Staking and Crypto Yield: Where the Number Actually Comes From

Risk-forward article, and the correction belongs first: a staking reward is not interest, and a crypto yield is not income. Interest is paid by a borrower under a contract. A dividend is paid from profits. A staking reward is newly issued units, or a share of fees paid by other users — nothing earned it and no enterprise generated it. Rewards are denominated in an asset that can fall by more than the reward in a day; assets are often locked and unwithdrawable; validators can be penalised by forfeiture; and where the arrangement is intermediated, a firm holds your assets and multiple such firms have failed, leaving depositors as unsecured creditors. This article explains the mechanisms and where each fails. It recommends nothing and names no protocol or platform.

Separating them is the whole task, because the percentage tells a reader almost nothing about what they would be exposed to. The consensus article already established the key fact for the first of them: on a proof-of-stake network, rewards are new issuance, so a holder who stakes is substantially receiving a transfer from holders who do not.

Four arrangements that share a percentage and nothing else

Protocol staking is the only one that is what it claims. Assets are committed to secure a network, and the protocol pays rewards from new issuance plus a share of transaction fees. The reward is real in units and is not a return on capital in the ordinary sense — the network generated no revenue, so a large part of what a staker receives is dilution paid by non-stakers, per the previous article. Three exposures: assets are often locked for a period or subject to an unbonding delay, so they cannot be sold during a fall; slashing forfeits part of the stake if the validator misbehaves or fails, which is a loss of principal for a reason having nothing to do with price; and the reward is denominated in the asset, so a 5% annual reward against a 40% price decline is a 37% loss. Delegated or intermediated staking is the same activity conducted through a service, because running a validator has technical and minimum-size requirements. This adds a counterparty, and the assessment is the four questions the custody article set out: segregation, the nature of the claim, jurisdiction and whether it covers this activity, and independent verification. Lending and "earn" products are a different thing entirely: a firm takes assets and pays a rate, and the rate is funded by lending them onward to someone paying more. So a depositor is a lender to that firm, exposed to whoever it lent to, on terms usually not disclosed. Several such firms have failed, and depositors discovered their position was an unsecured claim in an insolvency — a firm displaying a rate is not a bank and the deposit is not a deposit, whatever the interface resembles. And DeFi yield — liquidity provision and lending through protocols — carries the contract, oracle, design, composability, and governance risks that article set out, plus the divergence effect that can make providing liquidity worse than simply holding. The diagnostic question that cuts through all four: where does the money come from? If the answer is new issuance, the holder is being paid in dilution. If it is fees from other users, the yield depends on their activity continuing. If it is a firm lending onward, the exposure is to that firm and its borrowers. And if nobody can explain the source, that is itself the answer, because a rate with no identifiable payer is being funded either by new deposits or by the promoter — which is the structure the fraud article examines.

Why the percentage is the least informative number

Four reasons a headline rate misleads, each of which the sector rarely surfaces. It is denominated in a volatile asset. A rate quoted as an annual percentage carries an implicit and usually unstated assumption that the unit holds value. On an asset with the drawdowns the opening article described, the reward is a rounding error against the price move — and quoting an annualised percentage on such an asset invites a comparison with a savings rate that the arrangement cannot support. Compounding presentations inflate it. Rates are often shown compounded to an annual figure, which assumes uninterrupted reinvestment at an unchanged rate — an assumption that rarely holds and is never labelled. The lock-up is the real cost. An arrangement paying a rate while preventing withdrawal has sold liquidity, and liquidity is worth most precisely when a holder wants to exit. Any rate should be read as compensation for that, not as a free addition. And a high rate is information about risk, not opportunity. This is the pillar's most useful transferable idea and it is not crypto-specific: the credit-spread article established that extra yield is compensation for something, and the high-yield article that a higher rate marks a higher probability of not being paid. A rate far above what conventional instruments offer is a statement about the risk being borne, whether or not the arrangement is honest. Two things this article will state without qualification. First, the portal applies the same discipline here it applied to option premium and carry credits: money arriving is not income when it is payment for accepting a risk that has not yet materialised. That rule has now applied three times across three pillars, which is a sign it is a property of financial products rather than of any one asset class. Second, staking is a genuine technical mechanism performing a real function — securing a network — and the criticism here is of how the reward is presented, not of the mechanism. A reader who understands they are receiving dilution and fees in exchange for locking a volatile asset and accepting slashing and counterparty risk understands the arrangement. A reader who has been told they are earning interest does not.