Washington is ending its 30-year Food Insecurity Report. U-M research shows why the numbers still matter.
By Jon Meerdink
When the U.S. Department of Agriculture announced in September that it would end the country’s longest-running annual measure of food insecurity, it offered a striking justification: the numbers, the agency said, had “remained virtually unchanged.”
Nearly three decades of data collected at the University of Michigan tell a more complicated story.
For years, researchers at U-M’s Institute for Social Research have tracked not only how many American families struggle to put enough food on the table, but what happens to those families when federal policy shifts. Their central finding is not that food insecurity has held steady. It’s that it moves, often sharply, in response to the choices lawmakers make.
That distinction matters far beyond academia. When the federal government expanded food assistance during the COVID-19 pandemic, fewer families reported going without enough to eat. When those supports ended, more of them slipped back. Data collected the same way over time is how anyone with a stake in the problem, whether a voter, a food bank or a member of Congress, can actually see whether the safety net is working. Without it, they’re guessing.
“No one expected that the COVID-19 pandemic would see food insecurity fall, but it really proved that federal policy can make a huge difference,” said Noura Insolera, an assistant research scientist at ISR and assistant director of the Panel Study of Income Dynamics, one of the world’s longest-running household studies. “Allowing people to have an easier time applying for public safety-net programs and offering a higher amount made a real difference during the pandemic and a few years after. It was a very successful intervention.”
“No one expected that the COVID-19 pandemic would see food insecurity fall, but it really proved that federal policy can make a huge difference. Allowing people to have an easier time applying for public safety-net programs and offering a higher amount made a real difference during the pandemic and a few years after. It was a very successful intervention.”
What the long view reveals
The Panel Study of Income Dynamics, known as PSID, has followed the same American families since 1968, revisiting them annually through 1997 and every two years since, to ask about their income, health, program participation and, increasingly, whether they have enough to eat. That long arc lets researchers see patterns a single-year snapshot cannot, and it has produced more than 7,600 peer-reviewed studies, making it one of the most heavily used data resources in the social sciences.
Insolera’s work has drawn out those patterns. In research published in 2023, she found that chronic food insecurity, defined as families reporting across every wave of a multi-year period that they didn’t have enough to eat, had more than doubled in two decades, climbing from 2.1% of families in 1999–2003 to 4.5% in 2015–19. Among low-income families with children, the chronic rate reached 10.9%. This was happening even as the national economy grew.
“It’s not just that there are more instances of one wave of families saying they didn’t have enough to eat,” Insolera said, “but that it actually persists wave after wave.”
Then the pandemic scrambled expectations. Rather than spiking, food insecurity fell. In PSID’s nationally representative panel, the share of households reporting they didn’t have enough to eat dropped in the pandemic’s peak-support years, then climbed once that support lapsed, ending higher than before the pandemic began.
“We actually observed a decrease in reported food insecurity in 2021, which we think was due to the increase in SNAP benefits during the COVID-19 pandemic,” Insolera said. “But many of those benefits ended in 2023, and we saw a large increase in food insecurity in the PSID data.”
Researchers caution that no single program explains the swing. The 2021 improvement coincided with a whole package of temporary supports: expanded SNAP benefits, but also the enlarged Child Tax Credit and other pandemic-era relief. Hardship rose again as those supports expired one after another through 2022 and 2023. But the direction of the pattern is hard to miss. When the government widened the safety net, food insecurity fell; when it pulled back, food insecurity climbed. The rebound was steepest for the households with the least cushion: families with children, low-income households and SNAP participants.
Insolera’s most recent work extends the picture past 2023, and it shows food insecurity still higher than it was before the pandemic. It remains elevated, and highest among families with children and lower-income households.
Why the missing federal report matters
Since 1995, across administrations of both parties, the USDA had published an annual snapshot of food insecurity built on the Census Bureau’s December food-security survey, the official benchmark and the only one that produced state-by-state estimates. Its final edition, released in December 2025 and covering 2024, found that 13.7% of U.S. households, or 47.9 million people, were food insecure, among the highest levels in more than a decade. In ending the report, the USDA said the trend had stayed “virtually unchanged” and that it would rely instead on “more timely and accurate” data.
The “unchanged” framing holds up only on the narrowest reading. Year to year, 2024 was statistically flat against 2023. But step back and the official rate climbs steadily, from 10.2% in 2021 to 12.8% in 2022 to 13.5% in 2023 to 13.7% in 2024, the same upward march PSID picked up as pandemic supports fell away.
Insolera is careful not to overstate what PSID can do in the federal report’s absence. It is not a one-to-one replacement. PSID measures the nation as a whole, not individual states, and its two-year rhythm can’t match an annual federal series. The timelier federal data the September announcement pointed to, meanwhile, measures something different: whether a household had enough food in the last week, not the 12-month picture the discontinued report captured. What PSID offers is what the annual snapshot never could, the ability to watch the same families across decades and connect a policy change to a family’s next grocery bill.
“What we do is not a food insecurity survey,” Insolera said. “It’s a huge survey about health and wealth and income and program participation, among other topics. Income is where we started, but we want to have a broad view of things, with accessible data that allows for anyone to come in and do their own research.”
That breadth has real downstream reach. In one of her own studies, Insolera and colleagues followed children into adulthood and found that those who received both SNAP and WIC growing up had roughly four times the odds of improved food security in adulthood, compared with eligible children who received neither. It is concrete evidence of how the programs this data tracks can shape lives years later.
The stakes, Insolera argues, are not abstract. Food insecurity rarely begins with an empty cupboard; families often worry and stretch tight budgets for months before anyone skips a meal. Data collected consistently over time is what lets policymakers spot that pressure early and act before it deepens.
“The challenge now is maintaining the data sources that shine a light on these kinds of changes,” Insolera said. What her decades of numbers suggest, more than anything, is that the work is worth continuing. Policy can measurably change how millions of families live, and good data is how the country knows whether it’s helping.