Nor a Lender Be
When does a platform's origination become the investor's?
Polonius, it turns out, was worried about the wrong risk.
A non-U.S. investor wants loans, not a lending business. A platform (a peer-to-peer site, a fintech originator, a syndicate bank) finds the borrowers, runs credit, negotiates the paper, then sells the loan to the investor within 48 hours or has the investor fund it at closing. Is the platform’s origination work attributed to the investor, turning a passive holder into a lender taxed on a net basis under §882?1
The authors’ answer is refreshingly disciplined: two factors, and only two. Can the investor shape the loan terms? Can the platform commit the investor? Where both answers are “no,” the investor should win. Where both are “yes,” they suggest, with the tact of people who bill by the hour, that advisers “exercise caution.” Polonius, it turns out, was worried about the wrong risk.
01 The two-factor test
Factor 1: influence over terms. Di Portanova (Ct. Cl. 1982) is a neat transplant from oil and gas. A foreign holder of a 2.25 percent working interest was not in the operator’s business because it could not “play any significant role in or be able to exert any influence upon the management or operation of the unit.” Hendrickson (T.C. Memo 1987-566) stretched that to 22 percent. The lending analogy: a small slice of a big loan means little say over covenants, so the arranger’s origination is not yours. As your percentage climbs, so does the risk.
The §875 trap, flagged by the authors, to their credit. Di Portanova never asked whether the co-owners were partners. Cokes (91 T.C. 222) did, and a 42 percent working-interest holder under a unitization agreement was a partner in a U.S. business. Keep the investor group diffuse and not acting in concert; a joint-profit motive can undo the whole argument.
Factor 2: power to bind. The case review is the best part of the piece. Independence alone (the Taisei Fire legal-and-economic test) does not save you. In Adda, DeAmodio, Pinchot, and Lewenhaupt, courts attributed an independent agent’s U.S. activity to the foreign principal whenever the agent could commit the principal; nobody even argued the point. Cadwallader (13 T.C. 214) went the other way because the accept-or-reject decision was made in Manila (the lumber, helpfully, was also in Manila). Lesson: the power to bind “appears to be fatal”; its absence, with decisions made offshore, is the investor’s best fact.
The IRS view, and the cleverest move in the article. AM 2009-010 (the loan-origination GLAM) says power to bind is not required; anything done “on behalf of” the foreign investor counts. The authors read the GLAM narrowly by asking one question the GLAM never answers: did Origination Co. have other customers? One client means acting on its behalf; fifty means running its own shop. They then lean on YA Global (161 T.C. 173), where the Tax Court called the IRS’s attribution standard “almost certainly too broad.”
The policy backstop. Higgins (1941) has meant for 85 years that investing is not a trade or business. Funding a loan, they argue, is ministerial; the credit decision is the same whether the loan is new or seasoned. The elegant part: the customer goodwill belongs to the platform, which gets paid for it (fees, spread) and is taxed on it, so the fisc is not shortchanged. Congress, in the §864(b)(2) trading safe harbor, went further and exempted trading itself. Power to bind, the authors stress, “exists as a spectrum, not a binary determination.”
Numbers worth memorizing: 2 percent (securities), 10 percent (syndicated loans), 48 hours (the bank’s hold period), under 50 percent (forward-flow pool share), and 22 percent and 42 percent (the Hendrickson and Cokes bookends). Shakespeare gave us the title; the percentages are all EY.
02 Six market scenarios, mapped
| Scenario | Influence over terms | Power to bind | Read |
|---|---|---|---|
| A. Firm-commitment securities | None. Take-it-or-leave-it; about 2% of the issue. | None. Bank committed before any investor did. | Low. The market has assumed this for decades. |
| B. Firm-commitment syndicated loan | Comments to the bank only; about 10%; no direct negotiation. | None. Investor commits only after funding. | Low. Guidelines: 10% or less, no direct talks with the borrower. |
| C. Best-efforts syndicated loan | Same diffuse group, same small share. | “Soft.” Bank funds after investors signal; walking away has a cost. | Moderate. Riskier as your share grows. |
| D. Forward flow (credit box) | Weak. Fungible loans, random batches; buyer unknown at funding. | De facto. A standing credit box echoes Adda. | Moderate under 50% of the pool and free to reject. High on a must-buy. |
| E. Troubled-debt restructuring | Small holder’s input rarely survives the committee. | Soft. Holdouts get hurt; Reg. §1.1001-3 makes the loan “new.” | Low to moderate. Di Portanova helps; proposed §892 regs flag committee seats. |
| F. Trade receivables | Price only; nothing else to negotiate. | Ministerial. Investor approves each purchase. | Low if spread across buyers; untested if one buyer takes the batch (fn. 30). |
Swipe sideways for the other columns.
Exhibit 1. The article’s six fact patterns against the two factors. Positions are our reading of the article, not the authors’ chart.
Our take
Read it, then hand it to your credit team. The clearest map of the platform question we have seen, by people who know the deals. Its candor is a feature: the authors concede that no precedential case decides the independent-agent-without-power-to-bind question, and say so before reaching for policy. Most articles do it the other way round.
Build the file around Cadwallader. They are right that it is the investor’s best fact pattern. Credit approval and signature offshore, in substance as well as form; proof the platform shops each loan to more than one buyer; no power of attorney and no pre-funded account. Those facts do the work; policy is what we would argue after losing on them.
Forward-flow agreements need drafting, not just diligence. Keep a real right to reject; cap the pool share well under half; never sign a must-buy on every conforming loan. The Cottage Savings fungibility point (each loan is its own asset) is clever and, as the authors cheerfully admit, untested on nexus. Second argument, not first.
Watch the receivables corner. Footnote 30 is the sleeper (good of the authors to plant it): Reg. §§1.864-4(c)(5)(i)(C) and 1.904-4(e)(2)(i)(E) treat regular discounting of receivables as an active financing business. A single-buyer factoring program is not an investment because someone else wrote the invoice.
Source
Matthew A. Stevens and Huvie Weinreich, “Nor a Lender Be: Lending Platforms and Loan Origination,” Tax Notes Federal (Sept. 2, 2026), © Ernst & Young LLP. This note summarizes a third-party article for discussion; the quotations are the authors’, and readers should consult the article itself.
Important notes
This note addresses U.S. federal income tax only. The scenario readings are our own and are not a description of any client matter. Case and ruling citations are as reported in the article and should be checked against the primary sources before reliance.
-
References in this note to “§” and to the “Code” are to the Internal Revenue Code of 1986, as amended, and references to “Reg.” are to the Treasury regulations issued under the Code, each as in effect on the date of this note. “ECI” means income effectively connected with a U.S. trade or business. ↩
Download the note (PDF, 129 KB)
Client alerts go out about once a week.
This alert is for general information only and is not tax, legal, investment, or accounting advice. Consult your own advisers before acting. Carlos A. Schmidt, MBT, MBA, CPA · Managing Member · carlos@lapresallc.com · (917) 558-6393.