Farm cash flow stress testing, explained.

Why a lender reruns a farm's year with cheaper corn and a short crop before renewing its line

Where the stress test sits in a farm's renewal, from harvest records to committee

1. The year closes and the records come inVoss Family Farms sends its 2025 Schedule F (the farm page of its tax return), balance sheet and grain inventory. An analyst spreads five years into the bank's format.
2. The farmer signs next year's planAcres, yields, rent, inputs and insurance for 2027, signed October 2, 2026, with the loan payments that fall due.
3. The lender stress-tests the planPrice, yield, cost and rate shocks, break-evens and a two-year liquidity check. This is where the $3.90 corn break-even is found, or missed.
4. The officer writes the recommendationRenew as asked, renew with conditions, or decline; below 1.00 in the bank's downside, the chief credit officer reviews first.
5. Committee decides, and the year happensHarvest, the October insurance price and next fall's government payments decide what the farm really repays.

Step 3 is the subject of this guide. The farm, the bank and the figures in the example are fictional.

Insurance that pays little, payments a year late: stress test surprises

Repayment capacity, coverage, working capital and the RP guarantee

Fourteen terms are defined in the farm stress test on one page.

Base case, standard shocks and break-evens: the three parts of a farm stress test

In shortThe base case is the farm's own signed plan at the bank's prices. Standard shocks move one thing at a time by a fixed amount, then several together, and the same set is applied to every borrower. Break-evens run the question backwards and find the price, yield, rent or rate at which coverage meets the policy line and 1.00. A two-year test then asks how much working capital would be left.

Every stress test starts from a base case, and the base case decides how honest the rest is. The FDIC's examination procedures ask whether "Cash flow projections use supportable production yields and price assumptions" and whether "Operating income/expense projections realistically compare to historical averages", before they ask about stress.6

Farm Service Agency rules for guaranteed loans say the same in regulatory language: a standard eligible lender's projections must rest on "the applicant's proven record of production and financial management", and "Lenders must use price forecasts that are reasonable and defensible."7 In practice a lender caps plan yields at the farm's own average and uses its own price deck, a table of planning prices approved by credit administration, so that two officers testing two farms start from the same prices.

Three kinds of test, and what each answers

Sensitivity tests are easy to explain and easy to compare across borrowers. Their weakness is that they treat each risk alone, while a drought year brings a short crop and, often, a higher price, and a big national crop brings a normal yield and a low price. The combined downside tries to capture the second kind of year.

The two-year test addresses what a single year cannot: the FDIC's examination manual warns that carryover debt, the operating debt a farm cannot repay from one year's sales, "tends to contain a greater degree of credit risk".8 A farm that survives one bad year on its working capital may not survive two.

Which shock bites which kind of farm

What examiners, FSA and the 2025 farm law ask of a farm cash flow test, as of October 2026

In shortNo regulation sets a coverage number. Bank supervisors expect lenders to analyze each borrower's cash flow on supportable assumptions and to test it under worse ones, and they examine the bank against its own policy. FSA requires a feasible plan for guaranteed loans. The 2025 reconciliation law changed the commodity programs and crop insurance that every farm cash flow now contains, and the farm bill itself was still unfinished at the end of October 2026.

What the 2025 law changed in a farm's cash flow

How to calculate term debt coverage and working capital for a farm, line by line

In shortStart from net farm income, add back depreciation and the interest on term debt, add off-farm income, take off family living and income taxes. That is what the year leaves to pay term debt. Divide it by the year's scheduled principal and interest on term loans and capital leases. Then roll working capital forward: last year's balance plus or minus the year's margin.

These are the Farm Financial Standards Council measures that extension economists and most ag lenders use.21,22 Purdue's Michael Langemeier describes the coverage ratio simply: "divide capital debt repayment capacity by principal and interest on term debt", and a ratio above one means the payments can be covered.22 A replacement margin goes one step further and also subtracts the cash a farm needs to keep its machinery from wearing out, which he suggests measuring as depreciation plus 10 to 20 percent.22

A worked example: Voss Family Farms, 2027 plan

The fictional case used through this guide. Every line traces to a document: the signed crop plan, the 2025 spread, the debt statements.

The margin, $131,678, is the number a stress test spends. Each shock is a dollar amount taken out of it: a 10% drop in both crop prices removes $198,370, more than the whole margin, which is why coverage falls to 0.78. A 15% short crop removes $302,597 of revenue. A 200-point rate rise on an operating line that averages $792,000 removes $15,840. Writing each shock as dollars against the margin makes the result easy to explain to a committee and to the borrower.

Liquidity: the second half of the test

Working capital is the farm's cushion. Iowa State's extension economists note that working capital "measured at the beginning of the year is typically equal to about 50–70% of the farm's annual gross revenue" on the Iowa farms they track, and that "cash grain farms may need as much as 50%".21 Lenders set their own floors; many use a share of gross revenue so that the floor scales with the farm. A two-year test subtracts the stressed margin twice from today's working capital and compares what is left with that floor.

Illinois shows the same slide. Across 2,018 Illinois farms in the Farm Business Farm Management association, average working capital was $361,826 at the end of 2025 and the current ratio 2.49, against 3.52 in 2022; on grain farms, "working cash has decreased almost 40 percent from 2022 on the average farm to 2020 levels".4 A cushion that thin turns a one-year shortfall into a renewal problem quickly.

Where a lender's price deck comes from, and how big a price shock history supports

In shortLenders build next year's planning prices from futures for the harvest months, less the local basis, and check them against USDA's WASDE and university budgets. In October 2026 those sources pointed to about $5 corn and $12 soybeans for 2027 in central Illinois. History says a 10% price shock is gentle: corn's harvest price fell by more than 18% in a single year four times since 2006.

A bank's price deck is its planning price for each commodity next year. The usual starting point is the futures contract for the harvest month (December corn, November soybeans), less the typical difference between that futures price and the cash price local elevators pay at harvest, the basis.

The University of Illinois economists who publish the state's crop budgets build theirs the same way: "Price assumptions are based on current harvest futures for the 2027 crop, adjusted for typical basis in Illinois."9 USDA's monthly World Agricultural Supply and Demand Estimates (WASDE) give a second, national view of the season-average price.

The last two numbers together explain the mood of 2026: planning prices sit only just above the cost of growing the crop on rented land. The same authors warn that "a $0.50 move in corn prices, which is less than what has been experienced in harvest futures just over the past month, results in a more than $100 shift in returns."9 A deck set a little below the futures curve, as many banks prefer, already builds in some caution; the stress test adds the rest.

How big a shock history supports

Read against that history, a 10% price shock is a common year, and the bank downside of 7.5% is milder still, because it adds yield and rate moves on top. Two lessons follow. First, show the break-even price beside every shock, so the committee sees the room as a distance: Voss stops repaying 10.4% below the deck. Second, test the next year as well, because the big falls of 2013–2014 and 2023–2024 came in pairs.

Short crops and costlier inputs: sizing the yield and cost shocks from USDA's own data

In shortA 15% yield shock is roughly what the 2012 drought did to the national corn yield; on one farm, losses can be far deeper. Input costs moved even more: USDA's corn operating cost per acre rose 47% from 2021 to 2022 and is forecast near $473 for 2026 and 2027. A cost shock matters most to farms that rent their land, because their margin is thinnest.

National yields hide most of a single farm's risk, yet they show the scale. ERS's series puts the U.S. corn yield at 146 bushels an acre in 2011 and 118 in the 2012 drought, a 19% fall.24

On one farm a hailstorm or a dry July can take more than that, which is what crop insurance exists for. A sound plan yield is the farm's own average, capped there even when the borrower expects more; the FDIC's examiners look for "supportable production yields".6 The Voss plan uses 218 bushels against a five-year average of 219 and an insurance history of 221.

What it costs to grow an acre of corn

The gap between the two bars is what pays for land, machinery, the family and the debt. In 2025 it was about $322 an acre, while ERS put the total cost of an acre, land and machinery included, at $900.52.24 For 2027 ERS forecasts operating costs of $473.34 an acre and total costs of $952.09.25

How Revenue Protection puts a floor under a farm's revenue, and where the floor sits

In shortRevenue Protection guarantees the farm's history yield, times its coverage level, times the higher of the spring and harvest futures prices. It pays the difference only when the crop's value at the harvest price falls below that guarantee. A modest price drop on a normal crop usually stays above the guarantee and pays nothing; a deep drop or a short crop meets the floor. That floor is why a farm's coverage curve goes flat.

The definitions are in the policy's Basic Provisions, last printed in the federal regulations in the 2026 edition; since May 1, 2026 USDA publishes the policies on its Risk Management Agency site instead, with coverage unchanged.20 The Basic Provisions define the revenue protection guarantee as the production guarantee multiplied "by the greater of your projected price or your harvest price", and the harvest price as the price "used to value production to count".18 The coarse grains provisions settle a claim by subtracting the value of the production to count, at the harvest price, from the guarantee.19

For corn the projected price is the average of the December futures contract over February and the harvest price its average over October.12,2 In February 2026 those rules set corn at $4.62 and soybeans at $11.09.12

The 2025 example: a 10% price fall that paid little

In 2025 the corn projected price was $4.70 and the harvest price $4.22, 10.2% lower; soybeans fell 1.8%, from $10.54 to $10.35. University of Illinois economists worked out that at 80% coverage, corn yields had to be "more than 10.9% below APH" before Revenue Protection paid, and soybean yields more than 18.5% below.2 They wrote that payments in Illinois "are not expected to be widespread or large".2 A lender who had counted insurance as covering a 10% price fall would have been wrong in most of the state.

The floor on one farm

The curve has three parts. Above $4.16 the farm meets the bank's 1.25. Between $3.90 and $4.16 it repays, with less room than policy wants. Below $3.90 it no longer covers its term payments from the year, and below about $3.57 insurance makes up the corn revenue lost, so coverage stops falling at 0.69. Soybeans behave the same way at their own guarantee, $563 an acre: Voss's soybean price alone cannot push coverage under 1.00, because the floor catches it first.

ARC, PLC and ad hoc aid after the 2025 law: what to count in a farm cash flow, and when

In shortGovernment payments made up much of 2026 farm income: ERS forecasts $47.4 billion, of which $26.5 billion is ad hoc and disaster aid. ARC and PLC are written in law to 2031 and pay a year after the crop. A stress test should count ARC and PLC in the year the cash arrives, estimated at the stressed price where the program would pay more, and leave unenacted aid out.

ERS's September forecast puts net farm income at $158.4 billion in 2026, down 2.6% from 2025, with direct government payments up 69.8% to $47.4 billion.28,26

Of that, $26.5 billion is supplemental and ad hoc disaster assistance, mainly the Farmer Bridge Assistance Program and the American Relief Act's supplemental disaster assistance, and $15.6 billion is commodity program payments tied to prices or revenue, mostly ARC and PLC, up $13.4 billion largely because the 2025 law changed the 2025 crop year ARC and PLC programs.26 A Kansas banker told the Kansas City Fed that "Government ad hoc disaster payments are keeping conditions manageable for most producers."29

The 2025 crop's record payments, paid in October 2026

farmdoc estimates total 2025 payments at just over $13.2 billion against USDA's publicized $13.8 billion; either is a record. Corn base acres collect about $4.7 billion, soybeans $1.5 billion and wheat $2.9 billion, and "producers should receive by the end of October".3 For 2026 farms choose ARC or PLC again by December 11, 2026.3

Some summaries still quote an April 2026 reading of USDA's earlier forecast, with $624.7 billion of farm debt and $44.3 billion of payments.30 ERS's September 3, 2026 forecast replaced those figures.28

Interest-rate shocks on a farm operating line: what 2022–2026 rates teach about the right size

In shortMost farm operating lines float over prime, so a rate shock passes straight into interest expense on the average balance. Prime rose 5.25 points in about sixteen months in 2022–2023, fell back to 6.75%, and rose to 7.00% in September 2026. Farm interest expense is forecast at a record $33.4 billion in 2026, 76% above 2021.

Ag lenders in the Seventh Federal Reserve District reported average operating-loan rates of 7.63% in the second quarter of 2025, 7.08% in the first quarter of 2026 and 7.12% in the second.31,32 A farm with a $792,000 average balance pays $7,920 more interest a year for each percentage point. That is why a 200-point shock barely moves Voss (1.43 to 1.37) while a 10% price move takes it under 1.00. The farms that rates do hurt are those with large variable-rate term debt, big carried inventories financed on the line, or a carryover balance that has stopped revolving.

Against ERS's earnings before interest, taxes and capital consumption of $241.0 billion for 2026, interest takes about 13.8%, up from 9.3% in 2021 (our arithmetic from ERS's data).27 Farm debt is forecast at $605.1 billion and the sector's debt-to-asset ratio at 13.54%.28 The sector average is low; the farms a lender stress-tests hardest are the leveraged ones within it.

Weaker repayment, more renewals, thinner working capital: the Fed's 2024–2026 ag credit surveys

In shortRepayment weakened through 2025 and early 2026 and then eased a little. In the Chicago Fed's district, 38% of lenders saw lower repayment in the first quarter of 2026, renewals and extensions hit their highest level since 2020, and 17% of borrowers carried more carryover debt. By the second quarter the share seeing lower repayment had fallen to 27%. Lenders describe 2026 cash flows at or below breakeven, funded from working capital.

The Chicago Fed's May AgLetter quoted an Iowa lender: "cash flow projections for many operations are at or below breakeven for 2026." Lenders reported many borrowers drawing down working capital to cover the shortfalls, a repayment index of 63, down from a year earlier for the tenth quarter running, and a renewals and extensions index of 136, the highest since the second quarter of 2020.31 By August the repayment index was 73 and renewals 124, still worse than a year earlier but less so.32

Losses are still low, which is when stress tests earn their keep

Problem loans remain modest. At June 30, 2026, FDIC-insured institutions held $89.1 billion of farm loans, and at agricultural banks the noncurrent rate in the category the FDIC calls all other loans and leases, which holds farm production loans, was 0.68%.35 Farm Credit System nonperforming assets were $5.0 billion, 1.09% of loans.36

In Illinois, net farm income on Farm Business Farm Management farms recovered to an average $97,674 in 2025 from $18,358 in 2024, though returns to the operator's labor and management were still negative.37 Credit trouble in agriculture usually lags the income squeeze by a year or two, as working capital runs down; the surveys describe that middle stage.

Ten ways a farm stress test flatters the borrower, and the check that catches each

In shortMost errors make a farm look safer. Insurance counted where it would not pay, government money counted in the wrong year, yields above the farm's history, family living left low, and a one-year test where two bad years are the real risk. Each has a simple check, and most can be built into the template once.

Langemeier's 2016 case farm shows how fast the arithmetic turns. On a 3,000-acre west central Indiana farm with 2,250 acres cash-rented, a price 10% below futures took projected net farm income from about $48,000 to a loss of about $127,000, and made repayment capacity "worrisome".38 The more of a farm's costs are fixed in rent and payments, the more of a price move reaches net income.

Software and AI in farm stress testing: what the 2026 model guidance covers, and what stays with the lender

In shortReading the farm's records, rebuilding the cash flow and solving break-evens are arithmetic and fixed rules, which the April 2026 model risk guidance places outside its definition of a model. Generative AI is outside that guidance altogether and falls to the bank's own governance. What supervisors still examine is the analysis: supportable assumptions, documented shocks, and a person accountable for the recommendation.

SR 26-2, issued on April 17, 2026 with OCC Bulletin 2026-13 and FDIC FIL-15-2026, defines a model as "a complex quantitative method, system, or approach that applies statistical, economic, or financial theories to process input data into quantitative estimates", and excludes "simple arithmetic calculations, such as those found within spreadsheets, as well as deterministic rule-based processes".13 A footnote places generative AI, and AI that acts on its own, outside the guidance's scope, while the bank's own risk management "should guide the determination of appropriate governance and controls" for them.13

Three properties make an automated test checkable. Every line of the rebuilt cash flow should show its formula and the document it came from, so a committee member can follow the $440,714 to the plan, the spread and the statements. Prices should come only from the bank's approved deck or from a value an officer typed, labeled as such. And the software should test and draft while a named officer recommends, because the FDIC's guidance puts the analysis on the bank's management, whoever builds the spreadsheet.1

• Software reads documents and does arithmetic. Judging the farmer's management, a landlord's plans or a family's succession remains the officer's work.
• Insurance and hedges count only when the policy and contracts are on file.
• A shock set is credit policy. Its owner, its version and its approval belong to the chief credit officer or a committee.

Who rebuilds the farm spreadsheet for every what-if today, and where the analyst's day goes

In shortAn analyst spreads the farm, a loan officer gathers the plan, someone rebuilds the cash flow in a spreadsheet for each scenario, and the officer carries the result to committee. The rebuilding is where the day goes, and it happens again for every new question. Renewals bunch up between harvest and planting, when every farm's question is asked at once.

Five roles around one farm renewal

From renewal file to committee, re-keyed for each scenario

An illustrative day, most of it re-keying

The fall calendar adds pressure. Farm records for the year just ended arrive in winter; many renewals are decided between harvest and the spring input purchases, and committees ask new what-ifs as markets move. A question asked on Thursday, such as "what does $3.90 corn do?", often comes back the following week, after the committee has decided.

Spreadsheets, university tools or lending software

Voss Family Farms' $1.45 million line: stress-tested on Monday morning, signed for Thursday's committee

In shortThis is the example from the top of the page, inside the product. Six agents build Voss Family Farms' 2027 cash flow from its own records before the officer arrives, run the bank's standard shocks and solve the break-evens. Erin Kowalski, the ag loan officer, tries corn 15% lower in her meeting with Mark Voss, asks for conditions to be tested, writes her recommendation and signs the note for committee.

1

Every borrower going to committee, already tested

Monday morning, October 12, the renewal list holds nine borrowers and $12.67 million of operating lines. Eight are tested against the bank's standard shocks; one is waiting for its 2025 tax return. Seven pass on their own plans at 1.25, and six fall below 1.00 in the bank's downside, so each needs working capital or conditions to carry it.

Voss Family Farms is first: plan 1.43, downside 0.47, debt break-even at $3.90 corn, note ready to sign. A dairy below 1.00 in the downside is already with the chief credit officer.

2

The farm, built from its own records in 45 seconds

At 7:12 a.m. the agents read the farm's 14 documents. The spread reader restated five years: coverage of 2.31, 2.96 and 2.04 from 2021 to 2023, then about 1.00 in 2024 and 1.12 in 2025, and working capital down from $1,212,000 at the end of 2022 to $742,000, 36% of revenue.

The crop plan and insurance reader took the signed 2027 plan and the Revenue Protection policy; the debt schedule builder listed $309,036 of payments. The base case: 1.43 against the bank's 1.25.

3

In the meeting: corn 15% lower, and the curve crosses 1.00

The what-if screen shows the farm's 2027 cash flow on seven faders, with coverage drawn as a curve against the corn price. The corn fader already carries a red tick at $3.90 and an amber one at $4.16. Erin moves corn down 15%, to $3.70: capacity falls to $250,836 against $309,036, coverage to 0.81, a margin of −$58,200.

Below about $3.57 the curve turns flat at 0.69, where the insurance policy makes up the lost revenue. She saves "Corn $3.70" to the credit file, and it joins the comparison and the note.

4

Conditions, tested before they go in the note

In the bank's downside the farm covers 0.47, and working capital would fall to $413,230 after two such years, 23% of revenue, under the bank's 25%. Erin asks the assistant to add three conditions: 85% Revenue Protection, 35% of expected corn forward-sold at the co-op's $4.52 bid, and the combine note re-amortized over seven years.

They land as a new column: insurance pays $31,254, payments drop to $283,595, and coverage in the downside rises to 0.73. Two such years leave $588,838 of working capital, 32% of revenue. She adds a fourth condition: working capital of at least $500,000 at each year end.

5

A cited note, signed for Thursday

The note drafter wrote the request, base case, stress table, break-evens and liquidity, and the four conditions, with their sources. The recommendation is Erin's, in her words: renew with the four conditions, because the plan meets policy and, with them, working capital carries two downside years.

She signs at 10:35 a.m. with the meaning "Reviewed and recommended". The note goes to the credit file and the October 15 committee packet, and the activity record shows each agent's step and each person's action with the time.

The solution's design target is under a minute from a borrower's file to a full stress test, against about a day by spreadsheet; it is a target, to be measured on your own renewals.

Twelve screens behind the Voss answer: line-by-line working, a cattle farm with no insurance floor, break-evens and shock settings

In shortOne renewal is the story. These screens show the rest: the what-if on plan, the cash flow worked out line by line, the stress calculator, a cattle farm with no insurance floor, the break-even grid, the coverage chart, the crop plan and debt, the standard shocks and price deck, the bank's thresholds, the signature and the saved scenario.

Select any screen to open it larger. The bank, farms, people and figures are fictional; the price deck and insurance planning prices are the bank's sample assumptions; real market prices are in Part A.

Who does what in a stress-tested renewal: six agents compute and draft, the loan officer signs, the committee rules

In shortEach agent has one job and a limit it does not cross. None writes the recommendation or changes the bank's shocks, and none counts insurance or forward sales that are not on file. The loan officer writes the recommendation and signs; the committee decides; renewals below 1.00 in the bank's downside go to the chief credit officer first.

Agents that read spreads, shock plans and find break-evens

The officer's signature, the credit chief's review, the committee's decision

Every agent run, saved scenario and signature is on the borrower's activity record with who, what and when; threshold changes are versioned. Agent steps and people's decisions are kept apart.

Erin, Ruth and Mark: the fictional people on the screens

What the stress test reads from a renewal file and your credit policy, and the signed scenario note it leaves

In shortIt reads the documents the bank already collects at renewal, uses the bank's price deck, shocks and thresholds, and leaves behind the farm's base case, every scenario with its working, break-evens, a two-year liquidity view and a signed, cited note. The loan system and the credit file stay where they are.

The systems a farm stress test draws on

For each, check whether it has an API, a file feed, or only screens, and who owns it.

Your deck, shocks, thresholds and loan system stay yours

Three rules behind every coverage number

What a stress test rebuilt by hand costs an ag lender in hours, late answers and missed carryover

In shortStart from your own numbers, because nobody publishes the hours a farm stress test takes or how often projections are wrong. A few published figures frame them: what an analyst hour costs, how many lenders saw repayment weaken in 2026, and how much interest the farm sector now carries.

Clocking analyst re-keying and next-day committee answers

What a thin farm renewal costs that no calculator counts

Analyst hours released per renewal, and how many renewals pay for automated stress testing

In shortPut in your own numbers. The calculator starts from a published analyst wage and asks for the rest: how many renewals you test, the hours a test takes now and after, how many extra what-ifs each renewal brings, and what the tool and your testing cost a year. Every default marked illustrative is an assumption to replace.

Six places automated farm stress tests return value

The calculator: renewals, hours and the volume that covers the cost

At the defaults (eight hours today, one after, two extra what-ifs of two hours each), a renewal releases eleven hours. A bank testing 150 farm credits a year passes the 96 it needs; a bank testing 40 large, simple credits may not cover the cost, and should know that before buying anything.

How ready your shock set, price deck and renewal files are for automated farm stress testing

In shortPick the answer closest to your team today. Eight questions score 0 to 2; the last is for reflection. The meter adds up your score as you go, and each question says why it matters and what a strong answer sounds like.

What your readiness score says about the next step

Vendor questions that expose a weak farm stress test: insurance floors, payment timing, whose price deck, and evidence

In shortAsk every vendor the same questions, us included. The answers show how the farm's year is built, how insurance and government payments are counted, whose prices and shocks apply, what evidence you keep, and how you would measure it on your own files.

We found no published independent test of any farm stress testing tool's accuracy, as of October 2026. A side-by-side run on your own renewals is the fairest comparison, and the same run measures your current process.

A twelve-week side-by-side run on this season's renewals, ending in three numbers

In shortWrite down your shocks and measure your baseline, then run beside your current process on live renewals, then decide on three numbers. Renewal season is the right time: the files are fresh and every committee asks the question.

1. Write down your standard shocks and price deck, with an owner and a version, and the thresholds the committee uses.
2. Time three renewals from complete file to committee, and count how many went back for another scenario last season.
3. Choose 30 to 50 renewals across grain, livestock and mixed farms, with complete files and an approval path to share them.
4. Run in parallel for six weeks, with officers signing every note and analysts checking every difference.
5. Decide on three numbers: hours per renewal, questions answered in the meeting, and differences in coverage explained by cause.

Five numbers to record during the side-by-side run

Break-evens, crop insurance, ARC/PLC and coverage ratios: farm stress testing questions answered

In shortThe questions ag lenders, analysts and farm borrowers search for most, each answered on its own.