Page | 1
SHIFTING RELIGIOUS DEMOGRAPHICS - ANALYZING
MINORITY POPULATION TRENDS ACROSS COUNTRIES FROM
1950 TO 2015
The global economy is in a phase of churn and being closely watched
and analyzed by economists and policymakers everywhere. There are,
however, major demographic transitions that are also underway across
countries - but mostly going unnoticed by analysts ranging from economy-
watchers to democracy-watchers. These silent transformations have the
power to reshape societies and states. Ironically, of the four megatrends
whose cascading effects are bringing about these transformations -
demography, technology, economy and climate change3 - the forecasts for
demography are the most predictable. Shifting demographic trends are
aggravating economic disparity within and between countries, straining
governance and fuelling friction between states and people.
Political changes are mere symptoms of deeper structural changes
that are happening in societies due to a variety of transformations, of which
demographic evolution is an important component. The composition of
populations in countries around the world is changing along several axes
which are well-documented such as age, urbanization and migration levels.
This paper focuses on the changes in religious composition of populations
across countries around the world between 1950 and 2015 - an axis of
demographic change that has been under-studied. For our analysis, we use
the Religious Characteristics of States Dataset (RCS-Dem, 2017). The authors
of the RCS-Dem Dataset note in their 2017 paper, “Religious characteristics
emerge as significant state-level variables and rival other factors
conventionally thought to influence political processes and outcomes.”4
In this paper we provide a detailed cross-country descriptive analysis
of the status of minorities around the world measured in terms of their
changing share in a country’s population over 65 years between 1950 and
2015. For the
Page | 2
167 countries in our analysis, the average value for the share of the majority
religious denomination in the baseline year of 1950 is 75 percent. The Mean
of the distribution capturing the change in majority religious denomination
between 1950 and 2015 is -21.9. This means that globally the share of the
majority religious denomination has gone down by approximately 22
percent. In other words, on an average the world has become more
heterogeneous in the period under study.
Our study finds that of the 40 countries that experienced the biggest
changes, more than half are in Africa. Animism, as defined in the RCS-Dem
Dataset, was the majority religious denomination in 24 countries in 1950. By
2015, it was no longer the majority in any of these countries. Of the 94
countries that reported being Christian majority in 1950, 77 saw a decrease
in the share of the majority religious denomination. In the same period, 25
out of the 38 countries that reported being Muslim majority, saw an increase
in the share of the majority religious denomination. 33 out of 35 OECD
countries included in our analysis have witnessed a decline in the share of
the majority religious denomination, of which 30 have seen a significant
decline.
In keeping with the global trends of declining majority, India too has
witnessed a reduction in the share of the majority religious denomination by
7.81 percent. This is particularly remarkable given the wider context within
the South Asian neighborhood where the share of the majority religious
denomination has increased and minority populations have shrunk
alarmingly across countries like Bangladesh, Pakistan, Sri Lanka, Bhutan and
Afghanistan. It is not surprising, therefore, that minority populations from
across the neighborhood come to India during times of duress.
On the Indian subcontinent5, all the Muslim majority countries
witnessed an increase in the share of the majority religious denomination
except Maldives where the share of the majority group (Shafi’i Sunnis)
declined by 1.47 percent. In Bangladesh, there was an 18 percent increase in
the share of the majority religious group which is the largest such increase in
the Indian subcontinent. Pakistan witnessed an increase of 3.75 percent in
the share of the majority religious denomination (Hanafi Muslim) and a 10
percent increase in the share of total Muslim population despite the creation
of Bangladesh in 1971.
Among non-Muslim majority countries, Myanmar, India and Nepal saw
a decline in the share of the majority religious denomination. Myanmar
witnessed the steepest decline of the majority religious group in the region
with the share of Theravada Buddhist population declining by 10 percent in
5 SAARC plus Myanmar
Page | 3
the period under study. Of the three major religions in Nepal, the share of
the majority Hindu population declined by 4 percent, the share of Buddhist
population declined by 3 percent while the Muslim population increased by
2 percent.
Only Sri Lanka and Bhutan among non-Muslim majority countries have
witnessed an increase in the share of the majority religious denomination
between 1950 and 2015. In Sri Lanka, the share of the majority Theravada
Buddhist population increased by 5 percent while the share of Hindu
population (the next largest religious group) declined by 5 percent. In
Bhutan, the majority Tibetan Buddhist population increased by almost 18
percent while the Hindu population declined from 23 percent to 11 percent
in the same period.
In democracies, the change in political outcomes triggered by
demographic changes is particularly acute on two accounts. Firstly, they
redraw the contours of elections because demographic changes affect the
provisioning of public goods. Politics, especially in a democracy, is not an
exogenous variable but quite often an attempt to leverage these changes by
projecting a solution to the insecurities generated by changing demographics
or vouching to become a representative of the new demography.
By way of illustration, consider the increasing salience of demographic
changes in the politics and policies of Europe. In the 35 OECD countries
which were part of our study (25 of whom are in Europe), the share of the
majority religious denomination has decreased by 29 percent between 1950
and 2015. This is notably higher than the global average of a 22 percent
reduction in majority share. Such drastic societal changes over three
generations will inevitably lead to a political response, as we have witnessed
in the last cycle of elections across Europe.
Secondly, rapidly-changing religious demographics across countries is
also fuelling vigorous political debates on the treatment of minorities and
leading to emergence of shrill identity politics. Democracies run on the
principle of majority and therefore need legal and institutional protection for
minorities. The healthy tension between rule of numbers and rule of law
creates the equilibrium for a democracy to survive and thrive in a society.
Since majority-led coalitions get preference in governance, there will be
something akin to a “market failure” if democracies depend on elections for
protection of minorities. Therefore, democracies rely on a variety of tools for
the protection of minorities such as legal definitions for who constitutes a
minority, laws to secure their way of life, legal quotas and well-defined
Page | 4
immigration policies. Surprisingly, few countries have even defined
“minority” within their population and India happens to be one of them.
The modern architecture of minority protection within national and
international law was built post World War II and is now over seven
decades old. It is an opportune moment to take stock of the status of
minorities across countries.
The year 1950 is important as a baseline year for two major
reasons. Firstly, this was around the time that the international human
rights framework under the aegis of the newly created United Nations
began to take shape with minority rights and state responsibility for
protection of minorities being mainstreamed in international law.
Secondly, the seventy years period after World War II is considered a
golden epoch in human history with the number of democratic countries
increasing from 28 percent in 1950 to 63 percent in 20006.
Finally, as an ancillary point, the epoch starting with the 1950s is
also the period from which census data collection became more
scientific, timely and of high quality7. For example, India conducted its
first census post- independence in 1951 after setting up specialized and
permanent institutions like the Registrar General and Census
Commissioner, instead of the ad-hoc arrangements that used to be
undertaken for earlier exercises8.
The change in religious composition of a population is a complex
phenomenon because several factors contribute to it, such as migration,
conversion, differences in fertility rates and variation in territorial
boundaries as a result of political processes. We abstract away from the
causes of this change and instead focus on the share of the minority
population as a cumulative outcome measure of their well-being.
Our hypothesis is that the change in the proportion of the minority
population as a share of total population is a good proxy for the status of
minorities in a country over time. A society which provides a congenial
environment for the flourishing of minorities is more likely to witness an
increase or stabilization in their numbers over a period of three
generations. Conversely, a society which creates a hostile environment
for minorities and/or denies them access to public goods and services is
more likely to witness a decrease in their share of the overall total
population. The nature and extent of change in the share of the
minority population in a country is therefore a strong indicator of their
relative wellbeing within a country. This simple yet practical approach of
looking at the relative demographic trend of religious minorities, is a
Page | 5
good place to start an objective conversation on the track record of
different countries in creating a nurturing environment to ensure the
success of their minorities.9
There are several examples of the use of demographic information
to make conclusions about the cumulative impact of political and social
institutions on the target population. For example, there is a long and
well- established track record of the use of sex ratio as a sensitive
indicator of the status of women in society in development economics.
As a 2014 UNFPA Handbook titled Guide on the Gender Analysis of
Census Data notes, “Analyses of census data may uncover sex ratio
imbalances or unconventional population structures that are
symptomatic of growing inequalities in a country, region or
municipality.”10 In his landmark study, Amartya Sen estimated that 100
million women were “missing” and sex ratio is the important metric used
to estimate this number11. There is a rich and vast literature on this
phenomenon that has evolved through careful scrutiny of sex ratios over
time and geography. Unlike the gender studies which fundamentally
depend on the variations and deviations of sex ratio from the “natural
sex ratio” at birth to measure forms of discrimination, our analysis of the
share of minority population cannot rely on a “natural” measure. We,
therefore, scrutinize longitudinal trends within each country from 1950
onwards. We argue that the rate of change in the share of the minority
population is a measure of the cumulative impact of supportive policies
(or lack thereof) that affect these populations within a country.
Big decline in the relative proportion of the minority population
could indicate sustained relative discrimination against these groups.
Conversely, a big increase in the relative share of the minority population
within a country indicates an overall supportive environment for
minorities. The empirical results in this study make the limited but critical
point that religious demographics have experienced substantial changes
in most societies around the world between 1950-2015 and these
changes are an important indicator of the overall policy environment
affecting minorities within the countries.
I. RCS-DEM DATASET
The increasing salience of religion in helping individuals navigate
the modern world and its consequent impact on social interactions and
political processes has received limited and belated attention from
Page | 6
academics for a variety of reasons beyond the scope of our discussion in
this paper. However, one major reason that academic study on the
relationship between religion and outcomes within and between states
has been relatively neglected is the lack of credible, granular and timely
datasets on religious demographics for countries around the world.
Getting data on religious demographics remains a sensitive issue in
many countries. Only about one-third of the countries include questions
on religious beliefs in their national census and in several countries like
UK the question is voluntary12. In the USA, the national census (which is
mandatory for citizens) stopped asking questions about religion 1980
onwards after an injunction by the US Congress in 1976.13
This working paper uses demographic data from Religious
Characteristics of States Dataset Project - Demographics v. 2.0 (RCS-Dem)
published by Association of Religion Data Archives (ARDA) in the year
2019. RCS-Dem has more extensive and consistent data points on
religious demographics than other available sources. Notably, it records
population numbers for 97 religious denominations from the early 1800s
onwards (approximately 6.7 million data points).14
RCS-Dem has been prepared by two principal investigators - Dr.
Davis Brown (Nonresident Fellow, Baylor University Institute for Studies
of Religion) and Dr. Patrick James (Dornsife Dean's Professor of
International Relations, University of Southern California). We rely on
country-wise estimates of religious demographics for the years 1950 and
2015. The change in religious composition of the population between
these two years forms the basis of our analysis.
There are several other datasets available on religion. Prominent among
them include the following: World Religion Database (WRD), World
Religion Project (WRP), Pew Research Center's Religious Landscape
Studies (specifically for USA), Global Religious Futures Project (Run by
Pew Research project) and Atlas of Religion. Of the above-mentioned
datasets, WRD and WRP are the two leading databases on religion. They
were important sources in construction of the RCS-Dem Database,
especially WRD.
However, RCS-Dem has four distinct advantages over the WRD and
WRP to create a dataset that has unmatched richness in denominational
and temporal coverage. These advantages are explained in a 2018 paper
by the two creators of the dataset titled The Religious Characteristics of
Page | 7
States: Classic Themes and New Evidence for International Relations and
Comparative Politics15 and are briefly summarized below.
Firstly, RCS-Dem provides data points on an yearly basis as
compared to WRP which provided data only in five-year installments and
WRD which provides data for only seven years16 between 1900 and 2010.
Secondly, neither the WRD nor the WRP capture every variable for every
state. As a result, WRD’s observations can populate only 37.7 percent of
the Dataset covering all religious denominations for all countries and
WRP incorrectly codes missing data as “zero” leading to impractical
conclusions. Thirdly, in terms of historical depth WRD provides estimates
for the states in the year 1900, but only for states that exist today with
their present borders. Whereas, RCS-Dem using historical sources
provides data that is older than 1900s for more than 130 countries that
are or were independent. In certain cases, RCS provides data for
countries that don’t even exist today. Lastly, in terms of denominational
breadth WRD provides data for 44 religious denominations and WRP
provides for 31 religious denominations; whereas RCS-Dem provides data
for 97 religious denominations. More-over, RCS-Dem details a taxonomy
chart of the religious denominations also it introduces optional
composite categories for user’s benefit.
While constructing the dataset, the authors took the practical approach
of strictly defining religious denomination and simultaneously reported
on the non-religious segment to account for degree of secularization in a
country. They do not collect information on the intensity of religious
belief or the degree of participation of the adherents of a religious
denomination in social activities associated with religion (for example,
visiting places of worship on designated days). Twenty four sources
(primary and secondary) were used to construct this dataset. Source
citation is given at the granular level of denominations. The authors refer
to multiple resources and use subjective yet logical methods while
choosing the sources and to arrive at the estimates. Various
extrapolations and interpolations techniques have been used by the
authors to fill the gaps in the datasets.
Several academic studies on religion, status of religious minorities
and minority rights have relied on data from RCS-Dem.17 Among the
many, we highlight three research papers that have used the database to
prove important results from a policy perspective. Nilay Saiya (Nanyang
Technological University (NTU), Singapore) & Stuti Manchanda (NTU,
Singapore) in their paper (published in Journal of European Public Policy)
Page | 8
titled Do Burqa Bans Make Us Safer? Veil Prohibitions and Terrorism in
Europe show how even though the intended effect of burqa bans of
incorporating Muslim women into French culture and increasing public
safety; end up achieving the exact opposite by further isolating Muslim
women from the public, increasing stigmatization of Muslims and
increasing radicalization of Muslims. Seung- Whan Choi (Department of
Political Science at the University of Illinois at Chicago) in his paper
(published in Perspectives on Terrorism) titled Does Restrictive
Immigration Policy Reduce Terrorism in Western Democracies? talks
about how a total ban on immigration might not be the most optimal
way to address the problem of terrorism in western democracies.
Matthew Isaacs (Brandeis University, Waltham, MA, USA) in his paper
(published in Comparative Political Studies) titled Faith in Contention:
Explaining the Salience of Religion in Ethnic Conflict argues that
competition among religious leaders acts like a precursor for most ethnic
conflicts and determines the intensity and duration of that conflict.
