Appendix A: Methodology
Navigate this Page:
This study uses the standard demographic method of making population projections. Called the cohort-component method, it takes the age and sex structure of a population into account when projecting the population forward in time. This has the advantage of recognizing that an initial, baseline population can be relatively “young,” with a high proportion of people in younger age groups (such as Nigeria) or relatively “old,” with a high proportion of older people (such as Japan). Cohorts are groups defined by age and sex, such as females ages 15-19 and males ages 15-19. Components are the three ways in which populations grow or shrink: new entrants via births, exits via deaths and net changes from migration. Each cohort of the population is projected into the future by adding likely gains – births and people moving into the country (immigrants) – and subtracting likely losses – deaths and people moving out (emigrants) – year-by-year. The very youngest cohorts, those ages 0-4, are created by applying age-specific birth rates to each female cohort in the childbearing years (ages 15-49).
The cohort-component method has been in existence for more a century. First suggested by the English economist Edwin Cannan in 1895, then further improved by demographers in the 1930s and ’40s, it has been widely adopted since World War II. It is used by the United Nations Population Division, the U.S. Census Bureau, other national statistical offices and numerous academic and research institutions.58
For countries in which Muslim populations are large enough that fertility rates and other demographic data are available specifically for the Muslim portion of the population, this study’s projections were made using the Demographic (DemProj) module of the Spectrum Policy Modeling System developed for the U.S. Agency for International Development (USAID), which is designed to integrate with data from the United Nations. However, projections for the United States and 25 European nations with significant Muslim populations were made by the Age and Cohort Change project of the International Institute for Applied Systems Analysis (IIASA) following the same basic methodology but using IIASA’s own software. IIASA, an independent research center supported by more than 15 countries and based in Laxenburg, Austria, is a recognized leader in population projections and is collaborating with the Pew Forum on demographic analysis of major religious groups around the world. 59
For many countries with very small Muslim populations, data on the differences in fertility, mortality and migration rates between Muslims and non-Muslims are not available. In such cases, the percentage that Muslims comprise of the baseline (2010) population is carried forward to future years and applied to the country’s expected total population, which is projected using the cohort-component method. This assumes that the Muslim population in such countries is growing at the same rate as the country’s overall population. Additionally, for a few countries, this report uses cohort-component projections made by demographic consultants for this project or independent estimates by national statistical agencies.
The projections of the Muslim population for 2010, 2020 and 2030 are based on assumptions about patterns in births, deaths, migration and age structures – the main factors driving population change – which are detailed below. There may, however, be political, environmental or social events that affect fertility, mortality, migration and age structures but that are not captured in these projections.
This study also considered conversion to or from Islam. Because recent survey data do not indicate that conversion is having any clear impact on the size of Muslim populations, the report assumes that future conversions into Islam will roughly equal conversions away from Islam, either to other faiths or to no particular faith. (See discussion in the Conversion section.)
The Pew Forum’s population projections take into account several types of data. In some cases, statistics are available specifically for Muslim populations. In other cases, however, they are available only for the general population. Since many countries are conducting national censuses in 2010-11, more data is likely to emerge over the next few years, but a cut-off must be made at some point; this report is based on information available as of mid-2010.
Baseline Muslim Populations
To provide a current population baseline, the Pew Forum and its consultants used the best available sources to estimate the percentage of Muslims as a portion of each country’s population in 2010. In some cases, the best source is a census from several years ago, such as 2001 or 2005. However, this study does not simply carry the percentage of Muslims in 2001 or 2005 forward to 2010. Wherever sufficient data on the fertility and migration of Muslims populations in a particular country are available, they are used to project the earlier population figure to reach a 2010 estimate. This results in substantial differences, for some countries, between the 2010 Muslim population estimates contained in this report and the 2009 estimates published by the Pew Forum in its report Mapping the Global Muslim Population. (See the section below on “Differences Between this Report and the Pew Forum’s 2009 Report.”)
The standard measure of fertility in this report is the Total Fertility Rate (TFR), defined as the total number of children an average woman would have in her lifetime if fertility patterns did not change. More specifically, the TFR is calculated by adding up all the age-specific fertility rates for women in a particular country (or region) during a given year (or other time period). This study includes estimates of TFRs for Muslim women in all countries in which Muslims make up a substantial portion of the population. For countries in which fertility rates for various religious groups are not available, this study assumes that the TFR of Muslims is the same as the TFR in the general population. This applies to some countries in which Muslims are an overwhelming majority (constituting approximately 90% or more of the country’s population), such as Afghanistan and Morocco.
Fortunately, fertility rates specifically for Muslims (or for ethnic groups that are predominantly Muslim) are available in many countries where sizeable numbers of Muslims live as minorities, including India, Nigeria, Ethiopia, Russia, Tanzania and many countries in Europe and North America. In such cases, the study has used those differential rates for projection purposes.
In countries where Muslims are a minority, the Muslim TFR, in most instances, tends to be slightly higher than the rate among the general population. In addition, fertility rates of Muslims tend to be higher than average in countries where they are recent immigrants (though the Muslim TFRs are projected gradually to converge to the national level in those countries). Therefore, this study’s estimates of the Muslim population in countries that have small Muslim populations, and for which differential data are not available, are believed to be conservative.
Age and Sex Structure
In countries for which demographic data specifically on Muslims are available, the age and sex structure of Muslim populations are incorporated into the projection models. In countries for which differential age and sex data for Muslims are not available (usually countries in which Muslims make up a very small portion of the overall population), Muslims are assumed to have demographic characteristics similar to the country’s general population.
Life Expectancy at Birth
The study’s projections use United Nations assumptions about life expectancy gains. The study assumes that life expectancy at birth will improve somewhat for all populations by 2030 and that the greatest gains will be made in less-developed countries (where the majority of Muslims live). The study also assumes that in countries where Muslims live as a minority, life expectancy at birth for Muslims is similar to the life expectancy of the general population and is reflective of national standards of living. In addition, data on differences in life expectancy at birth among members of various religious groups within a country generally are not available.
Migration: Important Primarily in Europe and North America
Immigration is a key driver of Muslim population growth in Europe and North America. Muslim immigration to countries in more-developed parts of the world has risen steadily for decades as a result of evolving labor markets, changes in immigration laws, growing connections of immigrant families to communities abroad and increased globalization.
Projections for future immigration of Muslims to Europe were made in collaboration with IIASA, primarily using immigration data provided by Eurostat. 60 Muslim immigration has been an ongoing phenomenon for decades in a number of European countries. This study assumes that the annual flow of Muslim immigrants to European countries will decline in absolute numbers in the years ahead.
In North America, however, the number of new Muslim immigrants has risen steadily in recent years, even after accounting for a slight dip in immigration among Muslims and non-Muslims following the 9/11 terrorist attacks. As a result, the projections for the U.S. assume that overall annual immigration will increase slowly, by about 1% per year. Based on recent trends that show more Muslims are leaving Muslim-majority countries and immigrating to the U.S., this study assumes that Muslims will comprise a slightly larger share of all immigrants to the U.S. each year.
This study’s projections count only immigrants to the U.S. who receive permanent legal residency and do not include visiting family members, students or others who are in the U.S. temporarily or illegally. Therefore, the report’s projections on the number of Muslim immigrants can be considered conservative.61
Patterns of immigration to the U.S. have varied from year to year, including a spike in 1998- 2001, a sharp decline in 2002-2004 and a return to average increases in the last several years.62 Despite these ups and downs, there has been relatively steady growth in the last 70 years, contrasting with substantial long-term fluctuations that occurred in the 19th and early 20th centuries. While short-term fluctuations seem likely to continue to occur, the projections assume that those variations will be less important than the long-term trends.
The issue of illegal immigration to the U.S. has become highly contentious in recent years. However, there is no evidence that illegal immigration is a significant factor in Muslim population growth in the U.S., partly because the overwhelming majority of Muslim immigrants enter through airports rather than at land crossings, which are more difficult than airports to monitor.
This study made three alternative population projections – a high, medium and low estimate – for Muslims in each country. The country estimates were then added together to produce high, medium and low projections for the entire world. The three estimates are based on different assumptions about fertility, life expectancy at birth and migration of both Muslims and non- Muslims (where differential data are available). The main body of this report provides figures from the medium scenario for 2020 and 2030 because it is considered to be the most likely indicator of future population growth.
The assumptions about future changes in fertility used in this study largely follow those of the United Nations.
The medium fertility scenario assumes that the Muslim Total Fertility Rate in every country will eventually settle in a range of approximately 1.9 to 2.1 children per woman. Some countries are already in this range, and others are projected to reach this range by 2030-35. Accordingly, if a country has already dipped below 1.9, as Iran has, the medium projection assumes that it will eventually stabilize around 1.9.
The low fertility scenario assumes 0.5 fewer children per woman in 2020-25 and 2030-35 than the fertility rate in the medium scenario.
The high fertility scenario assumes 0.5 more children per woman in 2020-25 and 2030-35 than the fertility rate in the medium scenario.
Projections for countries with Muslim-specific fertility data, including the United States and many European nations, used similar fertility assumptions
Based on United Nations estimates, this study projects that life expectancy at birth will gradually increase in all countries. There is no high, medium or low assumption because each country, whatever its current economic condition, is assumed to be moving toward better living standards and, therefore, longer life expectancy at birth.
Based on United Nations estimates, this study assumes that in most countries the gap between the number of emigrants and the number of immigrants gradually will narrow. In Europe, for example, Muslim immigrant flows are projected to remain steady or decrease slightly. Based on historical patterns in U.S. immigration, however, the study assumes that the most likely scenario is a modest increase in the flow of Muslim immigrants to the United States over the next 20 years.
The Projected Global Muslim Population Scenarios
Unlike the projected number of Muslims, the proportion or percentage of Muslims comprising the world population is similar in all three growth scenarios. For example, the low scenario projects that Muslims will make up 26.29% of the world’s population by 2030. The medium scenario projects 26.36% and the high scenario projects 26.43%.
Some regions also show little difference between the high, medium and low scenarios. For example, the low scenario projects that Muslims will make up 7.9% of the European population in 2030. The medium scenario is 8.0% and the high estimate is 8.1%. Even if Russia – which has the largest number of Muslims of any country in Europe but whose landmass extends into Asia – is not included in the European estimates, the overall difference (in percentage terms) between the high and low scenarios is just 0.2 points.
While the three scenarios produce essentially the same projections at the global and regional level, there may be more noticeable differences at the country level, especially in countries with a large influx of Muslim immigrants, such as the U.S. The low scenario projects that Muslims will make up 1.54% of the U.S. population in 2030. The medium projection is 1.68% and the high projection is 1.89%. As in all cases throughout this study, the medium scenario is used as the best indicator of the future.
Definition of Muslims
This report seeks to provide the most up-to-date and comprehensive demographic estimates of the number of Muslims in the 232 countries and territories for which the United Nations Population Division provides general population estimates. In order to have statistics that are comparable across countries, wherever possible this study counts all groups and individuals who self-identify as Muslim. This includes members of the Sunni and Shia sects as well as Sufi orders and various smaller groups, such as the Ahmadiyya movement and the Nation of Islam, that may be considered heterodox by some Muslim authorities. It also includes Muslims who may be secular or nonobservant.
Differences Between This Report and the Pew Forum’s 2009 Report
This report builds on the findings from the Pew Forum’s 2009 report Mapping the Global Muslim Population, which acquired and analyzed about 1,500 sources of data – including census reports, large-scale demographic studies and general population surveys – to estimate the number of Muslims in every country. Some of those estimates have been revised to take into account new sources of data, such as the 2008 Nigerian Demographic and Health Survey, which were not available when the 2009 report was compiled.
In addition, this report improves on the methodology used to generate some of the 2009 estimates from older census and survey data. For the 2009 report, the number of Muslims in each country was calculated by multiplying the U.N.’s 2009 total population estimate for that country by the single most-recent and most-reliable estimate of the percentage of Muslims in the country’s population, based on the conservative assumption that Muslim populations are growing at least as fast as the general population in each country. In contrast, this new report uses estimates of the differential growth rates of Muslim populations in many countries where Muslims are a substantial minority, including India, China, Nigeria, United States, Canada and numerous European nations. For instance, the United Kingdom’s 2001 census found that 2.7% of the U.K.’s total population was Muslim, and that percentage was reflected in the Pew Forum’s 2009 report. This study, however, takes into account higher-than-average fertility among Muslims in the U.K. as well as additional Muslim immigration. Using the cohortcomponent method, it projects the U.K.’s 2001 Muslim population forward to 2010, which results in a revised estimate that Muslims comprise 4.6% of the country’s current population.
To illustrate trends, this report and its accompanying interactive maps (online) provide estimates for 1990 and 2000 based on national censuses, demographic and health surveys, and general population surveys and studies available for those years. These data points have not been altered retrospectively; no attempt has been made to correct or revise the historical figures in light of more recent data. However, Pew Forum staff did attempt to identify past overcounts or undercounts of Muslims. This evaluation involved two steps. First, any country or territory whose Muslim population size, as ranked in world order in 2010, was 15 or more places higher or lower than its world ranking in 1990 or 2000 was marked for further analysis. Staff then assessed whether the change was likely attributable to inconsistencies in the data sources rather than to an actual change in the size of the Muslim population. For instance, a 1988 report published by the Population Reference Bureau was used for the 1990 Muslim population of France. That report, based on an analysis of the few sources available at the time, estimated that there were slightly more than half a million Muslims living in France. In contrast, the 2010 estimate of 4.7 million Muslims in France – more than an eight-fold increase over the 1990 figure – is based on an analysis of a 2008 nationally representative survey of the French population. Given the limitations of the source used for 1990, compared with the strengths of the 2010 source, it seems more likely that the 1990 figure was an undercount than that the Muslim population of France grew eight-fold in 20 years.
All together, the sources for 1990 appear to have substantially understated the actual number of Muslims in Angola, Cyprus, France, Gabon, Mozambique and Ukraine, while substantially overstating the number in Colombia, Georgia, Mongolia, Panama, Taiwan and Vietnam. The 2000 estimates appear to have undercounted Muslims in Cyprus, France, Guatemala, Hungary, Slovakia and Vietnam, while overcounting the number in Laos, Lesotho and Taiwan. These likely undercounts and overcounts should be taken into consideration when looking at growth rates, particularly in the affected countries. The number of Muslims in Vietnam, for example, may not have dropped from 1990 to 2000 and then climbed rapidly from 2000 to 2010; rather, the 1990 figure was probably too high an estimate, and the 2000 figure may have been too low.
Discussion of Sources
Sources for this report include United Nations data, national censuses, demographic and health surveys, and general population surveys and studies. The specific source used for the Muslim population in each country is listed in Appendix B: Data Sources by Country. Readers should note, however, that general population surveys typically have smaller sample sizes than demographic surveys and are not designed to measure the size of small minority populations. This may lead to undercounts of Muslims in countries where they represent a small minority of the population and to overcounts where they represent the vast majority of the population. See below for more detail.63
With all sources, results may be affected by methodological decisions with respect to how the data are collected and managed. Social, cultural or political factors can also affect how answers to census and survey questions are provided and recorded.
United Nations and Other International Research Agencies
Data on fertility rate, age structure, life expectancy, migration and related factors come mainly from the United Nations Population Division. Differential demographic data on Muslims is taken from censuses, demographic and health surveys, and national statistical offices. Specific data sources are identified in each chart and table throughout the report.
For this study, Pew Forum researchers acquired and analyzed religious affiliation data from 81 censuses that were conducted since 1999, comparing more current sources of data with older census data on religious affiliation for an additional 103 countries as a cross-check. Religious affiliation questions from national censuses are the best source for estimating the number of Muslims because they generally cover the entire population and are conducted on a fairly regular basis. The chief limitation in using census data is that fewer than half of recent country censuses include a religious affiliation question. In addition, censuses typically are conducted only once every 10 years.
Where recent census data on religion are not available, religious affiliation questions from large-scale demographic surveys, such as Macro International’s MEASURE Demographic and Health Surveys, are the second-best source because of their large sample sizes, sampling frame and representative results at the province level. Though many fewer people are interviewed in a demographic survey than in a census, demographic surveys complete sufficiently high numbers of interviews to produce a generally accurate demographic profile of the country. The chief limitation of demographic surveys, for purposes of this report, is that they assume that children in a household or older members of a household have the same religion as the people interviewed, who are women and men in their reproductive years (ages 15-49 for women and ages 15-54 for men).
For this report, DHS data were acquired and analyzed for more than 60 countries, or nearly two-thirds of the countries where census data are lacking or are older than 1999. For most of the DHS surveys, both women and men are interviewed and Macro International provides the data in separate male-female data sets. Pew Forum staff pooled the female and male data sets in consultation with sampling experts at Macro International so that the combined data set retains nationally representative results. In countries where only females are interviewed, Pew Forum staff used those data to make the overall Muslim population estimate for the country.
General Population Surveys
Pew Forum researchers acquired and analyzed religious affiliation data from general population surveys for about 100 countries. In more than 20 of these countries, these surveys provide religious affiliation data where a recent census or demographic survey is lacking. Since general population surveys typically involve only 1,000 to 2,000 respondents, however, they provide less accurate numbers. This is especially true where the size of the Muslim population is quite small or Muslims live in concentrated locations that are not oversampled.
World Religion Database
Pew Forum researchers also used estimates from the World Religion Database (WRD), primarily for countries in which census and survey estimates are out of date, unavailable or lack sufficient coverage. Besides census and survey reports, WRD estimates also take into account other sources of information on religious affiliation, including anthropological and ethnographic studies as well as reputable statistical reports from religious groups themselves. The WRD is an outgrowth of the international religious demography project at Boston University’s Institute on Culture, Religion and World Affairs.
A Note on Country and Territory Designation
This report provides population estimates for 232 countries and territories. The word “territories” is used as a general term for geographical entities that are not recognized as countries by the United Nations but that have separate population estimates reported by the U.N. Population Division. Territories in this report include entities such as Hong Kong and Macau (special administrative regions of China), Greenland (an autonomous constituent country within the Kingdom of Denmark) and the Commonwealth of Puerto Rico (an unincorporated territory of the United States). For convenience, “countries” is often used in charts, tables and the text of the report as an umbrella term for countries and territories.
58 For a detailed explanation of the cohort-component method as well as a discussion of the accuracy of population projections, see Brian O’Neill and Deborah Balk, “World Population Futures,” Population Reference Bureau, September 2001. (return to text)
59 Forthcoming global projections by the Pew Forum and IIASA will use multistate projection modeling, which goes beyond traditional cohort-component analysis by building projection scenarios that take into account not just fertility, mortality and migration but also other predictors of demographic change – that is, other demographic “states” – such as education levels. Multistate projection modeling was developed at IIASA by the American geographer Andrei Rogers in the 1970s. (return to text)
60 Eurostat is the statistical office of the European Union, situated in Luxembourg. Its task is to provide the European Union with statistics that enable comparisons among countries and regions in Europe. (return to text)
61 Data are not available on the number of permanent residents who move out of the U.S. each year, including the number of Muslim permanent residents who leave. These losses may be partially offset by counting only legal, permanent residents as immigrants. (return to text)
62 See Jeffrey S. Passel and Roberto Suro, Rise, Peak and Decline: Trends in U.S. Immigration 1992–2004, Pew Hispanic Center, September 2005. (return to text)
63 Additional discussion and evaluation of these sources can be found in Brian J. Grim and Becky Hsu, “Estimating the Global Muslim Population: Size and Distribution of the World’s Muslim Population,” Interdisciplinary Journal of Research on Religion, Volume 7, Article 2, 2011. (return to text)