The Definition Game: Why the Same Fight Keeps Happening

This desk has now analyzed four separate products where the central controversy was not whether the product was useful, priced fairly, or well designed. It was whether the product met a definition written before it existed. Is an advance against earned wages a loan? Is a lease with a purchase option a credit sale? Is buying future receivables lending? Is a shared appreciation arrangement debt or equity? The recurrence isn't coincidence and it isn't primarily firms being clever. It's what category-based regulation necessarily produces when product design moves faster than definitions — and understanding it as structural rather than adversarial changes what you'd propose about it.

Why categories generate arbitrage

Consumer finance rules attach to things, not to effects. There are rules for credit, rules for leases, rules for deposits, rules for insurance. Each set applies to arrangements meeting a definition.

Definitions have to be written at some point in time, around products that exist at that point. Which produces the mechanism:

1. A category is defined around the products of its era.
2. The definition necessarily includes features that were typical of those products alongside features that were essential to the concern.
3. A new arrangement delivers a similar economic result without one of the typical features.
4. It falls outside the category, so the category's rules don't apply.
5. A dispute follows about whether the missing feature was typical or essential.
6. The dispute resolves, slowly, and by then there's a new arrangement.

Step two is where the whole thing lives. A definition drawn from observed examples can't cleanly separate what mattered from what merely co-occurred, because at the time of drafting nothing distinguished them — every credit product charged interest, so "charges interest" looked like part of what credit is rather than a fact about credit products as then constituted.

Our thesis: definitional arbitrage is a structural property of regulating by category, not a pathology of the firms doing it. Any category-based regime generates it. The frequency rises with the rate of product innovation, and the resolution rate is bounded by how fast legislatures and courts move — which means the gap widens over time by default.

A definition drawn from examples cannot separate what mattered from what merely co-occurred. Every credit product charged interest, so charging interest looked like part of what credit is.

The four cases

Product	Economic function	Missing element	The argument
Earned wage access	Money now against money later	No interest; repayment frequently not legally compelled; a tip or flat fee instead	Is a non-recourse advance against already-earned wages credit, or access to your own money early?
Lease-purchase	Goods now, payments over time, ownership at the end	Terminable at will; framed as a rental	Is an agreement most users complete a credit sale, or a genuine lease?
Merchant cash advance	Money now against future revenue	No fixed maturity; framed as a purchase of receivables	Does contingent repayment make it a purchase rather than a loan?
Shared appreciation	Money now against a share of future value	No interest, no payments, no fixed obligation	Is a claim on future appreciation debt or an equity interest?

The detailed treatments are in our analyses of earned wage access, lease-purchase, merchant advances, and shared appreciation.

A fifth case sits outside consumer finance and demonstrates the pattern isn't confined to it. Worker classification — the dispute in our classification analysis — has exactly this structure. Is a worker an employee or a contractor? The economic function is labour provided for payment; the disputed elements are control, integration, and independence; three separate legal tests apply simultaneously and can produce different answers on identical facts. The same structure appears wherever regulation attaches to a category rather than to an effect, which suggests the mechanism is general rather than a feature of financial regulation specifically.

The common shape

Every one of the four has the same economic core:

Value transferred to the consumer now, in exchange for more value later.

That is what credit is, in economic substance. And each product has it.

What varies is which conventional attribute is absent:

• No stated interest rate — the charge is a fee, a tip, a price differential, or a share.
• No fixed repayment obligation — repayment is contingent, voluntary, or terminable.
• No fixed maturity — repayment happens as revenue arrives or on a future event.
• A different legal form — a lease, a purchase, an equity interest.

Notice that each absent attribute is one that the disclosure regime depends on to function. Annualized cost requires a rate and a period. Without both, the standardized comparable number our disclosure analysis identifies as the only kind that works cannot be computed — so a product outside the category is also a product outside comparability, and that consequence is more damaging than the substantive rules it escapes.

That's a genuinely important observation. The debate is usually framed around rate caps and licensing. The larger effect is on the consumer's ability to compare at all, which is the mechanism our search analysis identifies as the one that disciplines prices. A product that can't be compared isn't priced by competition regardless of whether anyone caps it.

Why it isn't usually cynical

The framing that treats every such product as a deliberate evasion is both unfair and analytically wrong, and getting this right matters for what you'd propose.

Products are designed against constraints, and regulatory categories are one constraint among many. A designer optimizing for cost, speed, approval rate, and compliance burden will land outside a category sometimes without that being the objective — and when they do, they'll defend the position, because the alternative is accepting a cost their competitors don't bear.

Some of these products are genuinely novel. An advance against wages already earned, repayable only from the next deposit, with no recourse if it doesn't arrive, is not obviously the same thing as a loan. Reasonable people can hold that it's an income-timing service. Our shock absorption analysis argues it addresses a timing mismatch, which is closer to what the household actually needs than credit is — and that's an argument for the product's substance, not a defence of its classification.

The definitional question is frequently genuinely hard. A lease-purchase agreement that most users complete looks like a credit sale; one most users terminate looks like a rental. The same contract, and the answer depends on empirical behaviour that's contested — which is why courts in different states have reached different conclusions on similar facts.

The cynical cases exist too, and the tell is specific: a structure whose only apparent purpose is the classification. When a product's economics are identical to a regulated one and the differences exist solely in the paperwork, the arbitrage is deliberate. That's a distinguishable subset rather than the whole phenomenon.

Why substance-over-form is hard

The obvious remedy — regulate by economic function — is proposed constantly and adopted rarely. The reasons are better than they're usually credited.

Function is contested at the margin. "Anything that functions as credit" requires deciding what credit is. Money now for money later covers a lease, a layaway plan, a prepayment discount, an insurance premium paid in instalments, and an employer advance. A test broad enough to capture the arbitrage captures a great deal else, and drawing the line back in reintroduces the category problem one level up.

Uncertainty has costs borne by everyone. A bright line lets a firm know whether it's covered. A function test means finding out through litigation or supervision. That uncertainty is a fixed cost, and fixed costs bind hardest on small providers and small transactions — the mechanism in our selection analysis. A pure substance regime would predictably reduce the supply of exactly the small-dollar products that are scarcest.

Categories are administrable. Examiners can check whether a product meets a definition. Assessing economic substance requires judgment applied case by case, which is expensive and inconsistent.

Substance tests exist and are partial. Courts do apply them — the state decisions holding lease-purchase arrangements to be credit sales are substance reasoning, as is the recharacterization analysis applied to some receivables purchases. They work, slowly, in litigation, and they don't scale to prospective compliance.

So the honest position: the recurrence of definitional disputes is partly the price of administrability, and that's a real tradeoff rather than a failure to be corrected. A regime with no arbitrage would be one with so much uncertainty that fewer products existed at all — which is a cost paid by the households with fewest options.

Where these get resolved

An empirical observation with consequences.

These disputes resolve at state level and in courts, not through federal rulemaking. The pattern across all four cases:

• States legislate first — a number have enacted specific earned wage access statutes, lease-purchase statutes, and commercial financing disclosure requirements while federal treatment remained unsettled.
• Courts decide the cases in front of them, producing state-by-state holdings rather than a national rule.
• Federal action arrives late, if at all, and can be reversed — the volatility our classification analysis documents in worker classification, where a rule, an enforcement pause, and a new proposal followed within roughly two years.
• By resolution, the next product exists.

The consequence is a patchwork: a product treated as credit in one state and not in another, for years. Which produces a specific and underappreciated effect.

Fragmentation is itself a fixed cost. Complying with fifty regimes, tracking divergent definitions, and maintaining state-specific product variants costs money that doesn't scale with transaction size. Per our selection analysis, that raises the minimum viable loan size — so the definitional fight, whatever its merits, has a distributional consequence entirely separate from its resolution: it makes small transactions less viable while it's unresolved.

Which is worth stating plainly because both sides of these disputes tend to ignore it. The duration of the uncertainty is a cost, and it falls on the smallest transactions regardless of who eventually wins.

What the recurrence costs

Four costs, none of which is the one usually discussed.

Comparability, as above. The largest effect, and the one that undermines the price discipline that would operate even without substantive rules.

Uneven playing field during the gap. A firm inside the category bears compliance costs a functionally similar firm outside it doesn't. That advantage is temporary but can be years, and it distorts which products get built — capital flows toward the definitional gap rather than toward the better product.

Consumer confusion. A household cannot reasonably be expected to know that two products doing the same thing carry different protections, and the protections are invisible in the transaction.

Fragmentation costs, which fall on small transactions.

And the cost that isn't usually real: the products themselves are not necessarily worse. Several of the four have plausible claims to serving needs the regulated alternatives don't — the timing mismatch case for earned wage access, the no-fixed-obligation case for shared appreciation. Treating definitional escape as evidence of harm conflates two separate questions, and the analysis of whether a product is good should be conducted independently of whether it meets a definition. This desk's own treatments of these products reached mixed conclusions on the merits, which is the correct outcome if the two questions are genuinely separate.

The strongest objections

"This is too generous to firms exploiting gaps." The main objection, and it lands partly. Some structures exist solely for classification, and describing the phenomenon as structural risks providing cover. Our response is that the analysis identifies the cynical subset by a specific test — economics identical to a regulated product with differences confined to paperwork — and that treating every case as cynical makes that distinction impossible to draw. A framework that can't separate genuine novelty from deliberate evasion can't target enforcement.

"Substance-over-form works better than you allow." Reasonable. Courts apply it successfully, and tax law operates with substance doctrines at scale. Our qualification is about prospective application: a firm designing a product needs to know its treatment before launching, and substance doctrines answer well after the fact. They're a good backstop and a poor design constraint.

"The recurrence is manageable — just update definitions faster." Partly true and worth trying. Some jurisdictions have moved quickly on these questions. But the update cycle is bounded by legislative capacity, and the number of definitional questions grows with product variety while legislative capacity doesn't, which means faster updating narrows the gap without closing it.

Testable implications

1. New products should cluster just outside category boundaries rather than distributing randomly across the design space — the signature of the mechanism.
2. The interval between a product's emergence and its definitional resolution should be lengthening as product variety grows against fixed legislative capacity.
3. State-level resolution should precede federal action in most cases, and the state count adopting a position should predict eventual federal treatment.
4. Products outside a disclosure regime should show wider price dispersion than comparable products inside one, because comparability is what disciplines price.
5. Fragmentation should raise minimum viable transaction size, so the small-dollar end of a market should thin during periods of definitional uncertainty.
6. Capital should flow toward definitional gaps independent of product quality, which is testable by comparing funding into newly-created categories against their measured consumer outcomes.

The fourth is the most useful. If products outside disclosure regimes show systematically wider dispersion at comparable risk, that establishes the comparability channel as the dominant cost — and it would shift the policy target from substantive rules toward mandatory comparable disclosure applied by function, which is a narrower and more achievable version of substance-over-form than a full function test.

That's where we'd land. Extending comparable cost disclosure by economic function is much easier than extending substantive regulation by economic function, because it requires only computing a number rather than deciding which rules apply. A regime that says "if you transfer value now for more value later, publish an annualized cost" captures every case above without resolving any of the hard categorical questions — and it addresses the largest documented cost of the recurrence rather than the most argued-about one.

Frequently asked questions

---

This report presents an analytical framework and the authors' interpretation; it is not legal advice. The classification of any specific product under federal or state law is a legal question that varies by jurisdiction and continues to develop; nothing here should be relied on as a determination of how any particular arrangement is or will be treated. Consult qualified counsel.