You have Marine Bank v. Weaver (and FDIC insurance) to thank for that.

> We see, therefore, important differences between a certificate of deposit purchased from a federally regulated bank and other long-term debt obligations. The Court of Appeals failed to give appropriate weight to the important fact that the purchaser of a certificate of deposit is virtually guaranteed payment in full, whereas the holder of an ordinary long-term debt obligation assumes the risk of the borrower's insolvency. The definition of "security" in the 1934 Act provides that an instrument which seems to fall within the broad sweep of the Act is not to be considered a security if the context otherwise requires. It is unnecessary to subject issuers of bank certificates of deposit to liability under the antifraud provisions of the federal securities laws, since the holders of bank certificates of deposit are abundantly protected under the federal banking laws. We therefore hold that the certificate of deposit purchased by the Weavers is not a security.

Doesn't this just show that the "easy 4 question test" is not sufficient?

Or is that largely considered a mistake in interpretation nowadays?

That's where the 'decades of caselaw' bit comes in.

The "decades of caselaw" specifically includes instruments that "failed" the test but were still judged to be non-securities. So, it makes sense to push for a legislative/regulatory judgement.

The test still applies. So do the exemptions to it (some established by Congress, others by judicial action). There's no exemption for cryptocurrency staking pools.

Perhaps Coinbase's eventual appeals will establish one, but I doubt it.

This feels very hand wavy, to me. The test applies. Except where it was challenged and exemptions were codified.

Is the argument that it did apply in those situations before the exemptions?

> This feels very hand wavy, to me.

This is how the system works. "Code is law" does not apply to actual law; it's a long history of interpretation, wiggle room, precedent, legislative intent, and common sense being applied in various amounts. It works pretty well, all things considered; far more than trying to explicitly write down every little edge case.

Ah, completely agreed. But we can't lean on "it is easy, the Howey test explains all" without a huge caveat.

I will note that this particular path is treading on the "lets write down every edge case." As such, it would be better to have a slightly stronger guideline than that. Specifically, when the law is "this test, except for specific carve outs that all fail the test", that is a pretty bad law. Unless there is some unification of the carve outs other than "we decided these don't apply."