Fertility, mortality, migration, and population scenarios for 195 countries and territories from 2017 to 2100: a forecasting analysis for the Global Burden of Disease Study

Abstract

Background: Understanding potential patterns in future population levels is crucial for anticipating and planning for changing age structures, resource and health-care needs, and environmental and economic landscapes. Future fertility patterns are a key input to estimation of future population size, but they are surrounded by substantial uncertainty and diverging methodologies of estimation and forecasting, leading to important differences in global population projections. Changing population size and age structure might have profound economic, social, and geopolitical impacts in many countries. In this study, we developed novel methods for forecasting mortality, fertility, migration, and population. We also assessed potential economic and geopolitical effects of future demographic shifts.

Methods: We modelled future population in reference and alternative scenarios as a function of fertility, migration, and mortality rates. We developed statistical models for completed cohort fertility at age 50 years (CCF50). Completed cohort fertility is much more stable over time than the period measure of the total fertility rate (TFR). We modelled CCF50 as a time-series random walk function of educational attainment and contraceptive met need. Age-specific fertility rates were modelled as a function of CCF50 and covariates. We modelled age-specific mortality to 2100 using underlying mortality, a risk factor scalar, and an autoregressive integrated moving average (ARIMA) model. Net migration was modelled as a function of the Socio-demographic Index, crude population growth rate, and deaths from war and natural disasters; and use of an ARIMA model. The model framework was used to develop a reference scenario and alternative scenarios based on the pace of change in educational attainment and contraceptive met need. We estimated the size of gross domestic product for each country and territory in the reference scenario. Forecast uncertainty intervals (UIs) incorporated uncertainty propagated from past data inputs, model estimation, and forecast data distributions.

Findings: The global TFR in the reference scenario was forecasted to be 1·66 (95% UI 1·33-2·08) in 2100. In the reference scenario, the global population was projected to peak in 2064 at 9·73 billion (8·84-10·9) people and decline to 8·79 billion (6·83-11·8) in 2100. The reference projections for the five largest countries in 2100 were India (1·09 billion [0·72-1·71], Nigeria (791 million [594-1056]), China (732 million [456-1499]), the USA (336 million [248-456]), and Pakistan (248 million [151-427]). Findings also suggest a shifting age structure in many parts of the world, with 2·37 billion (1·91-2·87) individuals older than 65 years and 1·70 billion (1·11-2·81) individuals younger than 20 years, forecasted globally in 2100. By 2050, 151 countries were forecasted to have a TFR lower than the replacement level (TFR <2·1), and 183 were forecasted to have a TFR lower than replacement by 2100. 23 countries in the reference scenario, including Japan, Thailand, and Spain, were forecasted to have population declines greater than 50% from 2017 to 2100; China's population was forecasted to decline by 48·0% (-6·1 to 68·4). China was forecasted to become the largest economy by 2035 but in the reference scenario, the USA was forecasted to once again become the largest economy in 2098. Our alternative scenarios suggest that meeting the Sustainable Development Goals targets for education and contraceptive met need would result in a global population of 6·29 billion (4·82-8·73) in 2100 and a population of 6·88 billion (5·27-9·51) when assuming 99th percentile rates of change in these drivers.

Interpretation: Our findings suggest that continued trends in female educational attainment and access to contraception will hasten declines in fertility and slow population growth. A sustained TFR lower than the replacement level in many countries, including China and India, would have economic, social, environmental, and geopolitical consequences. Policy options to adapt to continued low fertility, while sustaining and enhancing female reproductive health, will be crucial in the years to come.

Funding: Bill & Melinda Gates Foundation.

Figure 1

Model fit for CCF50 CCF50-fitted trends are presented as a function of education across varying levels of contraceptive met need. Each point represents a single location-year of past data, and they are coloured by GBD 2017 super-regions. Education is measured in years of attainment (0–18 years), and contraceptive met need is measured on a scale of 0% to 100%, reflecting the portion of the female population whose need for contraception has been met. CCF50=completed cohort fertility by age 50 years. GBD=Global Burden of Diseases, Injuries, and Risk Factors Study.

Figure 2

Global life expectancy in the reference, slower, faster, fastest, and SDG pace scenarios, 1990–2100 The reference scenario is presented with 95% UIs, which are represented by the shaded area. Life expectancy was computed at birth, and values are reported in years. Past estimates are from GBD 2017. GBD=Global Burden of Diseases, Injuries, and Risk Factors Study. SDG=Sustainable Development Goal. UI=uncertainty interval.

Figure 3

Global and sub-Saharan African total fertility rates, 1990–2100 Past data reflect GBD 2017 estimates, with future results for the reference, slower, faster, fastest, and SDG pace scenarios. The reference scenario is presented with 95% UIs, which are represented by the shaded area. GBD=Global Burden of Diseases, Injuries, and Risk Factors Study. SDG=Sustainable Development Goal. UI=uncertainty interval.

Figure 4

Map of the year that the net reproduction rate falls below the replacement level Replacement is defined as a net reproduction rate of 1. Past estimates are calculated on the basis of GBD 2017 estimates and future estimates are calculated from the reference forecast. GBD=Global Burden of Diseases, Injuries, and Risk Factors Study.

Figure 5

Global population in the reference, slower, faster, fastest, and SDG pace scenarios, 1990–2100 The reference scenario is presented with 95% UIs, which are represented by the shaded area. Past estimates are from GBD 2017, and values are in billions. GBD=Global Burden of Diseases, Injuries, and Risk Factors Study. SDG=Sustainable Development Goal. UI=uncertainty interval.

Figure 6

Global population age structure in 2017 and in 2100 in the reference, slower, faster, fastest, and SDG pace scenarios Estimates for 2017 are from GBD 2017. Triangles indicate the mean age for each scenario. GBD=Global Burden of Diseases, Injuries, and Risk Factors Study. SDG=Sustainable Development Goal.

Figure 7

Population size in ten countries in the reference, slower, faster, fastest, and SDG pace scenarios, 1990–2100 The reference scenario is presented with 95% UIs, which are represented by the shaded area. Countries were selected on the basis of the most populous countries in 2017 and those projected to be the most populous in 2100. Past estimates up to 2017 are from GBD 2017, and estimates for 2018–2100 are means of each scenario. GBD=Global Burden of Diseases, Injuries, and Risk Factors Study. SDG=Sustainable Development Goal. UI=uncertainty interval.

Figure 8

Number of working-age adults from 1950 to 2100 in the reference scenario in the ten most populous countries in 2017 Working-age adults are defined as individuals aged 20–64 years. Past data are from GBD 2017. GBD=Global Burden of Diseases, Injuries, and Risk Factors Study.

Figure 9

Ranking the top 25 economies by total GDP in 2017 and the reference scenario in 2030, 2050, and 2100 Countries are coloured by GBD super-region. GBD=Global Burden of Diseases, Injuries, and Risk Factors Study. GDP=gross domestic product. *(province of China).