Pew Research Center has released a new study exploring how the sizes of seven major religious groups will change by 2070, if recent trends continue. We project that the populations of some of these groups will grow rapidly, while others will decrease over time.
In this Q&A, we speak with Senior Demographer and Associate Director of Research Conrad Hackett about how the Center makes its projections. The interview has been condensed and edited for clarity.
What is a projection?
A projection estimates what the future size of a population will be if certain assumptions prove correct.

Typically, demographers make projections by dividing a population into groups – or “cohorts” – based on age and sex, like women ages 15 to 19 or men ages 40 to 44. Then we estimate how many births and deaths will occur over time in each cohort. Projections may also include estimates of how many people will migrate from one country to another.
Our projections look at the sizes of specific religious groups in each country, so we also estimate how many people in a particular place will switch from the group in which they were raised to a different group.
Why is it important to estimate how different religious groups are changing in size globally?
About three-in-four people worldwide identify with a religion. But by 2070, the religious makeup of various geographic regions will look very different than it does now, since religious groups grow at different rates.
Surveys often measure the religious identity of adults at the country level. In some cases, national censuses also measure religion. However, without doing a lot of work to analyze this information, we wouldn’t know the big picture or what kinds of changes are happening at the regional and global levels. That work is the foundation of this project.
What are the main factors that affect whether a religious group is increasing or decreasing in number?
Globally, fertility differences have a major impact. Muslims have a total fertility rate of 3.1 children per woman, while Buddhists average 1.6 children per woman. Muslims’ fertility rate is enough to grow their population, while Buddhists’ rate isn’t. If recent trends continue, there will be fewer Buddhists in 2070 than there are today.
Fertility differences lead to differences in the age profiles of groups. Populations in which most people are younger tend to grow faster than populations in which most people are older.
In addition, some countries attract many migrants who bring a diverse mix of religions. For example, most people born in the United Arab Emirates are Muslim, but about one-in-ten people there are Hindu migrant workers from India.
In the United States and many European nations, religious switching is a cause of substantial religious change. The most common type of switching in the U.S. and around the world is from Christianity to no religion, but we account for less common types of switching in our projections, too.
You project that fewer people worldwide will be religiously unaffiliated in 2070 than are today. Does that mean most countries are getting more religious?
No. In most countries, the number of people who are religiously unaffiliated is actually expected to grow in the decades ahead. But this growth isn’t projected to add up to as many people as the decline in one country: China.
In 2020, two out of three religious “nones” in the world lived in China. But China’s population is expected to decrease significantly by 2070, including a drop of more than 360 million religious “nones.” In the rest of the world, the number of “nones” is expected to increase by about 180 million by 2070. So China’s population decline will have an outsized impact on the world’s religious composition.
How is the religious makeup of the United States expected to change by 2070?
The United States is one of the countries where the population of religious “nones” is expected to increase, from 30% in 2020 to 41% in 2070. Christians made up 64% of the total U.S. population in 2020, but that share is expected to fall to 48% in 2070.
Currently, people who identify as Jewish when asked about their religion are the third-largest U.S. religious group, making up 1.7% of all Americans. By 2070, we project Muslims will be the third-largest group at nearly 4% of the population.
What kinds of trends are easiest and most difficult to anticipate?
Most population change is the result of birth, aging and death, processes that generally lead to slow changes in a population. With some assumptions about fertility and mortality trends, we can model these slow processes very well. Generally, most 10-year-old Muslims will survive to be 15-year-old Muslims. Since we know that Muslims have the highest fertility and the youngest age distribution of all the groups we study, we can be quite confident that their share of the world’s population will increase.
“Most population change is the result of birth, aging and death, processes that generally lead to slow changes in a population. With some assumptions about fertility and mortality trends, we can model these slow processes very well.”
— Conrad Hackett, senior demographer and associate director of research
Within a country, the sizes of religious groups can change quickly because of migration or religious switching. However, it is impossible to know for certain what future migration and switching patterns will look like. In our main projection scenario, we assume that the recent migration and switching rates we’ve estimated will continue in the decades ahead. However, if migration or switching patterns change, our projections could be off.
Events like war, pandemics, economic crises and natural disasters can also alter the trajectory of populations. Anticipating such events and how they may affect religious groups is beyond the scope of our projections.
Were your previous projections on track?
Largely, yes! In 2015, we projected population change from 2010 to 2050. When we look back at the first decade of that period, the amount and direction of change we projected are similar to what actually happened in most cases.
However, we underestimated many changes in Christian and religiously unaffiliated populations from 2010 to 2020. For example, in the United States and the United Kingdom, the shares of Christians dropped and the shares of religious “nones” rose more than we anticipated. This was largely because switching patterns changed – net movement out of Christianity increased, but our projection model assumed fixed switching rates.
Do your projections account for “informal” religious identities, such as people who identify as culturally Jewish, or for spirituality?
Not directly. In most countries we studied, people answered a survey or census question like “What is your religion, if any?” We categorized them into seven groups: Christians, Muslims, Hindus, Buddhists, Jews, the religiously unaffiliated and people who belong to all other religions (such as Sikhs, Daoists and followers of traditional or folk religions).
However, people with “informal” religious identities can be included in these groups in various ways:
- The religiously unaffiliated category includes people who identify as culturally but not religiously Jewish. It also includes some people who engage in religious and spiritual beliefs and practices but do not formally identify with a religion.
- Each religious group includes people who do not adhere to all the beliefs and practices typically associated with the group.
In other words, our data reflects the complexity of social life.
In the U.S., immigration policy has changed immensely between the Biden and Trump administrations. How can you account for migration without knowing what countries’ policies toward immigrants and refugees will look like in the future?
It’s true that migration patterns could change with shifting policies or events like war and natural disaster. This could have a big impact on some local religious landscapes. But about 96% of the world’s people currently live in the country where they were born. And in the future, for the countries where most people live, we project that natural population change – the difference between births and deaths – will produce more religious change than migration.
