Investing in Farmland Through REITs and Funds

On July 11, 2018, a pseudonymous author calling himself Rota Fortunae published a piece on Seeking Alpha titled “Farmland Partners: Loans To Related-Party Tenants Introduce Significant Risk Of Insolvency — Shares Uninvestible.” Farmland Partners fell 39% that day and lost as much as $115 million of market value. No acre of the roughly 160,000 acres the company then owned changed hands, produced a worse crop, or lost a tenant. The dirt did what dirt does, which is nothing, and the asset that was supposed to be the least volatile real asset in the world lost two fifths of its value in an afternoon because it had been put inside a share.

The story has an unusual ending. The author, Quinton Mathews, was identified in 2020 and, in the settlement the company announced on June 21, 2021, retracted the report, acknowledged that its defamatory statements were false, and agreed to pay Farmland Partners a multiple of the trading profits he and his clients had made. That outcome is rare enough to be worth remembering, and it does not undo the lesson. What the shareholders owned in July 2018 was not farmland. It was a small-capitalisation listed security whose price was set by a stock market that had read an article.

This guide is about the gap between the asset and the wrapper. Farmland’s long record belongs to an appraisal-based institutional index whose constituents are pension separate accounts; the things an outsider can actually buy are two listed REITs under $11 a share, a set of private funds with seven-figure minimums, a non-traded REIT that spent a year in escrow and opened on September 1, 2026 with 1.16% of its $3 billion target raised, and platforms selling fractions of a single farm under a private-placement exemption. Investing in Farmland, the hub’s flagship, covers the asset itself; the siblings arriving alongside this one are Buying Farmland Directly, Investing in Timberland and Investing in Water Rights. This one covers the toll booth between you and the crop.

Every route into farmland is one of four legal objects, and the object decides your liquidity, your tax form, your fee stack and who marks your position.

A listed REIT is a corporation that owns farms and whose shares trade on an exchange. You own stock. Price is discovered continuously by whoever is trading that day, the company reports on a 10-Q, and your exit is a market order. There are exactly two of any size in the United States: Farmland Partners (FPI, New York Stock Exchange) and Gladstone Land (LAND, Nasdaq).

An institutional fund or separate account is a private vehicle run for tax-exempt investors, usually a pension plan, an endowment or a sovereign fund. You own a limited-partner interest or a beneficial interest in a managed account. Price is an appraisal the manager commissions, liquidity is a queue, and the minimum is measured in millions. Nuveen Natural Capital, Manulife Investment Management, PGIM Agricultural Investments and UBS Farmland Investors run the largest of them, and their properties are what the NCREIF Farmland Index measures.

A non-traded REIT is a corporation registered with the Securities and Exchange Commission as a reporting company whose shares do not trade. You buy at a net asset value the manager calculates, receive a distribution, and ask to be repurchased subject to a cap. This wrapper arrived in farmland when Nuveen registered the Nuveen Farmland REIT in 2025; it sold its first shares on September 1, 2026.

A fractional platform offering is a limited liability company formed to hold one identified farm, whose membership interests are sold under Rule 506(c) of Regulation D to verified accredited investors. You own an interest in an entity that owns a farm, not the farm. There is no market, the hold is the sponsor’s business plan, and the exit is a sale the sponsor arranges. AcreTrader and FarmTogether are the two best known.

A fifth door is not equity at all. Harvest Returns and Steward mostly arrange agricultural credit: you lend to a farm operator at a stated rate against the farm or its equipment. That is a credit position with a credit position’s risks, and the roughly 9.9% weighted-average annual return Harvest Returns self-reports on its private-credit offerings since 2019 is a credit spread rather than a land return. The distinction matters because platforms in both camps advertise against the same NCREIF chart.

Who is on the other side is worth naming. In the listed REITs it is another investor, and increasingly the company itself: both FPI and LAND have been net sellers of farms in 2024–2026. In the institutional funds it is a manager buying from farm families and selling to other institutions. On a platform it is the sponsor, which has already bought the farm and is reselling it to you at a price it set, with its fee inside that price.

The NCREIF Farmland Index is the number in every farmland pitch deck, and understanding what it is and is not is the single most useful thing an outsider can do before buying any wrapper.

The index is a quarterly composite of individual farm properties held on behalf of tax-exempt institutional investors, chiefly pension funds, and reported to the National Council of Real Estate Investment Fiduciaries by the managers who run them. It begins in 1991, and it is small: 1,023 properties worth about $16.1 billion at the end of 2024, roughly four thousandths of the $3.8 trillion of US farm real estate. Its long-run number is genuinely strong: about 9.8% a year from 1992 through 2025 on FarmTogether’s tabulation of NCREIF data, at a standard deviation near 6.8% against roughly 17.6% for the S&P 500 over a comparable span. Vendor tabulations published before the 2024 downturn put the since-inception figure nearer 10.15%, which is the same series measured to a kinder endpoint.

Then the qualifications, which are large enough to change what you should pay for access.

It is appraised, not traded

Properties are valued by periodic appraisal, not by transactions. Appraisals lag the market by quarters and smooth it in both directions, so the index’s low volatility is partly a measurement artefact and its Sharpe ratio should not be taken at face value. The same criticism applies to private real estate generally, and the sibling guide on syndications and private REITs works through what appraisal smoothing does to a fund’s reported record.

It is gross of the wrapper

The index measures returns at the property level, unlevered, and the fee distinction matters more than it sounds. Property-level operating costs are already inside it — the farm manager, the property tax, the insurance, the people who run the ground. What is not inside it is the layer above: the asset-management and investment-management fees a manager charges the investors whose money bought the farms, and with them a platform’s servicing charge, a selling commission and the corporate overhead of a listed REIT. It is what the farm earned after the cost of running the farm and before the cost of running the money, which is the same basis the sibling timberland guide works through for the NCREIF Timberland Property Index. Every route in this guide subtracts that second layer from the index before it reaches you, and section 9 measures how much.

Its unbroken record is over

The marketing line that farmland has never had a down year retired in 2024, when the index returned −1.03%: income of 2.49% against appreciation of −3.46%. 2025 was barely positive at +0.20% (3.05% income, −2.80% capital), with annual cropland posting its weakest year since inception at +3.52% and permanent cropland at −5.43% on AgIS Capital’s March 2026 tabulation. Q1 2026 printed −0.20% and the index turned positive again in Q2 2026, per Agri Investor’s August 2026 report; we could not obtain the Q2 headline total, but row-crop values rose 0.49% in the quarter and 0.62% year over year while permanent cropland fell 1.62% in the quarter and 7.12% over the trailing year.

The index is therefore a ceiling rather than a forecast: measured before every fee in this guide, on appraisals rather than trades, on properties bought at institutional scale by managers with first refusal on the best ground. Any wrapper projecting a return at or above the index is claiming to beat the benchmark net of its own costs, and the burden of proof sits with the sponsor.

Farmland Partners is the larger of the two listed vehicles by land area and the one whose history best illustrates what a stock market does to a farm.

The company went public on April 16, 2014, selling 3.8 million shares at $14.00, the low end of a $14–$16 range, for gross proceeds of $53.2 million. It grew by acquisition through the second half of the 2010s into a portfolio dominated by row crops — corn, soybeans, cotton, rice — leased to farm operators on cash rents. The 2018 short-seller episode described in the cold open cost it 39% in a day and produced litigation that ran to 2021.

What matters more for anyone buying today is the reversal of the growth story. On October 16, 2024 the company completed the sale of 46 farms totalling 41,554 acres to Farmland Reserve, Inc. for $289 million in cash, a gain of roughly $50 million and about 21% over the aggregate net book value of those farms. It applied $146.6 million of the proceeds to debt reduction. As of June 30, 2026 it owned approximately 70,100 acres in 11 states — Arkansas, California, Colorado, Illinois, Indiana, Louisiana, Missouri, Nebraska, South Carolina, Texas and West Virginia — and told investors it was evaluating further disposals.

The dividend, and what it is telling you

FPI has paid a small quarterly dividend and used year-end specials to distribute the proceeds of sales. It declared a $1.15 per share special dividend on December 13, 2024 after the Farmland Reserve sale, a $0.20 special on December 15, 2025, and raised the regular rate 50% for 2026, from $0.24 to $0.36 a share a year, declaring $0.09 for the quarter payable October 15, 2026. Read that sequence honestly: a REIT that sells a quarter of its land, pays down debt, hands back a large special, then raises the ordinary rate off a smaller base is returning capital, not compounding it.

The valuation argument

For the second quarter of 2026 the company reported net income of $3.1 million ($0.07 a share), adjusted funds from operations of $1.7 million ($0.04 a share) and net operating income of $7.1 million, up 2.9% year over year, and raised the low end of 2026 AFFO guidance to $0.31 against an unchanged $0.35 top. On our tape’s September 4, 2026 close of $10.65 that is an AFFO yield of 2.9% to 3.3% — approximately the gross rent yield on the land itself, which is what a well-run, lightly levered farm REIT should produce and no more.

The bull case is the discount. Estimates of net asset value per share published in 2025 and 2026 ran from a consensus near $14.64 to Raymond James at roughly $16, implying the shares changed hands at 67% to 73% of appraised land value. The bear case is that the discount prices a shrinking asset base and a forecast decline in funds from operations. Both rest on an estimate of what the farms would fetch, which nobody observes until a sale closes — and the one that did close, in October 2024, went at a 21% premium to book.

Gladstone Land is the older listed vehicle, the one built for income, and the one that shows what happens when a farm REIT concentrates in the crops that had the worst three years in the index’s history.

It priced its initial public offering on January 28, 2013 at $15.00 a share, raising $56.7 million gross including the over-allotment. It pays monthly rather than quarterly and, on its own count in the July 2026 distribution announcement, has made 161 consecutive monthly cash distributions since that IPO, at $0.0467 a share a month — $0.5604 a year — declared flat for both the second and third quarters of 2026. On our tape’s September 4, 2026 close of $9.83, that is a distribution rate of 5.70%.

As of the second quarter of 2026 it owned 142 farms, approximately 98,000 acres in 14 states, and about 56,000 acre-feet of water assets in California. The water is not incidental; it is a distinct, separately valued holding and one of the few ways a public-market investor gets direct exposure to Western water, which the sibling guide on water rights takes up properly.

Where the strategy hurt

Gladstone’s tilt is toward permanent crops — almonds, pistachios, citrus, berries — and permanent cropland was the worst part of the farmland market from 2024 onward. The NCREIF permanent cropland index returned −10.18% in 2024 (capital −11.77%, income 1.70%), with the almond index at −16.44% and pistachios at −16.5%; it fell a further 5.43% in 2025 and was still down 7.12% over the trailing year to June 2026, with California permanent cropland alone down 10.70% year over year in Scythe & Spade’s Q2 2026 regional review. A REIT concentrated there could not have avoided the drawdown, and did not.

The consequences showed up in the accounts. The company reported a Q2 2026 net loss of $13.5 million, or $0.32 a share, on revenue of $12.69 million, including a non-cash impairment of about $4.2 million on four Arizona farms written down to an agreed sale price. It has been selling: in July 2026 it completed the sale of two Florida farms totalling 617 gross acres for roughly $3.2 million, booking a small net loss of about $159,000 after closing costs, and management has said more may follow. It has also been re-tenanting vacant ground into alternative uses — solar, cattle, fallowing programmes — telling the August 2026 call that the three properties closest to resolution could add about $1.5 million of annual net operating income.

The capital structure is the risk

Gladstone Land carries a substantial stack of cumulative redeemable preferred stock sitting ahead of the common: Series B and Series C at 6.00% and Series E at 5.00%, each with a $25.00 stated value, after the Series D term preferred was redeemed in full on January 30, 2026. The company has not published a single headline figure for the total; our own arithmetic on the share counts in the Q2 2026 Form 10-Q equity rollforward puts it near 15.5 million preferred shares, about $385 million at stated value, and should be read as an estimate rather than a reported number.

That is ordinary REIT practice and it has a specific consequence: preferred dividends are a fixed claim on farm income, so a flat or falling rent roll is absorbed entirely by the common. It is why a 5.70% distribution rate on a portfolio whose underlying land yields well under 3% in cash rent is a levered number, not a free lunch, and why the common share price has been the more volatile of the two listed names.

The debt is fixed for now and resets soon. On the August 2026 call management said more than 95% of borrowings were at fixed rates, at a weighted average of 3.45% locked for an average of 2.3 years, with roughly $33 million maturing inside twelve months — but about $148 million of loans reset within the year, $130 million of it under the MetLife facility in January 2027. A book struck at 3.45% repricing into 2027 rates is a dated, disclosed event with a known size, and it is the backdrop to management’s stated plan to put disposal proceeds toward debt reduction and preferred repurchases rather than acquisitions.

The two listed REITs are the only farmland most people will ever own, so it is worth being exact about what they deliver and what drives their price.

Start with the cash. On our tape’s closes of September 4, 2026, FPI’s declared 2026 rate of $0.36 a share was a 3.38% distribution rate and LAND’s $0.5604 was 5.70%. Set those against the land itself: US cropland averaged $6,020 an acre in USDA’s July 2026 Land Values summary and cash rent averaged $160 an acre in the August 2026 Cash Rents survey, a gross yield of 2.66% before property tax, insurance and management. A listed REIT paying more than the gross rent on the land it owns is doing it with leverage, with preferred capital, with proceeds from selling farms, or with some of all three. That is not a scandal — it is the arithmetic of a levered income vehicle — but it does mean the distribution rate is not a farmland yield.

The ten-year record

Over the ten years to September 2026 Farmland Partners returned about 4.2% a year and Gladstone Land about 3.7% a year in total return, against roughly 12.6% a year for the S&P 500 over the same decade. That is the answer to the question this guide exists to ask, and it is not close. A decade of owning the listed wrapper paid you approximately its dividend and nothing else, while the index the wrapper is sold on was compounding at 9.8% a year over a longer and kinder span.

Two qualifications, and neither rescues the number. These are vendor calculations — FinanceCharts for Farmland Partners, PortfoliosLab for Gladstone Land, both read in September 2026 — rather than company disclosure, and they are the least authoritative figures on this page. And ten years is a short window that happens to contain the whole 2024–2026 soft patch, so the decade flatters nobody. Neither changes the shape of the finding. The wrapper’s realised decade sits several points a year below both the stock market and the index it is sold on, and the listed REIT is the one route in this guide that charges no explicit fee at all. The gap is not a fee stack. It is what a stock market does to a farm.

Why the price moves

A listed farm REIT is a small-capitalisation, income-paying, rate-sensitive equity, and that is the wrapper’s real signature. It reprices instantly on interest rates, on index-fund flows, on a short report, on a quarter’s guidance — none of which touches a bushel of corn. The 2018 episode is the extreme case; the ordinary case is that these two names move with small-cap REITs, not with the NCREIF index. If you buy them expecting the smooth, appraisal-paced ride the farmland literature describes, you have bought the wrong instrument, and you will sell it at the wrong time.

Our tape

Invest Alternative carries farmland in the composite through exactly these two tickers, which is itself the finding. Our farmland sleeve is still marked awaiting in the index — a 0.6% target weight, accruing history from EODHD REIT proxies — because the only daily-priceable farmland in existence is two small-cap shares. Over the five closes we hold from August 31 to September 4, 2026, FPI ran from $10.29 to $10.65, up 3.50%, and LAND from $9.18 to $9.83, up 7.08%; the IA Composite, provisional, stood at 100.271 on September 8, 2026. A week is not a return series and two securities are not a market; the point of quoting them is the opposite of a market call. A market-wide daily farmland price does not exist, and any publication that appears to have one is showing you either a quarterly appraisal series or a pair of REIT tickers.

The managers who own the farms in the NCREIF index run separate accounts and commingled funds for tax-exempt institutions, and knowing their shape tells you what the retail wrappers are imitating and where they fall short.

Nuveen Natural Capital, the farmland arm of TIAA’s asset manager and among the largest farmland managers in the world, reported $13.7 billion of assets under management across more than three million acres, 580 properties and 11 countries as of September 30, 2025, with a more recent company page giving $14 billion and 2.6 million acres. PGIM Agricultural Investments has served institutional farmland investors since 1989 and reports over $10 billion of assets under management and administration — but read the split before you read the headline: $9.3 billion of that is agriculture debt against $2.2 billion of equity, so the largest name in the table is mostly a lender. Its PGIM U.S. Agriculture Fund is sized at about $1.45 billion.

UBS Farmland Investors runs a commingled fund and separate accounts holding roughly 280,000 acres across 15 states and 30 crops, with approximately $2.5 billion of agricultural investments as of December 31, 2024. Manulife Investment Management, through the former Hancock Agricultural Investment Group, managed roughly 400,000 agricultural acres in the United States, Australia, Canada and Chile as of March 31, 2023.

Three features matter for any comparison. They are unlevered or lightly levered, which is why their reported volatility is low and why a retail vehicle using leverage should not be measured against them. They are open-ended or long-dated, with redemption queues rather than a fixed fund life, so the manager is never a forced seller. And they are the index’s own constituents: the NCREIF Farmland Index is a composite of what these managers report, so the benchmark and the products measured against it are drawn from the same population. NCREIF is a fiduciary body with a long record, but that is a reason to treat the index as an industry-reported figure rather than an independent audit.

The minimum is the barrier. These are qualified-purchaser and institutional products; the separate accounts start in the tens of millions and the commingled funds in the millions. For an individual the practical consequence is simple: the cheapest and best-run farmland in the country is managed by people who will not take your money, and every wrapper in the rest of this guide exists to sell you a version of it at a higher cost.

The Nuveen Farmland REIT sold its first shares on September 1, 2026, a year after the largest institutional farmland manager in the world registered it, and its terms are the template every competitor in retail farmland will copy: perpetual life, a monthly net asset value, and the share-class fee ladder Blackstone popularized in commercial property.

It filed a Form 10-12G with the SEC in 2025, targeting roughly $3 billion and offering exposure including California’s Central Valley. Its terms, from the registration statement, are the modern non-traded template. Shares sold during the escrow period at $20.00, and from the initial closing onward sell at the prior month’s NAV per share plus any upfront selling commission. It offers eight types of common shares — A-I, A-II, I, S, S-I, S-II, D and E — and the management fee runs from 0% to 1.25% of NAV a year by class: 0.95% for Type A-I and S-I, 1.15% for A-II and S-II, 1.25% for Type I, S and D, and 0% for Type E.

On top of that sit the distribution charges the class you are sold decides: an upfront selling commission of up to 3.5% of the transaction price on Type S, and an ongoing shareholder servicing fee of 0.85% of NAV a year on Types S, S-I and S-II and 0.25% on Type D, with none on A-I, A-II, I or E. A special limited partner holds a performance participation interest of 12.5% of total return, subject to a 6% hurdle, a high-water mark and a catch-up (9% on Class A-I). A share repurchase plan operates quarterly from the initial closing, capped at 5% of aggregate NAV in any calendar quarter.

Read those terms together and the product’s shape is clear. The flat fee is charged whatever happens. The performance fee only bites above 6%, which on farmland’s realistic 5% to 6% gross return means it will rarely be paid — the fee that matters is the flat one. And the repurchase cap is the liquidity: 5% a quarter is generous until everyone asks at once, at which point requests are prorated. The sibling guide on syndications and private REITs documents what that cap did at Blackstone’s BREIT from November 2022 and at Starwood’s SREIT in April 2026; the mechanism in farmland is identical and the underlying asset is less liquid.

A year in escrow, then $34.7 million

As of its Form 10-Q for the period ended June 30, 2026, the REIT had no shares outstanding and no management fee accrued — total assets of $1 thousand and total equity of nothing. A vehicle registered in 2025 by the biggest name in the asset class, targeting $3 billion, was still sitting in escrow more than a year later.

Escrow broke on September 1, 2026. The company reported the initial closing of its private offering and the commencement of operations, selling 1,732,960 common shares for approximately $34.7 million at the $20.00 escrow price; the same day it executed its advisory agreement with Nuveen Farmland Advisors and its board adopted the share repurchase plan. An initial portfolio is to be bought from a subsidiary of TIAA at an externally appraised value, after which a Nuveen affiliate intends to invest at least $25 million in Class A-I shares.

Read the number rather than the headline. $34.7 million is 1.16% of the $3 billion target, raised after a year in registration, and the sponsor’s own promised $25 million would be most of that again on top. That is a slower start than a $3 billion ambition implies, and it points where the earlier reading pointed: appetite for private farmland at institutional fee levels is not yet demonstrated, and the demonstration will take years rather than quarters. A vehicle this new also has no diversification, no operating history, no published NAV series and no repurchase queue to test, and the seller of its first assets is an affiliate of its sponsor.

This is the fastest-moving material in the guide, and it rests on two filings: the Form 8-K of September 1, 2026 and the Form 10-Q for the quarter ended June 30, 2026. As of September 10, 2026 the composition of that $34.7 million, the first published monthly NAV and whether the TIAA initial-portfolio acquisition has closed were all still unpublished. Those three readings, rather than the raise total, will establish whether retail capital wants private farmland at institutional fee levels.

Fractional platforms are the only route into a specific, identified farm for a five-figure cheque, and after roughly a decade of operation there is finally enough realised history to judge them on outcomes rather than projections.

The structure

Each farm sits in its own limited liability company. Shares are sold under Rule 506(c) of Regulation D to verified accredited investors, the platform collects rent from a farm operator and distributes it, and the exit is a sale of the farm in a stated window, typically five to ten years. There is no public market, resale is thin to nonexistent, and the hold is not optional.

The fee stacks

AcreTrader takes roughly 0.75% a year in servicing plus about 2% at closing, with a disposition fee at sale, and starts at about $10,000. FarmTogether charges roughly 2% upfront, 1% to 2% a year in management and a slice of net operating income, from $15,000 for a single-farm deal and $100,000 for its fund. Both sets of terms come from 2026 platform reviews rather than the platforms’ own fee pages, and terms are deal-specific — read the offering documents. Farmland LP, which converts conventional ground to organic and regenerative production, charges the private-equity shape: 2% a year plus 20% carried interest above a 6% preferred return.

What has actually been realised

AcreTrader has the fullest disclosure and it repays reading carefully. On March 30, 2026 it announced the disposition of 57 farmland assets across 13 states, funded between 2019 and 2024, generating over $135 million in investor distributions; it reports 75 realisations since inception and more than $189 million distributed in total. Disclosed net internal rates of return on realised deals span 9.4% to 30.3%, and the third-party tracker AltStreet, which verified 15 exits independently, put the average near 15.2%.

The Northland Organic Farm in Minnesota’s Red River Valley, funded on the platform in September 2020 and bought out in January 2026 by the operator exercising a purchase option, returned an 11% net IRR and a 1.8× multiple on an announcement dated February 25, 2026 — a useful calibration, because 1.8× over roughly five years is what an 11% net IRR looks like, and it sits at the bottom of the disclosed 9.4% to 30.3% range rather than near its middle.

Those are good numbers. Here is the qualification that decides how much weight to put on them: of 139 offerings AltStreet tracked, 71 were funded before 2022 and carry no public exit summary. Just over half the platform’s history is therefore invisible in the realised record, and the visible half is the half the sponsor chose to publish. Nothing here suggests the missing deals are bad. But an average IRR computed across only the deals with a press release is not a track record, it is a sample, and the direction of the selection bias is not ambiguous.

The platforms are becoming funds

The single-farm model is being replaced by pooled vehicles, which changes both the diligence you can do and the fee you pay. FarmTogether has run a Sustainable Farmland Fund alongside its single-farm deals. In 2026 AcreTrader went further, announcing in March the launch of the Proterra AcreTrader Farmland Fund LP, an open-ended vehicle structured as a private non-traded REIT, alongside a roughly 16,000-acre transaction; its initial Form D was filed February 19, 2026 with no capital raised disclosed at filing. Farmland LP closed fundraising on its $250 million-target Fund III in March 2026, by then holding four farms totalling 3,809 acres and more than $100 million of farmland, against $350 million and more than 19,000 acres firmwide across California, Oregon and Washington on the firm’s own reporting.

The trade is straightforward. A fund gives you diversification across farms and crops and a manager who can time dispositions; it takes away your ability to read one soil map and one lease before committing, and it usually adds a carry. If you cannot read a soil survey, the fund is the better product, and the single-farm listing was never giving you an informational edge.

Farmland’s gross cash yield is low enough that the wrapper’s annual charge should be compared to the rent rather than to the total return, because that is the comparison that decides whether you own an income asset or a leveraged bet on land prices.

The denominator first. US cropland cash rent averaged $160 an acre in 2026 against a cropland value of $6,020, a gross yield of 2.66%, per USDA NASS’s August and July 2026 releases. Property tax, insurance and farm management typically take a quarter of that, leaving roughly 2.0% net at the property before anyone in a wrapper is paid.

Now the wrapper. On a $25,000 position, the annual charges are these.

Set those against the roughly $500 of net property-level income a $25,000 slice of average cropland produces in a year. The Class S non-traded share consumes more than all of it. Farmland LP’s 2% consumes almost all of it, which is the honest logic of its strategy — the fund is selling an organic conversion that raises rent from about $300 to $775–$800 an acre on its own reporting, so the fee is being charged against a rent the manager claims to create rather than against the rent that was there. The platform’s 0.75% takes 38% of the net income before the entry and disposition fees. The listed REIT charges you nothing visible, and pays for its overhead out of the same rent inside the corporation, where you can read it in the income statement but cannot decline it.

The point is not that fees are outrageous. It is that on a 2.0% net yield a 1% wrapper fee is half your income, and arithmetic that works at private-equity return levels does not work here. The durable version of farmland ownership is the low-fee, long-hold one.

Farmland’s price is slow because it is appraised; the wrappers’ prices are fast because they are quoted, and the transmission mechanism in both directions is the interest rate.

For the land itself, the rate works through the capitalisation rate: a farm is a stream of cash rent, and the price a buyer will pay is that rent divided by a required yield. A one-point rise in the required yield on a 2.7% starting yield implies a very large fall in value — which is not what happened, because farm real estate is bought substantially by neighbouring farmers and families with no required yield at all, and because most transactions never come to market. Farmland’s rate sensitivity therefore shows up as a multi-year drift rather than a repricing: cropland values still rose 3.3% in USDA’s 2026 summary, to $6,020, while the Chicago Fed’s August 2026 AgLetter had Seventh District values flat year over year and down 3.7% after inflation.

For the wrappers, the rate works through the discount rate on a quoted security and through the cost of the debt in the capital structure. A listed farm REIT is a long-duration income equity: its dividend is small, its growth is slow, and its price therefore behaves like a long bond with equity risk stapled on. Gladstone Land’s preferred stack makes this sharper — the 6.00% Series B and Series C and the 5.00% Series E coupons are fixed claims, so when rates rise the preferred becomes more expensive relative to new issuance and the residual left for the common shrinks. That is why the common of the more income-oriented REIT has been the more volatile of the two.

For the private wrappers, the rate works through the appraisal with a lag of quarters, and through the exit. A platform farm bought in 2020 and scheduled to sell in 2027 is selling into whatever cap rate exists then; the sponsor’s projected IRR was built on the one that existed at purchase. That gap, not crop prices, is the largest single uncertainty in a fractional offering’s outcome, and it is the one the offering document quantifies least.

The tenant’s ability to pay sits underneath all of it. USDA’s September 2026 forecast put 2026 net farm income at $158.4 billion, down 2.6% from 2025 and 5.5% in real terms, with a record $47.4 billion of government payments inside the total. Cash rent is negotiated against that number, so a rent roll supported by government payments carries political risk into whichever wrapper holds the lease.

Two vehicles can both be called farmland and own assets with almost nothing in common, and three variables explain most of the difference in outcome.

Row crop against permanent crop

Annual row crops — corn, soybeans, wheat, cotton, rice — are replanted each season, leased on cash rent, and cheap to exit. Permanent crops — almonds, pistachios, citrus, apples, wine grapes — are multi-year plantings with large upfront development cost, higher yields when the market is good, and no ability to switch when it is not. Permanent crops led the index for nearly a decade before 2020 and then delivered the worst stretch in its history: −10.18% in 2024, with the almond index at −16.44% and pistachios at −16.5%, against annual cropland’s +5.66% in the same year. Annual cropland has out-returned permanent for six straight years through 2025.

Farmland Partners is the row-crop wrapper, Gladstone Land is the permanent-crop wrapper, and most platform offerings say which they are on the first page. A portfolio holding one of each is not diversified across farmland; it is long two different businesses that happen to share a soil profile.

Water

Irrigated ground in the West is worth what its water is worth, and the water is under legal pressure with fixed dates. California’s Sustainable Groundwater Management Act requires critically overdrafted basins to reach sustainability by 2040 and other basins by 2042; the Colorado River’s 2007 interim guidelines expire at the end of 2026, with the Interior Department’s August 2026 record of decision setting a framework for 2027 to 2036 and Lower Basin cuts locked for 2027 and 2028. Gladstone Land holds about 56,000 acre-feet of California water as a separate asset, which is a genuine differentiator and a genuine concentration. A wrapper that does not tell you the water right’s priority date, source and basin is not telling you what it owns.

Tenant and manager quality

In every wrapper the actual return depends on someone you will never meet: the operator who signs the lease and the asset manager who negotiates it. In a listed REIT that risk is disclosed in aggregate and diversified across hundreds of leases. On a platform it is one tenant on one farm, and the offering document’s description of the operator is the diligence you are relying on. Gladstone’s 2026 experience of vacant acreage moved to solar, cattle and fallowing is what tenant risk looks like when it arrives: not a default, just a lease that does not renew at the old rent.

Farmland rarely goes to zero; farmland wrappers can, and the failure modes are properties of the structure rather than of the soil.

The mark is made by the party paid on it

In every private wrapper — institutional fund, non-traded REIT, platform LLC — the value you see is an appraisal commissioned by the manager whose fee is a percentage of it. Its practical form is not fraud but lag: your statement is right about last year. Judge a private farmland manager on realised dispositions against the marks that preceded them, and on nothing else.

The exit is a permission, not a right

A platform interest has no market. A non-traded REIT’s repurchase plan is capped at 5% of NAV a quarter and can be modified or suspended by the board. An institutional fund has a redemption queue. In each case the ability to leave is a contractual permission that is most likely to be withdrawn in the conditions that make you want to use it. The only wrapper with a genuine daily bid is the listed one, and the price of that bid is the volatility described in section 5.

The listed wrapper imports equity-market risk

The 2018 Farmland Partners episode is the case: a 39% single-day decline and up to $115 million of market value erased by a pseudonymous article, later retracted and settled with a payment to the company in 2021. The unusual ending should not obscure the ordinary lesson, which is that a listed vehicle inherits every pathology of the equity market it trades in — short campaigns, index flows, tax-loss selling in December, liquidity that vanishes in a drawdown — none of which has any bearing on the crop.

Concentration hides inside diversification

A wrapper holding 142 farms can still be one bet. Gladstone Land’s permanent-crop tilt made it, in effect, a leveraged position on tree-nut economics through the worst tree-nut market in the index’s history, and 142 separate farms did not help. Read a portfolio by crop and by water basin, not by farm count.

The externally managed structure adds a party

Gladstone Land is externally managed by an adviser affiliated with its sponsor, a structure common in smaller REITs. It is disclosed, it is legal, and it means that the entity setting strategy is paid on assets rather than owning the equity alongside you. That is a reason to read the advisory agreement, not a reason to avoid the stock, but it belongs in the file.

Farmland’s tax treatment is one of the best in the alternatives universe, and most of what makes it good is unavailable through a REIT or a fund.

Listed and non-traded REIT shares

A REIT distribution arrives on Form 1099-DIV and splits three ways. The ordinary dividend portion is taxed at your marginal rate, reduced by the Section 199A 20% deduction for qualified REIT dividends, which the One Big Beautiful Bill Act (P.L. 119-21, July 4, 2025) made permanent for tax years beginning after December 31, 2025, holding the rate at 20% rather than the 23% the House bill proposed. The return-of-capital portion is not taxed on receipt but reduces your basis, deferring the tax to sale. The capital-gain portion, including any unrecaptured Section 1250 gain taxed at up to 25%, is passed through. For a 32%-bracket investor the effective rate on the ordinary slice is about 25.6%, plus the 3.8% net investment income tax where it applies. A REIT held inside an IRA generates no unrelated business taxable income, which is a real advantage over the partnership wrappers.

Platform LLCs and private funds

These issue a Schedule K-1. Rental income is ordinary, passive-activity limits under Section 469 apply to the losses, and you may acquire state filing obligations wherever the farm sits. In a self-directed IRA, a farm carrying a mortgage inside a partnership can generate debt-financed unrelated business taxable income under Section 514, which is the specific reason to check the leverage before putting a platform deal in a retirement account. The K-1 also arrives late, often after the April deadline.

What every wrapper forfeits

Three of farmland’s best tax features attach to the real property, not to a share in an entity that owns it.

• Section 1031. The 2017 tax law limited like-kind exchanges to real property, which is why farmland kept the deferral that art and wine lost. It also rewrote the statute: Section 1031(a)(1) now reaches only real property and Section 1031(a)(2) is the narrow exception for real property held primarily for sale, so the express exclusion of partnership interests that used to sit at Section 1031(a)(2)(D) is gone — and the result is the same, because a partnership interest is not real property and a REIT share is a security. Neither a platform LLC interest nor a REIT share can be exchanged. Only direct ownership — or a Delaware statutory trust interest, which under Revenue Ruling 2004-86 is treated as an undivided interest in the underlying real property rather than an interest in a business entity — qualifies. The hub’s 1031 guide walks the clocks and the intermediary.
• Section 2032A special-use valuation, which lets an estate value inherited farmland at its farm use rather than its development value, capped at a $1,460,000 reduction for 2026 deaths under Rev. Proc. 2025-32, requires qualified use by the decedent or family. Shares do not qualify.
• Conservation easements, the deduction under Section 170(h), are granted by the landowner. A fractional interest holder is not the landowner.

What the wrappers do keep is the stepped-up basis at death under Section 1014, which applies to shares and LLC interests as it does to land, and the federal estate exemption the same act set at $15 million per person from January 1, 2026, indexed and no longer scheduled to sunset. None of this is tax advice.

The clearest way to see what a wrapper costs is to give each one the same farm and the same decade and let the fee schedules do the rest.

The common assumptions

Every route below owns the same underlying economics, chosen to match what the asset has actually been doing rather than what it did in the 1990s: gross cash rent of 3.0% of land value, land appreciating 3.0% a year, property tax, insurance and management taking 25% of rent, a ten-year hold, and a 32% marginal federal bracket with the 3.8% net investment income tax applying. Ordinary REIT dividends get the 20% Section 199A deduction, an effective 29.4% all-in; long-term gains are taxed at 23.8%; state tax is excluded. Every figure is arithmetic on those assumptions, not a forecast.

Route one: the listed REIT

$25,000 buys shares less a 0.2% bid-ask spread on a small-capitalisation name, so $24,950 goes to work. The dividend is 3.4%, roughly FPI’s 2026 declared rate on our tape’s September 4, 2026 close, growing 3% a year with the land. Ten years of dividends total $9,725, taxed at 29.4% for $2,859, leaving $6,866. The shares compound at 3% to $33,531 and sell at another 0.2% spread for $33,464; the $8,464 gain costs $2,014. Total after tax: $38,315, or 4.36% a year — on the assumption doing the heavy lifting here, that the discount to appraised land value neither widens nor narrows. Close the 27% discount implied by a $14.64 NAV estimate and the answer is several points a year higher; double it and several points lower. That range is the listed wrapper in one sentence.

Route two: the platform single farm

A 2% closing fee means $25,000 buys $24,500 of land. Gross rent at 3.0% is $735 in year one, less 25% of property costs, less 0.75% annual servicing on asset value, leaving $367 of net cash in year one, all of it growing 3%. Ten years of net income totals $4,213, taxed as ordinary K-1 income at 35.8% for $1,508, leaving $2,705. The land compounds to $32,926 and sells with 5% of combined disposition and closing friction for $31,280; the $6,280 gain over a $25,000 basis costs $1,495. Total after tax: $32,490, or 2.66% a year.

Route three: the non-traded REIT, in two share classes

Assume the vehicle earns the same 5.25% gross after property costs. Class S, sold through a broker, takes the maximum 3.5% upfront selling commission, so $24,125 goes to work, and charges a 1.25% management fee plus an 0.85% annual shareholder servicing fee, leaving 3.15% net; the 12.5% performance participation is never triggered because 3.15% is below the 6% hurdle. Both charges are the Nuveen Farmland REIT’s own filed terms, read off its Form 10-12G registration statement rather than borrowed from a sister vehicle, so the worst outcome in this table rests on a schedule an investor would actually sign.

Assume 2.0% of NAV is distributed and the remaining 1.15% compounds. Ten years of distributions total $5,083, taxed at 29.4%, and NAV reaches $27,047. Total after tax: $30,148, or 1.89% a year. Class I, bought through a fee-only adviser with no commission and no servicing fee, pays 1.25% and nets 4.00%; distributions total $5,475 and NAV reaches $30,475, for $33,037, or 2.83% a year.

What the arithmetic says

Two results are worth carrying away. The first is the spread between Class I and Class S of the same fund holding the same farms: 94 basis points a year, or $2,889 over the decade, produced entirely by the door you walked through. The second is that the listed REIT wins here mainly because it charges no explicit fee and is assumed to track its assets; remove that assumption and the ranking is a function of the discount rather than of the wrapper’s merit. What the table establishes beyond assumption is the floor: on realistic farmland economics, a wrapper charging 2% or more a year cannot produce a return worth the illiquidity, whatever the projection says.

The mistakes in this market are made in the first month, and they are mistakes of sequence rather than of selection.

1. Decide whether you want farmland or farmland’s return profile. If you want the low correlation and the inflation hedge, the listed REITs will not give it to you and the private wrappers will only appear to, through appraisal lag. If you want a real asset that pays cash and holds value across decades, all four doors work and the question becomes cost.
2. Fix the size and the horizon before the vehicle. Assume private money is gone for seven to ten years and that a non-traded REIT’s repurchase cap will bind. A position you might need back inside three years belongs in the listed wrapper or nowhere.
3. Choose row crop or permanent crop deliberately. They are different businesses. Section 11’s 2024 numbers are the evidence; decide which you are buying and write down why.
4. Read the fee against the rent, not against the IRR. Divide the annual fee by expected net rent. Above one-third, you are buying appreciation.
5. Demand the full realised table from any platform — every offering funded before a stated cutoff, realised and unrealised, with money-weighted returns. Section 8 explains why the published average is not that.
6. Check the tax wrapper against the plan. REIT shares for an IRA, K-1s only if you can absorb late filings and state returns, and direct ownership if Section 1031 or Section 2032A is part of the plan.
7. Start with the liquid version. A year of holding the listed pair through a rate move teaches you more about your own tolerance than any offering document, and it costs one bid-ask spread to reverse.
8. Then diligence the specific asset: soil rating, water right with its priority date and basin, lease term and tenant, and the sponsor’s realised dispositions against the marks that preceded them.

These are the readings that would change the view in this guide, with the thresholds, so a reader in 2027 can check them in an hour.

• The NCREIF Farmland Index, quarterly. After −1.03% in 2024, +0.20% in 2025 and −0.20% in Q1 2026, four consecutive positive quarters with income above 3% would say the soft patch has ended. A second negative calendar year would make 2024 a regime rather than an event, and would justify a lower long-run assumption than the 9.8% history.
• The listed pair’s trailing ten-year total return. About 4.2% a year for Farmland Partners and 3.7% for Gladstone Land to September 2026 on vendor tabulations, against roughly 12.6% for the S&P 500. Recompute both each September. A trailing decade that clears 6% a year in two consecutive Septembers is the first evidence the listed wrapper delivers more than its dividend; one still under 5% in September 2029 is a verdict on the wrapper rather than a comment on the weather.
• Permanent cropland’s trailing-year return. Down 7.12% to June 2026. Back above zero for two consecutive quarters is the signal that tree-nut economics have cleared, and it is the single most important number for Gladstone Land.
• The Nuveen Farmland REIT’s raise. No shares were outstanding at June 30, 2026; the initial closing on September 1, 2026 sold 1,732,960 shares for about $34.7 million, or 1.16% of the $3 billion target. Three readings settle it, all of them on EDGAR: the first published monthly NAV, the size of the first repurchase queue once the plan is live, and whether the raise passes $300 million — a tenth of target — by September 2027. Below that, the wrapper has not found its buyer.
• Both listed REITs’ acreage. FPI at approximately 70,100 acres in 11 states and LAND at approximately 98,000 acres in 14 states as of Q2 2026, both after disposals. Continued shrinkage means these are liquidating vehicles priced on their discount; a return to net acquisition means the equity market has reopened for them.
• The discount to appraised net asset value. Estimates in 2025–2026 ran from about $14.64 to about $16 a share for FPI against a $10.65 close on our tape on September 4, 2026. Gladstone Land publishes its own estimated NAV per common share each quarter; the most recent figure we could retrieve was $17.59 at June 30, 2024, which predates the permanent-crop writedowns and should not be read as current — the number to look up is the one in the latest 10-Q. A discount that persists below 75% of a defensible NAV for a year, while the company sells farms at or above book, is the clearest arbitrage the sector offers — and its persistence is also the clearest evidence that the market disputes the appraisals.
• USDA net farm income and the government-payment share. $158.4 billion forecast for 2026 with $47.4 billion of payments inside it. A rising payment share is a warning about the durability of cash rents in every wrapper.
• Water deadlines with fixed dates. California’s SGMA sustainability dates of 2040 and 2042, and the Colorado River’s post-2026 framework covering 2027 to 2036 with Lower Basin cuts locked for 2027 and 2028. Any wrapper holding irrigated Western ground is exposed to these on a published calendar.

The positions in this guide do not depend on this quarter’s prints. They depend on the structure: the index is measured before fees, the private mark is made by the party paid on it, the exit is a permission in every wrapper but the listed one, and farmland’s best tax features do not survive being put inside a share.