Poverty is often represented as insufficient income, but contemporary research describes a more complex condition. It encompasses limited access to nutritious food, safe housing, education, healthcare, sanitation, energy, secure employment and the social resources required to withstand economic or environmental shocks. Consequently, poverty cannot be understood through a single monetary threshold.
The scale of the problem also depends on how poverty is measured. In 2025, the World Bank updated its international extreme-poverty line from $2.15 in 2017 purchasing-power-parity dollars to $3.00 per person per day in 2021 PPP terms. Under the revised methodology, approximately 838 million people were living in extreme poverty in 2022. This statistical increase did not mean that millions of people had suddenly become poorer; it resulted from updated prices, national poverty lines and household-consumption data. The revision demonstrates why poverty estimates must always be interpreted alongside their definitions and reference years. World Bank, 2025
Recent findings nevertheless point to a consistent conclusion: progress against poverty has slowed, deprivation is increasingly concentrated in fragile regions, and climate hazards, food costs and unequal access to public services are creating additional barriers to poverty reduction.
Measuring the Changing Geography of Poverty
Monetary poverty remains extensive
The World Bank estimated that approximately 831 million people—around one person in ten—were living below the new $3.00 extreme-poverty line in 2025. Nearly three-quarters of this population lived in rural communities. At the substantially higher threshold of $8.30 per person per day, designed to reflect living standards in upper-middle-income economies, an estimated 3.7 billion people were classified as poor.
These different thresholds are not competing answers to the same question. The international extreme-poverty line measures the minimum consumption level typical of low-income countries, whereas higher lines identify economic insecurity in countries where the cost and social requirements of participation are greater.
Global averages can also obscure the changing distribution of poverty. The largest historical reductions occurred in East and South Asia, while extreme poverty is now increasingly concentrated in sub-Saharan Africa and fragile or conflict-affected states. The World Bank estimated that 451 million people in fragile or conflict-affected situations experienced extreme poverty in 2024. Since 2022, these settings have contained a majority of the global population living in extreme poverty. World Bank Poverty and Inequality Update, 2025
Multidimensional poverty reveals overlapping deprivation
Income alone cannot establish whether a household has electricity, whether children attend school or whether safe drinking water is accessible. The global Multidimensional Poverty Index therefore measures ten indicators across health, education and living standards.
The 2025 index covered 6.3 billion people in 109 countries. It identified 1.1 billion people living in acute multidimensional poverty, more than half of whom were children. Nearly 740 million lived in middle-income countries, showing that national economic classification does not reliably indicate whether all population groups benefit from development.
Common deprivations included inadequate housing, poor sanitation, undernutrition and lack of clean cooking fuel or electricity. Importantly, the index identifies combinations of disadvantages at the household level. A family may simultaneously face inadequate nutrition, interrupted schooling and unsafe energy use, with each deprivation reinforcing the effects of the others. UNDP and OPHI, 2025
However, multidimensional indicators also have limitations. Survey years differ between countries, conditions can vary markedly within households, and standard indicators may not capture debt, personal security, digital exclusion or the quality of public services. No existing measure provides a complete and universally applicable description of poverty.
How Poverty Persists Across Generations
Childhood poverty is not equally persistent everywhere
Longitudinal research consistently associates childhood poverty with poorer health, educational disruption and greater economic insecurity in adulthood. Yet recent comparative evidence indicates that this persistence is strongly influenced by institutions.
A 2025 study in Nature Human Behaviour analysed administrative and panel data from the United States, Australia, Denmark, Germany and the United Kingdom. The researchers measured how the proportion of childhood spent in poverty predicted the proportion of early adulthood spent in poverty.
The estimated intergenerational persistence coefficient was 0.43 in the United States, compared with 0.21 in Australia, 0.16 in the United Kingdom, 0.15 in Germany and 0.08 in Denmark. In practical terms, spending an entire childhood in poverty in the United States was associated with a 43-percentage-point higher average poverty rate in early adulthood than experiencing no childhood poverty.
Employment, household employment and family structure mediated part of the relationship. However, differences in tax and transfer systems were particularly important. The authors estimated that adopting the protective effects observed in peer countries could reduce US intergenerational poverty persistence by more than one-third. The analysis was a decomposition rather than a randomized experiment, so it cannot establish the effect of any specific policy reform. It nevertheless shows that childhood poverty does not produce an identical or inevitable outcome across institutional contexts. Parolin et al., 2025
Unequal environments shape cognitive opportunity
Poverty may influence development through nutrition, environmental exposures, chronic stress, housing instability, school quality and access to learning resources. These mechanisms are difficult to isolate because they frequently occur together.
A 2025 study in Nature Computational Science applied Bayesian multilevel modelling and post-stratification to data from the US Adolescent Brain Cognitive Development cohort. Its purpose was to examine variation in associations between socioeconomic conditions and cognitive development while improving the representation of demographic groups that are often underrepresented in cohort datasets.
The study illustrates an important methodological development: average associations may conceal substantial differences between social and demographic subgroups. Combining cohort data with census information can generate more representative estimates, although modelling cannot remove all selection bias or transform observational relationships into proven biological mechanisms. Osayande et al., 2025
It would therefore be scientifically inaccurate to claim that poverty determines a child’s cognitive capacity. Current evidence instead suggests that unequal material and socio-spatial environments affect access to developmental opportunities. Individual outcomes remain heterogeneous, and the relative contributions of nutrition, stress, pollution, schooling and family resources are not fully known.
Poverty, Food Security and Mental Health
Food affordability is a central dimension of poverty
The relationship between poverty and nutrition extends beyond insufficient calorie consumption. A household may obtain enough dietary energy while remaining unable to afford fruits, vegetables, legumes, animal-source foods or other nutrient-dense products.
According to the 2026 State of Food Security and Nutrition in the World, approximately 645 million people experienced hunger in 2025. More broadly, about 2.7 billion people could not afford a healthy diet. This distinction is critical: hunger captures severe inadequate food consumption, whereas diet unaffordability identifies a much larger population exposed to nutritional constraints. FAO, IFAD, UNICEF, WFP and WHO, 2026
Food affordability depends on income, prices, transport, storage infrastructure, market concentration and agricultural productivity. Therefore, increasing household income may not be sufficient when nutritious foods remain physically unavailable or unusually expensive. Conversely, food-price policies that favour inexpensive calorie-dense staples may reduce hunger without ensuring dietary quality.
Further research is needed to determine which combinations of social protection, agricultural investment, transport infrastructure and food-market regulation can improve healthy-diet affordability across different regions. Policies effective in an urban middle-income setting may not transfer directly to isolated rural or conflict-affected communities.
Poverty and mental health form a bidirectional relationship
Poverty can increase exposure to debt, insecure housing, violence, hazardous employment and uncertainty about food or healthcare. These stressors can contribute to depression, anxiety and psychological distress. Mental-health conditions may, in turn, reduce employment continuity, educational participation and the capacity to manage administrative or financial demands.
A systematic review published in Scientific Reports in 2025 examined 17 studies combining poverty-reduction measures with psychological or psychosocial interventions. Cash and asset transfers were the most common economic components, while psychological support was frequently delivered by non-specialist professionals.
The review found indications that combined programmes can improve mental-health outcomes and, to a lesser extent, socioeconomic outcomes. Economic components appeared more consistently beneficial when they provided substantial or sustained support. However, the studies differed considerably in their populations, intervention designs, outcome measures and methodological quality. Only four discussed implementation difficulties in detail.
The researchers consequently could not determine an optimal intervention model or establish whether combined programmes consistently outperform well-designed single-component programmes. More comparative trials are required, particularly among children and adolescents and in settings where robust economic assistance can be paired with targeted mental-health care. Tanski et al., 2025
Climate Change as a Poverty Multiplier
Climate hazards overlap with existing deprivation
Low-income households generally have fewer savings, less insurance and more limited access to credit or public services. Their livelihoods may also depend heavily on agriculture or outdoor labour. Consequently, the same flood, drought or heatwave can have very different effects across socioeconomic groups.
The 2025 Global Multidimensional Poverty Index found that 887 million of the 1.1 billion people experiencing multidimensional poverty lived in regions exposed to at least one of four hazards: extreme heat, drought, flooding or air pollution. Approximately 309 million faced three or four of these hazards concurrently.
These figures identify geographic overlap, not proof that every person experienced an individual climate-related loss. Nevertheless, the results reveal where material deprivation and environmental exposure may combine to restrict adaptive capacity.
Subnational data expose effects concealed by national averages
An analysis published in Nature Climate Change in 2026 assembled a panel dataset covering subnational areas in approximately 130 countries. The researchers found that a 1°C increase in temperature was associated with a 0.629–1.178 percentage-point increase in poverty at the lowest poverty threshold used in the study. Their projections corresponded to an additional 62.3–98.7 million people living in poverty by 2030.
The estimated relationship was stronger in subnational data than in national aggregates. A national average may conceal severe effects in agricultural, remote or already impoverished regions. The analysis also estimated that a 1°C temperature increase was associated with a 1.3–1.9% increase in the Gini inequality index.
The effects were more pronounced in lower-income countries and economies with a larger agricultural share. Agricultural yields, soil moisture and vegetation conditions are plausible pathways, although they do not account for every mechanism connecting heat and poverty. Labour productivity, health, infrastructure damage, migration and food prices may also be important. Dang et al., 2026
As with other macro-level observational studies, these estimates remain sensitive to model specifications, available data and assumptions about future temperature exposure. Science cannot yet provide a precise forecast for every region. The results nevertheless support poverty-reduction policies that incorporate climate adaptation rather than treating environmental and socioeconomic risks separately.
What Does the Evidence Say About Reducing Poverty?
Economic growth is necessary but insufficient
Broad-based economic growth has historically contributed to major reductions in global poverty, particularly when it increased agricultural productivity and employment. However, aggregate growth does not automatically reach disadvantaged groups.
Access to healthcare, education, transport, electricity and digital infrastructure influences whether people can participate in new economic opportunities. The increasing concentration of poverty in fragile states also demonstrates that growth strategies depend on institutional capacity, physical security and functioning public services.
Gender and age must be considered as well. Children represent more than half of the global population in extreme poverty despite constituting approximately 30% of the total population. Women’s poverty risk varies across the life course and can be intensified by unpaid care work, reproductive responsibilities, discriminatory labour markets and restricted control over household resources.
Effective policy probably requires coordinated interventions
Recent research does not identify one universally superior anti-poverty programme. Cash transfers can provide immediate material relief and may improve psychological well-being, but their effects depend on payment size, duration, price levels and service availability. Education and training can strengthen long-term opportunity, but they may have limited effects where secure employment is scarce. Mental-health care can improve functioning, but it cannot compensate for persistent hunger or unsafe housing.
The strongest interpretation of current evidence is therefore not that every programme must be maximally complex. Rather, interventions should correspond to the mechanisms operating in a specific population. Where poverty involves simultaneous food insecurity, psychological distress and insecure employment, narrowly addressing only one outcome may leave the reinforcing cycle intact.
Researchers still need better long-term studies, standardized outcome measures and direct comparisons between cash-only, service-only and genuinely integrated programmes. Evidence is also limited for people living in conflict zones, displaced populations and many rural communities, even though these groups represent a growing share of global poverty.
Conclusion
Recent scientific studies portray poverty as a dynamic system of economic, social, health and environmental constraints. Monetary thresholds remain indispensable for global monitoring, but multidimensional measures reveal deprivations that income statistics alone cannot capture.
The latest evidence also challenges deterministic interpretations of poverty. Childhood deprivation is associated with adult poverty, but the strength of this relationship varies considerably between countries. Tax and transfer systems, employment conditions and public services can alter life-course outcomes. Similarly, poverty and mental illness may reinforce one another, yet research on integrated interventions remains too heterogeneous to define a universal treatment model.
Climate change is making poverty research still more urgent. Hundreds of millions of multidimensionally poor people already live in areas exposed to multiple climate hazards, and subnational analyses indicate that rising temperatures may increase both poverty and inequality.
Poverty reduction therefore requires more than raising average income. It involves expanding material security, affordable nutrition, health and education, protecting households against shocks and ensuring that institutions can reach populations facing overlapping disadvantages. Science has identified many of these mechanisms, but it has not established a single solution applicable to every society. Continued interdisciplinary research—and more representative data—will be essential for designing policies that produce durable rather than temporary exits from poverty.
Subscribe to our newsletter!
