Abstract
The recently released research paper on
religious minorities by the members of the Economic Advisory Council to the
Prime Minister of India has stirred up an unseemly storm. Much of the
controversy stems from how the authors compute the change in the shares of
Hindu and Muslim populations between 1950 and 2015. Instead of reporting the
more straightforward and precise percentage point difference in shares (a 4.35
percentage points increase in the share of the Muslim population), the authors
report the percentage change in the percentage shares (43% increase in the
share of the Muslim population). Predictably, the alarmist number of 43% has
been picked up in the media to demonise the Muslim population. I critique the
methodology used by the EAC-PM paper and also question the need to release this
paper in the middle of the Indian elections when tensions are already high. I
call this research clickbait because its primary purpose seems to be not so
much to illuminate the dynamics of population shares in India as to attract
media attention and provide talking points to those already antagonistic
towards minorities in India.
Introduction
The recently released report by Shamika Ravi,
Abraham Jose and Apruv Kumar Mishra titled 'Share of Religious Minorities A
Cross-Country Analysis (1950-2015)' has created a lot of controversy.[1]
Much of the controversy stems from the authors being members of the Economic
Advisory Council to the PM. If the authors had written the report in their
individual capacities, it might have made for an interesting discussion at a
seminar at an academic institution. Much of such discussion would have focused
on the contentious methodology employed by the authors to reach their
conclusion about the dynamics of religious populations in India, especially the
population of Hindus and Muslims. But, given the office the authors occupy, the
paper has had an outsized impact.
The Methodology
The EAC-PM paper uses data from the Religious
Characteristics of States Dataset (RCS-Dem),[2]
[3]
focusing on the time period 1950 to 2015. In fact, they focus on just two years,
namely, 1950 and 2015. They include in their analysis 167 countries, but
expectedly, all attention has been focused on India. There has been criticism about
why the RCS-Dem dataset has been used for India when far richer data would have
been available from Indian Census reports till 2011. [4]
The authors of the EAC-PM paper have justified their choice of the dataset to
ensure comparability across the 167 countries they study. I don't have much
objection to using the RCS-Dem dataset since we often use a non-Indian dataset
(e.g. World Bank or IMF) when doing a cross-country study.
The problem I have with the EAC-PM paper is
with their methodology, specifically how they measure the change in the share
of the religious populations between 1950 and 2015. They measure this as:[5]
Note that the authors measure shares in 1950
and 2015 as percentages. So, the above measure computes the percentage change
in the percentages of 1950 and 2015. The same formula as above can be used to calculate
the rate of change in share for any other (religious) group.
The above formula is a most unusual way of
computing the change in a population group's share. It is unnecessarily
complicated, which is perhaps why the paper has created such a controversy. A
far more straightforward and much less ambiguous way of measuring a change in
shares is to compute not the rate of change of share but the absolute
difference in the shares:
The Empirics
Using percentage points makes things so much more
straightforward than using the cumbersome approach of the authors of the
EAC-PM. Consider Table 1, which I created from data extracted from the RCS-Dem.
Table 1: Data on Muslim and Hindu Population of India
The EAC-PM paper has reported the change in the
percentage of the Muslim population from 9.84% in 1950 to 14.09% in 2015 as a 43.15%
increase (Note: the computation from the data in Table 1 is 43.19% due to
rounding off). The change in the Hindu population share has been reported as a
decline of 7.82%. These percentage changes in shares have created much
controversy, and the usual fears have been stoked to demonise the Muslim
population. The Population Foundation of India had to issue a call to the media
not to misreport the findings of the EAC-PM paper.[8]
The problem and the possibility of
misrepresentation stems from the confusing methodology employed by the EAC-PM.
The same change in the shares of Muslim and Hindu populations could have been
easily conveyed using the percentage point difference. The change in the share
of the Muslim population could have been stated far more clearly as an increase
of 4.35 (14.09% in 2015 minus 9.84% in 1950) percentage points. Reporting a 4.35
percentage points increase in the share of the Muslim population would have
been far less alarmist and less open to misrepresentation than reporting it as
a 43.15% increase in the share of the Muslim population.
This imprecise and confusing methodology employed
in the paper has also led to an incorrect understanding of the issue, even by
the Indian Prime Minister. In his interview with Arnab Goswami of Republic
World TV, the Prime Minister incorrectly stated that the Hindu population has
gone down by 7.8% and the population of "minorities" has gone up by
43% (see the segment of the interview at about the 58 minutes mark).[9] Clearly, the Prime Minister confused the changes in shares of Hindu and Muslim populations with the changes in the levels of populations of these two groups. If there is confusion in the understanding of these issues in the mind of the Prime Minister, one can well imagine the confusion of
the ordinary person on the street. Also, imagine the handle the reported
results of the paper have given to TV anchors who revel in anti-Muslim propaganda
on their channels. The authors of the paper have done a grave disservice to the
Prime Minister of the country, to the people of India, and to scientific
research.
There is no shortage of studies which show that
the rate of growth in population and fertility rates of all communities start
to dip with higher levels of education and income, and this is what is
happening to the Muslim population also.[10]
I present below two graphs to correct the distorted picture that the EAC-PM
study has put out.
There is clear evidence that the population of the two communities has diverged over the years. This figure shows how unnecessarily alarmist the conclusions of the EAC-PM paper were.
Fig. 2 shows that the growth rate of Hindu and
Muslim populations is gradually converging, and, given the difference between
levels of the population seen in Fig. 1, it is inconceivable that the Muslim
population will come even remotely close to the Hindu population.
Conclusion
The final point I wish to make is about the
timing of the release of the EAC-PM paper. The paper's authors have done an extreme
disservice to the social science discipline by not considering the impact of their
paper and the results they report on Indian society. That they chose to release
their paper in the middle of the Indian elections shows how disconnected the
authors seem to be from the atmosphere in which elections take place in India.
The question the authors must ask themselves is whether they could not have
delayed publishing their paper until after the elections had concluded. What
purpose was served by releasing the paper in an atmosphere of high tension?
And, given the widespread publicity their paper has received, have the authors
sought to rectify the egregious errors and misrepresentations that have been
rampant in the media? This is the least we could have expected from responsible
social scientists and those who advise the highest office in the country.
[1] Ravi, S.,
Jose, A. and Kumar Mishra, A.M. (2024) Share of Religious Minorities A
Cross-Country Analysis (1950-2015), EAC-PM Working Paper Series, EAC-PM/WP/29/2024,
[2] The
Association of Religious Data Archives (2019) Religious Characteristics of
States Dataset Project - Demographics v. 2.0 (RCS-Dem 2.0), COUNTRIES ONLY, https://www.thearda.com/data-archive?fid=RCSDEM2
[3] Brown, D., & James, P. (2019, February 10). Religious
Characteristics of States Dataset Project - Demographics v. 2.0 (RCS-Dem 2.0),
COUNTRIES ONLY.
[4] Mehrotra, S. and Singh, B. (2024) The EAC-PM's Paper on 'Muslim
Population' Is a Travesty of Research Practices, https://thewire.in/economy/pm-eac-muslim-population-increase-research
[5] See page 11 of the reference in note [1]. Their formula makes a specific reference to the majority population while I have made it more general and referred to any population group
[6] Government
of India (2023) Economic Survey 2022-23, https://www.indiabudget.gov.in/economicsurvey/doc/echapter.pdf
[7] Ordway,
D-M. (2022) Percent change versus percentage-point change: What’s the
difference? 4 tips for avoiding math errors, https://journalistsresource.org/home/percent-change-math-for-journalists/
[8] Population
Foundation of India (undated) Media must not misreport the study on population
by PM-EAC, https://www.populationfoundation.in/media-must-not-misreport-the-study-on-population-by-pm-eac-population-foundation-of-india/
[9] Republic
World (2024) PM Modi And Arnab: Nation's Most Watched Interview Of 2024, https://www.youtube.com/watch?v=A9jEEIpabhk
[10] See the
reference quoted in note [4] above
