All field notes

DeFi · Explainer

Before looking at the yield, ask who pays it

Beads flow from a jar along three channels into cups; the middle channel stops short at a gap.
Concept illustration · AI-generated

A percentage is a useful starting point for a calculation. It is a poor substitute for an explanation.

When a DeFi product displays a yield, I want to trace the income before comparing the number: who pays, why they pay, and what could prevent the money coming back. Those questions make lending and staking easier to understand without pretending they carry the same risks.

A borrower pays to use liquidity

In a lending market, suppliers make assets available and borrowers pay interest to use them. On Aave, borrowing is backed by collateral, and borrowing rates change with factors including utilisation—the proportion of supplied assets that is borrowed—and market parameters. Collateral can be liquidated when a position breaches the relevant requirements. Aave's borrowing guide explains that mechanism.

For a supplier, the borrow rate is not automatically the rate earned on the entire supplied balance.

Consider a deliberately simplified pool:

ItemIllustrative amount
Total supplied1,000 tokens
Total borrowed600 tokens
Annual borrow interest, held constant10%
Gross interest over one year60 tokens

Before protocol deductions, losses or other costs, 60 tokens spread across 1,000 supplied tokens is 6%. The unborrowed 400 tokens do not generate borrower interest in this example. If we also assume a 10% deduction from interest, suppliers receive 54 tokens: 5.4% of the initial supplied balance.

Those are invented inputs for a simple-interest calculation, not an Aave quote or forecast. Real balances and rates change; compounding, market rules and incentives can change the result.

A balance and an immediate withdrawal are different

Suppose a supplier in that fictional pool has a claim worth 500 tokens, but only 400 tokens are presently unborrowed. The pool cannot satisfy a 500-token withdrawal immediately from that available liquidity alone.

Aave's documentation states that withdrawals depend on available pool liquidity. Using a supply position as collateral can impose further constraints. Its withdrawal guide describes both conditions.

A delay is different from a permanent loss. Yet “my balance is still displayed” is not the same as “I can use the whole amount today”. If you need the money at a particular time, that distinction matters.

Staking pays for a different activity

Ethereum staking supports validators performing consensus duties. It is not a loan to a borrower. The responsibilities, rewards and penalties follow network rules; a pooled service introduces its own arrangements around those duties. Ethereum's staking overview explains the available approaches and their trade-offs.

Counting rewards in tokens also differs from measuring purchasing power. Suppose an illustrative holding grows from 10 to 10.4 tokens, while the price falls from $100 to $80. Its value moves from $1,000 to $832 before fees and taxes. More tokens can coexist with a lower money value.

That arithmetic is not a price forecast. It simply keeps the reward calculation and the asset-price calculation visible at the same time.

Read the income and the loss together

I would write down the following before treating two displayed rates as comparable:

  • Income: what activity generates it, and in which asset is it paid?
  • Deductions: which costs separate the displayed rate from the amount received?
  • Availability: what must happen before the position can be withdrawn or sold?
  • Loss: which failures can reduce the assets or their value?

For a vault that allocates into other positions, repeat those questions for each underlying activity. A simpler deposit screen does not make the underlying cash flows self-explanatory.

The lending-yield lesson lets you change a fictional pool's inputs. The native-staking lesson separates token rewards, price changes and validator duties. Use them to practise tracing the mechanism; neither supplies a live rate or a product recommendation.