Comment in Support of Proposed Rule, “Agricultural Foreign Investment Disclosure Act of 1978”

U.S. Department of Agriculture
Office of Homeland Security
1400 Independence Avenue SW
Washington, DC 20250

Re: Comment in Support of Proposed Rule, “Agricultural Foreign Investment Disclosure Act of 1978,” Docket No. USDA-2026-0001; RIN 0560-AI70; 91 Fed. Reg. 38315 (June 25, 2026)

To Matt Allen, Director of the Office of Homeland Security:

The America First Policy Institute submits these comments through its China Policy Initiative, which researches foreign adversary ownership of American agricultural land and agricultural infrastructure and has published on the matters addressed in this rulemaking. We write in support of the Department of Agriculture’s (hereafter "Department” or “USDA”) proposed revisions to the regulations implementing the Agricultural Foreign Investment Disclosure Act of 1978 (AFIDA), and we urge the Department to finalize the rule with its adversary-specific provisions intact. The proposed rule rests on a correct and long-overdue conclusion: The national security risk presented by foreign holdings of American agricultural land is a function of the identity, location, and control structure of the holder rather than a function of acreage, and ultimately of whether said holder answers to an adversary government.

I. The Rule Correctly Measures Risk by Identity, Location, and Control, Not Acreage

The most recent AFIDA annual report records foreign persons holding an interest in nearly 46 million acres of U.S. agricultural land (corresponding to approximately 3.6% of privately held agricultural land in the United States).1 Canadian investors alone account for roughly a third of that total,2 while Chinese primary-investor filers reported 247,659 acres—less than 1% of foreign-held acres.3 At first blush, it would seem that, as a small piece of an already small category of foreign holdings of U.S. farmland, adversary-linked holdings of U.S. farmland are negligible. In fact, this perspective misses several important points.

First, the aggregate numbers conflate categorically different phenomena. The 16 million acres attributable to investors from Canada—a treaty ally—present a fundamentally different risk profile than land held by entities ultimately answerable to the government of the People's Republic of China (PRC). A reporting system that treats these holdings identically is blind, not neutral. The proposed rule’s tiered structure corrects that oversight while avoiding the introduction of meaningful new burdens on investors from friendly nations.

Second, the PRC-associated total is dominated by a small number of corporate holders. That figure is established by the transactions themselves rather than by the ledger: WH Group’s 2013 acquisition of Smithfield Foods transferred more than 146,000 acres in a single deal, and a subsidiary of Guanghui Energy assembled roughly 130,000 acres in one Texas county beginning in 2015. Both networks sit at the top of USDA’s list of Chinese-owned holders. Although the precise totals are contested, and the Government Accountability Office (GAO) found that AFIDA’s data double-counted the largest PRC-associated holding, there is no dispute that PRC-linked entities own U.S. farmland. The policy question therefore concerns a handful of large corporate holders whose ownership chains terminate in Beijing's jurisdiction, not thousands of dispersed smallholders. Disclosure obligations calibrated to identity and control are the right instrument for that fact pattern.

Third, and most importantly, the recorded acreage is a floor, not a ceiling. The AFIDA report itself cautions that country-specific totals “should be interpreted as a minimum,”4 and the GAO found in 2024 that AFIDA data contained significant errors, including double-counting of the single largest holding associated with the PRC, and that USDA was not sharing the data with the Committee on Foreign Investment in the United States (CFIUS) member agencies on a timely or usable basis.5 A dataset that cannot reliably count the largest Chinese holding in the country is not one from which reassuring conclusions can be drawn. The proposed rule’s electronic filing portal, geospatial reporting, expanded beneficial ownership disclosure, and enforcement provisions are the necessary predicate for the data to be reliable.

II. The Record Confirms That Identity, Location, and Control Are the Operative Variables

Control of supply-chain chokepoints. Two CFIUS-cleared transactions show how modest land footprints can accompany very large positions of control. WH Group's 2013 acquisition of Smithfield Foods, then the largest Chinese acquisition of an American company, was financed in part by a $4 billion loan from the state-owned Bank of China collateralized by U.S. assets. The transaction transferred more than 146,000 acres of farmland along with processing plants and proprietary hog genetics. Smithfield-affiliated entities appear in the most recent AFIDA report at approximately 100,000 acres; the difference appears to reflect subsequent divestiture and restructuring, not a discrepancy in the record.6 WH Group retains approximately 87% of Smithfield's common stock today,7 and Smithfield accounts for roughly a quarter of American hog slaughter, the largest share of any packer.8 State-owned ChemChina's $43 billion acquisition of Syngenta, financed overwhelmingly by state institutions through a chain of offshore special-purpose vehicles,9 placed under ultimate Chinese jurisdiction a company that operates manufacturing and research facilities across the United States, licenses seed technologies to American farmers, and remains the primary producer of atrazine, one of the two most intensively used herbicides in the country.10 Syngenta's state-owned acquirer, ChemChina, appears on the Department of Defense's Section 1260H list of Chinese military companies operating in the United States.11 Yet in AFIDA's ledger, Syngenta registers as a scattering of modest research parcels. Acreage is not the appropriate metric to capture the influence of Syngenta on domestic agriculture.

The legal regime behind these ownership chains is troubling. China's National Intelligence Law of 2017 and Data Security Law create an affirmative obligation for firms under Chinese jurisdiction to assist state security organs on request,12 while China prohibits reciprocal foreign investment in its own agricultural sector—seed sales, technology licensing, acquisitions, and land ownership—on explicitly national-security grounds.13

Location, not acreage. In 2022, China’s Fufeng Group purchased approximately 370 acres 12 miles from Grand Forks Air Force Base, a parcel the Department of the Air Force assessed as “a significant threat to national security,”14 but one CFIUS reviewed and determined it lacked jurisdiction to reach. Whatever the merits of that determination, its consequence is that AFIDA disclosure was the only federal instrument that touched the transaction at all.15 Meanwhile, in Val Verde County, Texas, a subsidiary of China's Guanghui Energy assembled roughly 130,000 acres near Laughlin Air Force Base, the Air Force’s largest pilot-training installation, with a proposed wind project along low-level training routes. CFIUS, which did have jurisdiction in that case, found no unresolved concern, a determination Texas answered by enacting the Lone Star Infrastructure Protection Act of 2021.16 The fact that 370 acres was too much when located in the wrong place underscores that a regulatory regime organized around acreage thresholds and undifferentiated country tables is measuring the wrong things. The proposed rule’s reorientation toward the identity of the holder, the location and use of the land, and the structure of control tracks the actual determinants of risk.

Enforcement. The Grand Forks episode also surfaces another troubling finding: Fufeng did not file its AFIDA report until American media began asking questions. The project was ultimately stopped by action of the Grand Forks City Council, not by any federal instrument. What the statute supplied, late and only under media pressure, was a filing.17 The GAO found that USDA assessed a total of eight penalties for late or absent filings in the decade from 2012 to 2021, notwithstanding statutory authority to assess penalties of up to 25% of the property's fair market value.18 The proposed recurring accrual structure and the removal of routine downward adjustments are the minimum required to make the statute's existing penalty ceiling mean something.

III. The Department Should Finalize Its Key Provisions

A. The “foreign adversary” definition (proposed § 5100.2). The Department should finalize its proposed definition, which borrows the “foreign country of concern” definition of 42 U.S.C. § 19237(2). That definition reflects a bipartisan legislative judgment that substantially insulates the rule from arbitrariness challenges; is flexible, through the Secretary of State's designation authority; and is harmonized with the definitional family already governing semiconductor incentives, Department of Energy programs, and a growing body of state farmland statutes. The companion “Foreign Adversary Controlled Entity” definition is equally sound, and its “jurisdiction or direction” prong is essential: A definition stopping at formal ownership would invite exactly the Hong Kong and third-country intermediation that characterizes the Smithfield and Syngenta structures. The Department should retain the full breadth of the proposed language.

B. The lease exemption. The elimination of the lease exemption for foreign adversaries and Foreign Adversary Controlled Entities should be finalized without modification. Nearly every security-relevant use of agricultural land—from positioning monitoring equipment near sensitive installations to operating research plots—can be accomplished through a lease as readily as a purchase, and at a fraction of the cost. An adversary-controlled entity that can hold a nine-year renewable lease adjacent to a military installation without appearing in AFIDA data is a major blind spot. As to the residual sub-one-year exemption for other foreign persons, the Department's concern about circumvention is well founded. If retained, the exemption should be conditioned on the following: (1) an anti-stacking rule aggregating successive or substantially continuous leases of the same or contiguous parcels; (2) attribution of leases held through shell corporations or intermediaries to their beneficial owners; and (3) making a lease of any duration reportable whenever any entity in the lessee’s ownership chain is adversary-linked. If those conditions prove too complex to administer, the better course is the Department’s identified alternative: eliminate the exemption for all foreign persons.

C. Tiered penalties and high-risk factors (proposed § 5100.4). The tiered penalty structure (2.5% of fair market value accruing weekly for adversary-linked violators and 1.5% for others, capped at the statutory 25%) is lawful and sound. Graduating the accrual rate by the violator’s risk profile is an ordinary exercise of the Secretary's longstanding penalty discretion, and the differential is justified on the merits. For example, if a pension fund from an allied country files late, the government temporarily lacks a data point; when an adversary-controlled entity fails to file, the government loses the only systematic mechanism by which an acquisition like Fufeng’s comes to official attention. In response to the Department's request for additional high-risk factors, the final rule should apply the higher accrual rate to violations involving any of the following, regardless of the filer's formal designation: land within a defined radius of military installations (including those not yet listed in Appendix A) or other sensitive federal sites designated by CFIUS; use of shell corporations, nominees, or layered intermediaries that obscured beneficial ownership; prior AFIDA violations by the filer or affiliates under common beneficial ownership; ownership chains including entities on the Section 1260H list, the Entity List, or Treasury sanctions lists, or identified as state-owned or state-financed; holdings involving agricultural research, seed production, or biotechnology facilities; and false, misleading, or incomplete statements in a filing. Each factor is objective, administrable, and tied to a documented failure mode in the record.

D. The data elements that were declined. The data elements the Department considered but did not propose (proximity to sensitive sites, aggregate U.S. holdings, ties to adversarial governments, leadership and ownership detail, and leaseholder detail) should be adopted, at minimum for adversary-linked filers, as identified via expanded beneficial ownership disclosure, through a two-tier reporting form: a streamlined schedule for ordinary foreign persons and a supplemental schedule for adversary-linked filers. Each omitted element maps onto a documented gap. Proximity disclosure converts an analytic burden on the government into a certification obligation on the party best positioned to know, with a false certification constituting an independent violation. Aggregate holdings disclosure would have revealed the Val Verde assemblage as a single 130,000-acre position rather than a series of county-level transactions. Ties to adversarial governments are the very facts that distinguish a Syngenta from a Swiss family seed company. For the same reasons, the Department should adopt annual verification for adversary-linked filers: ownership structures of the kind documented above are dynamic, and a one-time snapshot ages quickly.

E. NAICS citation. The Department should cite sector 11 of the North American Industry Classification System (NAICS) codes in its entirety, supplemented by the enumerated extra-sector codes the rule identifies. Sector-level citation is both durable and gap-proof. An itemized list will silently fall out of date at the next NAICS revision, reintroducing the obsolescence that stranded the regulations on 1987-vintage Standard Industrial Classification (SIC) codes for a generation, and an enumerated list invites the argument that activity within the sector but outside a listed code is unreportable.

F. Solar and wind generation. The inclusion of solar and wind generation on agricultural land should be finalized. The provision prevents the conversion of reportable land into unreportable land through a change in use, at the very moment the change makes the holding more sensitive: energy infrastructure connects the holding to the grid, brings towers and telemetry, and, as Val Verde demonstrates, can place adversary-controlled installations along military training corridors. The final rule should make explicit that the provision reaches land under option, lease, or easement for future generation development, not only operating facilities.

G. Ownership thresholds. The reduction of the “significant interest or substantial control” threshold from 50% to 10% merits support, and the identified 5% alternative deserves serious consideration for adversary-linked interests. Control of a widely held enterprise is routinely exercised far below a majority; the securities laws have used 5% and 10% disclosure thresholds since 1968 on exactly this logic.

IV. The Amendments Are Faithful to AFIDA's Character as a Disclosure Statute

Some will object that identity-based tiers smuggle a regulatory regime into a reporting statute. The objection has AFIDA backwards. Nothing in the proposed rule prohibits any foreign person from acquiring, holding, or leasing a single acre; every provision concerns what must be reported and what follows from the failure to report. Disclosure, moreover, is the informational substrate for every other lever of national security policy: The CFIUS process, the farmland statutes now enacted in a majority of states, and congressional oversight all depend on it. When the data is wrong, every downstream instrument malfunctions at once. Finally, Congress itself armed this disclosure statute with a penalty of up to 25%of fair market value, among the most severe civil penalties in the United States Code. The anomaly of the last four decades is not the present proposal but a regime in which that penalty was adjusted downward as a matter of course and assessed eight times in a decade.

V. Conclusion

The record before the Department supports each of the rule's adversary-specific provisions: two CFIUS-cleared acquisitions that placed a quarter of American hog slaughter and a leading position in seeds and crop protection under ultimate Chinese jurisdiction; a 370-acre parcel that threatened a strategic installation while 130,000 acres passed federal review; and a decade of near-zero enforcement atop unreliable data. We urge the Department to finalize the rule with the foreign adversary definition, the tiered penalty structure, and the elimination of the adversary lease exemption intact; to extend the declined data elements and annual verification to adversary-linked filers; to adopt sector-level NAICS citation; to retain the solar and wind provision; and to harden the residual lease exemption against circumvention.

Respectfully submitted,

Adam Savit

Director, China Policy Initiative

America First Policy Institute