Food bought for home in American cities cost 33% more than it did at the start of 2019. During the preceding seven and a half years, the increase was 6.4%. The national comparison is stark; what it does inside a kitchen depends on income, rent, transport, electricity, benefits and which foods a household cannot surrender. [1]
AP's July 25 report, retained here as explicit backfill, combines government series with interviews in four urban areas. It advances the paper's earlier warning that inflation anxiety is not itself an operating result. These families describe operating results. Four-area interviews still cannot measure national prevalence.
That boundary protects the people in the story from being turned into a statistic they were never sampled to produce. Their substitutions are evidence of possible mechanisms and real experiences. They are not a poll of all households.
The same index, different kitchens
The lowest-earning fifth of households spent 33% of pretax income on food in 2024, AP reports. Ground beef reached $6.82 a pound, 79% above early 2019. SNAP enrollment stood at 37 million in April, down 12% from a year earlier. [1]
Those figures describe different populations and measures. A food-price index tracks prices. An expenditure share reflects income as well as spending. SNAP enrollment reflects program participation, not a direct count of hunger, skipped meals or people who remain eligible. Adding them into one deprivation total would manufacture a denominator.
The household accounts show the adjustments behind those aggregates: buying less meat, replacing preferred foods with cheaper versions, choosing shelf-stable staples, shopping among stores and weighing food against other bills. [1] Such choices can preserve calories while changing variety, nutrition, time and pleasure. A cheaper basket is not necessarily an equivalent basket.
Price increases also accumulate differently across a menu. A household can postpone a durable purchase, but food returns to the list every week. Substitution therefore repeats: one less costly item replaces another until the remaining choices collide with allergies, preferences, preparation time and what children will eat. The national index cannot display that narrowing.
Clinical empathy begins with that tradeoff rather than a lecture about budgeting. A parent who changes protein, produce or portion size may already be comparing every advertised solution. Bulk buying requires cash and storage. Driving to another store requires time, fuel and transport. Cooking from scratch requires equipment and hours. Advice that ignores those inputs simply moves a cost out of view.
Texture is not prevalence
AP interviewed families in four urban areas. Their accounts give the national series a human scale, but they cannot tell us how often Americans made one substitution, which regions faced the largest burden, or whether the same patterns persisted among rural households. [1]
A compatible national answer would require household expenditure microdata linked to income, geography and quantities. It would distinguish buying fewer units from switching products, eating less from wasting less, and losing nutrition from finding cheaper equivalents. It would also need real wages, food-bank use, debt and health outcomes.
Cause requires similar restraint. Prices reflect labor, transport, weather, commodity markets, disease, tariffs, retailer strategy and other costs. Household pressure also reflects wages and nonfood bills. The interviews do not authorize one explanation for every changed grocery list.
SNAP deserves particular care. A fall in enrollment can reflect income changes, eligibility, renewals, denials, administrative churn, migration or revised records. It does not by itself show that need fell, nor does participation establish whether a household had enough food. [1]
No authorized X post was recovered. Online stories about a cart costing twice as much or a family thriving through clever substitutions therefore remain unmeasured discourse, not a national series. The divergence is not between social media and arithmetic so much as between one national number and the unequal lives concealed within it.
AP makes those lives visible without claiming representativeness. [1] The responsible next step is to retain both halves: the government series tells us prices moved dramatically; the interviews show several ways families absorbed the move. Neither tells us how the whole country ate.
The question for policy is not only whether inflation slows. A lower rate of increase leaves the higher price level in place. Families need wages, benefits and budgets that catch up, and researchers need compatible records showing quantities, nutrition and health. Until then, cheaper food is a coping action, not evidence that the affordability problem has been solved.
-- NORA WHITFIELD, Chicago