The Annual State of USDA Rural Development Financing: 2026A record year, a delinquency cycle, and the repricing of rural credit

MMCG Invest, LLC. Feasibility Research

On January 19, 2026, USDA Deputy Secretary Stephen Vaden launched a dashboard called Lender Lens. Without much ceremony, it made the entire commercial guaranteed loan portfolio of USDA Rural Development available for download at the individual loan level: every Business & Industry loan, every Community Facilities guarantee, every REAP and Water & Waste Disposal credit, with borrower names, lender names, sector codes, and delinquency status attached, refreshed monthly (1)(2).

Four months later that transparency produced consequences. On May 12, 2026, Secretary Brooke Rollins removed ten lenders from the OneRD guaranteed lending platform, including several of the program's historically largest originators. USDA cited roughly $620 million in delinquent loans concentrated in their portfolios, approximately 47% of all of Rural Development's delinquencies (12).

Those two events frame the year in USDA rural lending. This is a system that just posted the largest Business & Industry program level in its 53-year history, $3.5 billion in FY2025, at a credit subsidy rate of 0.20%, the second-lowest on record (5)(6). The same system was carrying more than $1.6 billion of delinquency on a $14 billion guaranteed book as of May 31, 2026. It spent the fall working through a 42-day government shutdown. Its farm bill authorization has been extended three times. Its flagship energy grant program has been frozen, thawed, and finally suspended pending new rules. Its staff base is roughly a third smaller than it was eighteen months ago (10)(13)(23)(40).

None of this is contradictory once you look at the mechanics, and the mechanics are the point of this report. We wrote it for the people who underwrite against these programs: lenders, secondary-market buyers, sponsors, and the analysts who advise them.

This is the first edition of what MMCG intends to publish annually. It is built on the primary record: the Lender Lens loan-level file and the Rural Data Gateway, the OMB Federal Credit Supplement, the Federal Register, Congressional Research Service and GAO reports, congressional testimony, and USDA's own budget documents. This year the sources disagreed with each other more often than usual. Where they do, we say so, and we identify which source we treat as authoritative.

I. The number everyone gets wrong

The figure most often quoted about USDA rural lending this year will also be the one most often quoted incorrectly.

The $3.5 billion FY2025 Business & Industry number is an appropriated allotment. It is the ceiling on guarantee authority that Congress made available, and it is the highest in the program's history in both nominal and inflation-adjusted terms (5). It is not federal spending, and it is not the volume of loans actually made. Core B&I obligations have run between $1.6 and $1.9 billion annually in recent years: $1.799 billion in FY2024, $1.862 billion in FY2023. The OMB Federal Credit Supplement records FY2025 guarantee commitments of $3.228 billion, an unusually high year that came closest to exhausting the record allotment (5)(6)(7).

The distinction matters because the federal cost of the program is neither of those numbers. Under Federal Credit Reform Act accounting, the cost is the credit subsidy: expected defaults net of recoveries and fee income, expressed as a percentage of loans guaranteed. For B&I in FY2025 that rate was 0.20%, down from 2.29% a year earlier (6). At that rate, $3.5 billion in guarantee authority requires roughly $7 million in appropriated subsidy. The leverage works out to something like 500 to 1. A decade ago the subsidy rate was 5.11% (5). The path since then (2.32% in 2019, roughly zero in 2021, 2.29% in FY2024, 0.20% in FY2025) is the quantitative basis for the program's reputation as budget-efficient, and the reason B&I survived an administration budget request that proposed to gut it.

B&I Program Level vs. Core Obligations & Use Rate, FY2015–FY2025. Combo chart: allotment bars, core-obligation bars, use-rate line peaking at 98.24% in FY2024.

The forward curve is less flattering. The FY2027 Federal Credit Supplement, released April 3, 2026, shows the B&I subsidy rate rising to 1.09% for FY2026 and 1.93% for FY2027. It also carries a startling FY2027 commitment placeholder of just $85.2 million, which reflects the administration's proposed program reduction rather than any market forecast (7). Readers should treat the FY2025 rate of 0.20% as a cyclical low, not a new equilibrium.

B&I Credit Subsidy Rate, FY2015–FY2027. Line chart with FY2026–FY2027 shown as budget estimates; annotation at the FY2025 trough.

II. How the machine is built: OneRD, six years in

The architecture underneath all of this dates to October 1, 2020, when USDA consolidated four separate guaranteed loan regimes (B&I, Community Facilities Guaranteed, REAP Guaranteed, and Water & Waste Disposal Guaranteed) under a single regulation, 7 CFR Part 5001, known as the OneRD Guarantee Loan Initiative (3)(4). One application form, one annual fee notice, one lender-approval framework. For B&I, a flat 80% guarantee replaced the old tiered structure of 80% up to $5 million, 70% from $5 to $10 million, and 60% above.

Six years in, OneRD reads as a genuine simplification that also concentrated risk. A single approval opened all four programs to a lender at once, which lowered the barrier for nonbank and specialty originators, and the market responded. Average B&I loan size rose from $2.75 million in 2015 to $8.1 million in 2024 while annual loan counts fell from 309 to 176 (5). Fewer and larger loans, originated increasingly by a small set of high-volume specialists. The FY2020 lender report, the last static annual report USDA published before Lender Lens replaced it, shows the shape of that market: Live Oak Bank, Greater Nevada Credit Union, North Avenue Capital, Crestmark/Pathward, BOM Bank, United Community, Byline (5). Several of those names return in Section IV under worse circumstances.

The fee machinery adjusts each year by Federal Register notice. For FY2026, published at 91 FR 11272 and effective October 1, 2025, USDA reintroduced size tiering for the first time since consolidation. B&I loans under $5 million now carry an 85% guarantee. Loans of $5 to $25 million keep 80%. The standard 3.0% upfront guarantee fee and 0.55% annual retention fee are unchanged (14). Reduced-fee tiers at 1.0% upfront remain available for qualifying loans under $5 million in persistent-poverty, high-unemployment, and tribal communities, capped at 12% of program authority. A 90% guarantee exists only in the high-cost isolated Alaska carve-out (14)(15). We raise this because a competing reading has circulated in analyst commentary this year, claiming a 90% guarantee at a 1% fee for all sub-$5 million loans. That reading is wrong. The Federal Register table is unambiguous.

FY2026 OneRD Fee & Guarantee Schedule. Styled table: B&I <$5M / $5–25M / Reduced Fee / Alaska; CF; REAP; WWD — guarantee %, upfront fee, retention fee.

The statutory foundations are stable even where the funding is not. B&I sits under Section 310B of the Consolidated Farm and Rural Development Act (7 U.S.C. 1932(a)), Community Facilities under Section 306 (7 U.S.C. 1926), REAP under Section 9007 of the 2002 farm bill (7 U.S.C. 8107) (4)(16). One recurring underwriting trap deserves a permanent footnote. The rural population thresholds differ by program: B&I and REAP use 50,000 or less, while the Community Facilities direct and grant programs use 20,000. Conflating the two is one of the most common eligibility errors we see in loan files.

III. Business & Industry: the flagship, examined

B&I accounts for roughly half of Rural Business-Cooperative Service obligation dollars, which makes it the natural center of this report. The decade's numbers come from the Summit LLC economic assessment prepared for the National Rural Lenders Association and entered into the House Agriculture Committee record in September 2025. It remains the best loan-level series publicly available, built from a FOIA extract covering FY2000 through FY2024 and validated against the Rural Data Gateway (5).

Core obligations rose 108% from 2015 to 2024. Total core obligations for the decade came to $13.4 billion, with 60% of that arriving in the last five years. Use rates, meaning obligations as a share of allotment, climbed from 70% in FY2015 to 90.83% in FY2023 and a record 98.24% in FY2024. USDA has said FY2024 would have hit 100% but for year-end de-obligation recapture (5). Demand caught up with supply, which is exactly why the $3.5 billion FY2025 allotment mattered.

Where does the money go? Hospitality leads, and has led every year since 2015. Accommodation and food services, meaning hotels, resorts, RV parks, and restaurants, absorbed roughly $4 billion over the decade and 32.06% of FY2024 obligations. Manufacturing runs a consistent second at roughly $2.3 billion and 18.48% in FY2024. The volatile sectors are the interesting ones. Mining and oil and gas spiked to 15.76% of FY2023 obligations before falling back to 5.66%. Utilities swung from 3.44% to 10.54% in a single year (5). The program bends with administration priorities, which is a feature for policymakers and a modeling variable for the rest of us. A third of the portfolio in a cyclical, discretionary-spending sector is also a concentration worth pricing.

B&I Sector Mix, FY2024. Donut with decade-trend sparkline for hospitality and manufacturing shares.]

Geographically the program tilts South. North Carolina ($957 million), Louisiana ($926 million), and Texas ($838 million) lead cumulative FY2015 to FY2024 obligations. Alaska, Wyoming, and Louisiana lead per capita, and the urban Northeast trails. The texture inside the table is instructive. Missouri booked 269 loans for $481 million, a community-bank market of smaller credits. Texas booked 97 loans for $838 million, a large-project market (5).

B&I State Distribution, Cumulative FY2015–FY2024. Choropleth with top-10 table; per-capita toggle.

Credit performance has been cumulatively strong and recently stressed. Lifetime gross losses of roughly $501.6 million against $13.4 billion obligated over the decade work out to about 3.7%, roughly half the comparable SBA 7(a) cumulative charge-off experience of about 7.5% (5)(30). Losses concentrate in the pre-2012, 2013, and 2015 vintages. The FY2021 to FY2024 average ran just under $30 million a year. That history is what the 0.20% subsidy rate was pricing. What it was not pricing is the subject of the next section.

IV. The reckoning: delinquency, desk audits, and the ten

The first public signal came in a letter. Between February 18 and 20, 2026, Rural Business-Cooperative Service Administrator J.R. Claeys wrote to the program's more than 775 OneRD lenders. The active guaranteed balance exceeded $12 billion. Delinquent loans exceeded $1 billion. The agency had paid approximately $300 million in repurchases and losses over the prior twelve months. The letter cited 7 CFR 5001.132, the lender-removal authority, and warned that negligent underwriting or servicing would draw corrective action (11).

On May 12, USDA acted. Ten lenders were removed: Bank of Montgomery, Byline Bank, Celtic Bank, Community Bank & Trust of West Georgia, Genisys Credit Union, Greater Nevada Credit Union, North Avenue Capital, Optus Bank, U.S. Eagle Federal Credit Union, and ReadyCap Commercial. Their portfolios, per USDA's release, held approximately $620 million in delinquent loans, about 47% of Rural Development's delinquent stock (12). Lender Lens data shows North Avenue Capital carrying the single largest 90-plus-day exposure at $168.5 million. Greater Nevada Credit Union was second at $165.9 million, roughly a third of its OneRD book. Optus Bank's smaller portfolio ran a 57% delinquency rate (13). By May 31 the outstanding OneRD book had grown to $14 billion, with about $1.6 billion delinquent (13).

A few things are worth separating here.

The numbers themselves are a time series, not a contradiction. Twelve billion and a billion-plus in February; fourteen billion and $1.6 billion at the end of May. Delinquency is running 8 to 11% of balance, elevated for a program whose historical norm sat below 5%, and it rose within the snapshot window.

The overlap between the removal list and the historical league table is the structural finding of the year. North Avenue Capital was USDA's number-one traditional B&I lender in FY2020. Greater Nevada Credit Union was number one including CARES-era volume. Byline, Celtic, BOM, and ReadyCap were fixtures of the top tier (5)(12). The originate-to-distribute specialists who built the modern B&I market are now, disproportionately, the institutions outside it. And the delinquency was not spread evenly. It sat in fewer than fifteen of 760 lenders, against a program-wide average exposure of roughly $2 million per lender (13).

This is also contested. ICBA and America's Credit Unions wrote to Secretary Rollins on June 26, 2026, challenging the process and its transparency. At least three removed lenders have signaled appeals. Senator Joni Ernst has asked SBA to examine the same lenders' 7(a) books (13). The $620 million and 47% figures are USDA's assertions, not adjudicated fact, and we present them that way.

Our read is that this is a genuine inflection rather than a crisis. The guarantee is a full-faith-and-credit obligation of the United States, so holders of guaranteed paper are insulated. The stress lands on originator economics, on secondary-market appetite for marginal paper, and on processing timelines. The leading indicators worth watching are the FY2027 subsidy re-estimate (a move back above roughly 2% would raise the appropriated cost of a given program level about tenfold from FY2025 levels), delinquency crossing $2 billion, and the disposition of the appeals. Underwriting had tightened well before the removals, for what it is worth. OCC data show B&I approval rates falling from 89.1% of submissions in FY2021 to 52.7% in FY2023 (30).

The OneRD Portfolio Under Stress. Time series: active balance and delinquent balance (Feb → May 31, 2026), with the removed-lender share ($620M / ~47%) broken out.

V. Community Facilities: the largest program nobody talks about

If B&I is the flagship, Community Facilities is the fleet, and it is far larger than most market participants realize because almost all of it is direct federal lending rather than guarantees. In FY2024, Congress supported roughly $2.8 billion in CF direct loan authority with essentially no loan subsidy, against $650 million in guaranteed authority and a mere $5 million in competitive grants (16)(17). Realized FY2023 obligations ran $1.08 billion in direct loans, $183 million in guarantees, and roughly $149 million in grants, a 6-to-1 direct-to-guaranteed ratio (17). Terms run to forty years at up to 100% of project cost. Three interest-rate tiers apply: a poverty rate set at 4.5%, plus intermediate and market rates adjusted quarterly. There is no prepayment penalty, and the rate locks at the lower of approval or funding. The quarterly market rate swung from 4.125% to 5.25% within calendar 2024 before settling near 4.75%. On a forty-year amortization that swing moves feasibility conclusions, and it is one reason we time CF applications the way bond issuers time pricing (17).

The evidence base for the program's marquee use, rural hospitals, matured this year. USDA ERS report ERR-344 (Rupasingha and Cho, January 2025) is the first rigorous matched-sample study of CF's effect on hospital survival, and its findings are striking. CF-recipient nonmetro hospitals were 94% less likely to close six years after funding than matched non-recipients, 90% less likely at eight years, and 88% at ten (18). The context makes the finding urgent rather than academic. Between 2005 and 2023, 146 nonmetro hospitals closed or converted away from inpatient care, 81 of them outright shutdowns, with 44% of closures in the Southeast (18). Chartis's February 2026 assessment puts 417 rural hospitals at risk of closure and 41.2% of rural hospitals operating in the red (19).

There is a mismatch buried in the regional data that we consider the analytical heart of the rural health finance question. The Southeast absorbed 64 of 146 closures but received only 23% of rural CF hospital dollars over 2000 to 2020. The Plains received 38%, or $1.8 billion. The Rocky Mountain region, with zero closures, received 9% (18). Capital is not flowing to where the closure risk is. CF hospital financing is also overwhelmingly debt rather than grants: 99.8% loans and 0.2% grants out of the $5.7 billion invested in hospitals over two decades (18). For distressed Southeastern systems that cannot service new debt, the instrument itself is part of the constraint.

The near-term story is congressional. FY2026 appropriations made Community Facilities the single largest line-item increase in the entire Agriculture bill: $659 million in congressionally directed spending, meaning earmarks, against $5 million of competitive money the year before (8)(9). The pipeline for 2026 and 2027 CF grants runs, in practice, through Members' offices. We advise sponsors accordingly.

VI. REAP: the boom, the freeze, and the cliff

No program had a stranger three years than the Rural Energy for America Program. The Inflation Reduction Act's Section 22002 appropriated $2,025,450,000 for REAP, roughly forty times the program's permanent $50 million annual farm-bill baseline. It doubled the maximum grant share from 25% to 50% and set the money to run through September 30, 2031 (20). Deployment was fast. By USDA's count, 8,012 IRA-funded clean-energy projects drew $1.3 billion. By Brookings' accounting, just over $1.5 billion in total REAP grant obligations went out between October 2022 and March 2025, equal to 68% of all grant value in the program's history, with the average grant rising from $31,641 in the first Trump administration to $138,506 in the Biden years (20)(21). Roughly three-quarters of projects involved solar (22)(39).

Then, on January 20, 2025, Executive Order 14154 froze IRA disbursements. The freeze captured approximately $911 million in already-obligated REAP grants, about 75% of the program's IRA funding, affecting more than 4,800 recipients. Most of them, as Brookings noted with some irony, sit in Republican congressional districts (21)(35). Because REAP is reimbursement-based, farmers who had already spent their own money went unpaid, and litigation followed. Cultivate KC v. USDA had cross-motions for summary judgment pending as of late 2025 (24). USDA released obligated funds in late March 2025, paired with a 30-day window for recipients to voluntarily strip DEIA and climate elements from their projects (37). The FY2026 application window was cancelled in June 2025. Restrictions on ground-mounted solar over 50 kW and on foreign-manufactured panels followed in August. On April 15, 2026, USDA rescinded the FY2025 through FY2027 funding notice outright, pending a rewrite of the REAP regulation (23).

Two clarifications, because the market keeps getting both wrong. First, the One Big Beautiful Bill Act did not rescind REAP's IRA money. Its agriculture rescissions hit IRA conservation and forestry accounts, not Section 22002. The statutory text is unambiguous despite loose secondary reporting (33). Second, the loan-guarantee channel stayed open throughout. It is the grant program that is suspended.

The budget arithmetic has turned regardless. The REAP guaranteed subsidy rate flipped from negative 0.88% in FY2025, a program that made money for the Treasury, to positive 4.54% in FY2026 and 3.78% in FY2027, on commitments collapsing from $2.0 billion to $100 million (6)(7). The cliff is structural. Roughly half the IRA money is obligated, no new grants have been awarded since January 10, 2025, and when the IRA funds run out the program reverts to its $50 million baseline. As one of REAP's original drafters put it, no IRA-like opportunity is coming again soon (39). Our house position for feasibility work is blunt: model no new REAP grant until a replacement notice publishes, model reimbursement risk explicitly on executed agreements, and treat the guaranteed loan as the surviving instrument.

VII. Broadband, power, and water

The infrastructure programs completed a handoff this year. USDA's ReConnect broadband program closed its books, for now, at $5.54 billion cumulative across five rounds: $692.81 million, $927.69 million, $1.62 billion, $1.8 billion, and $476.45 million, plus $17 million in directed spending, with no round currently open (25)(26). The deployment role has passed to NTIA's $42.45 billion BEAD program, restructured in June 2025 under the "Benefit of the Bargain" policy notice to remove the fiber preference and require technology-neutral, lowest-cost awards. NTIA projects roughly $21 billion in savings against the original allocation, and Texas's award alone fell from $3.31 billion to about $1.26 billion in the re-run round (27). Whatever one thinks of the policy, the practical guidance is uniform. Rural broadband capital in 2026 through 2028 flows through state BEAD offices, not USDA, and it is now open to fixed wireless and satellite.

The electric side splits in two. The legacy RUS Electric Loan Program, cooperative lending under the 1936 Rural Electrification Act, runs at negative subsidy rates (negative 1.61% on Treasury-rate loans and negative 4.30% on FFB loans in FY2025) and had its loan authority raised to $7 billion for FY2026. It is among the most fiscally sound credit programs in the federal government (7)(9). The IRA-funded New ERA program ($9.7 billion, with roughly $9 billion obligated to cooperatives by January 2025) and PACE ($1 billion appropriated, with more than $1.6 billion in partially forgivable loans obligated) went through the same freeze, review, and conditional release cycle as REAP, and OBBBA rescinded their unobligated residuals (33). Obligated awards appear protected. Execution and timeline risk on revised project scopes is real.

Water & Waste Disposal remains what it has always been: the only federal program exclusively dedicated to rural water and wastewater, with $1.6 billion obligated across 705 projects in FY2024, forty-year terms, and grants up to 75% of cost for the poorest communities, now with its guaranteed side folded into OneRD (17). It was the second-largest earmark recipient in the FY2026 bill.

VIII. The politics of the money

Every figure above sits inside a policy structure that is, at this writing, held together with extensions. The 2018 Farm Bill has now been extended three times, most recently through September 30, 2026, inside the same November 12, 2025 law that ended the 42-day government shutdown, the longest in U.S. history, and enacted full-year FY2026 agriculture appropriations (9)(10)(34). Some trade coverage says 43 days. CRS counts 42 full days, October 1 through November 11, and we use CRS. The One Big Beautiful Bill Act functioned as a partial farm bill in July 2025, rewriting the commodity and nutrition titles by reconciliation, but it left the Rural Development title untouched. The House Agriculture Committee's follow-on farm bill, marked up beginning February 2026, remains stalled over SNAP (33)(34).

The appropriations outcome is the clearest data point on where power over rural credit actually sits. The administration's FY2026 request proposed cutting agriculture discretionary spending 21.5%, zeroing Community Facilities grants and most Rural Business-Cooperative Service programs as duplicative, and cutting Rural Development staffing by a third. Congress enacted $26.6 billion, a 1.3% increase, with Rural Development Title III up 19.1% to $4.089 billion, B&I loan authority at $1.75 billion, and the earmark-driven CF surge described above (8)(9). The FY2027 request proposes another 19% cut. Two cycles of evidence suggest Congress will again decline.

Beneath the appropriations fight, the agency itself is being rebuilt. Rural Development lost roughly 36% of its staff in 2025 through the deferred-resignation program and other separations, and loan and grant functions are consolidating into Dallas-Fort Worth and St. Louis hubs (40). We now build 30-to-45-day processing buffers into every USDA-dependent closing timeline, and we recommend clients do the same.

Then there is the paradox that gives this field its peculiar politics. REAP grants flowed 64.9% to Republican-held House districts under the first Trump administration and 66.7% under Biden. The distribution is structural, a function of rural geography rather than partisan targeting, yet Republican budget requests repeatedly propose eliminating the programs and congressional Republicans repeatedly restore them (21). Persistent-poverty targeting adds its own wrinkle. The statutory 10-20-30 set-aside directs at least 10% of designated funds to persistent-poverty counties, but GAO found Rural Development missed that threshold in at least one year for six of ten designated accounts (32). Rural capital, in short, is politically protected but imperfectly aimed.

IX. The capital markets layer: where USDA paper actually trades

The least-covered dimension of this market also determines whether it functions: the secondary market for guaranteed portions. It is real, it is profitable, and it is structurally thinner than its SBA cousin, which has consequences.

SBA runs a mandated, standardized system with a single fiscal transfer agent, a pooled certificate market, and deep liquidity. USDA mandates neither. A USDA guaranteed portion leaves an originator's balance sheet one of two ways: an SBA-style one-time premium sale through broker-dealers via a tri-party assignment agreement, or a sale to Farmer Mac, where, in USDA's own words, the lender continues to earn the full spread between the note rate and the Farmer Mac base rate rather than a one-time premium (28)(29)(30). Farmer Mac's on-balance-sheet USDA securities stood at roughly $2.40 billion at year-end 2024, essentially flat within a company growing toward $33 billion in total volume. That is a run-off book, not a growth engine (28). There is no USDA analog to SBA pool certificates, and no published USDA premium series. SBA benchmarks (premiums averaging around 110% of par in 2025, with a 50% program user fee on any premium above 110) are the closest available proxy, and we label them as such wherever they appear (30)(31).

Why originate at all, then? Because the economics stack three ways. The guaranteed portion sells at a premium; Live Oak disclosed roughly a 7% average premium on about $322 million of guaranteed sales in a single 2025 quarter. The originator capitalizes a servicing asset earning 40 to 100 basis points annually. The retained unguaranteed strip is a yielding asset (31). Meanwhile the guaranteed portion carries a 0% or 20% risk weight as a full-faith-and-credit U.S. obligation, sits outside the bank's legal lending limit, and can earn CRA credit (30)(38). The March 2026 bank-capital re-proposal leaves that sovereign-guarantee treatment intact (38). Add B&I's lifetime loss experience of about 3.7% against 7(a)'s roughly 7.5%, and the investment case for seasoned guaranteed paper is straightforward (5)(30).

The program-selection question, B&I versus SBA 7(a), comes down in our practice to three variables: size, geography, and speed. B&I lends to $25 million, or $40 million for value-added agricultural cooperatives, with no small-business size standard and no credit-elsewhere test, on negotiated fixed or variable rates with terms to thirty years on real estate. 7(a) caps at $5 million, prices off Prime, requires a size standard, and clears in days under delegated authority with a far more liquid exit (16)(30). The FY2026 move to an 85% guarantee on sub-$5 million B&I loans deliberately narrows 7(a)'s advantage at the small end (14). For rural projects above $5 million, B&I has no real federal competitor.

The 2026 stress reprices this market at the margin, not at the core. Holders of guaranteed paper are protected by the guarantee. What changes is the origination side. Several of the most active distributors are gone, secondary buyers are scrutinizing originator quality by name (Lender Lens makes that trivially easy now), and the spread between top-tier and marginal paper should widen. For institutional buyers, our guidance is to favor post-2015, and especially post-2021, vintages, to demand a liquidity premium over comparable SBA paper, and to read the monthly Lender Lens file the way one reads servicer reports.

[EMBLEM 10 — B&I vs. SBA 7(a) vs. SBA 504: Program Comparison Matrix. Guarantee %, max size, geography, size standards, rates, terms, fees, secondary liquidity, loss experience.]

X. What we are watching into FY2027

Five markers will tell us whether 2026 was an inflection or an aberration.

The FY2027 subsidy re-estimate for B&I. The FCS estimates of 1.09% and 1.93% for FY2026 and FY2027 already price some normalization. A re-estimate above roughly 2% would mean the delinquency cycle has reached the budget, raising the cost of every dollar of program level about tenfold from the FY2025 low (7).

The delinquency curve. $1.6 billion on $14 billion at May 31. Crossing $2 billion, north of 14% of balance, would move our assessment from contained inflection to something requiring a harder word (13).

The lender appeals and the trade-group challenge. If removals are reversed or the process is found deficient, the enforcement regime that Lender Lens enabled loses its teeth. If they stand, expect the desk-audit era to be permanent (13).

The REAP replacement rule. A published NOFO restores grant-based capital stacks to rural energy feasibility. Continued silence runs the clock toward the $50 million baseline (23).

And the farm bill. A genuine Title VI reauthorization could hard-wire mandatory baselines, or it could codify the current restrictions. The extension expires September 30, 2026. We are not holding our breath, and neither should sponsors with 2027 closings (34).

The larger point of this first annual report is that USDA Rural Development financing in 2026 is bigger, cheaper to run, more transparent, and more stressed than at any point in its history, and that each of those conditions helped produce the others. The record allotments drew in the volume. The volume drew in the specialists. The specialists' underwriting is now being marked to market in public, monthly, at the loan level. For borrowers with sound projects in eligible geographies, the machinery still works, and works at scale. For everyone underwriting against it, the era of taking the plumbing of the guarantee ecosystem for granted is over.

Methodology and data note

This report is built on primary sources wherever they exist. Program-level obligations and portfolio data come from the USDA Rural Data Gateway (a rolling ten fiscal years, project-level for most programs) and the Lender Lens dashboard (loan-level OneRD portfolio, launched January 19, 2026, refreshed monthly) (1)(2)(36). Credit subsidy rates come from the OMB Federal Credit Supplement volumes accompanying the FY2025 and FY2027 President's Budgets. These are budget estimates rather than audited actuals, and re-estimates appear separately (6)(7). Fee and guarantee schedules come from the annual OneRD Federal Register notices, most recently 91 FR 11272 (14). The FY2015 to FY2024 B&I loan-level series derives from the Summit LLC economic assessment entered into the House Agriculture Committee record. Its programmatic data trace to a FOIA extract validated against the Rural Data Gateway, and we cite its macroeconomic impact claims as association analysis rather than official statistics (5). The February 2026 portfolio figures come from the RBCS Administrator's letter to lenders, and the May 2026 figures from USDA's removal announcement and subsequent trade reporting. Where snapshots differ, we present them as a dated series (11)(12)(13). Delinquency attributions to individual removed lenders reflect Lender Lens data as reported and remain subject to pending appeals. Figures marked as provisional in the text should be re-verified against the monthly Lender Lens file before being cited onward.

MMCG Invest, LLC is a national commercial real estate feasibility consulting firm serving SBA 7(a), SBA 504, USDA, and conventional lenders across more than thirty asset classes.

Author: Michal Mohelsky, J.D., Principal, MMCG Invest, LLC

Request a Feasibility Study → Book a Meeting: https://calendar.app.google/EJzWEz3GCqLY2jU86

Michal Mohelsky, J.D. | Principal | mmcginvest.com 

Contact: michal@mmcginvest.com

Phone: (628) 225-1125

Disclaimer: This report is provided for informational purposes only and does not constitute investment advice. Data presented herein is derived from proprietary MMCG databases and third-party sources believed to be reliable; however, MMCG Invest makes no representation as to the accuracy or completeness of such information. Figures from third-party industry databases have been independently verified and, where appropriate, adjusted to reflect MMCG's proprietary analytical methodology. Past performance is not indicative of future results.

Sources

(1) USDA, "USDA Launches Lender Lens Dashboard to Promote Data Transparency," press release, January 19, 2026. (2) USDA Rural Development, Lender Lens: OneRD Loan Portfolio, rd.usda.gov/rural-data-gateway/onerd-loan-portfolio. (3) OneRD Guaranteed Loan Regulation, Final Rule, 85 Fed. Reg. 42518 (July 14, 2020). (4) 7 C.F.R. Part 5001 (Guaranteed Loans), as amended through December 2025. (5) Summit LLC, "USDA B&I Guaranteed Loan Program: Economic Assessment 2025," prepared for the National Rural Lenders Association; House Agriculture Committee witness document, September 18, 2025. (6) OMB, Federal Credit Supplement, Budget of the U.S. Government, FY2025 (BUDGET-2025-FCS), Tables 1–2. (7) OMB, Federal Credit Supplement, Budget of the U.S. Government, FY2027 (BUDGET-2027-FCS), April 3, 2026, Tables 1–2. (8) Congressional Research Service, R48564, "Agriculture and Related Agencies: FY2026 Appropriations," updated February 17, 2026. (9) P.L. 119-37 (H.R. 5371), Division B (FY2026 Agriculture Appropriations) and Division E (farm bill extension), enacted November 12, 2025. (10) Congressional Research Service, R48765; R41759; RS20348 (funding-gap duration). (11) USDA Rural Business-Cooperative Service, Administrator J.R. Claeys, letter to OneRD lenders, February 2026. (12) USDA, Release No. 0065.26, removal of ten OneRD lenders, May 12, 2026; USDA OneRD Revocation FAQ. (13) American Banker / Asset Securitization Report, OneRD lender-removal coverage and May 31, 2026 portfolio figures; ICBA and America's Credit Unions letter to Secretary Rollins, June 26, 2026. (14) OneRD Annual Notice of Guarantee Fee Rates, FY2026, 91 Fed. Reg. 11272 (March 9, 2026; effective October 1, 2025). (15) OneRD Annual Notice of Guarantee Fee Rates, FY2025, 89 Fed. Reg. 53041 (June 25, 2024). (16) Congressional Research Service, R47438, "Federal Credit Assistance and Grant Programs for Rural Businesses." (17) Congressional Research Service, R48462, "Rural Community Facilities: A Guide to Programs," March 19, 2025; CRS R46471. (18) Rupasingha, A., and J. Cho, "Federal Assistance and Rural Hospital Closings: The Impact of the USDA Community Facilities Program," USDA ERS, ERR-344, January 2025. (19) Chartis Center for Rural Health, "Rural Health State of the State," February 2026. (20) Inflation Reduction Act, P.L. 117-169, §22002; IRA Tracker (Columbia/EDF), Section 22002 program page. (21) Pipa, A., et al., "The Rural Energy for America Program primarily benefits Republican congressional districts," Brookings Institution, April 2025. (22) National Sustainable Agriculture Coalition, REAP funding-freeze and deployment analyses, 2025–2026. (23) Notice of Rescission of Funding Opportunity for the Rural Energy for America Program, 91 Fed. Reg. 20090 (April 15, 2026). (24) Cultivate KC v. USDA, No. 1:25-cv-00737 (D.D.C., filed March 13, 2025). (25) USDA, ReConnect Loan and Grant Program, cumulative program page (accessed 2026). (26) Congressional Research Service, R47017, "USDA's ReConnect Program: Expanding Rural Broadband." (27) NTIA, BEAD Restructuring Policy Notice ("Benefit of the Bargain"), June 2025; NTIA administrator statements on projected savings, December 2025. (28) Federal Agricultural Mortgage Corporation, Form 10-K FY2024 (Table 18) and Q3 2025 results, November 3, 2025. (29) Congressional Research Service, IF11595, "Farmer Mac and Its Board Members." (30) OCC, Community Developments Insights: USDA Rural Development Business & Industry Guaranteed Loans, June 2025. (31) Live Oak Bancshares, Inc., Forms 10-K, FY2024–FY2025. (32) GAO-21-470, "Areas with High Poverty: Changing How the 10-20-30 Funding Formula Is Applied Could Increase Impact in Persistent-Poverty Counties." (33) One Big Beautiful Bill Act, P.L. 119-21 (H.R. 1), July 4, 2025, §10201 and related rescission provisions; American Farm Bureau Federation and CALT analyses. (34) Congressional Research Service, R47659 (farm bill extensions); IF12115 (programs without baseline). (35) Executive Order 14154, "Unleashing American Energy," 90 Fed. Reg. 8353 (January 20, 2025). (36) USDA Rural Development, Rural Data Gateway, "Rural Investments" data tables and Data Dictionary. (37) USDA, "USDA Delivers on Rural Energy Commitments, Provides Path for Applicants to Support U.S. Energy Independence," March 25, 2025. (38) OCC, Federal Reserve, FDIC, Regulatory Capital Rules re-proposal, 91 Fed. Reg. (March 27, 2026), Docket R-1888. (39) Environmental Law & Policy Center analyses via Canary Media, REAP coverage, 2025–2026. (40) USDA reorganization announcement, July 24, 2025; NSAC analysis of Rural Development staffing separations, 2025.