Buying Farmland Directly

A farmland auction in the Corn Belt takes about forty minutes and is over before an equity trade would have cleared compliance. The parcel is described, the survey and the soil map are on the wall, the auctioneer opens at a number nobody accepts, and then two or three neighbours who have farmed within sight of the field for thirty years bid it up in $100-an-acre increments until one of them stops. In April 2026 an Iowa farm in Sioux County with a CSR2 soil rating of 91.4 sold for $26,600 an acre — about two and a half times Iowa’s 2026 statewide cropland average of $10,700, and the equivalent of $291 per CSR2 point; a Floyd County farm rated 85.0 sold for $131 a point in the same market. Same state, same crop, same year, and more than twice the price for a point of measured corn suitability, because the second farm did not have the first farm’s neighbours.

That is how most of this asset changes hands. Roughly 79% of Iowa farmland trades at auction and about 21% through a private listing, by the tally of brokers who run those sales, which means the price you pay is set in public, in a room, by people who know the ground better than you do. It also means almost none of it is for sale: farmland turnover in Illinois has clustered between 1.3% and 1.6% of acres a year since 2003, by a farmdoc study published in April 2026. You are not buying into a market with a bid and an ask. You are waiting for a specific field to become available and then competing for it against the person who already farms it.

This guide is what happens after the gavel: how an acre is valued, how it is leased, what it costs to finance and carry, how it is taxed, what water and foreign-ownership law do to who may buy it and what it is worth, and what the arithmetic looks like on 80 acres held for a decade. Investing in Farmland, the hub’s flagship, covers the asset class as a whole and its long-run record; Investing in Farmland Through REITs and Funds, Investing in Timberland and Investing in Water Rights are the packaged versions of the same exposure. This one is the unpackaged version, where there is no manager between you and the tenant.

Six consecutive annual increases have taken average US farm real estate to $4,500 an acre, and the number that matters to a direct buyer is not the average but the spread beneath it.

USDA’s National Agricultural Statistics Service publishes the Land Values summary every July and the Cash Rents summary every August, and the two together are the closest thing this market has to an index. The 2026 Land Values summary, released July 31, 2026, put average farm real estate — cropland and pasture combined — at $4,500 an acre, up 3.4%; cropland at $6,020, up 3.3% and above $6,000 for the first time; and pasture at $2,000, up 4.2%. That is a sixth straight annual rise, and farm real estate is now nearly 44% above its 2020 level. It is also slower than 2025, when cropland rose 4.7% to $5,830.

The state detail is where a buyer actually lives. Iowa cropland averaged $10,700 an acre in 2026, up 3.9%; Illinois $10,200, up 3.6%; Nebraska $6,960, up 2.4%. On the all-land measure Iowa led at $10,100, followed by Ohio at $9,650 and Illinois at $9,250. Against a national cropland average of $6,020, the Corn Belt trades at a 70% to 80% premium, and within a single Iowa county the range between an 85-rated field and a 91-rated field is wider still. Nobody buys the national average. What the national average tells you is the direction of the tide.

Who sells, and why

Farmland comes to market for three reasons and almost never for a fourth: an estate settlement, a retirement without an heir who farms, or a debt problem. Investors trading in and out are a rounding error, which is why turnover sits near 1.5% and why an auction date is set by a family calendar rather than a market view. Supply therefore does not respond to price the way it does in housing — a 40% run in values from 2020 to 2026 produced no wave of listings, because the owners are not price-sensitive sellers.

More than a third of non-operator landlords are 75 or older, and less than 5% of owned farmland is expected to be sold or gifted within five years, by the American Farm Bureau Federation’s July 2026 survey work: the demographic case and its own refutation in one sentence. What that land is expected to do instead is pass inside the family — 10% into a trust, 15% written into a will. The transfer is coming; it is not coming quickly.

The income side went the other way

Cash rents did not follow values up in 2026. The NASS Cash Rents summary, published in August 2026 from a survey sent to roughly 242,000 operations, had average US cropland rent falling $1 to $160 an acre, a 0.6% decline; irrigated cropland held at $244, non-irrigated fell $1 to $146, and pasture rose $1 to a record $16.50. Rent is still about 15% above its 2020 level, but values are 44% above theirs, and the arithmetic of that divergence is the single most important number for a direct buyer: $160 divided by $6,020 is a 2.66% gross yield, before property tax, before insurance, before management, and before the tenant has a bad year.

The NCREIF farmland index has compounded at roughly 9.8% a year since 1992, and the gap between that number and what a single owner earns is the most important thing in this guide.

The index is an institutional composite: hundreds of properties across dozens of states and a dozen crops, appraised rather than traded, held by managers with agronomists on staff. It posted a positive total return in every year through 2023, fell 1.03% in 2024 (income +2.49%, capital −3.46%), managed +0.20% in 2025, and printed −0.20% in the first quarter of 2026 before turning positive again in the second. Those figures are tabulated from NCREIF by vendors and managers rather than published free by the index itself, and the flagship guide on this hub treats them at length.

Three things separate that record from yours. Diversification is the first: the index owns row crops in Iowa and almonds in Kern County, and its 2024 result blended annual cropland’s +5.66% with permanent cropland’s −10.18%. You will own one farm, in one county, with one crop mix and one tenant. Appraisal smoothing is the second: institutional properties are marked by appraisers on a schedule, which damps the reported volatility to about 6.8% standard deviation against 17.6% for the S&P 500 by one vendor’s tabulation. Your farm is marked when you sell it, and the mark is whatever two neighbours will pay on a Tuesday in February. Cost structure is the third: the index return is reported after property-level expenses but the institutional owner spreads legal, management and diligence costs over thousands of acres, while you spread them over eighty.

What survives all three caveats is the shape of the return, and it is worth being precise about it. Income has been the larger and steadier of farmland’s two components over the long run; appreciation has been the lumpier one, and in 2024 and 2025 it was negative. A direct owner buying at 2026 prices is buying a 2.66% gross coupon, perhaps 2.0% net, plus whatever the land does. If the land does 3% a year, the total is about 5%. If the land does nothing, as Seventh District values in nominal terms did in the year to June 2026, the total is the coupon.

What our tape shows

Invest Alternative runs a small composite of alternative-asset proxies and a handful of single-name series; the farmland entries are the two listed farmland REITs, priced daily. On our tape Farmland Partners closed at $10.65 and Gladstone Land at $9.83 on September 4, 2026 (source: our radar store, eodhd proxies, five observations from August 31). Our provisional IA Composite read 100.271 on September 8, 2026, up 5.74% over thirty days and 0.29% over a year. These are our marks on our method, not market-wide index levels, and they measure something different from the acre: the listed vehicles are equities, they move with rates and the stock market, and both companies have been selling farms. The useful signal in them is the reference price. When a REIT that owns row-crop ground can sell a farm above the value carried inside the company while its shares trade near $10, the market is telling you that packaged farmland is cheaper than unpackaged farmland. The companion guide on farmland REITs and funds is where that trade lives.

Soil productivity, drainage, water and access explain most of the difference between two fields in the same township, and each of them can be checked for free before you make an offer.

Soil, and the indexes that price it

Two rating systems do most of the work in an American farmland negotiation. NCCPI, USDA’s National Commodity Crop Productivity Index, is the national standard: an NRCS model that scores a soil map unit from 0 to 1 on its inherent capacity to grow commodity crops without irrigation, built from SSURGO soil properties, topography and climate. It is free through the NRCS Web Soil Survey, it covers the whole country, and because it is national it lets you compare a field in Indiana with a field in Missouri. CSR2, Iowa’s Corn Suitability Rating, runs from roughly 5 to 100 and is the number Iowa buyers actually quote; above about 83 is top-tier ground.

The reason to learn one of them is that the market prices in points. Individual Iowa auction results reported through 2026 show dollars per CSR2 point ranging from about $131 to $291 on farms rated between 80 and 92 — a Sioux County farm at 91.4 points fetching $291 a point, a Black Hawk County farm at 90.8 fetching $227, a Worth County farm at 88.2 fetching $175, a Floyd County farm at 85.0 fetching $131. Dividing the price by the points is the single most useful normalisation in the business, because it strips out soil quality and leaves the part of the price that is location, access, neighbours and emotion. If you cannot explain why the farm you want is trading 60% above the recent per-point comps in its county, you are the explanation.

Both indexes measure potential, not the field in front of you. NCCPI does not know about a broken tile line, a compaction layer from a wet harvest, or twelve years of drawn-down potassium. That is what the soil test is for, and the soil test has a tax consequence covered in the tax section below.

Drainage, water and access

Tile drainage is the improvement that most reliably converts into both yield and price in the eastern Corn Belt. Sub-surface perforated pipe pulls excess water off a field so it can be planted on time and rooted deeply; intensive pattern installation has been costed at roughly $800 to $1,000 an acre, with North Carolina extension pricing at 40-foot spacing with a fabric filter closer to $1,394 an acre, and published yield gains generally in the 10% to 15% range with higher claims on the worst-drained soils. For a buyer the questions are whether the field is tiled, at what spacing, when, and whether there is a map. An untiled field on a poorly-drained soil series is a capital project you are inheriting, and it should come out of the price.

Water is the difference between a $6,000 acre and a $16,000 acre in the West and the difference between an asset and a liability by 2040 in the wrong basin. Irrigated cropland rents for $244 an acre against $146 for non-irrigated, a 67% premium, and that premium is entirely a bet on the water right surviving. The diligence is specific: source, priority date, quantity, the district and the basin’s plan. Paper water is not wet water, and the water section below is about what has happened to the difference.

Access and shape are the unglamorous residual: road frontage, field shape, distance to the nearest elevator, whether the parcel is landlocked behind a neighbour’s ground, whether a pipeline or transmission line crosses it, and whether a wind or solar lease has decades left to run. Any of these can move the price 10% and none shows up in a soil index.

An auction is a price-discovery mechanism designed for the seller, and in the Corn Belt it is the default rather than the exception.

Roughly four in five Iowa farms sell at auction, and the format matters to the buyer’s arithmetic. The seller sets the date, the terms and the reserve, publishes a book with the survey, soil map, FSA records and tax bill, holds an open house, and sells on a fixed afternoon whether or not the buyer’s financing has cleared. Bidding is usually per acre, sometimes by tract with a right to combine, and the contract is signed at the ring with a deposit of around 10% and a closing set 30 to 60 days out. There is no inspection contingency and typically no financing contingency. Every piece of diligence a buyer wants to do has to be done before the auction, on a farm they may not win, at their own cost.

Private treaty — a listing with a broker, an offer, a negotiation — is the minority route and it is the friendlier one for an outsider. It allows contingencies, a longer look, and a conversation with the seller about the lease and the tenant. It is also where farms go when the seller does not want a public price, which sometimes means a quiet deal and sometimes means a farm the neighbours already declined.

Transaction costs on the way in are modest by the standards of other alternatives and are mostly the seller’s. Land brokerage commissions are quoted across a wide band — auctioneer fees commonly around 1% to 5% of the sale, general land brokerage often 5% to 10%, and specialist farmland brokers across a wider range still, all per 2026 industry sources rather than a published schedule — and are conventionally paid by the seller, though at auction a buyer’s premium of a few percent is increasingly added to the hammer price. The buyer’s own costs are title work, an owner’s title policy, a survey where the legal description needs one, recording fees, an appraisal if a lender requires one, and legal review of the lease and any easements. Budgeting 1% of the purchase price for the buyer’s side is realistic on a clean Midwestern parcel; a farm with a partial survey, a shared drainage district, or an unrecorded lease will cost more.

Read the FSA file before you bid

Every farm has a file at the county Farm Service Agency office, and it carries the farm number, the tract numbers, cropland acres as USDA measures them, base acres and program yields, any CRP contracts with their expiry dates, and the wetland and highly-erodible-land determinations. Two things in that file routinely change a price. CRP contracts pay a per-acre rent for taking ground out of production and run 10 to 15 years; buying a farm with six years of CRP left means buying six years of a different, lower, government-set income and a re-entry cost when it ends. Wetland determinations under the conservation-compliance rules can make part of a field undrainable without losing program eligibility on the whole farm. Neither will be in the auction book in a form that draws your attention to it.

There are three ways an owner gets paid from a farm, and choosing between them is choosing how much of the crop’s risk you keep.

Cash rent is a fixed dollar amount per acre, usually paid half in spring and half in autumn, occasionally all in advance. The tenant takes all of the yield and price risk and all of the upside; you take the credit risk of one operator. It is the arrangement most non-farming owners want and the one the national $160 figure describes.

Crop share splits the harvest, classically 50/50 on the grain with the landlord paying a matching share of seed, fertiliser and chemicals, and the tenant paying all of the machinery and labour. Shares of 60/40 and 2/3–1/3 in the tenant’s favour are common where the landlord contributes less. The owner takes real commodity risk and gets real commodity upside, and — importantly for the tax section — a crop-share landlord who materially participates is running a farm business rather than collecting rent, with consequences for self-employment tax and for the passive-activity rules.

Flexible cash is the hybrid, and it is where a growing share of the market has settled. A base rent is set low and a bonus is calculated from actual yield and actual price, so the owner captures part of a good year without underwriting a bad one. Iowa State’s flexible-lease guidance describes the common formulation as rent equal to some share of gross crop revenue, most often in a 25% to 40% band; Iowa’s 2017 ownership and tenure survey found flexible leases at 18% of all cash-rent leases in the state, and practitioners describe the share as having grown since. The design question is the floor: a flex lease with no minimum is a crop share wearing a cash-rent suit.

Rent is sticky, in both directions

Rents move less than crop prices and much less than land values, because they are annual contracts between people who expect to deal with each other again. Iowa’s 2026 statewide average cash rent came in $1 lower at $270 a crop acre, a 0.4% decline, in Iowa State’s May 2026 survey — $9, or 3.2%, below the $279 high the survey recorded in both 2023 and 2024 — with district results ranging from −$5 to +$2. The Chicago Fed’s August 2026 AgLetter had Seventh District cash rents down 3% for 2026, a second consecutive annual decline after increases from 2021 through 2024. Meanwhile USDA’s September 3, 2026 forecast put 2026 net farm income at $158.4B, down 2.6% nominally and 5.5% in real terms, with production expenses up 4.5% to $492.8B and a record $47.4B of direct government payments inside the total. A tenant whose margin depends on a record government payment is not a tenant who will absorb a rent increase, and the 2026 rent prints say the landlords have worked that out.

The trap in the lease you inherit

Buying a farm does not usually buy you the right to choose the tenant. A farm lease runs with the land, and several Midwestern states go further, providing that a farm tenancy renews automatically for another crop year unless written notice of termination is served by a statutory date. Iowa’s is the strictest and the best known: under Iowa Code section 562.6, a farm tenancy carries over into the following crop year on the same terms unless written notice is served on or before September 1, in which case the tenancy ends the following March 1. It applies to oral leases as well as written ones and to multi-year leases as well as one-year leases, and section 562.7 prescribes how notice must be served — certified mail being the usual route.

Other states set different dates and different service rules, so confirm the specific state with counsel before you close, because getting it wrong costs you a full year of control. The practical consequence is that a farm bought in October comes with next year’s tenant already installed at last year’s rent. If the tenant is good, that is a gift. If the tenant is the reason the farm is for sale, it is a year.

A professional farm manager is the difference between owning land and having a second job, and the fee is a straight deduction from a yield that starts under 3%.

The service is concrete: the manager markets the ground to tenants and negotiates the lease, collects the rent, inspects the fields, approves and monitors fertility and drainage work, handles the FSA and crop-insurance paperwork, files the landlord’s share of the accounting, and produces an annual report. The large firms — Farmers National Company, Hertz Farm Management and their regional peers — manage millions of acres between them and also broker sales, which is worth knowing when your manager recommends a sale.

Fees are quoted as a percentage of gross income and cluster, per extension guidance, at 5% to 10% with about 8% typical, with cash-rent management at the lower end (commonly 5% to 7%, scaled down for larger farms) and crop-share management at the higher end because there is more to do. Some firms use a hybrid: a per-acre base of a few dollars plus a percentage of net crop income. On the national average acre — $160 of gross rent — an 8% fee is $12.80, or 8% of the gross yield and rather more than that of the net.

Whether to pay it turns on distance and lease type. An owner who lives an hour away, holds one cash-rented farm and knows the tenant can plausibly self-manage. An owner in another state, or with a crop-share or flex lease that requires someone to verify the yield and the price, is buying a service they cannot replace. What a manager does not do is make the asset perform: the fee comes out of the same 2.0% net yield as everything else, and two farms plus two managers is not diversification.

Farm real estate debt in 2026 costs more than farmland yields, which turns leverage from an accelerator into a subsidy the borrower pays.

The lenders are three. The Farm Credit System — a network of borrower-owned cooperative associations funded in the capital markets — is the largest holder of farm real estate debt and the default first call for a land loan, offering fixed terms out to 20 and sometimes 30 years and paying patronage dividends that reduce the effective rate. Commercial and community banks with agricultural desks compete on relationship and speed, more often at shorter fixed periods with a balloon or a repricing. USDA’s Farm Service Agency is the government lender of last resort and of first entry, making direct loans, guaranteeing bank loans, and running a beginning-farmer down-payment programme and a joint-financing arrangement in which FSA takes a subordinate piece at a subsidised rate alongside a commercial first mortgage.

Rates in the second quarter of 2026 sat close to 7%. The Chicago Fed’s August 2026 AgLetter reported average nominal rates in the Seventh District of 6.79% on farm real estate loans, 7.12% on farm operating loans and 7.14% on feeder-cattle loans, each up slightly from the first quarter. The Kansas City Fed’s second-quarter survey had Tenth District rates flat on the quarter, with loans above $100,000 averaging just under 7% and smaller loans a touch more, and noted that rates remained above their historical norms even after coming down from the 2023 peak. FSA’s direct farm-ownership rate was 6.00% for August 2026, with joint-financing farm ownership at 4.00% and the beginning-farmer down-payment loan at 2.00%; FSA resets these monthly, so check the current month before you rely on one.

Down payments and the negative carry

Farmland does not qualify for the low-down-payment programmes that residential property enjoys. Lender guidance in 2026 clusters around a minimum of 20% down, 25% to 30% where the property does not include a primary residence, and 35% as a common practical requirement on investment farmland; beginning-farmer programmes go far lower, with FSA down-payment loans at 5% and some Farm Credit associations financing up to 95% for qualified beginning farmers.

Now put the two numbers together. Borrow 70% of a $6,020 acre at 6.79% and the interest alone is $286 on that acre in year one, against $160 of gross rent. Leverage on farmland at 2026 prices does not produce income; it consumes it, and the loan is serviced out of the borrower’s other earnings while the equity waits for appreciation. That is not automatically a bad trade — it is the same trade a farmer makes when buying the neighbouring field, and it works if land compounds faster than the after-tax cost of the debt — but it should be named for what it is. A leveraged farmland purchase in 2026 is a financed bet on land prices with a rent offset, not an income investment.

Farmland is a perpetual income stream, so its value is rent divided by a discount rate net of growth, and a change in the discount rate moves the price far more than a change in the rent.

The mechanism is worth doing in numbers because it explains almost everything the Fed surveys reported in 2026. If an acre earns $160 of rent and the market capitalises it at 2.66%, the acre is worth $6,020. Hold the rent constant and move the capitalisation rate to 3.00%, and the acre is worth $5,333 — a fall of 11% from 34 basis points. Move it to 3.5% and the acre is worth $4,571, down 24%. This is why farmland values are more sensitive to the bond market than to the corn market, and why the decline in farm loan rates from their 2023 peak, which the Kansas City Fed has repeatedly credited with supporting land values, mattered more to prices than two soft harvests did.

The Kansas City Fed put the relationship precisely in its second-quarter 2026 survey: the capitalisation rate on nonirrigated cropland had slipped slightly below the ten-year moving average yield on ten-year US Treasuries, while the ten-year moving average of annual land appreciation remained more than 1.25 percentage points above current bond yields. In plain terms, buyers were accepting a current yield no better than a long bond and being compensated by an expectation of appreciation drawn from the past decade. That is a coherent position and a fragile one: it holds exactly as long as appreciation continues.

What the district surveys actually said in 2026

The two Federal Reserve district surveys are the fastest read on this market, and in 2026 they agreed about row crops and split over grazing land. The Chicago Fed, covering Illinois, Indiana, Iowa, Michigan and Wisconsin, reported in its August 2026 AgLetter that Seventh District farmland values in the second quarter were flat year-over-year, their slowest growth since the fourth quarter of 2024, with Illinois and Iowa up and Indiana and Wisconsin down — and down 3.7% adjusted for inflation, the largest real decline since 2016. Only 5% of responding lenders expected values to rise in the third quarter; 81% expected them stable and 14% expected a fall.

The Kansas City Fed, covering the western plains, found the average value of Tenth District nonirrigated cropland up about 1% year-over-year in the second quarter and irrigated cropland up about 4%, with ranchland up more than 7% to new record highs and credit conditions deteriorating gradually from a modest base. Row-crop ground on the plains, in other words, was doing roughly what it was doing in the Corn Belt. The grazing land was the exception.

The reconciliation is geography and cattle. Corn Belt row-crop ground is priced off grain margins that were compressed in 2025 and 2026; plains ranchland is priced off a cattle cycle that was not. A national average of 3.3% conceals both.

Property tax, insurance and management take roughly a quarter of gross rent on typical Midwestern cropland, and property tax is the largest and the strangest of the three.

Use-value assessment

Farmland in most agricultural states is not taxed on what it would sell for. It is taxed on what it can produce, under a use-value or agricultural-use assessment regime, and the mechanics differ enough between states to change the carry materially. Iowa assesses agricultural land on a five-year average landlord crop-share income model — production, yields, prices and government programme payments, less expenses — capitalised at a rate fixed in statute at 7% by Iowa Code section 441.21, to produce a county productivity value per acre. Illinois sets an equalised assessed value per acre for each soil productivity index at 33.33% of an agricultural economic value, itself derived by capitalising net return to land at a moving average of the Federal Land Bank mortgage rate; permanent pasture is then assessed at a third of the cropland figure for its index and other farmland at a sixth, and no index’s value may move more than 10% in a year. The Department of Revenue publishes the table annually.

Two consequences follow. First, the assessed value of an acre bears almost no relation to its market value, and it moves on a lag: an income-based assessment capitalised at 7% is anchored to farm profitability from up to five years ago. Second, the tax bill is the one carrying cost that can rise while your rent falls, because the assessment model is averaging a different set of years than the lease is.

The rest of the stack

Insurance on bare cropland is a liability policy rather than a property policy and is inexpensive; a farm with buildings, grain storage or a house is a different exercise. Management is the 5% to 10% of gross discussed above where you use a manager. Maintenance and capital is the item most first-time owners omit: tile lines fail, terraces silt in, waterways need reshaping, fences and field entrances need work, and a tiling project runs $800 to $1,000 an acre. Prudent underwriting sets aside something every year for it rather than treating it as a surprise. Fertility is the tenant’s cost under a cash lease and yours under a crop share, and a departing tenant on a cash lease has every incentive to mine the soil in the final year, which is why an entry and exit soil test belongs in the lease.

Taken together, a quarter of gross rent is the working assumption on a professionally managed Midwestern cash-rented farm: roughly $160 of gross becomes roughly $120 of net, and the 2.66% gross yield becomes about 2.0% net.

Farmland’s tax treatment is the strongest genuine advantage direct ownership has over every other alternative asset in this series, and four provisions do most of the work — though the most lucrative of the four is not available to every owner.

Section 180 and the fertility deduction

IRC Section 180 is an election available to a taxpayer engaged in the trade or business of farming, and that phrase settles who gets it before any soil is tested. A landlord who collects a fixed cash rent and does not materially participate is generally not in the business of farming for these purposes and cannot take the deduction; a crop-share landlord, or a cash-rent owner who materially participates, generally is. The passive, professionally managed, cash-rented farm that the worked example below prices is precisely the case that does not qualify. The lease you sign and the participation you actually undertake therefore decide whether the largest first-year tax item here exists for you at all, which is another reason the lease form is settled with the preparer before closing rather than after.

For the buyer who does qualify, the election converts part of the purchase price into a current deduction. Someone who acquires farmland and did not farm it in the immediately preceding crop year may elect to expense the value of residual excess soil fertility — the phosphorus, potassium and micronutrients present above a baseline at acquisition — rather than capitalising all of it into non-depreciable land. The deduction is claimed against the buyer’s income, most usefully in the year of acquisition, and it requires parcel-specific soil tests taken after closing and before any new fertiliser goes on, plus a certified agronomist’s calculation of the excess above baseline. A sample drawn before the deed changes hands documents the seller’s farm rather than yours, and area-wide fertility data does not meet the bar either. The IRS does not require it, but the strongest documentation is an explicit allocation of purchase price to excess fertility written into the purchase contract, which means the time to think about this is before you sign, not at the next tax filing.

Two things follow that the marketing around this deduction leaves out. The first is that it is a deferral rather than a saving: the deducted fertility comes back as ordinary income when the farm is sold, so at a constant marginal rate the deduction and its recapture cancel and the ten-year return is the same with the election as without it. What the owner buys is the use of the money in between, which on a large allocation is worth having and is not a tax cut. The second is the audit risk. Practitioners at the Iowa State Center for Agricultural Law and Taxation and the University of Illinois Tax School have written extensively on the conditions; the deduction is real, it is fact-intensive, it is routinely missed by buyers whose accountants do not do farm work, and it is routinely claimed by owners who do not qualify.

Depreciation, and what the land itself is not

Land is never depreciable. The improvements on it usually are. Drainage tile, irrigation systems, wells, fences, grain bins, single-purpose agricultural structures and general farm buildings each carry their own recovery period, and under the general depreciation system in IRS Publication 225 the ranking is not intuitive: drainage tile and most land improvements run 15 years, grain bins and agricultural fences 7, single-purpose agricultural and horticultural structures 10, and general-purpose farm buildings — the machine shed, the barn — the longest at 20. Confirm each asset’s class with a farm tax preparer before you file, because the allocation is where the money is. The One Big Beautiful Bill Act permanently restored 100% bonus depreciation under Section 168(k) for qualifying property acquired and placed in service after January 19, 2025, and the Section 179 expensing limit for 2026 is $2,560,000, phasing out dollar-for-dollar above $4,090,000 of qualifying property placed in service.

The mechanism that matters to a buyer is allocation. A farm purchase is one price for a bundle: land, tile, fences, a well, perhaps a bin and a machine shed. Every dollar allocated to a depreciable improvement in a defensible way is a dollar recovered against income instead of sitting in basis for thirty years. Every dollar so allocated is also subject to depreciation recapture at sale, taxed at ordinary rates on personal property and at a maximum 25% on real property, so the benefit is a deferral and a rate arbitrage rather than a free lunch.

The exchange, the surtax and the rate

Farmland is real property, so it remains eligible for a Section 1031 like-kind exchange, the deferral that art, wine, cars and watches lost when the 2017 tax law confined 1031 to real estate. A farm can be exchanged into another farm, into commercial property, or into a Delaware statutory trust interest, with the 45-day identification and 180-day closing clocks run through a qualified intermediary. Both clocks start on the day the relinquished property transfers — the 45 days are inside the 180, not added to it — and the exchange period ends at 180 days or the due date of that year’s return including extensions, whichever comes first. Neither deadline can be extended for a weekend or a holiday. The hub’s dedicated 1031 guide covers those mechanics and their failure modes. Farmland is also one of the commonest replacement properties in an exchange, which is a quiet source of bid in the market: a retiring landlord selling an apartment building has 180 days and a strong motive to buy dirt.

Absent an exchange, a farmland gain held more than a year is an ordinary long-term capital gain at a top federal rate of 20%, not the 28% collectibles rate, plus the 3.8% net investment income tax for passive owners above the income thresholds. Whether the surtax applies is a function of participation: rent from a cash lease to an unrelated tenant is generally passive investment income; a materially participating crop-share arrangement is a trade or business, which changes both the surtax analysis and the self-employment exposure. That question is worth settling with a preparer before the lease is signed, because the lease form drives the answer.

The estate case

The reason American farmland is held for generations rather than traded is the combination of a large exemption and a step-up. The One Big Beautiful Bill Act set the federal estate-tax exemption at $15 million per person and $30 million per couple from January 1, 2026, indexed and no longer scheduled to sunset. Stepped-up basis at death resets the heirs’ basis to fair market value, extinguishing the income tax on a lifetime of appreciation. And Section 2032A special-use valuation allows qualifying farmland in a qualifying estate to be valued at its farm use rather than its highest and best use, capped at a reduction of $1,460,000 for 2026 deaths under Revenue Procedure 2025-32, up from $1,420,000 in 2025 — with a recapture period if the heirs stop farming or sell.

Irrigated ground carries a 67% rent premium over dryland, and in three regions the legal basis of that premium is actively being reduced.

The Ogallala, the High Plains aquifer running under eight states, supplies roughly 30% of the groundwater used for US irrigation and is being pumped faster than it recharges across most of its southern and central extent. In the groundwater management area covering southwest Kansas, water levels fell 1.52 feet between January 2024 and January 2025, a larger drop than the 1.43 feet of the year before. Kansas has moved from monitoring to mandate: districts had to identify their priority areas of concern by July 1, 2024 — areas with less than fifty years of usable lifetime being the benchmark practitioners argued for — consult the water-right holders inside them, and file action plans with the state’s chief engineer by July 1, 2026, with the state stepping in where they did not. A farm whose value rests on an irrigation right in a district about to cut allocations is a farm whose value rests on a document that is being rewritten.

California’s SGMA is further along and more explicit. The Sustainable Groundwater Management Act requires critically overdrafted basins to reach sustainability by 2040 and other priority basins by 2042, and the State Water Resources Control Board has begun putting failing basins on probation. The Tulare Lake subbasin was designated probationary in April 2024 and the Tule subbasin in September 2024, the first two of six San Joaquin Valley subbasins whose groundwater plans the state found inadequate.

Probation carries state reporting and fees — a $300 annual filing fee per well plus $20 per acre-foot pumped — and those obligations only became live in 2026: the Kings County Superior Court enjoined the Tulare Lake designation in September 2024, the Court of Appeal reversed that injunction in December 2025, first extraction reports fell due May 1, 2026, and in April 2026 the board exempted pumpers drawing 20 acre-feet a year or less. The Public Policy Institute of California projects that SGMA implementation could take 500,000 to 1 million acres of San Joaquin Valley farmland out of irrigated production by 2040 — idled, retired or repurposed. Buying permanent crops in a probationary basin means buying a tree that needs water every year in a place that is pricing and rationing it.

The Colorado River waters about 5.5 million irrigated acres, and its operating rules expire at the end of 2026. The seven basin states failed to agree a replacement, and the Interior Department adopted its own decision framework in August 2026 covering 2027 through 2036, locking in the Lower Basin’s proposed reductions for 2027 and 2028 and revisiting the rules every two years thereafter. Acreage dependent on that water is priced today on a rulebook that does not yet exist.

Water also has an observable price now, which is new and useful. On our tape the Nasdaq Veles California Water Index printed $566.37 per acre-foot on September 7, 2026, up from $535.91 on September 1 — our reading of the series, on our method, six observations. The companion guide on water rights treats that market directly; the point for a farmland buyer is narrower. When water trades separately and at a price, the water component of an irrigated farm’s value can be estimated rather than assumed, and a seller who will not give you source, priority date and quantity in writing is a seller who has done that estimate and would prefer you did not.

Foreign ownership of US farmland is about 3.6% of privately held agricultural land, and the law governing it has changed more in three years than in the previous forty.

What AFIDA actually requires

The federal statute is the Agricultural Foreign Investment Disclosure Act. It does not prohibit anything; it requires disclosure. Any foreign person holding an interest in US agricultural land must file Form FSA-153 with the county Farm Service Agency office within 90 days of the transaction, and failure to file, or a late or inaccurate filing, carries a civil penalty of up to 25% of the fair market value of the land. USDA reported roughly 46 million acres of foreign-held interests through December 31, 2024, up 1.3 million on the year, of which 10.6 million acres sit under long-term wind-farm leases rather than crop farms. Canada is the largest holder at about a third of the total; Chinese-linked interests held under 250,000 acres at end-2024, roughly 0.02% of US farmland.

The rules are tightening on two tracks. Federally, USDA published a proposed rule on June 25, 2026 that would expand AFIDA’s scope, move oversight to USDA’s Office of Homeland Security, and replace the flat penalty with an accruing one: starting on the 91st day after a report was due, an initial $250 and then 2.5% of fair market value every seven days for foreign adversaries and adversary-controlled entities and 1.5% for other foreign persons, capped in aggregate at 25% per violation. Comments closed on August 10, 2026. An online filing portal opened in January 2026. At state level, the National Agricultural Law Center counted 28 states with restrictions specifically touching foreign ownership of agricultural land by the end of 2025, with 25 states considering further bills in 2026 and seven amending existing laws; counting more broadly across real property, natural resources and land near military installations, the Committee of 100’s 2026 tracking puts roughly 36 states with some restriction, and states with no restriction or a permissive regime have halved since 2022.

For a US buyer this is diligence rather than prohibition, but it is diligence that is easy to fail. The exposure arrives through the capital stack, not the buyer’s passport: an LLC or partnership with a single foreign member, a foreign trust beneficiary, or a fund with an offshore feeder can bring both the AFIDA filing obligation and, in a restricting state, an outright ownership problem.

Anti-corporate-farming statutes

Eight states maintain anti-corporate-farming statutes limiting which entity forms may hold or lease farmland at all: Iowa, Kansas, Minnesota, Missouri, North Dakota, Oklahoma, South Dakota and Wisconsin. This is a separate and much older body of law than the foreign-ownership rules — Kansas passed the first in 1931 — and every one of them carries exemptions for corporations and LLCs held by a small number of related family members, which is why most farm entities are unaffected in practice. The strictest versions have not survived: Nebraska’s Initiative 300 and South Dakota’s Amendment E were both struck down by the Eighth Circuit as violations of the dormant Commerce Clause, so a general prohibition of that severity no longer operates anywhere.

The sequence that works is to settle the ownership entity and check both the foreign-ownership and the entity-form rules of the specific state before you make an offer, not after the title company raises it.

Farmland has no forgery problem and no custodian to fail, so the losses come from title, from the tenant, from the rulebook and from the rate cycle — and three of the four are visible before you sign.

Title, severances and the rights that left before you arrived

The deed conveys the surface; it does not necessarily convey what is under it, over it or across it. Mineral rights are commonly severed in the plains states and can be worth more than the surface. Wind and solar leases run for decades and can encumber field layout long after the signing bonus is spent — AFIDA counted 10.6 million acres of foreign-held US farmland sitting under long-term wind leases at the end of 2024, a reminder of how much cropland already carries an energy encumbrance. Drainage-district assessments, shared tile easements, access easements and unrecorded field-road agreements are all common and none is exotic. The fix is an owner’s title policy plus a lawyer who reads the exceptions schedule rather than the summary page.

The tenant, and the year you cannot get back

The credit risk in a cash lease is one operator, and the recovery mechanism is slow. A tenant who cannot pay the autumn instalment leaves you with an unpaid rent, a crop you may have a landlord’s lien against depending on the state, and — in most Midwestern states — an automatically renewing tenancy unless notice was served by the statutory date. The 2024 through 2026 stretch is the relevant stress test: USDA’s September 2026 forecast has net farm income falling for a second year with a record $47.4B of government payments inside it, which is what a squeezed tenant base looks like in aggregate.

The rulebook moving under the asset

The clearest recent case is water. California put the Tulare Lake and Tule subbasins into SGMA probation in 2024, litigation held the first of them up until December 2025, and the reporting and the fees — $300 per well a year and $20 per acre-foot pumped — landed on pumpers in 2026; Kansas required all five groundwater management districts to file action plans by July 1, 2026; the Colorado River’s operating rules expire at the end of 2026 with an Interior framework adopted in August 2026 in place of a seven-state agreement. Ownership law moves too: the Fufeng corn-mill purchase of 370 acres near Grand Forks Air Force Base, bought for $2.6M in 2022 and killed by a 5–0 city council vote in February 2023, is the case that produced much of the state legislation now on the books.

The cycle, which is the one that actually costs money

Farmland’s reputation for never losing money ended in 2024, when the NCREIF index posted its first negative annual total return, −1.03%, followed by +0.20% in 2025 and −0.20% in the first quarter of 2026. The mechanism is the capitalisation arithmetic above, and the vulnerable buyer is the leveraged one buying at a 2.66% rent yield with 6.79% debt. Nothing in this section is exotic. It is the ordinary risk of owning a single illiquid parcel financed at a negative carry, and it is why the position should be sized as a decade.

Every fee, carry cost and tax on a real purchase, held ten years and sold, with the arithmetic shown.

The assumptions

Eighty acres of average US cropland at USDA’s 2026 national figure, bought for cash, cash-rented to an established local operator, professionally managed, held ten years and sold. Rents flat at the 2026 average, land compounding at 3% a year — roughly the 2026 print, and below the six-year average. Federal tax only; state income and transfer taxes are ignored and will reduce the result. This is arithmetic on stated assumptions, not a projection.

Buying

Purchase price: 80 × $6,020 = $481,600. Buyer’s closing costs — title search, owner’s policy, survey, recording, legal review of the lease, appraisal — at 1% of price: $4,816. Total basis: $486,416.

Carrying

Gross cash rent: 80 × $160 = $12,800 a year, a 2.66% gross yield on price. Deduct a farm management fee at 8% of gross ($1,024), property tax, insurance and a maintenance reserve at a combined 17% of gross ($2,176) — a quarter of gross rent in total. Net rent: $9,600 a year, a 1.99% net yield. Over ten years that is $96,000 of net rent, taxed as ordinary income as it arrives; at a 32% marginal federal rate, about $30,700 of tax, leaving roughly $65,300 after tax.

Section 180, for the owner who qualifies

The base case above is the passive cash-rent landlord, and that owner takes no Section 180 deduction at all. Run the same purchase for a buyer whose arrangement puts them in the trade or business of farming — a crop share, or material participation — and year one carries an election on a documented residual-fertility allocation. A conservative allocation on eighty acres of good ground supported by a soil test is commonly in the tens of thousands of dollars; at, say, $30,000 deducted against a 32% marginal rate, the tax deferred in year one is about $9,600, the whole of that year’s net rent again. That figure illustrates the mechanism on assumed numbers; it is not a promised deduction, and the allocation has to be supported by the test and the agronomist.

Deferred is the operative word. The $30,000 returns as ordinary income when the farm sells, at the same 32% in this example, so the $9,600 gained in year one is the same $9,600 paid in year ten. Both legs appear in the arithmetic below and they cancel exactly, which means the compound return is identical with the election and without it. What the qualifying owner has bought is ten years of the use of $9,600.

Selling

Land value after ten years at 3% compounding: $481,600 × 1.03 to the tenth = $647,200. Selling costs at 5% for brokerage and closing: $32,360. Net proceeds: $614,840.

Gain: $614,840 − $486,416 = $128,424. Federal capital gains at 20% on that gain is $25,685, and the 3.8% net investment income tax adds $4,880. Total federal tax at sale for the passive landlord: about $30,565.

The farming owner’s bill at sale is a different bill, and it is worth keeping the two apart. That owner does not pay the $4,880 surtax, because a gain on property used in a trade or business in which the owner materially participates is outside net investment income; but the $30,000 of fertility comes back as ordinary income and adds $9,600 at the same 32% rate. Total for the farming owner: about $35,285. Two owners, two bills, and the difference between them is not the soil test.

The result

Take the passive landlord first, because that is the archetype this guide has followed throughout. After-tax sale proceeds: $614,840 − $30,565 = $584,275. Add the $65,300 of after-tax rent collected along the way and the trip returns $649,575 on $486,416 invested over ten years, a compound after-tax return of about 2.9% a year. Before federal income tax the same trip returns $96,000 of net rent plus $614,840 of net sale proceeds — $710,840 on $486,416, or about 3.9% a year. The gap between the two is the whole federal tax bill, and it is smaller than it looks because so much of the return is a capital gain.

The farming owner ends the same decade at $654,455: $579,555 of after-tax proceeds, the same $65,300 of rent, and the $9,600 of year-one fertility benefit whose recapture is already inside the $35,285. That is about 3.0% a year, a tenth of a point better than the landlord — our subtraction on the two runs above. And the whole of that tenth of a point is the 3.8% surtax the farming owner escapes, not the Section 180 election, whose two legs cancel to the dollar. The most-marketed line in farmland tax moves this example by nothing at all; the participation question moves it by $4,880.

Two routes remove most of what is left. Executed through a Section 1031 exchange rather than a sale, the whole bill — $30,565 or $35,285, whichever owner you are — is deferred; held to death under the $15M exemption, the capital-gains component is extinguished by the step-up. The arithmetic also leaves out two items that should be named: state income and transfer taxes, and the 3.8% surtax on the rent itself for a passive owner above the thresholds. Both push the after-tax number down.

What the same money does elsewhere

Two comparisons are worth making honestly. The same $486,416 in a listed farmland REIT buys about 45,700 shares of Farmland Partners at our tape’s September 4, 2026 close of $10.65, or about 49,500 shares of Gladstone Land at $9.83, with daily liquidity, no tenant, no soil test and no county assessor — and with equity-market volatility, no 1031, no step-up on a specific parcel and a management fee inside the vehicle. The REITs and funds guide sets out that trade properly. And the same money in a ten-year Treasury yielding about 4.8% in early September 2026 pays more than twice the farm’s 2.0% net yield with none of the work, which is the honest benchmark that the appreciation assumption has to beat. Direct farmland is not a yield trade at 2026 prices. It is a real-asset, tax-advantaged, multi-decade holding whose case rests on inflation linkage, the exchange and the step-up.

The sequence below is ordered so that the cheap, reversible steps come before the expensive, irreversible ones.

1. Choose the county before the farm. Pick one or two counties you can reach in a morning, and learn them: the dominant soil series, the prevailing CSR2 or NCCPI range, the local elevators and processors, whether the ground is tiled, and the water situation. Depth in one county beats breadth across five.
2. Read a year of the primary documents. USDA’s Land Values and Cash Rents summaries (both released between late July and August), the Chicago Fed AgLetter and the Kansas City Fed Ag Credit Survey each quarter, and your state extension service’s annual land-value and cash-rent surveys. Those five sources are most of what a farm manager knows about the macro, and all of them are free.
3. Build a comparable-sales file. Attend auctions without bidding, record the price, the acres, the soil rating and the date, and compute dollars per productivity point on each. Twenty entries takes a season and is worth more than any appraisal you will buy.
4. Interview farm managers and brokers before you need one. Ask what they manage, what they charge, whether they also broker, and what happened to the rents on their farms in 2024 through 2026. Get the fee schedule in writing.
5. Get the money arranged in advance. Auctions have no financing contingency. Have a Farm Credit or bank pre-approval, know your down-payment requirement, and know the rate you would actually pay; if you qualify as a beginning farmer, understand the FSA down-payment and joint-financing programmes before you need them.
6. Do the diligence before the auction, on a farm you may not win. Pull the soil map and NCCPI from Web Soil Survey, pull the FSA file for base acres, CRP contracts and wetland determinations, read the existing lease, confirm the statutory lease-termination date, verify any water right by source, priority date and quantity, and settle the ownership entity against the state’s entity and foreign-ownership rules.
7. Line up the tax work before closing, not after. Instruct a farm-experienced preparer, settle whether your intended lease puts you in the trade or business of farming, engage the agronomist and negotiate the purchase-price allocation into the contract. Then draw the fertility samples after closing and before any fertiliser goes on, which is the sequence Section 180 requires and the one a rushed buyer gets backwards.
8. Set a price and stop at it. Write your maximum dollars-per-point on a card before the auction opens. The most expensive farms in America are bought in the last two minutes by people who did not do this.

Five readings, each with a threshold and an as-of date, would change the view set out above.

The rent-to-value ratio, published each August

As of the 2026 USDA summaries it stands at 2.66% ($160 against $6,020). Below 2.4% the asset is being priced on appreciation alone and a direct purchase at scale is hard to justify on income; above 3.2% the income case returns and leverage starts to work again.

Farm real estate loan rates in the Fed district surveys

As of the second quarter of 2026 the Chicago Fed reported 6.79%. A sustained fall below 6% would reopen the leveraged bid that has been absent since 2022; a move back above 7.5% would compress values through the capitalisation channel faster than rents could offset.

Real district land values

As of the second quarter of 2026 the Chicago Fed had Seventh District values flat nominally and down 3.7% in real terms, the largest real decline since 2016, with only 5% of lenders expecting a third-quarter rise. Two more quarters of negative real prints with lender expectations still skewed to “stable” would mark a genuine repricing rather than a pause.

Net farm income and the government-payment share

As of USDA’s September 3, 2026 forecast, 2026 net farm income is $158.4B with a record $47.4B of direct government payments inside it. If payments fall materially without crop receipts rising to replace them, rents follow within one or two lease cycles, and rents are what values are capitalised from.

The water rulebooks, all of which have dates

The Colorado River’s post-2026 framework governs 2027 through 2036 and is revisited every two years. California’s probationary-basin list — Tulare Lake and Tule, with four more San Joaquin Valley subbasins under inadequate-plan findings — will lengthen or shorten. Kansas groundwater management districts filed action plans by July 1, 2026, and the allocation cuts follow from them. Any acre whose price includes an irrigation premium should be re-underwritten as those documents land.