Crypto Yield Explained: Where APY Comes From and Who Pays It

Crypto yield comes from a payment source: borrowers paying interest, traders paying fees, a network rewarding validation, an issuer passing through asset income, or a project distributing tokens. An APY compresses those arrangements into a percentage, but the percentage does not identify the payer, the currency of payment, or the risks attached to earning it. The strongest analysis starts with the cash flow and ends with the rate. A yield can be mathematically correct on a dashboard while saying very little about the holder’s eventual return in dollars or their ability to withdraw the principal.

Follow the Payment Before Comparing the Percentage

For any yield product, write a plain sentence: “This position earns payments because this party pays for this service.” If that sentence cannot be completed, the economic mechanism is still unclear.

In lending, a borrower pays for temporary access to capital. In market making, traders pay for execution against supplied liquidity. In staking, participants receive protocol-defined compensation for performing or supporting consensus duties. In a token incentive program, a sponsor distributes an asset to attract behavior, such as deposits or trading activity.

These are different businesses. A borrower-funded return depends on demand for credit and the ability to manage losses. Trading fees depend on transactions occurring where the position supplies usable liquidity. An incentive depends on a distribution budget and the value of the token distributed. Combining them into a single APY can obscure the variable most likely to change.

The table is a classification tool, not a claim that a product has only one source. A position may receive all of these indirectly through a strategy. The more layers it contains, the more useful it becomes to separate the components before adding them.

A Borrower’s Rate Is Not a Depositor’s Rate

Lending illustrates why the rate displayed next to one side of a market does not automatically belong to the other side. If only some deposited capital is borrowed, interest paid by borrowers must be spread across a larger deposit base. Reserves, fees, and other allocations can create additional differences.

For a deliberately simplified hypothetical pool, assume deposits of $1 million, borrowing of $800,000, a constant 6% annual borrower rate, and a 10% allocation of borrower interest away from suppliers. Annual borrower interest would be $48,000. Suppliers would receive $43,200 before their own costs, equivalent to 4.32% of the $1 million deposit base.

This is a model for understanding the denominator, not the formula of every protocol. It ignores compounding, changing balances, incentives, losses, and implementation details. Its useful result is structural: 6% paid on borrowed funds and 4.32% earned on total supplied funds can coexist without contradiction.

Aave’s supply documentation identifies utilization and governance parameters as drivers of supply rates. A live rate is therefore an observation of a changing market. Treating it as a contractual promise for the next year would require terms that a variable-rate pool does not establish merely by displaying an annualized number.

Demand also deserves interpretation. Borrowing to finance productive activity, borrowing to hedge, and borrowing to collect a temporary reward can all produce interest payments today. Their persistence after an incentive ends may differ. The current payer is visible; the durability of that payer’s motive requires further analysis.

APR and APY Answer Different Arithmetic Questions

APR usually expresses an annualized rate without incorporating the particular reinvestment effect shown by APY. APY expresses a compounded annual result under specified assumptions. Product definitions still need checking: a familiar abbreviation does not guarantee consistent treatment of fees, rewards, or compounding across interfaces.

For a hypothetical 12% nominal annual rate compounded monthly, the calculation is:

APY = (1 + 0.12 ÷ 12)^12 − 1 = approximately 12.68%.

That result assumes the rate remains constant and each monthly payment is reinvested on the stated basis. It says nothing about whether rewards can actually be reinvested without a transaction charge, whether they arrive in the same asset, or whether the strategy remains available.

Frequency matters most when reinvestment is feasible and costs are small relative to the payment. If harvesting a tiny reward requires a fixed network cost, more frequent compounding can reduce net wealth. The mathematical advantage of an extra compounding event must exceed the cost of creating it.

Annualization introduces a separate issue. A high rate observed for a short interval can be extended into an annual number, but the extension is an assumption. A short burst of trading fees does not establish twelve months of identical activity. Readers should ask for the measurement window as well as the formula.

Separate Cash Income From Token Rewards

Consider a hypothetical $10,000 position advertised with a 10% simple annualized reward estimate: 6% from payments in the principal’s stable unit and 4% from a promotional token. Assume the position runs for a full year, the first component stays constant, and the promotional allocation is 200 tokens initially marked at $2 each.

The initial presentation values the two components at $600 and $400. But suppose those 200 reward tokens are worth $0.50 each when sold. Their proceeds are then $100, making gross receipts $700. If total transaction and service costs are $80, net receipts are $620, or 6.2% of the original $10,000.

Nothing about that arithmetic requires the original dashboard to have fabricated the starting token price. The problem is that a marked reward estimate and realized proceeds answer different questions. The calculation also excludes losses on principal, taxes, interruptions, and changes in the base payment rate. Those exclusions should be explicit before anyone treats 6.2% as an expected result.

Promotional tokens can retain value or appreciate. A subsidy may also help a useful market acquire users before fee income grows. The analytical objection is not that incentives are inherently worthless. It is that their market value, duration, and funding source should remain visible instead of becoming an undifferentiated yield number.

There is another distinction between income earned by a protocol and income paid to a particular position. The mechanics of protocol revenue and tokenholder value determine whether fees reach suppliers, operators, a treasury, tokenholders, or some combination. Owning a token with a connection to a busy protocol does not by itself establish a right to its revenue.

More Tokens Can Still Mean Less Wealth

Yield is commonly quoted in the asset deposited or the asset rewarded. Total return in another currency depends on the exchange rate between those assets and that currency.

If a hypothetical position increases its token balance by 8% while the token’s dollar price falls by 30%, its dollar-value change before costs is:

(1 + 0.08) × (1 − 0.30) − 1 = −24.4%.

Subtracting 30% from 8% would give the wrong result because the price change applies to the enlarged token balance. This is a basic but consequential distinction between earning units and preserving purchasing power.

Staking makes the distinction particularly clear. Ethereum’s staking explanation describes rewards for validator participation alongside penalties and different participation arrangements. Those mechanisms concern the operation and compensation of the network. They do not guarantee a positive return in a holder’s chosen spending currency.

Stable-denominated yield narrows one source of volatility but does not remove every source of principal risk. The token’s conversion mechanism, reserves, and access conditions still matter. A useful analysis of how stablecoins hold their peg belongs upstream of a yield comparison that assumes the deposited unit remains worth one dollar.

Fee Income Must Be Compared With the Position That Earned It

Liquidity provision is a good example of income that cannot be evaluated separately from the underlying position. A market maker earns fees while allowing traders to change its asset inventory. Counting only the fees can ignore what happened to the inventory during the same interval.

Uniswap’s fee documentation explains fee allocation within its pool designs, including the importance of active liquidity in concentrated-liquidity arrangements. The exact allocation depends on the version and configuration. A displayed pool rate is therefore not automatically the realized rate of every position carrying the pool’s name.

The relevant comparison includes what the supplied assets would have been worth if held outside the strategy. The analysis of impermanent loss and inventory risk explains why a position can collect fees and still underperform that benchmark. Underperformance against holding is also different from an absolute loss in dollars; a clear report should say which comparison it uses.

Treasury-linked products have a different reference point. Their gross income may come from underlying government securities, but the tokenholder receives the economics of a particular fund or claim after its expenses and conditions. Tokenized Treasury risk analysis separates the underlying asset from the structure through which the holder accesses it.

These examples reveal a broader rule: compare complete positions, not isolated income streams. Two products can quote the same APY while exposing their principal to substantially different assets and exit conditions.

Capacity and Leverage Can Change the Result

A rate earned on a small amount may not scale. Additional deposits can spread a fixed reward budget across more capital, lower utilization, or place more liquidity in competition for the same trading activity. Entering a strategy can change the very rate used to justify entering it.

A hypothetical reward budget of $100,000 per year represents 10% of a $1 million deposit base but only 2% of a $5 million base, assuming the budget and token valuation remain unchanged. No decline in protocol activity is needed for the percentage to fall. The denominator alone can do it.

Leverage introduces another denominator problem. Borrowing to enlarge a yielding position may raise the estimated return on the user’s own capital when the spread is positive. Borrowing costs, collateral-price changes, and forced liquidation can reverse that result. An advertised return on equity is not comparable with an unleveraged return on deposited assets until the funding and loss mechanisms are included.

The detailed interaction between collateral, oracles, and DeFi liquidation matters more than a simple yield-minus-borrow-rate calculation. A strategy can have a positive expected carry over a year and still be unable to survive an adverse move next week.

What Would Make the Yield More Credible?

A high yield can be legitimate compensation for scarce capital, difficult operational work, inventory exposure, or another service buyers value. A low yield can still sit on top of weak controls. The number alone cannot establish either safety or fraud.

More useful evidence would show who paid the income, how much of it reached the position, how the denominator was measured, and what costs were deducted. It would separate externally paid fees from token distributions, report realized rather than merely marked rewards where possible, and identify the period over which the rate was observed.

A practical review can then vary one assumption at a time. What remains if promotional rewards end? What if utilization falls? What if reward-token prices halve, transaction costs rise, or withdrawals take longer? The purpose is not to forecast all of those events. It is to identify which assumption carries the result.

Finally, write the return in two forms: units of the relevant asset and value in the currency used to judge the outcome. State the holding period and the cost basis. That simple discipline makes an APY useful again—as an output of understandable assumptions, rather than a substitute for understanding the position.