Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →From 2015 through 2024, actual dryland county yields averaged above USDA Risk Management Agency (RMA) projected yields for corn, soybeans and wheat, but below projections for upland cotton. In a follow-up comparison, a five-year moving average that excluded the lowest yield was generally more accurate for dryland corn, soybeans and wheat than the RMA projections. These are retrospective results from selected counties—not a forecast of future accuracy or a guarantee about any farm’s yields or insurance payments.
What the study measured
The April 1, 2026 analysis compared RMA county yield projections with actual RMA county yields for crop years 2015–2024. It included counties with actual RMA yield data for every year since 1991 and weighted county averages by each included county’s share of insured acres in 2024. The reported results therefore describe this selected, insured-acre-weighted sample, not all U.S. counties or individual farms. farmdoc daily’s April 1 analysis reports the following average differences:
| # | Preview | Product | Price | |
|---|---|---|---|---|
| 1 |
|
Elements of Forecasting | $295.92 | Buy on Amazon |
| 2 |
|
Forecasting for the Pharmaceutical Industry: Models for New Product and In-Market Forecasting and... | $81.67 | Buy on Amazon |
| 3 |
|
Bankruptcy, Credit Risk, and High Yield Junk Bonds | $77.34 | Buy on Amazon |
| Crop | Counties in sample | Actual dryland yield relative to RMA projection, 2015–2024 |
|---|---|---|
| Corn | 1,907 | 7.6% above |
| Soybeans | 1,549 | 4.8% above |
| Wheat | 1,487 | 1.9% above |
| Upland cotton | 463 | 8.8% below |
“Above” and “below” describe the average comparison between actual county yields and projected county yields over the period; they do not mean every county or crop year had the same result. The analysis also does not establish why projections differed from outcomes.
How alternative yield projection methods compared
The April 15 follow-up compared four approaches for dryland corn, soybeans and wheat. Its measure is actual RMA yield divided by projected yield, minus 100%; values closer to zero indicate a smaller average difference under that measure. The figures below are average percent differences over 2015–2024, as reported by farmdoc daily’s comparison of yield projection methods.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitches#1 Best Overall
| Crop | RMA | Ordinary least squares (OLS) | Five-year Olympic moving average | Five-year moving average excluding the minimum |
|---|---|---|---|---|
| Corn | 7.2% | 5.7% | 7.0% | 2.4% |
| Soybeans | 6.2% | 5.1% | 5.5% | 1.6% |
| Wheat | 5.2% | 4.6% | 7.2% | −0.6% |
What each method does
- RMA: The agency’s regression approach, adjusted to capture changing yield growth rates.
- OLS: An ordinary least squares regression trend.
- Five-year Olympic moving average: A five-year average that removes both the highest and lowest observations.
- Five-year moving average excluding the minimum: A five-year average that removes the lowest observation but retains the highest. The authors’ rationale is that low yields may depart farther from average than high yields.
Both regression and moving-average approaches incorporate trends. For the moving averages, the authors used a one-year lag because RMA county yield updates are released in June. For example, the 2024 projection uses county yields reported for 2018–2022.
Among these three dryland crops and four methods, the moving average excluding the minimum had the smallest absolute difference for each crop. The follow-up says the RMA projections were not the most accurate of the methods examined for any of the five crops in its broader comparison. The approach did not lead uniformly across every crop and production type; the article notes exceptions for cotton and rice payment comparisons. These period-specific results do not establish which method will perform best in future years.
Rank #2
Why the rating period matters
The follow-up reports that RMA projections fit the 1997–2014 actual yields used to rate the initial 2015 Supplemental Coverage Option (SCO) offering accurately, but their accuracy declined notably in the 2015–2024 comparison. That contrast is important when interpreting the results: the evaluation period differed from the historical period used for the initial rating. The analysis does not establish the cause of the change.
What the simulated insurance payments show—and do not show
The April 1 article also simulated payments under ECO 95% Revenue Protection (RP) for 2015–2024. Average payments as a percentage of projected insurance revenue were 2.1% for soybeans, 2.3% for corn, 3.2% for wheat and 5.0% for upland cotton. These are modeled outcomes for that coverage and period, not observed indemnities across all policies, producers or insurance products.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Rank #3
- Used Book in Good Condition
The authors say the payment patterns were consistent with the expected direction of yield differences, other factors equal, while emphasizing that projection accuracy is only one influence on payment rates. A separate illustration recalculated ECO payments as if each county’s projected yield equaled its average actual yield from 2015–2024. In that counterfactual, the range between the highest and lowest crop-average payment rates narrowed from 2.9 to 1.4 percentage points. It is a sensitivity illustration that relies on knowing future yields—not an available forecasting procedure or a proposed replacement method.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How to apply the findings to a coverage decision
These county-level comparisons are useful for understanding how projection methods behaved in the study period, but they cannot determine what a particular farm will yield or whether a producer will receive an indemnity. Area-based insurance results depend on the policy and its program rules as well as county outcomes. For example, USDA RMA says that, for GRIP, final county revenue is based on final county yield and harvest price and is used to determine whether an indemnity is due; the agency says final revenues for corn and soybeans are released by April 16 following the crop year. That is GRIP-specific information, not a rule to apply to every crop-insurance plan. See USDA RMA’s GRIP Final County Revenues page, and check current program details before making a coverage decision.
Quick Recap
- Use the study to compare historical county-level projection performance, not as a promise of a future result.
- Do not treat an average across selected counties as a forecast for an individual farm.
- Check the specific policy’s coverage terms and county data; the simulated ECO 95% RP payment figures do not describe all insurance outcomes.
- For a farm-specific coverage decision, consult a licensed crop-insurance agent who can review the policy and applicable program details.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
