Built to Miss the Definition
A legal definition is a boundary drawn in words, and words have edges you can measure. Once you can measure an edge, you can build right up to it. That's not a loophole being exploited by accident. It's a structure engineered against a spec, the same way a bridge is engineered against a load rating.
Take something like the futures exemption for onions, on the books since the late 1950s. Congress banned futures trading in onions specifically, by name, after a market manipulation scandal. The ban didn't touch potatoes, or pork bellies, or anything else. If you wanted to build a market instrument that captured onion-adjacent price risk without becoming an "onion futures contract," you weren't defying the law. You were reading its exact perimeter and building a shape that traces it without crossing it. That's a design brief, not a dodge.
The instinct is to call this evasion and leave it there, as if the whole project were an attempt to defeat the law's purpose while pretending not to. But purpose isn't what a definition enforces. The definition enforces its own text. A statute that says "futures contracts in onions" is a claim about a specific mechanism, specific settlement terms, specific market structure. Miss any one of those elements deliberately and precisely, and you haven't broken the rule. You've built something the rule doesn't describe.
That precision is the whole discipline. Anyone can ignore a law. Building something that sits one degree outside its stated terms, while still doing almost everything the banned thing does, requires you to understand the definition better than the people who wrote it. You have to know which clause is load-bearing and which is decorative. You have to know whether "contract for future delivery" hinges on settlement timing, on physical delivery, on the counterparty structure, or on all three at once. Get that wrong and you've built either the banned thing with a new coat of paint, which is fraud, or something so far outside the definition it doesn't do the job you needed, which is a wasted structure. The target is a specific ring around the boundary, not "somewhere else."
This is why regulatory arbitrage deserves to be called engineering rather than cleverness. Engineering is the practice of designing to a constraint you didn't choose and can't argue with, only measure. A bridge doesn't negotiate with wind load. A structure built to sit outside a legal definition doesn't negotiate with the statute's intent either, because intent isn't the thing being tested. The text is the thing being tested. You design to the text the way you design to a tolerance, and the tolerance here is the gap between what the definition says and what it was trying to say, which is very often not the same gap at all.
That gap is also why these structures are fragile in a way ordinary engineering isn't. A bridge that meets its load spec stays meeting it; physics doesn't rewrite the load spec after the fact. A statute can be amended the moment the gap gets noticed, and often is, once the structure built around it becomes visible enough to embarrass someone. So the engineering has a shelf life built in from the start, which is unusual for engineering. You're building to a spec you know is going to be rewritten specifically to close the gap you found, and the only question is how long the window stays open before it does.
None of this settles whether any particular instance is good for the world. A structure that sits precisely outside a pollution definition is a different animal from a fund structure that sits precisely outside a securities definition, and both are different from onions again. What stays constant across all of them is the shape of the work: read the boundary as written, not as intended, and build exactly to its edge. That's not a moral stance. It's a spec, and somebody read it more carefully than the people who wrote it.