
Murshidabad	Female	Muslim	383	1	463	0.2	471
Murshidabad	Female	Hindu	295	1.4	339	1.9	374

2. East Bengal Demography  in '000s
District	Sex	Religion	Population 1931	%Growth Rate	Population 1941	%Growth Rate	Population 1951
Rangpur	Male	Muslim	955	1.1	1068	1.4	1224
Rangpur	Male	Hindu	395	0.9	430	-3.3	308
Rangpur	Female	Muslim	140	1.1	160	1.1	159
Rangpur	Female	Hindu	600	0.6	657	-2.9	490
Bogra	Male	Muslim	461	1.6	540	0.6	575
Bogra	Male	Hindu	95	0.6	100	-1.7	84
Bogra	Female	Muslim	445	1.5	518	0.4	540
Bogra	Female	Hindu	83	0.5	87	-1.2	78
Pabna	Male	Muslim	570	1.7	674	0.2	689
Pabna	Male	Hindu	168	1.6	197	-4.1	131
Pabna	Female	Muslim	542	1.7	640	-0.1	636
Pabna	Female	Hindu	164	1.3	186	-3.9	127
Dacca	Male	Muslim	1168	2.2	1453	1.6	1699
Dacca	Male	Hindu	569	2.1	699	-4.8	435
Dacca	Female	Muslim	1125	2.1	1388	0.9	1513
Dacca	Female	Hindu	556	1.7	661	-4.9	405
Mymensingh	Male	Muslim	2034	1.7	2410	0.4	2518
Mymensingh	Male	Hindu	619	1.2	696	-3.4	496
Mymensingh	Female	Muslim	1894	1.7	2255	0.1	2277
Mymensingh	Female	Hindu	555	0.8	601	-2.9	451
Bakerganj	Male	Muslim	1078	2	1322	1.4	1516
Bakerganj	Male	Hindu	416	1.8	497	-2.9	373
Bakerganj	Female	Muslim	1027	1.9	1245	1	1382
Bakerganj	Female	Hindu	396	1.5	462	-3	344
Tippera	Male	Muslim	1210	2.4	1541	0.4	1604
Tippera	Male	Hindu	384	1.7	457	-2.4	360
Tippera	Female	Muslim	1147	2.3	1435	0.3	1482
Tippera	Female	Hindu	366	1.5	423	-2.1	342
Noakhali	Male	Muslim	671	3.3	931	0.8	1006
Noakhali	Male	Hindu	187	1.2	212	-1.2	188
Noakhali	Female	Muslim	668	2.7	873	0.4	909
Noakhali	Female	Hindu	179	1.1	201	-1.7	170
Chittagong	Male	Muslim	641	2.4	817	1.6	957
Chittagong	Male	Hindu	195	1.8	234	-0.2	230
Chittagong	Female	Muslim	686	1.4	788	0.5	830
Chittagong	Female	Hindu	197	1.3	224	-0.9	205

---

[i] http://www.voiceofdharma.org/books/tfst/appii1.htm mentions 4.45 million Hindus and not 5.3 million. Even that source is unreliable. It also quotes an article in the Weekly Holiday (The Missing Population, Holiday, Weekly, Dhaka, 7 January 1994) by Mohiuddin Ahmed, which is unreliable, stating that some 2 million Hindus had left Bangladesh between 1974 and 1991.

[ii] See, e.g., comments made in the book “Development Economics on Trial: The Anthropological Case for a Prosecution” by Polly Hill, Cambridge University Press (1986), p. 44; see also the works of Dewey and Charlesworth on Indian statistics as referenced in the above book. See also the article: “The Poor Quality of Official Socio-economic Statistics Relating to the Rural Tropical World: With Special Reference to South India” by Polly Hill, Modern Asian Studies, 18, 3 (1984), pp. 491-514, http://www.jstor.org/discover/10.2307/312264?uid=3739864&uid=2129&uid=2&uid=70&uid=4&uid=3739256&sid=21103548611091

[iii] http://iussp2005.princeton.edu/papers/52236, The Demographic Impact of Partition:

Bengal in 1947, Hill K1, Seltzer W2, Leaning J 3, Malik SJ3, Russell SS4, 1 Department of Population and Family Health Sciences, Johns Hopkins Bloomberg School of Public

Health; 2 Department of Sociology, Fordham University; 3 Program on Humanitarian Crises and Human Rights, Francois-Xavier Bagnoud Center on Health and Human Rights, Harvard School of Public Health; 4 Center for International Studies, Massachusetts Institute of Technology.

[iv] http://iussp2005.princeton.edu/papers/52236, op. Cit.

[v] http://iussp2005.princeton.edu/papers/52236