Wittgenstein Center (WIC) Population and Human Capital Projections - 2023

THIS IS VERSION WIC3.005 (internal version V15, projection chain run 14–16 April 2025; earlier versions were V14 (WIC3.004), V13, V12 and V11; the 2023 Zenodo versions were labelled WIC3.001 and WIC3.002). V15 is the vintage served by the Wittgenstein Centre Human Capital Data Explorer (2020 base) and the basis of the long-run extension to 2300 (https://doi.org/10.5281/zenodo.18596542). Note on the SSP Scenario Explorer. The population data served by the IIASA SSP Scenario Explorer ("Basic Elements 2025", model IIASA-WiC POP 2025) are the 2025-rebased update of these projections, which treat the SSP2 path from 2020 to 2025 as a nowcast and project every pathway from that common 2025 population. Those data are published separately (https://doi.org/10.5281/zenodo.21738149 for 2025–2100 and https://doi.org/10.5281/zenodo.21738637 for 2100–2300). SSP2 is the same trajectory in both families; the other pathways differ. Short Abstract We update the population and human capital components of the Shared Socio-Economic Pathways (SSPs) at the global level, considering the most recent baseline information. While the long-term assumptions based on extensive analysis and expert solicitations remain unchanged, we modify only the trend component. The first set of SSPs was based on demographic data through 2012. The population structures by age, sex, and education of the base year (2010) have since changed for most countries, mainly due to more recent data and reliable information. The mortality situation has improved in many countries affected by HIV/AIDS and among children in countries with higher mortality. The impact of COVID-19 on demographic trends must be addressed. In many countries, fertility rates have fallen faster than expected. International migration has been irregular and volatile as usual. These changes are reflected in the new update, with some improvements in operationalization. Changes in version "V15": V14 was produced by the projection chain run on 16–28 January 2025, after the sex codes of the education distributions of 37 countries had been corrected (see the V14 note below). V15 is a complete re-run of that chain on 14–16 April 2025, from the estimation of the education-specific fertility and mortality differentials, through the mortality variants, to every scenario. The re-run was initiated by a correction to the education inputs: attainment progression ratios that had been capped at 90 percent in the reference scenario were corrected, with the excess redistributed across the younger cohorts of the same country, and the attainment targets were regenerated on 15 April 2025. The same run carried a correction to the education-specific mortality differential for a small number of countries. Because the differentials are re-estimated in the chain, the education-specific fertility and survival inputs also changed slightly. The 2020 base populations are identical to V14. Measured against V14: sex ratios at birth are identical; education-specific fertility differs in 17 countries (Taiwan the largest); survival ratios differ only in the low-mortality variant for China, Mongolia, the Republic of Korea, the Netherlands and Singapore; the education attainment inputs for ages 15–34 shift slightly in most countries. World population in 2100 changes from 8.007 to 7.991 billion under SSP1, 8.034 to 8.019 under SSP5, 13.100 to 13.086 under SSP3, and by less than 0.02 percent under SSP2 and SSP4. Under SSP1 and SSP5 the only country changing by more than one percent is China, 2.4 percent lower in 2100; under SSP2 and SSP4 no country changes by more than one percent. Under SSP3, where educational expansion stalls, the corrected education inputs carry furthest through fertility: 16 countries change by more than one percent in 2100, twelve of them among the 37 corrected countries, the largest being Zimbabwe (+5.9 percent), Timor-Leste (-4.8), Nicaragua (-4.1) and China (-3.1). Details and a table are in README_V15.md. Two additions to the file layout: the PROJresult files carry region_level and region_label columns so that the 200 countries can be separated from the 28 UN aggregates (filter region_level == "country" before summing; the other values are subregion, area and world); flows (births, deaths, imm, emi) are recorded as missing in 2100 because no period starts there, and the newborn row (agest = -5, which holds the births of the period and the deaths among them, not a stock) exists for 2020–2095 only. Changes in version "V14": We found a bug affecting the education distribution sex (interchanged) in 37 countries (with UN country code): 32 Argentina, 44 Bahamas, 48 Bahrain, 64 Bhutan, 76 Brazil, 108 Burundi, 132 Cabo Verde, 156 China, 158 China, Taiwan Province of China, 170 Colombia, 214 Dominican Republic, 218 Ecuador, 226 Equatorial Guinea, 288 Ghana, 296 Kiribati, 360 Indonesia, 392 Japan, 400 Jordan, 410 Republic of Korea, 414 Kuwait, 496 Mongolia, 528 Netherlands, 558 Nicaragua, 583 Micronesia (Fed. States of), 591 Panama, 626 Timor-Leste, 630 Puerto Rico, 634 Qatar, 643 Russian Federation, 662 Saint Lucia, 702 Singapore, 764 Thailand, 784 United Arab Emirates, 788 Tunisia, 792 Turkey, 840 United States of America, 894 Zambia. The impact on the total population is minimal.