CONCEPTUAL CHALLENGES - DEFINITION OF
“RELIGIOUS MINORITY”
There is no internationally accepted definition of a “minority” or “religious
minority”. This makes cross-country comparisons difficult. As John Packer
(first Director of the Office of the OSCE High Commissioner on National
Minorities, later Professor of International Conflict Resolution & Director,
Human Rights Research & Education Centre, University of Ottawa) notes18:
“It has been correctly observed that international law supposes
the existence of minorities both in general and specific types.
However, while the existence of human beings and states are
‘axiomatic’ in international law, the existence of human groups is
problematic. Conceptually, international law struggles with the
definitions of actors beyond the ‘State’; indeed the problem of
defining actors has always troubled political theory in general and
international relations in particular…[While] the catalogue and
content of individual human rights has become relatively clear,
the specificity of protections for groups, particularly minorities,
has remained largely uncertain. Paramount among this
uncertainty has been the very definition of ‘the’ or ‘a minority’ to
whom any rights may accrue.”
For example, Article 18 of the Universal Declaration of Human Rights
(UDHR) provides for religious rights. It was further elucidated in International
Covenant on Civil and Political Rights 1966 (ICCPR) and the 1981 UN General
Assembly Declaration on the Elimination of All Forms of Intolerance based on
Religion or Belief (1981 Declaration). None of these instruments however
provide a definition of the term “religion”.
Similarly, international law instruments and scholars have been unable
to reach a consensus on a comprehensive working definition of the term
“minority”. UDHR, ICCPR, 1981 Declaration, 1992 UN Resolution and
Declaration on the Rights of Persons Belonging to National or Ethnic,
Religious and Linguistic Minorities, Council of Europe’s 1995 Framework
Convention for the Protection of National Minorities - none of these
instruments define the term “minority”.
Therefore, there is no internationally accepted definition of the term
“religious minority” and any attempts to define the term creates categorical
disagreements
which
brings
research
on
inter-country
comparisons
and
18 John Packer, On the Definition of Minorities, in The Protection of Ethnic and Linguistic Minorities in Europe 23, 24-27
(John Packer & Kristian Myntti eds., 1993)
objective conversation to a standstill. Instead, the task of recognising and
defining which religious community is considered as a religious minority is
left to the individual countries.
However, the existence of religious minorities is a question of fact, not
law. Therefore, we have to make efforts to work around these conceptual
challenges. In the absence of a uniformly accepted definition of religious
minority, we have to reconfigure our analysis through the study of the
change of the majority population in a country. Therefore, we restrict our
attention to the study of changes in the proportion of the majority in the
population, and based on that analysis we make inferences about
demographic changes of religious minorities.
I. METHODOLOGY
This study captures the distribution of the rate of change in share of
the majority religion for 167 countries from 1950 to 2015. We decided to
focus on only those countries that had a majority religion (more than 50
percent share of the total population) for 1950. Since we want to study the
demographic changes among the religious minority over the last 65 years;
we decided to restrict our analysis to the definition of majority (and by
extension minority) as defined in 1950. Countries for which the dataset did
not label a majority religious denomination in 1950 were excluded from our
analysis.
For the majority religious denomination as defined in 1950, we first
note its share of the total population in 2015. Next, we look at the rate of
change in share of population for the religious majority (as defined in 1950)
from 1950 to 2015. This is the key metric based on which we do our main
analysis and minor refinements. In other words, we study the distribution of
the rate of change in the share of the majority religion in countries around
the world by focusing on the following measure:
● Rate of Change in Share = [(Share of majority (as defined in 1950) in
2015 – Share of majority in 1950) / Share of majority in 1950] * 100
We look at percentage change in the share of the majority population
from 1950 to 2015 and make assessments about the quality of minority
protection. To clarify, we are not going into the composition of the minority
population or the mechanics of demographic change i.e. the factors leading
to either reduction or increase in share of majority religious denomination.
Our main hypothesis is that a reducing majority i.e. a negative rate
of change in the share of majority population (and tautologically an
increase in the share of the minority population) from 1950-2015 means
that we can safely conclude that there exists an overall congenial
environment for increasing diversity in the country.
Finally, it is important to note that the “Rate of Change in Share'' is a
measure of the change in majority share, which is a number greater than 50
percent in the baseline year of 1950 (as a definition). In the 167 countries
selected for analysis, the average value of the share of the majority religious
denomination calculated in the baseline year of 1950 is 75.08 percent while
the possible range is 50-100 percent by definition. The range for share of the
majority religious denomination in the endline year of 2015 however is 0 to
100 percent. This is why the rate of change in share of the majority religious
denomination is also reported to be negative for several countries.
Composite Categories: The RCS-Dem Dataset has introduced
composite categories, such as Western Christian, Eastern Christian, East
Asian Complex and Buddhist Complex. The purpose of this categorisation is
to group those denominations which have some fundamental commonalities
whose numbers may be cumulatively added to touch the majority threshold
of 50 percent. These composite categories include the following19:
●Western Christians: Combined Catholic (less non-Latin-rite Catholic),
Extended Protestant and Liminal Christian (less Other Liminal
Christian), less double-affiliates
●East Asian Complex: Combined Shintoist, Confucianist, Taoist, Chinese
Folk, less double affiliates.
●Buddhist Complex: Combined Buddhist, Buddhist Syncretic, Shintoist,
Confucianist, Taoist; not including Chinese Folk Religionist
●Liminal Categories: This term designates groups of denominations of
Christianity and Islam that identify as Christian and Muslim,
respectively, but which depart significantly from the theology or canon
of the other major branches.
A. Process of Assigning Majority ID in the RCS-Dem Dataset
In the RCS-Dem Dataset, the smallest factor greater than 50 percent is
designated as the majority religious denomination in the dataset and
labelled as Majority ID. The Majority ID could be assigned to three categories
of denominations (see Table 2 in the Appendix):
1. A subgroup of a religion which in itself crosses the 50 percent
threshold. For example, in the Czech Republic (1950), Latin-Rite Roman
Catholics constituted 76.35 percent of the total population while the
Christian population was 93.75 percent. Since the subgroup “Latin-Rite
Roman Catholic” was big enough to cross the 50 percent threshold, it
was assigned the Majority ID.
2. A religion (including all its subgroups) crosses the 50 percent threshold.
For example, in Estonia (1950), “Christianity” was labeled as Majority
19 As defined in the Codebook of RCS-Dem 2.0, please see: https://www.thearda.com/ARDA/pdf/originalCodebooks/RCS
%20Demographics%20v2.0%20Codebook.pdf
ID20. This is because its subgroups - Extended Protestants (31.8
percent) and Orthodox Christians (25.43 percent) were not enough to
touch the majority threshold of 50 percent alone. But the next highest
category “Christianity” - which includes these two subgroups and
other denominations which are part of the same religious family -
totalled to approximately 57 percent.
3. A composite category invented by the creators of the RCS-Dem Dataset
crosses the 50 percent threshold. By way of illustration, for Palau in
1950 the dataset assigns Majority ID to a composite category called
“Western Christians”. This composite category includes Catholics,
Extended Protestants and Liminal Christian branches which
individually constituted 20.23 percent, 35.64 percent and zero percent
of the population of Palau in 1950. Their combined numbers were
required to cross the majority threshold of 50 percent.
B. Note on the Diversity Being Measured in This Study
Minority is a composite category and within this label there are
different rates of changes of various denominations. However, our paper is
not focussed on how the composition of the minority group is changing
within a country. Our intent is simply to make an assessment on whether an
environment fosters/conducive to diversity - we are not getting into the
composition of that diversity.
Therefore, in this study a change in Majority ID from one religion to
another religion is treated equivalent to a change in Majority ID from one
denomination within the same religion to another. Let us illustrate with the
case studies of Lebanon and Saudi Arabia both of which are labeled as
countries where the share of the majority population reduced between 1950
and 2015.
Lebanon was a Christian majority country in 1950 with “Eastern
Christian” as the Majority ID, constituting 51.65 percent of the population. In
2015, Lebanon became a Muslim majority country with Majority ID being
“Muslim” with approximately 61 percent of the population. The Eastern
Christian population meanwhile came down to 31 percent by 2015.
Next, consider Saudi Arabia’s case which is also labeled as a country
where the majority religious denomination has decreased its share because
20 Unspecified Christians (zero percent), Catholics (0.12 percent), Extended Protestants (31.87 percent), Orthodox
(25.43 percent), Other Liminal Categories (0 percent). Liminal Categories (0.01 percent)
the 1950 majority denomination (Hanbali sect with 58 percent of the
population) was reduced to 35 percent in 2015. Therefore, the Majority ID
shifted to the bigger denomination of “Sunni Muslims” which constituted 82
percent of the population in 2015.
This is even as Saudi Arabia remains a Muslim-majority country with
muslims constituting 99.8 percent of the population in 1950 and 92 percent
of the population in 2015. Thus, for the purposes of this study, even a shift in
Majority ID from Hanafi to Sunni Muslims is a measure of diversity in a
Muslim- majority country because our metric “rate of change of majority
denomination” is indifferent to change in either denomination or religion.
C. Note on Legal Entity as the Unit of Analysis
It is also important to note that the number of countries has increased
from around 102 to 195 in the period under study. The unit of our analysis is
not a fixed geographical area but a legal entity recognized as a country in
2015 in the database, without regard to its political and administrative
boundaries. If there is corresponding data available for this entity in 1950
(irrespective of whether it was recognized as a separate country or had
different borders) it is included in our study.
Sometimes it is the change in borders and not direct government
policy that explains demographic shifts. Increase in the share of minority
population by addition of territories to a country is different from increase in
minority population due to the government's protection of minorities by law
and policy. However, the change in borders of a country is never random but
always driven by political factors. Countries go through boundary changes for
a variety of reasons and we look at the cumulative impact of these changes.
Even the authors of the RCS-Dem Dataset understood the challenge of
making like-to-like demographic comparisons given the turbulent global
politics and constantly changing borders or countries. Accordingly, the RCS-
Dem Database also keeps ”State” as the unit of its analysis instead of a fixed
geographical area. The creators of RCS-Dem Dataset report demographic
data for a given country “using its contemporaneous, not contemporary,
borders…as they actually existed in the observed year.”21
21 Brown, D., & James, P. (2018). The Religious Characteristics of States: Classic Themes and New Evidence for
International Relations and Comparative Politics. Journal of Conflict Resolution, 62(6), 1340-1376. See:
https://doi.org/10.1177/0022002717729882
For example, Bangladesh did not exist as a separate country in 1950
but the corresponding demographic data for East Pakistan is available in the
RCS- Dem Database and therefore Bangladesh forms a part of the study.
Similarly, the boundary of India in 2015 is markedly different from 1950 and
that accordingly reflects in the demographic data. However, the database
does not have corresponding historical data for the present-day borders of
the Federal Republic of Germany which did not exist as a separate country in
1950 but was formed in October 1990 after the unification of East Germany
and West Germany. Therefore, the Federal Republic of Germany is excluded
from our analysis.
II. ANALYSIS AND SUMMARY
We begin our analysis by listing in Table 1 below all the religious
denominations/groups in the RCS-Dem Database that had a majority in at
least one country in the world in 1950 and 2015. Please note that the RCS-
Dem Dataset defines a “majority” as the narrowest religious category that
has more than 50 percent share of the population. There are a total of 37
religious denominations which have a majority in at least one country. This
gives the reader a flavor of the diversity and granularity (and consequently
the challenge) of inter-country comparisons on religious groups. Please note
that we have classified the religious denomination according to major
religion based on the tree structure for various religions as given in the RCS-
Dem Codebook22.
Table 1: All Religious Denominations with Majority in at least One Country
S.No Religion Denomination Major Religion S.No Religion Denomination Major Religion
1 Christian Christian 19 Muslim Muslim
2 Western Christian Christian 20 Sunni Muslims Muslim
4Latin-Rite Roman
Catholics Christian 21 Hanafi Muslim
5 Protestant Extended Christian 22 Maliki Muslim
6 Protestants and Anglican Christian 23 Shafii Sunnis Muslim
7 Protestant Christian 24 Hanbali Muslim
8 Lutherans Christian 25 Shia Muslim
9unspecified Reformed
Protestants Christian 26 Zaydis Muslim
10 Presbyterial Reformed
Protestants Christian 27 Khariji Muslims Muslim
11 Congregational Reformed
Protestants Christian 28 Hindus Hindus
12 Adventist Christian 29 Buddhists Buddhist
13 Wesleyans Christian 30 Theravada Buddhists Buddhist
14 Anglicans Christian 31 Mahayana Buddhists Buddhist
15 Oriental Orthodox Christian 32 Tibetan Buddhists Buddhist
16 Eastern Orthodox Christian 33 East Asian Religious
Complex
East Asian
Religion
17 Christian Syncretics Christian 34 Chinese Folk Religionists Chinese Folk
18 Jews Jews 35 Animists Animists
36 Not Religious Not Religious
37 Unspecified Not Religious Not Religious
22 Supra note 15
As we see from the table, there are sixteen different subcategories of
Christianity that are a majority in at least one country. Similarly, there are
eight subcategories of Islam and three denominations of Buddhism that
constitute the majority population in at least one country.
A. Countries Excluded from Analysis
In our analysis, we exclude countries that did not have a majority
religious denomination in the baseline year of 1950 and countries where the
data was simply missing. There are 22 such countries in the database. These
countries are listed in Table 3 of the Appendix. In the table, we have
captured the status of their majority religion in the endline year of 2015 for
the sake of completeness.
We discuss the countries excluded from analysis in this section.
Depending on the reason for exclusion, these 22 countries may be divided
into three distinct categories:
1. Countries where no religious denomination crossed the majority
threshold in 1950 (baseline year) and 2015 (endline year)
2. Countries with missing data for 1950 (baseline year)
3. Countries with no majority in 1950 (baseline year) but a majority
religious denomination in 2015 (endline year)
1. Countries Without a Religious Majority in both 1950 and 2015
There are some countries like Mauritius, Singapore and Nigeria where
no religion has been able to cross the 50 percent majority threshold in 1950
(baseline year) and 2015 (endline year). However, this does not mean that
they have not experienced changes in the religious composition of the
population.
Mauritius
Mauritius, for example, has seen a steady Hindu population between
47-49 percent whereas there has been a slight reduction in Christian
population from 1950 (35.82 percent to 30.19 percent) and a corresponding
increase in Muslim population from about 15 percent to 18 percent.
Singapore
In Singapore, the three largest religious categories in 1950 were East
Asian Religious Complex (36.99 percent), Buddhist Complex (33.89 percent)
and Muslims (18.59 percent). Whereas in 2015, Christianity rose to be the
second largest religion with the top three categories being Buddhist Complex
(30.92 percent), Christians (20.20 percent) and Muslims (14.64 percent).
Nigeria
Nigeria is another country where no religion crosses the majority
threshold in the period under study but the country has witnessed a
stunning rise of Christianity from 11.54 percent in 1950 to becoming the
largest religious category at 48.85 percent. Numbers suggest that this has
happened largely at the cost of Animism whose adherents have reduced
from 43.72 percent in 1950 to 7.4 percent in 2015.
2. Countries With Missing Data
The next category of countries excluded are those for which the RCS-
Dem Dataset reports missing data in 1950 (baseline year) for a significant
proportion of the population after conducting robustness and due diligence
checks on the underlying data sources. As a result, no religious group was
able to touch the majority threshold of 50 percent in 1950 (baseline year).
However, we can make rudimentary attempts to solve the puzzle of
“Missing” entry by observing other analogous sources like the World Religion
Project (WRP). We share the data points from the WRP database for the
following three countries to make the case that the social reality in these
countries may be more complicated and nuanced than the “Missing” label
suggests.
South Korea
In South Korea, the Majority ID in 1950 was assigned as “Missing”.
However, the dataset suggests that the demographic distribution was
Buddhist Complex (27.88 percent), East Asian Religious Complex (12.56
percent) and Unknown (66.34 percent). In 2015 the majority ID was again
assigned as “Missing” since the breakup was as follows: Unknown (40.37
percent), Christian (34.72 percent) and Buddhist Complex (22.73 percent).
Despite the missing data for a significant share of the population, it is
noteworthy that the share of Christianity in the population increased from
5.61 percent in 1950 to 34.72 percent in 2015. Whereas the share of East Asian
Religious Complex in the population decreased from 12.56 percent in 1950 to
0.24 percent in 2015.
If we refer to the closest analogous source - the World Religion Project
(WRP) - they label 63.86 percent of the population as adherents of Animism
in 1950 which decreases to 1.049 percent in 2010; whereas Christianity
increases from 8.99 percent in 1950 to 28.59 percent in 2010 in their
database.
Vietnam
Similarly, in Vietnam, the Majority ID in 1950 was assigned as
“Missing” because the demographic distribution was as follows: Unknown
(39.80 percent), Buddhist Complex (33.92 percent) and Indigenous
Religionist (10.71 percent). In 2015, the Majority ID was again assigned as
“Missing” because the demographic distribution was Unknown (51.10
percent), Not Religious (17.98 percent) and Buddhist Complex (10.75
percent).
On the other hand, the World Religion Project (WRP) makes very
different assessments about the religious demographics of Vietnam. In its
dataset, the percentage of Buddhists increased from 10.99 percent in 1955
to 49 percent in 2010 and the “Not Religious” percentage decreased from 70
percent in 1955 to 36.4 percent in 2010. This may be explained by the under-
reporting of religious activity and affiliation during the Cold War.
USSR/Russian Federation
In 1950, USSR had the Majority ID as “Missing” and the demographic
distribution was as follows: Christians (40.25 percent), Not religious (40.11
percent) and Muslims (10.90 percent). In 2015, the Russian Federation which
was the successor state of USSR also had the Majority ID as “Missing”
because the demographic distribution was Christians (48.20 percent),
Unknown (30.97 percent) and Muslims (10.61 percent). The share of “Not
Religious” in the population decreased from 40.11 percent from 1950 to 8.20
percent in 2015. Whereas the share of “Unknown” increased from 6.83
percent in 1950 to
30.97 percent in 2015.
The World Religion Project on the other hand reports that the
percentage of Christians in USSR/Russian Federation increased from 34.11
percent (1950) to 74.23 percent (2010), thereby crossing the majority
threshold. Further, the percentage of “Not Religious” decreased from 48.13
percent (1950) to 14.04 percent (2010).
Both Vietnam and USSR were communist countries and the dataset
hides a more complex social reality. For example, in both these countries one
notes a drastic decrease in the percentage of population labeled as “Not
Religious” after the end of the Cold War. This may suggest that there was an
under-estimation and under-reporting of religious institutions and beliefs in
these societies during the heyday of Communism.
On the contrary, the change in Majority ID in China from “Chinese Folk
Religionists” in 1950 to “None” in 2015 may be explained by the Chinese
Communist Party’s official policy of atheism and severe restrictions which
discourage citizens from religious practices and academics from studying
religion in contemporary China. As a result, large sections of the population
may be reluctant to formally identify themselves with a religion and credible
data may not be available due to restrictions imposed by the Chinese
government on survey research23.
Malaysia
Malaysia, which got its independence in 1957 from British rule is
assigned a Majority ID of “Missing” for the year 1950. The RCS Dataset notes
its demographic distribution as follows: Muslims (45.20 percent), Unknown
(43.08 percent) and Hindu (8.70 percent). However, unlike South Korea,
Vietnam and Russia, the Majority ID in 2015 was assigned to “Sunni
Muslims” who accounted for approximately 56 percent of the population.
The next two largest categories were Christians (9.29 percent) and Buddhist
Complex (8.69 percent). The Hindu population declined to about 6.27
percent in 2015 from
8.7 percent in 1950.
3. Countries Without Religious Majority in 1950 but With a Religious
Majority in 2015
The third category of countries that were excluded from our analysis
are those which did not have a Majority ID in 1950 (baseline year) but by
2015 (endline year) one religious denomination increased its share to attain
the Majority ID. In the three cases discussed below - Armenia, Kazakhstan
and Mongolia (all of whom were under the control of the USSR during the
Cold War) - this may be explained to a large extent by the widespread under-
reporting of religious beliefs and state-supported suppression of religion in
the Soviet sphere of influence.
Armenia
In 1950, the Majority ID was assigned as “None” since the demographic
distribution was as follows: Christians (45.08 percent), Not Religious (44.33
23 Measuring Religion in China (August 2023), Pew Research Center, See:
https://www.pewresearch.org/religion/2023/08/30/measuring-religion-in-china/
percent) and Muslims (9.57 percent). In 2015, Christians constituted 95.69
percent of the population of which an overwhelming proportion was that of
the Oriental Orthodox denomination (93.21 percent). On the other hand, the
population defined as “Not Religious” was 3.23 percent. Therefore, we can
observe a 108 percent rise in the Christian population between 1950 and
2015. On the other hand, the share of “Not Religious” decreased from 44.33
percent in 1950 to 3.23 percent in 2015.
Mongolia
Mongolia has historically been a Buddhist country. However, officially
it saw a dramatic rise in its Buddhist population as per the RCS-Dem Dataset
between 1950 and 2015. One reason could be the under-reporting and state
suppression of religious beliefs during the Cold War when it was a one-party
socialist state supported by the USSR. In 1950, the Majority ID was assigned
as ”None” because the demographic distribution was reported as follows:
Not Religious (44.86 percent), Unknown (38.43 percent) and Buddhist
Complex (7.71 percent). In 2015, the Majority ID was assigned as “Tibetan
Buddhist” with 54.27 percent share of the population. The distribution of
“Not Religious” (18.61 percent) and “Unknown” (18.43 percent) fell
substantially from 1950. The rise in the Buddhist population from about 7
percent in 1950 to 54 percent in 2015 and corresponding decrease in the
share of “Not Religious” from 44.86 percent in 1950 to 18.61 percent in
2015 and the share of “Unknown” from
38.43 percent in 1950 to 18.43 percent in 2015 provides circumstantial
evidence for our intuition about the under-reporting of religious belief
during one-party socialist rule.
Kazakhstan
Kazakhstan presents a story similar to Mongolia about the suppression
of religion and religious beliefs under Soviet rule. Even though it has been an
Islamic country for several centuries, in 1950, the Majority ID was assigned
as “None”. This is because as per the RCS-Dem dataset the demographic
distribution in Kazakhstan in 1950 was: Not Religious (43.23 percent),
Muslims (39.95 percent) and Christians (15.83 percent). However, in 2015,
Hanafi Sunnis constituted 75.77 percent of the population and were
accordingly assigned the Majority ID whereas the “Not Religious” population
unsurprisingly fell from 43.23 percent to 2.11 percent after the Soviet
influence disappeared. The Christian population meanwhile witnessed a
modest growth as it rose from 15 to 21 percent.
Eritrea
In 1950, the majority ID was assigned as None, the demographic
distribution was Muslims (49.88 percent), Indigenous Religionist (26.90
percent) and Christians (22.34 percent). In 2015, the majority ID was
assigned as “Shafii Sunnis” since they constituted 50.42 percent of the
population by themselves. The share of Christians increased from 22.34
percent in 1950 to
46.29 percent in 2015. Whereas the share of Indigenous Religionists
decreased from 26.90 percent to 0.62 percent in 2015.
Guyana
Guyana is another country which had no majority religion in 1950 as per
the RCS-Dem Dataset. The demographic distribution was as follows:
Christians (49.50 percent), Hindu (31.99 percent) and Muslims (8.07
percent). By 2015, the Christian population increased its share to become
almost 65 percent of the population with “Western Christian” composite
category attaining the Majority ID with 50.4 percent share. The Hindu
population on the other hand fell from about 32 percent to 23 percent in
the period of 65 years.
B. Key Result – Changes in Minority Population for Selected Countries
The result of our main analysis is shown in Figure 1 which
represents the change in majority religious denomination between 1950
and 2015 for the 167 selected countries. In Figure 1, the countries coded
in green are those where the majority population has reduced and the
countries coded in red are those where the majority population has
increased in the period under study. The intensity of shading is indicative
of the quantum of such change. The distribution of this change is also
presented in Figure 2 through histogram and density plot. Table 4 in the
Appendix reports the exact data for the 167 countries under analysis.
