Help Me--and the U.S.--Define Poverty
Updated: Oct 18
I find it ironic that in an age of surveillance, we can’t see the bare necessities.
Before Mollie Orshansky revolutionized our social policy, she was answering letters from families all over the country. Post-WWII life was tough and people were asking the USDA economist how they could afford to feed their families. Mollie would mail back budget pamphlets and low-cost meal plans. She knew that life.
As a kid, Mollie stood with her mom in relief lines for surplus food.¹ Poignant then, that she ended up using the cost of food to derive our official poverty line, shaping social policy for the next 60 years and counting.
It happened little by little, then all at once. After moving from the USDA to the Social Security Administration, Mollie became the go-to person on what it costs to get by. Then President Johnson declared a War on Poverty, and Washington needed a way to count the poor. Mollie’s math was ready. Having pored over survey data, she knew families spent about a third of their income on food. So she took the cheapest of USDA’s four food plans and multiplied it by three. Fall below that in pre-tax cash income, and you had too little.² Adjust for family size, update for inflation each year, and Houston, we have poverty.
But times have changed, and what we call poverty hasn’t. Today we spend about a tenth of our income on food, while child care, housing, and health care cost more.³ Rather than wait for a new poverty line, some agencies have rewritten it themselves. SNAP sets its income cutoff at 130% of the official line.⁴
Our Supplemental Poverty Measure, first published in 2011, was supposed to fix this.⁵ It improves the limitations of using pre-tax cash income by counting benefits as well and it looks at more expenses than just food. But through the looking glass of the SPM, some wealthy Boomers are poor.
Somewhere in the US, a Boomer is getting new knees for pickleball, and we label them poor for it. The SPM subtracts out-of-pocket medical bills from their income, but it doesn’t count the nest egg that actually pays them. On paper, their retirement income is small and their bills are big, so they look broke. In 2012, elderly SPM poverty was 14.8%. Without the medical subtraction, it would have been 8.4%.⁶ Using income causes other headaches.
Some income we can't even track. Roughly half of food stamp and Temporary Assistance for Needy Families (TANF) dollars go unreported in the Census survey. When benefits vanish from the data, income-based measures makes poor families look much worse off than they are.⁷
I try to live by the mantra K.I.S.S. Keep it simple, stupid. Here, that means skipping to the end. Use consumption, not income, to define poverty. Rather than chase down every source of income and deal with missing data, just read the receipts. Paychecks, SNAP, savings: all of it shows up in what you buy.
Between 1980 and 2022, consumption poverty fell from 33.8% to 6.0%. The official rate fell just 1.5 points.⁸ Our world's a lot better than we’re told.
This isn’t fringe. The BLS is building its own consumption measure.⁹
Consumption poverty changes the how many and who of poverty. The consumption-poor are less educated and less likely to own a home than the income-poor. They’re less likely to be elderly or living alone, and more likely to be married parents with kids.¹⁰ If we want to get aid to people who need it, it helps to know who they are.
Here’s the catch. Consumption data isn't great either. Even the researchers behind consumption poverty say not to look at it alone.¹¹
Consumption data comes from the Consumer Expenditure (CE) Survey, which asks about 10,000 homes to remember what they bought over the last three months. The sample is too small to say much about any one state, so it can’t tell Mississippi from Manhattan. The income survey isn’t doing much better: its response rate this year was 61.3%.¹²
Mollie created a lens for us to see our most vulnerable. We let it get dusty. Here’s how we clean it.
Make the data talk to each other. The IRS knows what we earn. The SNAP office knows what it paid us. They just don’t talk to the survey. When researchers linked tax and benefit records to the Census survey and counted every benefit, 55% fewer people were poor. For single parents, the survey said deep poverty rose for two decades. The linked data said it fell.¹³ That’s the irony. The data exists. We just don’t connect it.
Then make the surveys worth answering. When response is required by law, people answer: the ACS gets 97%. Our voluntary income survey gets 61%. Short of a mandate, pay people, and let them snap a photo of their receipts. That’s what the National Academies told BLS to do more than a decade ago.¹⁴
Private data can help, but it can’t lead. Banks see spending in real time, but 5.6 million households have no bank account, and two-thirds of them live on cash.¹⁵ Card data goes blind exactly where we need to see.
Mollie showed that you can bring a head and a heart to civil service, and that you’re better for it. Let’s get serious about cleaning our poverty lens. Neglect has splintered the conversation about who needs help and what that help looks like. We’re lone rangers by birthright, but some national problems demand that our government have quality information about us. Flock cameras? Nope. Basic measures of well-being? Absolutely.
Define poverty yourself. The dashboard below uses consumption data to show who’s poor as you move the line, and which groups are overrepresented. It sets the line as a share of median spending, so your numbers won’t match the fixed-line figures above.
Notes
Gordon M. Fisher, “Remembering Mollie Orshansky—The Developer of the Poverty Thresholds,” Social Security Bulletin 68, no. 3 (2008). Immigrant background: MacTutor History of Mathematics, “Mollie Orshansky.” Quote: Sewell Chan, “Mollie Orshansky, Statistician, Dies at 91,” New York Times, April 17, 2007.
Fisher, “The Development and History of the U.S. Poverty Thresholds,” U.S. Dept. of Health and Human Services (ASPE); John Cassidy, “Relatively Deprived,” The New Yorker, April 3, 2006.
USDA Economic Research Service, “Food Prices and Spending” (2024 data); Meyer & Sullivan, Annual Report on U.S. Consumption Poverty: 2018, Wilson Sheehan Lab for Economic Opportunities (LEO), University of Notre Dame.
Congressional Research Service, “Supplemental Nutrition Assistance Program (SNAP): A Primer on Eligibility and Benefits,” R42505, September 2025.
U.S. Census Bureau, Poverty in the United States: 2025 (P60-290), September 2026.
Korenman & Remler, “Rethinking Elderly Poverty: Time for a Health Inclusive Poverty Measure?” NBER Working Paper 18900 (2013); Center for Retirement Research at Boston College, Working Paper 2015-29.
Meyer, Mok & Sullivan, “The Under-Reporting of Transfers in Household Surveys: Its Nature and Consequences,” NBER Working Paper 15181.
Han, Meyer & Sullivan, Annual Report on U.S. Consumption Poverty: 2022, povertymeasurement.org.
Bureau of Labor Statistics, “Building a Consumption Poverty Measure: Initial Results” (2022).
Meyer & Sullivan, “Dimensions of Progress: Poverty from the Great Society to the Great Recession,” as summarized by the University of Notre Dame.
Jeehoon Han, quoted in Luke Lattanzi, The Baylor Lariat, 2024.
Bureau of Labor Statistics, Consumer Expenditure Surveys, bls.gov/cex; U.S. Census Bureau press release, September 15, 2026 (2026 CPS ASEC weighted response rate).
Corinth & Meyer, “Poverty in the United States Using the Comprehensive Income Dataset,” American Economic Association (AEA) 2020; Corinth, Meyer & Wu, “The Change in Poverty from 1995 to 2016 Among Single Parent Families,” NBER Working Paper 29870.
PolitiFact, “Americans must answer U.S. Census Bureau survey by law,” January 9, 2014; National Academies, Measuring What We Spend: Toward a New Consumer Expenditure Survey (2012).
FDIC, 2023 National Survey of Unbanked and Underbanked Households.
