Three days. That is how long the SEC's Division of Corporation Finance let its own crypto buyback guidance stand before amending it.
On Friday, September 25, the division published an updated set of FAQs on crypto assets. One answer addressed buyback programs. When an issuer announces that it will repurchase its own non-security token, does that announcement function as a promise โ a promise to perform the "essential managerial efforts" that Howey's fourth prong asks about? The staff said the risk was low so long as the system was functional.
By Monday, September 28, the answer had changed. The division added five words: "and has no central party." Now the relief requires two conditions, not one. The network must be functional, and it must have no central party. The SEC published a side-by-side comparison of the two versions alongside the change, which is itself an unusual act of transparency โ an agency documenting the speed of its own second thoughts.
Fractures in the ledger reveal what hype obscures. The fracture here is not the revision. Revisions are healthy. The fracture is that the revision introduced a term the document never defines, into a test that determines whether an issuer's most routine capital-allocation decision is a securities-law event.
I have spent twelve years reading token supply schedules and regulatory text side by side, and I have learned to read them the same way. Look for the load-bearing term, then look for its definition. When the load-bearing term is undefined, the document is not a rule. It is a placeholder. Placeholders get filled by whoever litigates first.
Context: What the FAQ Is, and What It Is Not
Start with the instrument. The crypto FAQs are a product of the Division of Corporation Finance, not the Commission. They express staff views. The division says so explicitly in its own disclaimer โ the FAQs "have no legal force or effect." That sentence is not boilerplate. It is the entire legal architecture of the document, and it is the reason a three-day revision is legally unremarkable and strategically enormous.
The FAQs build on the interpretive release the SEC issued on March 17. That release set out the agency's current thinking on when a crypto asset is offered or sold as part of an investment contract. The analytical spine is the Howey test, unchanged since 1946: an investment of money, in a common enterprise, with an expectation of profit, derived from the efforts of others. The fourth prong โ "efforts of others" โ is where crypto has always lived or died. A token whose value depends on the ongoing managerial exertions of a founding team looks like a security. A token whose value depends on the operation of a sufficiently autonomous network does not.
The buyback question is a narrow application of that prong. Buybacks are interesting because they are the moment when a token issuer behaves like a public company. A repurchase program is a capital-return mechanism. It tells holders that the issuer is willing to convert treasury or revenue into demand for the token. That is a signal. Signals can be promises. And a promise of managerial effort is exactly what Howey's fourth prong is looking for.
The staff's original answer, published September 25, drew the line at functionality. If the crypto system was functional, an issuer's announcement of a non-security buyback program "would not constitute a representation or promise to undertake essential managerial efforts." One condition. Functional.
The revised answer, published September 28, draws the line at functionality plus decentralization. Both conditions must hold. The rest of the answer is unchanged. And the unchanged part is important: on a network that is not yet functional, the staff still say a buyback announcement could count as such a promise "if the issuer presents the buyback as creating yield or return for token holders."
That asymmetry tells you what the staff is actually worried about. It is not the buyback. It is the marketing of the buyback. A network that is not yet functional and an issuer that frames repurchases as yield is a network whose token value depends on the issuer's continued willingness to buy. That is a common enterprise with managerial effort embedded in the cash flows. The staff is not regulating the transaction. It is regulating the promise inside the transaction.
The revision arrived after pushback. Miles Jennings, general counsel and head of policy at a16z crypto, had raised concerns about the September 25 wording. Journalist Eleanor Terrett reported the update on X and noted that Jennings had said the earlier text could "empower an issuer to announce the buyback program without that announcement then creating an investment contract." In plain terms, the original draft was permissive enough to create a loophole โ an issuer could use a buyback announcement to signal confidence and support the token price without triggering the disclosure and liability regime that normally attaches to that kind of forward-looking statement.
On Monday, Jennings praised the revision. "The narrowing of Friday's guidance will bolster its durability," he wrote, "while guarding against attempts to misconstrue it by those that wish to circumvent securities laws. Pragmatic and clear rules are all crypto needs to succeed, and the SEC just delivered again."
Durability is the operative word. A permissive staff position that invites abuse is a position that invites rescission. A narrower one survives. Jennings is not praising the substance of the rule so much as its survivability โ and that is a sophisticated read of how soft law works.
Context matters here too. The guidance lands in a legislative vacuum. The Clarity Act, the market-structure bill that would have given crypto a statutory framework, collapsed. Two days later, the SEC introduced its innovation exemption. When Congress stops legislating, agencies legislate by memo. Staff FAQs become the de facto rulebook, even though the division tells you in writing that they are not a rulebook at all.
The industry is already behaving as if buybacks are the new tokenomics. Ethena proposed a buyback program in late August. It is not alone. More crypto projects are using protocol revenue to repurchase their own tokens, converting fee streams into float reduction. The practice is spreading because the market rewards it, and the market rewards it because it looks like the mature, equity-like behavior that institutional allocators say they want to see.
So the SEC narrowed the safe space on Monday. The price of the affected assets barely moved. That is the first sign that something structural is being mispriced.
Core: The Economics and the Legal Mechanics of a Buyback
A buyback announcement is a forward-looking statement about future demand. Read that twice, because it is the whole game. When an issuer says it will repurchase tokens, it is not merely describing a plan. It is telling the market that a specific, identifiable actor will be a standing bidder. That bid changes the expected return distribution of the asset. It is, functionally, a yield commitment denominated in the issuer's own balance sheet.
Under Howey, the fourth prong asks whether expected profits come from the efforts of others. A standing bid funded by a founding team's treasury or revenue is an effort. It is ongoing. It is managerial. The issuer decides when to buy, how much, and at what price. Those are discretionary decisions that affect the token's success. That is the definitional core of an investment contract.
The staff's revised answer is a narrow carve-out from that logic. If the system is functional and has no central party, the issuer's statement cannot create a new investment contract, because โ in the words of a neighboring FAQ โ no one "has control of the functional crypto system that would allow them to take any action which would affect the failure or success of the crypto system."
Read that sentence as an economist rather than a lawyer, and it says something precise. The relief depends on the absence of a control point. Not the absence of a foundation. Not the absence of a team. The absence of an actor who can affect the system's success. That is a systems-architecture test, not a corporate-form test. It is the closest the SEC has come to saying that decentralization is a property of the code and the validator set, not of the org chart.
The Equity Analogue the SEC Did Not Import
This is where my financial-engineering training makes me uncomfortable with the framing on both sides.
Public companies buy back stock all the time. Since 1982, Rule 10b-18 has provided a non-exclusive safe harbor for issuer repurchases, specifying manner, timing, price, and volume conditions. The safe harbor exists because Congress and the Commission recognized that a repurchase is a market-manipulation risk if unconstrained โ and simultaneously a legitimate capital-allocation tool if constrained. The safe harbor is the compromise. It lets companies buy back stock without fear of manipulation liability, provided they follow the rules.
Crypto has now been handed something that resembles a safe harbor in spirit but not in structure. The FAQ is not a safe harbor. It is a staff interpretation. It has "no legal force or effect." It offers no reliance protection. An issuer that structures a buyback around the FAQ is not protected from an enforcement action; it is merely arguing that the staff's own published reasoning supports it. In litigation, that is persuasive authority. It is not a defense.
The asymmetry is stark. An equity issuer that follows Rule 10b-18 has a defined safe harbor. A token issuer that follows the crypto FAQ has a sentence in a document the agency says has no legal force. The former is a legal right. The latter is a conversation.
And the disclosure regime that surrounds equity buybacks โ periodic reporting, insider trading rules, Regulation M, the entire apparatus that makes a repurchase legible to the market โ does not exist for tokens. There is no 10-Q. There is no 10b5-1 plan. There is no requirement to disclose daily repurchase volumes. A token issuer can announce a buyback, execute it through a multisig wallet, and never file anything again. The market prices the announcement and never sees the execution.
The Undefined Term Is the Guidance
Now the term. "No central party." What is it?
The buyback answer does not say. The neighboring FAQ gestures at it โ control over the system's failure or success โ but gestures are not definitions. Compare this to the SEC's own history. The November 2019 Framework for "Investment Contract" Analysis of Digital Assets introduced the concept of an "active participant" and asked whether that participant's efforts were "essential" to the token's value. The Hinman speech in June 2018 suggested that a digital asset could be "sufficiently decentralized" such that it was no longer a security. Neither "active participant" nor "sufficiently decentralized" was ever defined with operational precision. Both became litigation battlegrounds.
The Commission never adopted the 2019 Framework as a rule. It was staff guidance, and it faded into citation. "No central party" risks the same fate. It is a fourth-generation restatement of an old idea โ decentralization as a spectrum โ expressed in a form that sounds like a bright line but functions like a mood.
Here is the analytical move most readers will miss. The "no central party" condition is not a description of the relief. It is a de facto decentralization screen, and it is far narrower than the market is treating it.
Run a first-principles inventory of the top hundred tokens by market cap. How many belong to systems that are both functional and lack any party that could affect the system's failure or success? A handful. Bitcoin is the clean case โ no issuer, no foundation with unilateral control, no upgrade mechanism that a single actor can force. Beyond that, the list thins fast. Most large-cap tokens have foundations, core development companies, or treasury multisigs whose decisions demonstrably affect protocol outcomes. Most have upgrade keys. Most have fee switches that a governance vote can flip.
Under the revised text, a buyback announcement from any of those issuers is not covered. Not because the buyback is unlawful โ but because the condition for relief has not been met. The issuer is back to arguing the old Howey analysis without the staff's comfort. And it is back there with no way to know, ex ante, whether a court will agree that its foundation is disqualified.
The Unlock Cliff Meets the Buyback Bid
There is a supply-side dimension to this that the legal commentary is missing entirely, and it is the reason I keep returning to tokenomics before I return to case law.
Most tokens launched since 2021 carry a vesting schedule with a cliff. Team, investors, and advisors hold locked allocations that unlock on a calendar the issuer controls. Unlock events are mechanically dilutive. They add supply into the float on a date the market can see coming. Issuers hate them because they cap price appreciation. Holders hate them because they are the most predictable source of downward pressure in the asset class.
A buyback is the cleanest available offset. Announce a repurchase program, run it into the unlock window, and the visible dilution is absorbed by the visible bid. The float looks stable. The narrative looks confident. The team looks aligned.
But understand what has actually happened. The issuer has converted treasury assets into its own token in order to offset the selling pressure created by its own insiders. That is not capital return to holders. It is a circular transfer โ buyback demand meeting unlock supply, with the price impact smoothed and the underlying dilution unchanged. The insider who sells into the bid has received liquidity from the treasury. The retail holder who reads the announcement as a bullish signal has received nothing but a narrative.
I catalogued a version of this in 2017, when I audited more than forty ICO whitepapers and found that the emission schedules which looked most generous were the ones most dependent on a founding team's continuous intervention. The dependence was disguised as decentralization. The grants were disguised as community. Now the buyback is the new disguise. Complexity is often a disguise for fragility, and the repurchase program is the moment the disguise slips.
Ethena and the Revenue-Backed Repurchase
Ethena's late-August proposal is the useful case because it sits precisely on the fault line. Ethena is a synthetic-dollar protocol whose economics run through a centralized operational layer โ risk committees, custodians, exchange counterparties โ even where the token governance is distributed. A fee switch that routes protocol revenue into token repurchases converts that revenue into a demand program. The announcement of that program is exactly the kind of forward-looking statement Howey cares about.
I am not making a legal conclusion about Ethena. I am making a structural observation about the category. The projects most eager to announce buybacks are frequently the projects whose revenue depends on an operational entity that controls the flow. That operational entity is the central party. The buyback is evidence of it, not independent of it. The cleaner the repurchase mechanism, the more visible the hand that runs it.
There is also a counterparty question the guidance never touches. Who sells into the bid? In a bull market, the marginal seller is often the vesting team, the foundation, or an affiliated market maker. Without disclosure, you cannot tell whether the buyback is absorbing third-party supply or recycling treasury assets between related wallets. The FAQ regulates the announcement. It says nothing about the execution.
Three Days of Soft Law
The revision's speed is itself data. A three-day turnaround on a published FAQ means one of three things: the original text was inadequately reviewed, or internal pressure materialized fast, or external pressure materialized faster. Probably all three. Jennings' public concern and Terrett's reporting created a feedback loop, and the staff closed it before the weekend ended.
That is not a criticism of the staff. It is an observation about the regime. Soft law is fast in both directions. It can expand in a Friday FAQ and contract on a Monday. That flexibility is a feature in a fast-moving industry. It is also a risk: anything published without the procedural weight of notice-and-comment can be withdrawn without it too. The durable rulemaking โ the kind that carries reliance protection โ takes years. The FAQ took a weekend.
I watched a version of this dynamic in January 2024, when I built a dataset correlating the first week of spot Bitcoin ETF inflows with Grayscale's outflows and institutional rebalancing cycles. The finding that stuck with me was a roughly 48-hour lag between on-chain flow signals and price discovery in the token relative to the equivalent move in equities. Market-structure changes propagate through intermediaries โ custodians, authorized participants, compliance desks โ before they propagate into price.
Regulatory guidance works the same way. Traders read an FAQ and react. Legal teams read an FAQ and re-underwrite. The re-underwriting is slower and larger. The 48-hour lag I measured in ETF flows is the lag that governs this revision. The price impact will not show up on the day it was published. It will show up in the term sheets, engagement letters, and tokenomics decks that get rewritten over the next two quarters.
Liquidity, Not Utility
Strip the legal framing and look at the flow. A buyback is a liquidity event. It removes supply from the float, concentrates ownership, and creates a persistent bid. In the liquidity-first framework I have used since modeling stablecoin anchors during DeFi Summer, buybacks matter not because of what they say about a protocol's utility but because of what they do to the marginal supply-demand balance. My simulation work in 2020 quantified how stablecoin pegs acted as the primary liquidity anchor and produced a 15% error margin in standard valuation models that ignored them. The lesson holds: the market clears on flow, not on features.
Two structural facts about that flow deserve attention in a bull market.
First, buybacks funded by revenue and buybacks funded by treasury are economically distinct and legally indistinguishable under the FAQ. A revenue-funded buyback is a genuine capital-return mechanism. A treasury-funded buyback is a balance-sheet transfer: the issuer converts one asset โ stablecoins, ETH, BTC โ into another, its own token. If the token is overvalued relative to the treasury asset, that transfer destroys value for remaining holders. If it is undervalued, it creates value. The FAQ treats them the same. The market should not.
Second, buybacks are pro-cyclical. In an expansion, revenue is high, buybacks are large, float shrinks, price rises, and the mechanism looks self-validating. In a contraction, revenue falls, buybacks slow or stop, the standing bid disappears, and price discovers the absence. The bid was never a floor. It was a fair-weather participant. Anyone who watched the reflexive drawdowns of the 2022 cycle knows how quickly a support mechanism becomes a warning sign.
The Global Liquidity Frame
Zoom out. The buyback trend is downstream of something larger: the global liquidity cycle. Buybacks are funded from revenue and treasuries, and both expand when liquidity expands. M2 growth, dollar strength, and stablecoin dominance set the pace at which issuers can afford to convert balance-sheet assets into token demand. Stablecoin dominance โ the metric I have treated as a leading indicator since 2020 โ is the cleanest read on whether that fuel is present.
When stablecoin supply grows, buybacks get funded, floats shrink, and the narrative of mature tokenomics writes itself. When stablecoin supply contracts, the same programs are quietly paused, and the tokens that relied on them trade as if the bid never existed. The guidance does not change that dependency. It only changes the legal framing around it.
This is why I keep insisting that the buyback is a bull-market phenomenon before it is a regulatory phenomenon. The SEC's FAQ describes what is permitted. The liquidity cycle determines what is possible. Right now, the cycle is accommodating, and the market is reading accommodation as structural improvement. Those are different things.
Contrarian: The Consensus Read Is Backwards
Here is where I part company with the room.
The consensus interpretation of the September 28 revision is bullish โ a clarification that makes buybacks safer, a signal that the SEC is building a workable framework, and a green light for the revenue-to-repurchase trend. Jennings' praise is being read as validation. The narrowing is being read as precision.
I read it the other way. The narrowing reduced the set of networks that can rely on the relief, and the phrase "no central party" excludes most of the market. The guidance did not get more generous. It got narrower in a way that will only become visible when the first issuer discovers that its foundation, its upgrade multisig, or its fee-switch governance disqualifies it. The compliance community will have this realization months before the market does.
And there is a deeper problem: the instrument offers no reliance protection. "No legal force or effect" is not a footnote. It means the comfort is rhetorical. A sophisticated issuer's counsel will notice this and price it. The compliant path โ waiting for actual rulemaking โ remains slow and expensive. The aggressive path โ announcing a buyback under a staff FAQ โ remains exposed. When both paths are unattractive, capital does what it always does: it waits. The guidance may chill activity as much as it enables it.
There is one more contrarian reading worth stating plainly. A buyback is often the evidence that a central party exists, not the proof that it does not. The act of announcing, funding, and executing a repurchase program is a managerial act. It requires an entity with a treasury, a mandate, and discretion. The cleaner the buyback, the more visible the hand that runs it. The staff's test requires the absence of that hand. So the very projects that most want to buy back tokens are the projects least likely to satisfy the condition that makes buybacks safe to announce.
The chart is the symptom, not the disease. The symptom is the buyback. The disease is that most token projects have no legitimate, durable capital-return mechanism that does not depend on an identifiable operator. Equity markets solved this with dividends, buybacks, and a disclosure regime that makes both legible. Crypto has imported the buyback without the disclosure regime, the safe harbor, or the periodic reporting. It has imported the gesture and left the architecture behind.
I saw the mechanism-design version of this in 2026, when I led the macro-strategy work on autonomous agent liquidity provision โ modeling ten thousand agents drawing on decentralized credit lines, backtesting for systemic stability under high-frequency conditions. The question we kept hitting was the same one the SEC is circling: when an algorithm executes a repurchase, where is the promise? An autonomous executor that buys back tokens on a deterministic schedule is not making a representation. It is running code. The Howey analysis assumes a human decision-maker whose conduct can be characterized as an effort. Remove the human and the fourth prong loses its object.
The SEC's "no central party" condition was written for the world of foundations and core teams. It does not map onto a world of autonomous executors. That gap is where the next decade of regulatory ambiguity lives, and neither the September 25 nor the September 28 version of the FAQ so much as acknowledges it.
Takeaway: What to Watch, and the Question Nobody Is Asking
Three things will determine whether this guidance is a framework or a footnote.
First, the definition. If "no central party" is never operationalized, it will be resolved by enforcement โ and enforcement resolves definitions one defendant at a time. Watch for the first action that turns on the phrase. The staff's neighboring FAQ gives the test โ control over the system's failure or success โ but the boundaries will be set by whoever is charged first, not by whoever drafts best.
Second, codification. The March 17 interpretive release and these FAQs are staff positions. If the Commission adopts them as a rule, reliance protection follows and the analysis becomes durable. If it does not, the entire structure is one election cycle away from revision. Three days is the warning shot.
Third, the disclosure gap. A buyback regime without reporting standards is a signal without an audit trail. If revenue-funded repurchases become the standard tokenomics model โ and the Ethena proposal suggests they might โ the market will need a disclosure convention that does not currently exist. Volume, timing, funding source, counterparty. None of it is mandated. All of it is material. Solvency checks precede sentiment recovery, and disclosure precedes both. The industry wants the legitimacy that buybacks confer. It has not yet built the infrastructure that makes buybacks legible. Until it does, every repurchase program is a statement of confidence issued without an income statement attached.
So here is the question I would put to the people celebrating this guidance. If a functional network with no central party is the only place a buyback announcement is safe, and if a buyback announcement is the clearest evidence that a central party is directing the network's economics โ then which came first, the decentralization or the buyback?
And if the answer is "the buyback came first," then the SEC did not clarify the rule on Monday. It clarified the exception. The rule is still the rule, undefined term and all. Consensus is a lagging indicator of truth, and the market has not yet repriced the difference.