How Poverty Is Measured — and Why the Method Matters

Nobody agrees on exactly how many Americans are poor. In 2024, the official count was 35.9 million people: an official poverty rate of 10.6%, according to the Census Bureau's report on poverty in the United States¹. The Supplemental Poverty Measure (SPM), a separate and arguably more accurate count run by the same agency, put the 2024 poverty rate at 12.9%⁶. Two official government measurements of the same year produce a 2.3 percentage point gap. Understanding why requires answering the questions people actually have about poverty measurement, and why the methodology is not a technical footnote but a policy-critical decision.

How was the official U.S. poverty line set?

In 1963, an economist at the Social Security Administration named Mollie Orshansky was trying to measure child poverty. She knew, from a 1955 USDA Household Food Consumption Survey, that low-income families spent roughly one-third of their after-tax income on food. So she did something simple and practical: she took the Agriculture Department's "economy food plan" (the minimum-cost diet considered nutritionally adequate), calculated its annual cost, and multiplied by three.

For a family of four in 1963, that came to roughly $3,100 per year. That was the first U.S. poverty threshold.

The Census Bureau's history of the poverty measure and the Social Security Administration's tribute to Orshansky both document that her formula was adopted as the official federal poverty measure in 1969 through OMB Statistical Policy Directive 14²⁴. Since then it has been updated each year for inflation, and only for inflation.

This creates two problems that have grown more serious over sixty years. First, the food-budget-times-three multiplier made sense in 1955 when food was roughly a third of household budgets. Today, housing and healthcare collectively dwarf food spending for most families, which means the multiplier dramatically underestimates the share of income consumed by non-food necessities. Second, the threshold does not vary by geography: the same dollar amount defines poverty in rural Mississippi and in San Francisco, where rental costs are four times higher.

The 2024 official poverty threshold for a family of four was approximately $31,812: enough to cover basic food, but increasingly out of sync with what it actually costs to maintain a minimally adequate standard of living in high-cost metro areas.

What is the Supplemental Poverty Measure, and how is it different?

The Supplemental Poverty Measure, first published by the Census Bureau in 2011 following a National Academy of Sciences methodology panel, addresses the original measure's most serious limitations in two directions at once.

On the income side, it adds what the official measure ignores: SNAP food assistance, Medicaid and CHIP health coverage, housing vouchers, and the Earned Income Tax Credit all count as income. It also subtracts income taxes, payroll taxes, out-of-pocket medical expenses, and work-related costs like childcare and commuting, because a family that earns $35,000 but pays $8,000 in taxes and $6,000 for childcare has substantially less disposable income than the gross figure suggests.

On the threshold side, the SPM uses a more current measure of spending on food, clothing, shelter, and utilities by families near the 33rd percentile, updated to reflect actual modern consumption patterns.

The result in 2024: the official rate was 10.6%; the SPM rate was 12.9%¹⁶. Counterintuitively, the SPM is higher in recent years because while it captures the poverty-reducing effect of transfers (which the official measure misses), it also correctly counts the medical and childcare costs that pull households below a realistic subsistence threshold.

The SPM also makes it possible to evaluate individual programs: by comparing poverty rates with and without specific transfers, researchers can measure how much Social Security, SNAP, or the EITC actually reduces poverty. This is information the official measure cannot produce³.

What is the difference between absolute and relative poverty?

This distinction trips up most discussions of poverty statistics, particularly international comparisons.

Absolute poverty is defined by a fixed real-terms threshold: you are poor if your income falls below a specific purchasing-power level, regardless of what others earn. The U.S. official measure and the World Bank's international line are both absolute measures. If median incomes rise while the bottom decile's income stays flat, absolute poverty can fall even as the gap between rich and poor widens.

Relative poverty defines poverty as a fraction of median income, typically 50% or 60% of median household income, as used by OECD members. Within a relative measure, a country whose entire income distribution rises by 10% shows no change in poverty, because everyone's income rose proportionally. Relative poverty captures social exclusion: the inability to participate in the normal consumption patterns of your society, rather than bare subsistence.

The two measures can diverge sharply. In the U.S., the absolute poverty rate fell substantially between 1959 and 1970 as economic growth raised incomes broadly. But relative poverty (the gap between the bottom and the median) changed much less. A family in 1970 whose income was 45% of the median was, by relative poverty standards, in poverty; by absolute standards, better fed and housed than their 1959 counterpart. Both statements are true simultaneously.

For policy purposes, absolute poverty matters most when the question is survival: whether people can eat, find shelter, and access basic healthcare. Relative poverty matters most when the question is social mobility, educational opportunity, and civic participation, and whether the economic gap between the bottom and the middle is so wide that it becomes self-perpetuating.

How is global poverty measured?

The World Bank's international poverty line provides the standard for global comparisons. For decades it was set at $1.90 per day in 2011 purchasing-power-adjusted dollars, chosen to approximate the typical national poverty line of the world's poorest countries. In 2022 this was updated to $2.15 per day in 2017 PPP terms, reflecting updated purchasing power parity calculations.

In June 2025, the World Bank updated its international poverty line again to $3.00 per day, incorporating new national consumption data and 2021 PPP estimates. Within the previous $2.15 threshold, approximately 700 million people (roughly 8.5% of the global population) were living in extreme poverty as of 2024, according to the World Bank Poverty, Prosperity, and Planet Report⁵.

The $2.15/$3.00 line measures extreme deprivation: survival. The World Bank also reports poverty at higher thresholds: $3.65/day for lower-middle-income countries and $6.85/day for upper-middle-income countries, the latter affecting 44% of the global population by the most recent estimates. Which line is appropriate depends on the question: survival poverty in low-income nations, or the broader social exclusion that makes meaningful participation in a middle-income economy impossible.

What actually causes poverty?

Poverty is not one condition with one cause. Census Bureau Current Population Survey data and decades of poverty research identify several distinct pathways.

Inadequate wages. Many households in poverty have at least one employed adult. The problem is not absence of work but the wage level that work pays. At $7.25/hour federal minimum wage, a full-time, full-year worker earns $15,080: below the poverty threshold for a single adult with one child. Low wages produce working poverty; the cause is not behavioral but structural (wage levels, occupational segregation, education-to-wages mismatch).

Insufficient or unstable work hours. Retail, food service, and care work frequently involve involuntary part-time scheduling, irregular hours, and last-minute shift cancellations. A worker who would earn $28,000 in a stable full-time job may earn $18,000 in a volatile part-time one: the poverty comes from schedule instability, not from low hourly wages alone.

Household structure. Single-parent households have poverty rates roughly three times those of married-couple families. This is not a values judgment; it is arithmetic: one income covers the same fixed housing, childcare, and utility costs as two. The poverty gap for single mothers with children is the most persistent demographic pattern in poverty data across decades.

Barriers to work. Disability, chronic illness, lack of affordable childcare, transportation gaps, housing instability, and criminal records create concrete barriers to stable employment. These are not excuses but measurable constraints: families cycling in and out of poverty often do so as these barriers appear and resolve, not because of changes in effort.

Geographic concentration. Poverty clusters geographically: in rural regions with collapsed industries, in concentrated urban neighborhoods with depleted tax bases, in communities where low-income households are isolated from job networks, quality schools, and the social capital that employment requires.

What do the official numbers miss?

Several things matter for understanding living standards that neither the official nor the supplemental measure fully captures.

Household wealth (savings, home equity, financial assets) buffers households against income volatility. Two families with identical incomes live very different lives if one has $40,000 in savings and the other has zero. Income poverty statistics say nothing about wealth poverty.

In-kind support from family networks: many households officially counted as poor receive substantial non-governmental support through shared housing, food sharing, and childcare by family members. This support is real but invisible in official statistics.

Access to public goods matters too. A household in a well-resourced school district with nearby parks, libraries, and transit infrastructure has a higher effective standard of living than the same income in a resource-depleted one.

The one thing to remember

Poverty measurement is not a technical detail; it is the lens through which anti-poverty policy is designed, evaluated, and funded. The official U.S. poverty measure, built from a 1963 food-budget formula, systematically undercounts the poverty-reducing effect of the programs that actually work (SNAP, Medicaid, EITC) while using thresholds that no longer reflect actual consumption costs. The Supplemental Poverty Measure addresses both limitations, and the fact that it shows a higher rate than the official measure in 2024 is not a criticism of government programs; it is evidence that the programs work, and that the people the official measure still counts as poor face genuine cost burdens the original formula was never designed to capture.