After the world population increased more than 400% over the 20th century, population growth has slowed considerably: The fastest world population growth rate was already reached in the late 1960s, and it has been falling since. While the world population increased by 2% annually in the late 60s it has now slowed to an increase of just about 1%.
There are three primary determinants of global population growth: mortality, fertility, and population momentum. The global improvement falling mortality – seen in increasing life expectancy around the world and falling child mortality in every country – contribute to an increase of the world population. The decline of fertility rates on the other hand, – the number of children per woman – reduces population growth. The global average fertility rate was 5 children per woman until the end of the 1960s and has halved since then.
The UN projections for the global population growth rates, which have been produced since the 1950s, have a good track record in projecting the size of the global population as we will show in the last part of this entry.
# Empirical View
# The size of populations – UN Population Projection by country and world region until 2100
The visualisation below shows the change of the total population since 1950 and the UN population projection from 2015 to 2100. You can add any country to this chart.
# The growth rate of populations – UN Population Projection by country and world region until 2100
Global population growth has slowed down markedly since the peak in the 1960s. The following map shows the growth rate by country at the peak of population growth in 1968.
# Births and deaths
# The Three Drivers of World Population Growth: The Fertility Rate, Life Expectancy, & Population Momentum
Population growth is driven by three different factors: mortality, fertility, and population momentum.
# The World Fertility Rate
The visualization below shows the total fertility rate by the level of development and includes the UN projections through 2100. Until 1950 the fertility rate in the ‘more developed regions’ had already declined to less than 3 children per woman. Then in the 1960s the fertility rate in the ‘less developed regions’ started to fall and another decade later the fertility rate in the ‘least developed regions’ followed this decline.
The fertility rate of the world was still at 5 children per woman until the mid-1960s. Since then the fertility rate has halved and is today below 2.5 children per woman.
# The fertility rate by world region
Fertility is driven by the socio-economic development of the population, especially the status and wellbeing of women. Accordingly, the early developing populations started the transition much earlier than world regions that remained poor for a longer time. Europeans and Northern Americas reached the replacement rate at a time when fertility only started to decline in Asia and Latin America. Africa, where poverty and lack of education persisted even longer, was the last continent to experience a decreasing fertility rate.
# The World Life Expectancy
As health is rapidly improving around the world, life expectancy is also increasing rapidly. You can read more about life expectancy at the our life expectancy data entry.
# What matters for the size of the world population is fertility not mortatlity
Mortality is even inversely correlated with population growth
Where fertility is high population growth is high. This correlation is as strong as correlations in social science can ever be.
# Projections of the World Population
Shown below is the increase of the world population since 1750 combined with the latest projections of the UN Population Division.
The Medium Variant is the projection that the UN researchers see as the most likely scenario. The High- and Low-variants simply assume that the total fertility rates in each country are 0.5 higher and 0.5 lower than the Medium variant by the end of this century in every country.
The Constant Fertility Scenario is an illustrative scenario that plays out how the world population would change if fertility rates remained constant. It is obviously not intended to be a realistic scenario.
The History Database of the Global Environment (HYDE) – the data source of the population density estimates in the following maps – visualised the estimates of population density on an interactive world map here.
# Correlates, Determinants & Consequences
As shown above, changes in life expectancy and fertility rates determine population growth. The data entry on past world population growth discusses the demographic transition as the central concept that explains population growth.
# Data Quality & Definition
# Institutions that publish population projections
The most widely discussed projections are those published by the United Nations, the first of which were published already in 1951. The UN projections are called ‘assessments’ and a new update is published in their World Population Prospects series every two years. The scenario which the UN researchers see as the most likely scenario is the Medium Variant projection. But in addition to their main Medium projection the UN Population Division are additionally publishing a High and Low variant, which simply assume that the total fertility rates are 0.5 higher and 0.5 lower than the Medium variant by the end of this century in every country.
But there are also a number of other institutions that are preparing their own projections of the world population. Global population projections are also published by the US Census, the Population Reference Bureau (PRB), and by the closely related Austrian research centers IIASA and the Wittgenstein Centre. The World Bank also published projections for some time but has stopped doing so in the mid–90s.
# The WC-IIASA projections: How investments in education matter for the global population in the 21st century
A set of influential projections is published by IIASA and the Wittgenstein Centre, which I refer to as the WC-IIASA projections.1
Projections of the global population are uncertain and much of the uncertainty comes from the fact that we do not know which investments the world will make in those systems that determine mortality and fertility – most importantly in education, as we have just seen.
The WC-IIASA projections differ from the work of the United Nations in a number of fundamental ways. The UN projections are taking into account the empirical data on each country’s demography and are building projections based on this quantitative information. In contrast to this the WC-IIASA projections are also taking into account the qualitative assessments of 550 demographers from around the world which the WC-IIASA researchers have surveyed to gather their ideas on how the population change in different parts of the world will play out. They then combine the country specific expertise of these researchers with similar quantitative information that the UN and others rely on as well.2. The work by WC-IIASA is highly respected among demographers and key publications by the researchers are regularly published in the scientific journal Nature.3
The WC-IIASA projections are taking into account the demographic structure of the educational attainment of the population. While other projections are only structuring the demographic data by sex and age-group, the WC-IIASA data is additionally breaking down the population data by the level of highest educational attainment of different parts of the population. This information on educational attainment is then used for both the output of the model – so that population projections for each country of the world by highest educational attainment are available (also on Our World in Data). And crucially the information on education is also used as an an input into the model, so that the impact of different future scenarios for education on both mortality and fertility can be modeled explicitly.
The level of highest educational attainment is categorized in a system that aims to capture the structure of populations across the different country-specific educational systems. These categorizations are based on the the International Standard Classification of Education (ISCED), which was designed by the UNESCO to make education statistics comparable across countries. WC-IIASA breaks down the educational structure into the following 6 categories and the table summarizes how the six categories are defined, how they correspond to ISCED 1997, and the main allocation rules the researchers used.
# Categories of educational attainment used by IIASA-WC and how they correspond to the ISCED levels4
For children younger than 15 years old no educational attainment information is available as most of them are still in the process of education.
The four scenarios for global education by WC-IIASA
Projections of the global population take into account how the fertility rate will change in each country over the coming decades. The WC-IIASA projections are particularly helpful for the discussion here as they are the only projections that break down the demographic projections by the educational level of the populations and then model how different educational scenarios would affect the fertility rate in countries across the world. This then allows comparisons of how education matters for the size and distribution of the future population of the planet.
The researchers developed 4 different basic scenarios and a larger number of combinations based on these scenarios:
Constant Enrollment Numbers (CEN): This is the researcher’s most pessimistic scenario. Here it is assumed that no more schools are being opened in any place in the world so that the absolute number of people reaching a particular educational level is frozen at the current number. This means that enrollment rates are declining when the population size increases.
In practice the WC-IIASA researchers almost always consider CER as the most pessimistic scenario and only rarely discuss CEN.
Constant Enrollment Rates (CER): This is another pessimistic scenario. While in the CEN scenario the absolute number of enrolled students stagnates, the assumption in the CER scenario is that the rate of enrollment stagnates. In this scenario the most recently observed rates of educational enrollment are frozen at their current rate and no further improvement in enrollment is assumed.
This will still result in further improvements of adult education because in many countries the younger cohorts are better educated than the older ones. But in the longer run this scenario also implies stagnation.
Fast Track (FT): This scenario is the most optimistic one and here it is assumed that countries follow the most rapid education expansion achieved in recent history which is that of South Korea.
Global Education Trend (GET): This is the middle scenario and here the researchers assume that countries will follow the average path of educational expansion that other countries already further advanced in this process have experienced. In this scenario the researchers project the medium future trajectory based on the experience of all countries over the past 40 years
The researchers write: “The GET scenario is moderately optimistic, and can be considered as the most likely.”5
# The track record of the UN projections
# Global population growth
In a 2001 paper, Nico Keilman assessed the projections of the global demographic changes that the UN published between 1951 and 1998.6 The next graph shows the increase of the world population (solid black line) along with the UN projections published between 1950 and 1980. With the exception of the projection “1950 I”, which relied on poor data – especially for China – the forecasts are remarkably accurate. Even the 1950 III forecast that was made with better information on China after the 1953 Census is not far off for the population size half a century later.
Nevertheless it is interesting to see that over the first decades the UN underestimated the population growth and for the last period they overestimated the world population. Why this happened becomes clearer if we look at the projections for the world fertility rate and global life expectancy.
Estimates of the world population compared with the UN forecasts, since 1950 – Keilman (2001)7
# Global average fertility rate and life expectancy
Keilman also studied the two drivers of population growth separately and this can help to explain why the UN first underestimated the population growth and then overestimated the population.
As shown in the following graph he finds that the UN underestimated the rapid fall of world fertility.
And as shown in the graph below the UN underestimated the rise of global life expectancy.
The UN were too pessimistic in both aspects as Keilman notes: The “projection makers have been too pessimistic about future mortality” and “predicted life expectancy levels that were too low on average – much too low in many cases”.
On a regional level the study finds that “not surprisingly, problems are largest in pre-transition countries, and especially in Asia”.
Estimates of the world fertility rate compared with the UN forecasts, since 1950 – Keilman (2001)8
Estimates of the world life expectancy compared with the UN forecasts, since 1950 – Keilman (2001)9
# Data Sources
Data on population growth – including projections – can be found in the data section of the entry on world population growth.