POLICY • NEWS • MARKETS
AG POLICY & MARKETS DAILY
MONDAY, AUGUST 17, 2026 | SPECIAL REPORT & ANALYSIS
POLICY ANALYSIS | USDA STATISTICS
One Farmer’s Blunt Message to NASS: The Survey System Is Wearing Out Its Sources
Response rates will not recover until USDA cuts burden and restores trust.
Analysis · August 17, 2026
One farmer is not a representative sample. But one producer’s unusually candid response to the National Agricultural Statistics Service’s (NASS) survey problems identifies several reasons farmers may increasingly ignore questionnaires — or complete them with less care than the statistical system requires.
The farmer is not arguing that agricultural statistics have no value. He sees crop yield forecasts as a legitimate market-facilitating function that would otherwise be left largely to private companies, commodity analysts and trading firms. His complaint is about the bargain USDA presents to the people supplying the underlying information.
Farmers have often been told they should answer NASS surveys because USDA programs benefit agriculture. That is partly correct at the broader institutional level. NASS data are used by the Farm Service Agency and Risk Management Agency for program administration, yield benchmarks, price elections and other purposes. But the Agricultural Yield survey has a more specific function: It gathers farmer estimates of expected yields for monthly crop production forecasts. The survey program is voluntary, and completing an individual yield questionnaire is not a condition for receiving farm program assistance.
That distinction matters. Producers are not repaying USDA for a subsidy when they complete a yield survey. They are donating time, business knowledge and agronomic judgment to create a public information product.
The Benefit Is Real — but Indirect
USDA can make a strong public interest case for that product. Independent government estimates prevent national supply information from becoming the proprietary property of a few large grain companies, trading houses and private forecasting services. NASS describes its estimates as an official, independent baseline available to all market participants.
But equal publication does not necessarily mean equal benefit. Everyone may receive the number at the same moment, but commercial traders, algorithmic funds and large market information firms generally possess greater capacity to analyze and trade it immediately. Farmers may eventually benefit through more transparent bids, crop-insurance calculations and better-informed advisers, but that connection is less visible.
| THE EXCHANGE | THE PRODUCER WHO FILLS OUT THE FORM | THE PROFESSIONAL MARKET USER |
| What it costs | Stops work, locates records, estimates uncertain outcomes, meets a deadline | Nothing; the number arrives ready to use |
| When the benefit arrives | Later and indirectly — through bids, insurance calculations and better-informed advisers | Immediately, at the moment of release |
| Capacity to act on release | Limited; not a direct user of the Crop Production report | Analysts, algorithmic funds and information firms can trade it at once |
| Visibility of the benefit | Low — the answer disappears into a national data product | High — a priced, tradable input |
Table 1. Concentrated costs, diffuse benefits: why the bargain looks worse from the tractor seat than from Washington. Source: Ag Policy & Markets Daily analysis of the producer’s critique and NASS statements at USDA’s April 2026 Data Users’ Meeting.
NASS officials have acknowledged the problem. At USDA’s April 2026 Data Users’ Meeting, a NASS official said farmers generally are not direct users of the Crop Production report. Instead, they rely on advisers and information services whose analysis ultimately rests on NASS data. The official conceded that this indirect connection makes it more difficult to explain why farmers should complete the surveys.
The farmer’s perception that speculators capture the immediate benefit may not tell the whole story, but it cannot be dismissed as merely a communications failure. From his perspective, the costs are concentrated while the benefits are diffuse: The producer must stop working, locate records, estimate uncertain outcomes and meet a deadline, while the resulting information disappears into a national data product used most visibly by market professionals.
A Short Survey Can Still Be One Survey Too Many
USDA frequently responds to complaints by shortening questionnaires, improving online reporting and stressing that a particular yield survey takes only a few minutes. The farmer’s response explains why that may not solve the problem.
Farmers do not experience each NASS instrument in isolation. They experience the cumulative stream of envelopes, telephone calls, reminders and online requests arriving from the same agency.
NASS says it conducts hundreds of surveys each year covering production, economics, demographics and environmental issues. No individual operation receives all of them, but some producers—particularly larger or more diversified farms—can be selected repeatedly.
A two-page yield questionnaire may be simple by itself. But it arrives against a backdrop that can include lengthy surveys requiring acreage records, input purchases, labor information, landlord arrangements and financial data. The producer described opening questionnaires running 15 to 20 pages, realizing that records would have to be retrieved and setting the form aside for later. By the time “later” arrives, the reporting window may have closed.
That is a different kind of burden than USDA’s estimated minutes per questionnaire. The current federal approval for the Agricultural Surveys Program covers roughly 1.29 million annual responses and 206,959 burden hours — an average of about 9.6 minutes per response. But an average across many different questionnaires conceals the concentration of burden on repeatedly selected operations and the much larger amount of record-retrieval time required by complex surveys.
Figure 1. The burden a producer feels is a sequence of tasks; the burden USDA counts is an average. Source: Ag Policy & Markets Daily, from the current Agricultural Surveys Program clearance (about 1.29 million responses, 206,959 burden hours) and the producer’s description of 15- to 20-page questionnaires.
For a farmer, burden is not merely the number of questions. It includes stopping another task, determining which records are needed, locating those records, interpreting USDA terminology and deciding whether an answer is sufficiently reliable.
USDA Is Sometimes Asking for a Forecast, Not a Fact
The producer’s strongest objection concerns questions about expected harvested acreage and yield. To an analyst in Washington, those may look like two straightforward numerical entries. To a producer, they are conditional judgments about events that have not happened.
A field intended for grain today could be grazed later. Hail could reduce harvested acreage. Rain during the next several weeks could add bushels, while heat or disease could remove them. A producer may know planted acreage precisely but have no defensible point estimate for final harvested acres or yield.
“I don’t own a crystal ball,” the farmer said.
That does not mean NASS should stop asking for expectations. Forecasting requires informed judgments before final outcomes are known. But questionnaires should recognize the uncertainty embedded in those answers. Requiring a single number can create an illusion of precision and make the respondent feel that USDA is demanding a fact that does not yet exist.
| WHAT THE FORM ASKS FOR | WHAT THE PRODUCER ACTUALLY KNOWS | WHAT THE FORM COULD ALLOW INSTEAD |
| Expected harvested acreage, as one number | Planted acreage precisely; but a field intended for grain could be grazed, and hail could cut harvested acres | A range, or a “not yet reasonably knowable” response |
| Expected yield, as one number | Rain in the next several weeks could add bushels; heat or disease could remove them | A range with a confidence level |
| An implied commitment to the estimate | Intentions that may still change with weather, price and feed demand | Whether the harvest intention has changed, and which risks could still alter it |
| No question about the basis of the number | Whether the figure is a measured field estimate, a crop-insurance appraisal or a general impression | Ask the producer to identify which of the three it is |
Table 2. Fact versus forecast: what a single required number conceals. Source: Ag Policy & Markets Daily analysis based on the producer’s objections and the Agricultural Yield survey’s expected-yield questions.
Forms could instead permit ranges, confidence levels or a “not yet reasonably knowable” response. NASS could ask whether the producer’s harvest intention has changed, which risks could still alter it and whether the reported yield represents a measured field estimate, a crop-insurance appraisal or a general impression.
Such answers would be more complicated to process, but they would more honestly represent what the farmer knows.
Careless Answers May Be More Dangerous Than Missing Ones
The most serious part of the farmer’s response was not his willingness to ignore a survey. It was his suggestion that some producers may enter arbitrary numbers simply to complete it.
A missing questionnaire is visible. NASS knows that it did not receive the response and can pursue follow-up contacts or account for nonresponse statistically. A plausible but careless number can be harder to detect because it enters the system looking like a genuine producer estimate.
| A MISSING QUESTIONNAIRE | A CARELESS NUMBER | |
| Visible to NASS? | Yes — the agency knows the response never arrived | No — it looks like a genuine producer estimate |
| Available remedy | Follow-up contacts; statistical adjustment for nonresponse | None; the value is already inside the estimate |
| Effect on other respondents | None directly | Corrosive — conscientious producers question why they should retrieve records |
| Effect on the estimate | Manageable and measurable | Unmeasured error carried forward into the published forecast |
Table 3. The two failure modes are not equally detectable, which is why a falling response rate understates the risk. Source: Ag Policy & Markets Daily analysis.
The farmer also described a corrosive peer effect: If a producer believes neighbors are simply making up answers, the incentive to spend time producing an accurate response declines. One conscientious respondent begins to question why he should retrieve records and struggle with uncertainty when others may be scribbling down guesses.
Whether that belief accurately describes other farmers is almost beside the point. Once producers lose confidence in the seriousness of the respondent pool, voluntary cooperation can become a collective-action problem. Each farmer concludes that his own effort is unlikely to improve the final estimate enough to justify the cost.
NASS maintains that response rates alone do not determine statistical quality. At the April data meeting, agency officials said past studies found little evidence of substantial nonresponse bias. Those are important qualifications. But the agency also acknowledged that participation has declined and that maintaining good response remains essential.
Both propositions can be true: NASS may still produce credible estimates, while the system supporting those estimates is becoming less durable.
Increasing the Sample Could Deepen the Problem
USDA’s immediate response to falling participation has included contacting more producers. In April, the Office of Management and Budget approved a NASS request to increase sample sizes specifically to counter declining response rates. The change raised approved annual responses in the Agricultural Surveys Program by nearly 103,000 and added more than 16,000 burden hours.
Figure 2. The April 2026 clearance change: a larger sample to defend the same number of usable reports. Source: Ag Policy & Markets Daily calculations from the Agricultural Surveys Program information-collection approval; “before” values are the approved totals less the increase.
That is statistically understandable. When the percentage of respondents falls, a larger initial sample may be needed to obtain the same number of usable reports.
Institutionally, however, it risks creating a feedback loop. Lower participation leads to larger samples, larger samples lead to more contacts, and more contacts reinforce the sense that agriculture is being over-surveyed. USDA could end up solving a response-rate problem by expanding one of its underlying causes.
Figure 3. The institutional feedback loop the farmer describes: each turn of the cycle makes the next round of cooperation harder to obtain. Source: Ag Policy & Markets Daily analysis.
NASS Should Budget Farmer Time as Carefully as Federal Money
The farmer’s proposed field hearings should be taken seriously. USDA needs to hear not only whether individual questions are understandable, but how the entire survey system fits into the operating calendar of a commercial farm.
A redesign should begin with a respondent-level burden budget. NASS should know how many times each operation has been contacted across all surveys, how many minutes and record searches have been requested and whether those requests coincide with planting, harvest, calving or tax deadlines. Producers should receive an annual contact calendar, greater ability to choose reporting periods and protection from repeated selection for record-intensive surveys.
USDA should also prepopulate everything it already knows. NASS’s August crop briefing showed that acreage and yield estimates now integrate FSA-certified acreage, RMA-insured acreage, satellite imagery, field observations and farmer survey data. Administrative and remotely sensed information cannot fully replace timely producer knowledge, but it can reduce repetitive questioning and reserve farmers’ time for information only they can provide.
| PROPOSED CHANGE | WHAT IT IS MEANT TO FIX | DIFFICULTY |
| Field hearings on the survey system as a whole | USDA hears about individual questions, not about the calendar of demands on a farm | Low |
| A respondent-level burden budget | Burden is tracked per questionnaire, not per operation across all surveys | Moderate |
| An annual contact calendar and choice of reporting periods | Requests that collide with planting, harvest, calving or tax deadlines | Moderate |
| Protection from repeated selection for record-intensive surveys | Concentration of burden on larger and more diversified operations | Moderate |
| Prepopulation from FSA, RMA, satellite and field observations | Repetitive questioning about facts USDA already holds | Moderate to high |
| Compensated panels for the most demanding recurring surveys | Uncompensated professional time with a real opportunity cost | High — needs funding |
| Retire the “you owe it because agriculture gets assistance” argument | A rationale that is inaccurate for a voluntary survey and corrodes trust | Low |
Table 4. A redesign agenda implied by the producer’s critique, with a rough judgment of what each step would require. Source: Ag Policy & Markets Daily analysis.
For the most demanding recurring surveys, USDA should consider compensated producer panels, some observers note. Payment would not be practical or necessary for every short questionnaire, but farms repeatedly asked to retrieve financial records or provide monthly forecasts are supplying a valuable professional service. Compensation would recognize that modern producers are business operators whose time has an opportunity cost.
Most importantly, USDA should stop relying on the argument that farmers owe survey responses because agriculture receives government assistance. The stronger and more accurate message is that independent agricultural statistics are a public utility—and that USDA intends to treat the people producing the raw information as partners rather than an unlimited source of free labor.
Analysis: What to Watch Next
The next test is procedural, and it is already scheduled. Sample expansions and questionnaire changes move through Paperwork Reduction Act clearance, and every renewal opens a public comment window. Watch whether the next Federal Register notice for the Agricultural Surveys Program reports burden the way farmers experience it — contacts and hours per operation — rather than only as an average per questionnaire. That single change in accounting would make the concentration problem visible to OMB, and it costs nothing.
The second thing to watch is money. A contact ledger, prepopulated forms and paid producer panels are all investments in respondent relations at a moment when NASS has been trimming rather than adding. Reduced-cooperation costs are easy to defer because they show up as a slow erosion in data quality rather than as a line item. If Congress wants credible acreage and yield numbers, maintaining the respondent base is far cheaper than rebuilding it.
Third, watch the mix. As FSA-certified acreage, RMA data and satellite imagery carry more of the estimate, the share genuinely coming from producer questionnaires falls. That shift is defensible, but it deserves to be stated plainly. Farmers asked to keep answering are entitled to know which questions still move the number and which have been superseded by records USDA already holds.
Finally, treat the nonresponse-bias reassurance as perishable. Studies finding little bias were conducted when cooperation was higher and when the producers who still respond looked more like the ones who do not. As participation drifts down and concentrates among the most cooperative operations, that finding needs re-testing rather than repetition. The most useful metric for USDA to publish is not the response rate at all; it is the number of times the average operation was contacted last year — and whether that number is finally falling.
Bottom line
The farmer is not saying NASS data do not matter. He is warning that they matter too much to depend indefinitely on an over-contacted, uncompensated and increasingly skeptical respondent base.
The farmer concluded: “I don’t think this is about whether USDA/NASS has a division that loves nerding out on statistics… and whether those folks decided to retire early. They can throw all the Bayesian psychobabble they want at it… but it’s more fundamental than that. And I don’t think that conversation is happening.”
NASS cannot remind its way out of that problem. It must redesign the relationship.
AG POLICY & MARKETS DAILY | POLICY ANALYSIS | USDA STATISTICS — MONDAY, AUGUST 17, 2026
