Well, let's just not measure food insecurity then

In September 2025, the Trump administration said it will no longer release the Household Food Security Reports. The USDA also put relevant employees on indefinite leave.

This post provides context to the published press release based on the USDA’s own 2025 Information Collection Request renewal for the Food Security Supplement to the Current Population Survey, hosted on Reginfo.gov.¹

These redundant, costly, politicized, and extraneous studies do nothing more than fear monger.

Is the survey redundant?

From the Supplemental Statement A of the Information Collection Request Renewal:

If these data were not collected, USDA would not be able to continue to monitor food security annually, the research described above could not be completed, and USDA would need to develop or find other means to evaluate the effectiveness of the nutrition assistance programs.

It is true that the Food Security Supplement is not the only survey which includes the Food Security Module. Yet USDA 2025 Information Collection Request points out that:

none of these surveys [National Health and Nutrition Examination Survey, Survey of Income And Program Participation, National Center for Education Statistics, National Health Interview Survey, and the American Housing Survey] include the full complement of follow-up questions on the frequency and duration of food-insecure conditions nor the food spending and full range of Federal and community food and nutrition assistance program participation that are included in the CPS Food Security Supplement and are important factors to be examined in conjunction with food security. (USDA IRC, linked above)

Is the survey costly?

From Section 12b and 14 of Supporting Statement A FSS 2025:

The annualized cost of the respondents’ time spent answering the supplement questions is estimated to be $197,229.

The cost of the December 2024 Food Security Supplement was $1,058,692.

Note that this 2024 price is more costly than the previous submission, in part due to a one-time cost to conduct a nonresponse bias study, as well. Even ignoring that, it is surprising to me that $1 million per year is considered too expensive to understand food systems in the U.S., even as Trump and donors are happy paying $300 million for a new ballroom at the White House.

The USDA also just spent tens of millions of dollars in May for government food purchasing contracts as part of food assistance programs such as TEFAP and CSFP.

Is the survey politicized?

Three aspects here: 1) that the survey reflects a set of values, 2) that the survey methodology is designed to benefit a political group, or 3) that the outputs of the survey are used in political arguments.

On 1), since it is a government dataset, it is a political decision to collect this data. Notably - until now, I didn’t think that this is a partisan decision.

Embedded in the Household Food Insecurity survey (and measure) is a value judgement that individuals should have enough to eat of foods they want without worry.

Given the structure of the survey module, researchers are able to provide additional assessments of food security. This might be the case if the argument is that individuals not having foods they want is not a priority, but instead just having food that meets their basic caloric intake.

The more contemporary concern seems to be whether the Department of Agriculture should serve the needs of the producer or the consumer; this action shifts the focus back onto producers, which is in line with what Secretary Brooke Rollins promotes in her testimony.

For example, take these Agriculture Censuses can be accessed through the National Agricultural Statistics Service (NASS), which are a core task of the USDA historically. These are arguably just as political as any other; from the main page: “Through the Census of Agriculture, producers can show the nation the value and importance of agriculture and can influence decisions that will shape the future of U.S. agriculture.” It is notably entirely focused on producers (farmers, ranchers).

On 2), I don’t think that the data collection process or survey instrument unfairly reflect one political preference. You can read it here from the USDA and judge for yourself.

Some criticisms (that I’ll go over a bit more later) include questioning if the binary thresholds of food insecure or not should be deprioritized in favor of very low food security, if the recall period increases those reporting food insecurity, if people are underreporting government assistance, if anxiety over sourcing food is a helpful policy indicator, and if people underreport children going hungry. But none of these are explicitly political, and instead reflect limitations in survey design and delivery.

On 3), undoubtedly, people (reporters, politicians, op-ed writers) politicize this dataset.

Obviously, any dataset can be politicized; and I would argue that the Trump administration has repeatedly demonstrated a willingness to politicize anything it can for political gain.

For example, see the recent BLS firing, ignoring conservative political violence, vaccinations, hurricanes or egg prices or the announcements accusing Democrats of gender mutilation on agency pages. My point here is that the Trump administration has a pattern of attacking politically inconvenient datasets by arguing they are designed to hurt his political beliefs.

This dataset is not unusually susceptible to being politicized in one direction or another relative to other government datasets. From a political usage perspective, this data can be used to support consumer or producer-side interventions. For example, Senator Reed (D-Rhode Island) and Senator Justice (R-West Virginia) proposed the Strengthening Local Food Security Act of 2025, an approach to encourage local food purchasing to improve food security and support local food distributors.² And famously, as you’ve probably seen, Feeding America relies on these estimates in much of their publicity materials to demonstrate the scale of the problem.

Is the survey extraneous?

The second mission area listed on the U.S. Department of Agriculture’s About page is to “harness the Nation’s agricultural abundance to reduce food insecurity and improve nutrition security in the United States.” The Economic Research Service leads the program now and is authorized to do so.

It is true, like the USDA mentions, that the 2018 Farm Bill mentions food insecurity in the context of the Gus Schumacher Nutrition Incentive Program (unrelated to the Household Food Security report, but shows adoption of the measurement concept), funding public-private partnerships to implement pilot projects on addressing food insecurity, and is used as a benchmark for a series of microgrants in areas with high food insecurity.

For 30 years, this study—initially created by the Clinton administration as a means to support the increase of SNAP eligibility and benefit allotments—failed to present anything more than subjective, liberal fodder.

This is untrue. The survey was developed in response to H.R.1608 National Nutrition Monitoring and Related Research Act of 1990 which calls for a plan to support nutrition monitoring, specifically including food consumption measurements (Section 3 Definitions, 7B). This act was signed into law by President George H.W. Bush (a Republican).

Notably, prior to this law, food insecurity as defined has roots within a task force report released under President Reagan, which highlighted a need for a clear measure.

Development to create a questionnaire occurred through 1992 through 1995, with the survey first being launched in April 1995.

Is the survey subjective?

The USDA has regularly reviewed the validity of the survey. Their 2023 report is publicly available, alongside many reports on the methodology listed on the USDA website itself. The technical documentation is available to read. You can also access the replicate weights for the survey.

I have not seen any comprehensive arguments that attack the entire methodology of the survey, but I want to highlight a few articles I read about how to improve the survey and its limitations. Key concerns I found when exploring focused on measuring the severity and persistence of food security.

There was an analysis by the AEI writing partners of Rachidi and Rourke from 2024 on the Household Food Security report. They suggest that perhaps the food insecurity module does not reflect the most vulnerable families with respect to the association between food insecurity and economic hardship. A quote:

Because the FSS does not allow us to monitor the quantity and types of food that households purchased, we cannot directly assess how certain households can spend such high amounts on food while also reporting food insecurity.

Nonetheless, the distribution of food expenditures is approximately the same for food-insecure households as it is for food-secure households, raising questions about whether food insecurity primarily measures food-related anxiety, food preferences, or food-related hardship. (13)

For example, Gundersen (2023) suggests emphasizing the “resource gap”, or how much more money a household would need to spend weekly to meet its needs, alongside the binary food secure & insecure assignment.

A practical limitation of the survey that it does not measure household food security over time for respondents. To address this, some preprints in arXiv (Lee et. al 2025) introduces a “probability of food security” longitudinal household measure, and (Lee, 2025) also implemented this PFS measure and did not identify any significant effect on SNAP participation on estimated food security status, so something to keep an eye on.

Lastly, of course, the methodology (and the questions) will affect results. Rachidi and Gundersen in 2024 attribute some increase in food insecurity in 2022 to a change in the survey’s methodology, which the USDA raised in a 2023 report. The same paper also cites underreporting in government benefits received in the Current Population Survey, but does not fully explain the impact, so worth a follow up.

Trends in the prevalence of food insecurity have remained virtually unchanged, regardless of an over 87% increase in SNAP spending between 2019 - 2023.

Has the prevalence of food insecurity changed?

Virtually is a key word in this sentence, for sure. A rollercoaster ends up at the same place it started. It is trivial to find these estimates; see the USDA’s prevalence rate chart here.

I put together the below chart to highlight years with statistically significant changes in the prevalence rates also.

So this statement just ignores the statistically significant changes in food insecurity in the period of 2019 to 2023; for example, the increase from 10.2% in 2021 to 12.8% in 2022 reflected an additional 3.5 million households experiencing food insecurity.

Why was there an increase in SNAP spending from 2019 to 2023?

Two events explain the large increase in allotted spending for SNAP.

First, the Trump administration signed into law the 2018 Farm Bill, which included language that later allowed the USDA under the Biden administration to update payment distribution amounts to account for changes in prices over time.

Secondly, in 2020 the COVID-19 pandemic and resulting shutdowns caused large temporary job losses, contributing to a spike in SNAP participants. Sharp price increases in food (and many other categories of goods) due to the COVID-19 pandemic have persisted through 2025, potentially keeping SNAP participation high.

SNAP is designed specifically to alleviate economic crises, and spending on the program will rise during and after recessions and depressions.

Not going to spend much time on this since so much information is available online. If interested:

• SNAP Spending Rose and Fell With Pandemic-Era Changes to Benefit Amounts
• See Economic Linkages section of the USDA Key Statistics page

The relationship between trends in prevalence of food insecurity and spending more on SNAP?

The USDA’s statement seems to suggest that increased spending on SNAP should affect trends in the prevalence of food insecurity; specifically, the implication here is that the food insecurity rates should be lower because of the increased spending.

This is either an error in statistical literacy, or worse, an intentional attempt at deception.

We observe the prevalence of food insecurity with the effect of the increased spending on SNAP; so the counterfactual is then what the prevalence of food insecurity would be without the increased spending on SNAP. In other words, we would expect the food insecurity rates to be higher than recorded if not for increased SNAP spending.

The Supplemental Nutrition Assistance Program is a safety net, not a ladder. Research by Ratcliff et. al. (2014) indicates SNAP “reduces the likelihood of being food insecure by roughly 30% and reduces the likelihood of being very food insecure by 20%.” Gundersen et. al. (2017) “find that SNAP reduces the prevalence of food insecurity in households with children by at least six percentage points” through a causal study of partial identification effects. Wu et. al. (2025) find that “during the COVID-19 pandemic, food insecurity decreased among SNAP participants in most racial and ethnic groups but did not decrease among non-SNAP participants in any group,” with a similar finding from Brady et. al. (2023)

USDA will continue to prioritize statutory requirements and where necessary, use the bevy of more timely and accurate data sets available to it.

Is the survey a statutory requirement?

An Agri-Pulse article cites a release stating that “this non-statutory report…is unnecessary to carry out the work of the Department”. Note I wasn’t able to find a copy of the version of the release provided to news outlets, so I am relying on partial quotes here.

Core point is I don’t know, but I haven’t seen anyone challenge the government on this, so will take them at their word.

I tried to read the current code describing the Department of Agriculture.³

There shall be at the seat of government a Department of Agriculture, the general design and duties of which shall be to acquire and to diffuse among the people of the United States useful information on subjects connected with agriculture, rural development, aquaculture, and human nutrition, in the most general and comprehensive sense of those terms, and to procure, propagate, and distribute among the people new and valuable seeds and plants. (7 U.S. Code 2201)

The Secretary of Agriculture shall procure and preserve all information concerning agriculture, rural development, aquaculture, and human nutrition which he can obtain by means of books and correspondence, and by practical and scientific experiments, accurate records of which experiments shall be kept in his office, by the collection of statistics, and by any other appropriate means within his power; he shall collect new and valuable seeds and plants; shall test, by cultivation, the value of such of them as may require such tests; shall propagate such as may be worthy of propagation; and shall distribute them among agriculturists; and he shall advise the President, other members of his Cabinet, and the Congress on policies and programs designed to improve the quality of life for people living in the rural and nonmetropolitan regions of the Nation. (7 U.S. 2204 (a))

My read of section 7 U.S. Code 2204(h) (again I do not regularly read laws, so just saying this as an average citizen) also seems to indicate that the food insecurity survey would be supported, directing the secretary to “evaluate the manner in which local food systems a) contribute to improving community food security; and b) assist populations with limited access to healthy food.” I am unsure how the Secretary can evaluate how effective a food system is at feeding people without understanding which populations have limited access to healthy food.

But I guess the argument is that while it is allowed, it is not specifically requested by name, so I’ll leave it at that.

Instead, I spent some time reading about the early history of the USDA to learn how the role of the department evolved to focus on food security.

The USDA has a 90+ years of history focusing on ensuring people have food to eat. Small scale food distribution programs began in the 1930s.⁴ A Senate history on the Agriculture Committee notes that by 1939 the government distributed food to 12.7 million people through the Agricultural Adjustment Act of 1933. School Lunch programs started in the 1940s.⁵ The Senate History notes that in the 1950s, “The Congressional focus, however, became what to do with significant surplus food production. The answer was to distribute it to not only our own poor but the world’s poor as well.”

A major expansion of food assistance programs took place in the 1960s, establishing the Food Stamps Act in 1964.⁶ From an archived USDA Employee News Bulletin from January 13th, 1965, a quote from then-Secretary Freeman: “Food is for raising the sights of children everywhere above their bellies, so they may fully exploit the potentials of their minds and their hearts” (page 2).⁷

This is where I paused my reading, but one interesting point from a 2012 USDA video sharing the perspective of former USDA secretaries, the USDA had a role in disaster relief for Hurricane Katrina, “delivering tons and tons of food to people who desperately needed it”. My key point here, in case I’m losing it, is that the USDA has historically acknowledged that building effective food systems is not contained to analyzing farm productivity.

Does the USDA have more timely and accurate data sets?

Well, they certainly didn’t mention any in the News Release. The USDA - just 8 months ago - also said they don’t. No one has brought up any since.

Does acknowledging any of this matter?

If solving the problem is not a priority, collecting information about it certainly won’t be.

The OMB Extension Request has a single public comment attached to it. The USDA specifically noted this and described it as “concerns about the accuracy and reliability of statistics produced using the Current Population Survey and Food Security Supplement” (Section 8).

I was naive enough to think that perhaps here I would find an insightful analysis on the limitations of the data that the USDA referenced in the cancellation Press Release. Perhaps there were methodological errors or inaccuracies that were not resolved.

The comment?

Subject: work for your food - stop the giveaways

the us census has to start saying no to all these surveys. they are not qccurate. they are not necessary. thiey are just giveaway surveys for a fat cat bureaucracy that wants to make up a reason for not being sunshined to extinction. it is time to shut down this survey. it is wasteful. it is not accurate. it is wasteful.

- FRN Public Comment