Viewing the Global South through Population Migration: Deep Learning Reveals New Patterns in Four Decades of Population Movement

Introduction

Population migration is a core driver of global demographic change, profoundly affecting labor markets, social policies, and economic landscapes. For a long time, global migration data has primarily relied on high-quality statistical systems in developed countries, while migration flow data from the Global South—including developing regions such as Africa, Asia, and Latin America—has been severely lacking, leading to a "receiving-country bias" in international research. A 2026 study published in Nature (Gaskin & Abel, 2026) is the first to use deep learning models to construct a dataset of annual migration flows covering 230 countries and regions from 1990 to the present, opening a new window for research on population migration in the Global South.

Data Dilemma: The "Migration Blind Spot" of the Global South

Traditional migration statistics rely on census data, administrative records, and survey data. The migrant stock data published every five or ten years by the United Nations and the World Bank can only provide static snapshots and are concentrated in developed economies. For example, Europe's QuantMig project covers only 30 European countries, while annual flow data for most countries in Africa, South Asia, and Latin America is virtually nonexistent. This data imbalance has led research to focus on receiving countries, neglecting the complex dynamics unique to the Global South, such as South-South migration and intra-regional mobility.

Deep Learning Bridging the Gap

The key innovation of this study lies in integrating fragmented multi-source data—including official statistics, census stock data, net migration estimates, and past flow reconstructions—into a unified deep learning framework. By using a deep recurrent neural network (RNN) and incorporating covariates such as geography, economy, culture, and politics, the model can not only reproduce long-term migration trends but also capture flow changes triggered by short-term shocks such as conflicts, famines, and natural disasters. Compared with traditional five-year estimates, the new annual data significantly reduces uncertainty while preserving detail.

Key Insights for the Global South

1. The Emergence of South-South Migration

Annual flow data reveals, for the first time, large-scale population movements within Africa, within Asia, and between Southeast Asia and the Middle East. For example, seasonal labor migration among West African countries and the circular migration of South Asians to Gulf countries—flows previously smoothed over or overlooked by five-year data—can now be tracked on an annual basis. This provides an empirical foundation for understanding regional economic integration, such as the African Continental Free Trade Area.

2. Labor Markets and the Youth Dividend

The Global South has a large youth population, and migration is a key channel for labor reallocation. The new data can more accurately estimate the flows, stocks, and integration speed of youth migrants, helping emerging economies assess the balance between brain drain and remittance inflows. For instance, India exports a large number of skilled talents to OECD countries each year while simultaneously receiving low-skilled laborers from neighboring countries; annual data can refine the temporal patterns of this bidirectional flow.

3. Climate- and Conflict-Driven MigrationClimate change and conflict are causing sudden population movements in Africa, the Middle East, and elsewhere. Previous models have struggled to disentangle the effects of environmental factors from political factors on an annual scale. A new dataset, by aligning migration flows with annual covariates such as conflict, drought, and flooding, enables researchers to quantify the causal effects of extreme events on migration. For example, how the 2022 drought in the Horn of Africa accelerated migration from Somalia to Ethiopia and Kenya can now be modeled more accurately.

4. Policy and Investment Risk Assessment

For international investors and sovereign risk analysts, population migration is an important indicator of economic resilience, labor supply, and social stability. The new dataset provides the spatiotemporal distribution of annual migration flows globally, helping to identify labor pressures in rapidly urbanizing areas, the demand for real estate and infrastructure driven by cross-border migration, and changes in the tax base due to population outflow.

Uncertainty-Guided Data Collection

The study also identifies regions worldwide with the highest uncertainty, which are precisely where data collection is most urgent. The variance in migration flow estimates is large for landlocked African countries, Central Asia, and parts of Central America, signaling that international organizations and statistical agencies should prioritize establishing better entry and exit registration systems in these areas.

Conclusion

This research is not only a breakthrough in technical methodology but also a cornerstone for perfecting population studies in the Global South. Annual migration flow data will push emerging market research from "reliance on extrapolations from developed-country cross-sections" toward "precise analysis based on South-South flows." When the population dynamics of the Global South are re-illuminated, we can truly understand the micro-level push and pull factors behind the shift in the center of global economic growth.