Service, Stake, and the Curious Case of Misclassification
Why Proof-of-Work Is Labour, Proof-of-Stake Is Ownership, and How Blind Peer Review Momentarily Allowed Reason to Prevail
Keywords:
proof-of-work, proof-of-stake, legal classification, securities law, Howey test, staking, mining, service provision, equity holding, capital income, tax law, financial regulation, collective investment schemes, blockchain law, regulatory arbitrage, consensus mechanisms, digital assets
There is, I am told, a certain propriety expected when one announces that one’s work has been accepted for presentation at a conference—particularly after blind peer review, that most charming institutional fiction whereby ideas are judged without the inconvenient interference of personality. It is a system I have come to admire deeply. One submits arguments stripped of identity, and for a brief, luminous moment, merit is allowed to masquerade as the decisive factor. Fortunately, in my case, it appears to have worked.
The paper—Service Provision or Equity Holding? The Legal Classification of Proof-of-Work and Proof-of-Stake Consensus Mechanisms —has been accepted, and will shortly be presented, examined, and, one hopes, misunderstood in all the usual productive ways that accompany any serious attempt to clarify a confused domain.
The confusion, of course, is not accidental. It is cultivated.
One of the more persistent habits of modern regulatory thought is the tendency to treat unlike things as though they were identical, and then to express great surprise when the results are incoherent. In the present case, the objects of this intellectual flattening are proof-of-work and proof-of-stake—two mechanisms that, while often grouped together under the same technological banner, bear about as much resemblance to one another as a contract of employment does to a share certificate.
And yet, regulators, commentators, and assorted enthusiasts persist in treating them as variations of the same phenomenon.
This paper exists to end that particular indulgence.
At its core—though one must always be cautious when speaking of “cores,” as they tend to conceal more than they reveal—the argument is disarmingly simple. Proof-of-work establishes a relationship of service provision. Proof-of-stake establishes a relationship of ownership. The former is contingent, competitive, and external. The latter is constitutive, proportional, and internal. To confuse the two is not merely an analytical error; it is a category mistake of the most fundamental kind.
Naturally, once one states the distinction plainly, it becomes rather difficult to unsee.
In a proof-of-work system, the participant—let us call him, with appropriate historical sobriety, a miner—does not need to own the asset he helps to secure. He brings to the network something entirely external: computational effort, energy expenditure, capital equipment. He performs a task. That task is defined, measurable, and verifiable. If he succeeds, he is rewarded. If he fails, he is not. The relationship is transactional, almost brutally so.
One might even say it is honest.
The protocol, in this arrangement, behaves rather like a particularly austere employer issuing a unilateral contract: perform this task, and you shall be paid. There is no guarantee of success, no entitlement to reward absent performance, and no residual claim upon the system once the task is complete. It is labour, albeit of a highly mechanised sort, and its compensation is properly understood as payment for services rendered.
There is, in this, a certain moral clarity.
The miner earns what he receives. His profit derives not from possession, but from action. He may, if he chooses, retain the tokens he earns and speculate upon their future value—but that is an entirely separate activity, conceptually distinct from the act of validation itself. To conflate the two is to confuse a wage with an investment, and while such confusion is not uncommon, it is rarely defensible.
Proof-of-stake, by contrast, operates on an altogether different principle.
Here, one does not arrive empty-handed. One must already possess the asset in order to participate. Indeed, participation is inseparable from possession. To validate is to stake; to stake is to commit capital; and to commit capital is to place oneself in a position of entitlement to yield.
This is not labour. It is not service. It is ownership deployed.
The validator in a proof-of-stake system earns not because he has done something, but because he has something. His return is proportional to his stake, and his stake is a function of his prior holdings. The system rewards possession with further possession, in a manner that would be instantly recognisable to anyone familiar with equity, dividends, or interest-bearing instruments.
One might dress it in the language of participation, governance, or even community, but the underlying structure remains stubbornly financial.
It is, in essence, a yield on capital.
The distinction, therefore, is not merely technical. It is ontological. In one system, the participant stands outside the asset and interacts with it instrumentally. In the other, the participant is bound to the asset in a relationship of necessity. One can mine without owning; one cannot stake without owning. That single difference is sufficient to reconfigure the entire legal character of the activity.
And yet, as I have noted, the prevailing tendency is to ignore it.
Regulators, confronted with the unruly complexity of digital systems, have adopted the expedient solution of collapsing all consensus mechanisms into a single analytical category. It is a strategy of admirable efficiency and catastrophic consequence. By treating proof-of-work and proof-of-stake as interchangeable, they produce classifications that are simultaneously overinclusive and underinclusive—capturing what should be excluded, and excluding what should be captured.
The paper proceeds, with what I hope is a tolerable degree of restraint, to examine the implications of this distinction across three domains: securities law, tax law, and financial regulation.
Each, in its own way, reveals the cost of conceptual imprecision.
Consider, first, securities law.
The canonical test—at least in the United States—is that articulated in SEC v. W.J. Howey Co. (1946), which asks whether there is an investment of money in a common enterprise with an expectation of profits derived from the efforts of others. It is a test of admirable durability, though its application has not always been accompanied by equal clarity.
Applied to proof-of-work, the result is almost embarrassingly straightforward. The miner does not invest in a common enterprise; he expends resources to perform a task. His return is not derived from the efforts of others, but from his own computational work. There is no pooling of capital, no passive expectation of profit, no reliance upon managerial activity. There is, instead, a competitive process in which success is contingent upon performance.
To call this a security is to drain the term of meaning.
Proof-of-stake, however, presents a rather different picture.
Here, the participant commits capital—tokens—to the system. He expects a return proportional to that commitment. His return depends, in significant part, upon the continued functioning of the network, the efforts of developers, and the activity of other participants. The structure begins to look uncomfortably like an investment arrangement, particularly when one considers staking pools and delegated mechanisms, where capital is aggregated and management is effectively outsourced.
The resemblance to traditional securities is not perfect, but it is sufficiently close to warrant serious consideration.
And yet, the dominant regulatory impulse is to treat both mechanisms as though they fall on the same side of the line. It is, one might say, an example of legal analysis conducted with the lights dimmed.
The consequences extend, naturally, into taxation.
If one begins with the incorrect premise that all validation activity is fundamentally the same, one is inevitably led to treat all rewards as belonging to the same category of income. But the distinction between service income and capital income is not a trivial one. It affects rates, timing, deductibility, and compliance obligations. It shapes behaviour and incentives. It determines, in no small part, how individuals and entities structure their participation.
Proof-of-work rewards, as the paper argues, are most properly understood as income from services. They arise from the performance of a task, involve the expenditure of resources, and are analogous to business income. The miner incurs costs—electricity, hardware, maintenance—and receives compensation. It is, in essence, a productive activity.
Proof-of-stake rewards, by contrast, are income from capital. They accrue to the holder of an asset by virtue of ownership and commitment. They resemble dividends or interest, not wages. The staked tokens function as capital at risk, and the reward is the return on that capital.
To treat these two forms of income as identical is to ignore their economic substance.
It is also to invite arbitrage, confusion, and, inevitably, litigation.
The third domain—financial regulation—introduces yet another layer of complexity.
Here, the concept of the collective investment scheme becomes particularly relevant. A structure in which capital is pooled, managed, and used to generate returns for participants is, in most jurisdictions, subject to a specific and often stringent regulatory framework. It is designed to protect investors, ensure transparency, and impose accountability upon those who manage other people’s money.
Proof-of-stake, especially in its pooled forms, begins to resemble such schemes with disconcerting fidelity.
Participants contribute capital. That capital is aggregated. The validation process is conducted, often by a subset of operators. Returns are distributed proportionally. Fees are extracted. The entire arrangement bears a striking resemblance to managed funds, albeit expressed in the language of protocols rather than prospectuses.
Proof-of-work, once again, stands apart.
Mining pools do exist, of course, but they aggregate computational effort, not capital. Participants contribute hash power, not tokens. The pool coordinates activity and distributes rewards, but it does not manage an investment. It is closer to a cooperative of service providers than to a financial intermediary.
To regulate both structures under the same framework is to mistake labour for capital, and coordination for management.
It is, in short, to misunderstand the system one purports to govern.
The purpose of the paper, therefore, is not merely descriptive but corrective.
It proposes a differentiated regulatory framework—one that aligns legal classification with economic reality. Proof-of-work should be treated as service provision, subject to the ordinary rules governing business activity. Proof-of-stake should be analysed, with appropriate nuance, as capital deployment, with staking arrangements evaluated for their resemblance to securities and collective investment schemes.
This is not a radical proposal. It is, in fact, a rather conservative one.
It asks only that we apply existing legal categories with a degree of precision. That we resist the temptation to flatten distinctions for the sake of convenience. That we acknowledge, however reluctantly, that not all technologies are created equal, and that some require more careful thought than others.
One might even say it is an appeal to intellectual honesty.
Of course, such appeals are rarely fashionable.
There is a certain comfort in ambiguity. It allows regulators to maintain flexibility, commentators to maintain relevance, and participants to maintain plausible deniability. Clear distinctions, by contrast, impose obligations. They require decisions. They expose inconsistencies. They force one to choose.
And choice, as any serious thinker knows, is the most burdensome of freedoms.
It is perhaps for this reason that the paper has benefited from blind peer review.
Had my name been attached from the outset, it is entirely possible that certain readers might have found reasons—perfectly sincere, no doubt—to object. Bias, after all, is the one commodity that remains in inexhaustible supply. But stripped of identity, the argument was allowed to stand on its own merits, to be evaluated without the usual adornments of reputation or prejudice.
One cannot help but find that reassuring.
It suggests, however faintly, that ideas may still prevail when given the opportunity to do so unencumbered. That substance may, on occasion, triumph over sentiment. That the system, flawed though it may be, retains within it the possibility of fairness.
Or at the very least, the appearance of it.
The forthcoming presentation, then, is less a debut than a continuation.
The arguments have been made, examined, and accepted in one forum. They will now be subjected to another. There will be questions, objections, perhaps even the occasional moment of comprehension. It is, in other words, precisely what one would hope for.
For the issue at hand is not trivial.
The classification of consensus mechanisms is not merely an academic exercise. It determines how systems are regulated, how participants are taxed, how risks are allocated, and how rights are defined. It shapes the development of the entire field. To misclassify is to misgovern. To misgovern is to distort.
And distortion, once introduced, has a remarkable tendency to persist.
It embeds itself in policy, in practice, in expectation. It becomes the background against which future decisions are made. It is, in many ways, the most dangerous form of error: the one that no longer appears as such.
The task, therefore, is to intervene early, to clarify before confusion becomes doctrine.
If the paper succeeds in that—if it persuades even a small number of readers that proof-of-work and proof-of-stake are not, in fact, interchangeable—then it will have achieved something of value. Not a revolution, perhaps, but a correction. And in matters of law, corrections are often more significant than revolutions.
They endure.
And so, with a mixture of anticipation and mild amusement, I prepare to present it.
The arguments are ready. The distinctions are drawn. The consequences, I suspect, will take rather longer to unfold.
But then, as with all worthwhile endeavours, the point is not to conclude the conversation, but to begin it properly.