DeFi Lending and Borrowing

DeFi lending markets like Aave and Compound are pooled money markets run by smart contracts: depositors supply assets and earn a floating interest rate, borrowers post crypto collateral worth more than their loan and pay a floating rate, and an algorithm sets both rates from supply and demand. No credit checks, no loan officers — and no mercy either: if your collateral falls too far, the contract sells it automatically.

How the Pool Model Works

A DeFi money market is not peer-to-peer matching. All deposits of a given asset — say USDC — go into one shared pool, and all loans of that asset come out of it. When you deposit, you receive interest-bearing receipt tokens (aTokens on Aave, cTokens on Compound) representing your share of the pool; your balance grows continuously as borrowers pay interest.

Interest rates are set by an on-chain formula driven by utilization — the fraction of the pool currently lent out:

• At low utilization (plenty of idle USDC), borrowing is cheap and deposit rates are low.
• As utilization rises, both rates climb gradually.
• Past a target level — often around 80–90% — the borrow rate turns sharply upward, sometimes to double or triple digits annualized.

That kink is deliberate. Depositors can only withdraw idle funds; if 100% of the pool were lent out, withdrawals would halt until loans were repaid. The punitive rates near full utilization push borrowers to repay and attract fresh deposits, restoring liquidity. It usually works, but "usually" is worth remembering: in stressed markets, utilization can pin near 100% and depositors may briefly be unable to exit.

The spread between what borrowers pay and depositors receive covers the protocol's reserve fund and the share of the pool sitting idle. Typical stablecoin deposit rates in calm markets have historically been low-to-mid single digits, spiking when demand for leverage runs hot. Any rate you see is a snapshot, not a promise.

Overcollateralization: Loans Backed by More Than They're Worth

Because the contract knows nothing about you, it cannot price your trustworthiness. Instead, every loan is secured by collateral worth more than the debt.

Each collateral asset has a maximum loan-to-value (LTV) — say 80% for ETH — and a slightly higher liquidation threshold — say 82.5%. Deposit $10,000 of ETH and you can borrow up to $8,000 of stablecoins; liquidation logic kicks in when your debt exceeds 82.5% of your collateral's current value.

Protocols summarize your safety as a health factor: liquidation threshold value divided by debt. Above 1.0 you are safe; at 1.0 you are liquidatable. Borrow $5,000 against $10,000 of ETH at an 82.5% threshold and your health factor is (10,000 × 0.825) / 5,000 = 1.65 — comfortable. If ETH drops 40%, collateral is worth $6,000, and the factor falls to (6,000 × 0.825) / 5,000 = 0.99 — you are over the line.

Why would anyone lock $10,000 to borrow $5,000? Common reasons: unlocking spending power without selling an asset (and without triggering a taxable sale in many jurisdictions — verify your own rules); borrowing stablecoins to deploy elsewhere; or leveraging, by borrowing against ETH, buying more ETH, and re-depositing. That last use is popular and dangerous — it stacks liquidation risk deliberately.

Liquidation: The Automatic Margin Call

When a position's health factor drops below 1.0, anyone — in practice, bots watching every position — can call the protocol's liquidation function. The liquidator repays a chunk of your debt (often up to half per transaction) and receives an equivalent slice of your collateral plus a bonus, commonly around 5–10% depending on the asset. That bonus is the liquidator's incentive and your penalty.

Concretely: you owe $5,000 USDC against ETH collateral now worth $6,000, and your health factor slips below 1.0. A bot repays $2,500 of your debt and takes about $2,625 of your ETH (with a 5% bonus). You keep the borrowed USDC and the remaining position, but you have permanently lost the bonus, plus you were forcibly sold ETH at what is often a local price low.

Three practical implications:

• There is no phone call and no grace period. Liquidation happens the moment the on-chain price crosses the line, at 3 a.m. on a Sunday if that's when the market moves.
• Prices come from oracles. Protocols read prices from oracle networks such as Chainlink, not from any single exchange. A brief wick on one venue usually won't liquidate you, but genuine fast crashes will — and oracle failures are themselves a known (if rare) risk category.
• Cascades are real. In sharp downturns, mass liquidations sell collateral into a falling market, deepening the fall and triggering further liquidations. During extreme congestion, gas prices spike, which can make topping up collateral slow and expensive exactly when you need it.

Sensible borrowers leave a wide margin: many target a health factor of 2 or higher for volatile collateral, meaning ETH must roughly halve before trouble starts. They also set alerts on their position and know in advance which asset they would use to repay or top up in a crash.

What Depositors Are Actually Risking

Lending on a blue-chip money market is among the tamer things in DeFi, but "tamer" is not "safe." A depositor's risks:

• Smart contract failure. The pool is code holding billions; bugs and exploits, though rarer on mature audited protocols, are never impossible.
• Bad debt. If collateral crashes faster than liquidators can act — a sudden depeg, a collapsing token accepted as collateral, oracle trouble — loans can end up worth more than their backing. That shortfall lands on the protocol and, past its reserves, on depositors. Isolated or exotic collateral markets are where this most often happens; major protocols mitigate it by capping and segregating risky assets.
• Liquidity crunches. As noted, near-full utilization can delay withdrawals.
• Rate variability. The APY that attracted you can drop to near zero as conditions change.

The mitigations are unexciting and effective: prefer the largest, longest-running markets; prefer main pools of major assets over exotic isolated markets; understand that deposit yield is bounded by what borrowers genuinely pay; and treat anything advertising a dramatically higher "lending" rate as a different, riskier product wearing the same name.

Key Takeaways

• DeFi lending is pooled: depositors earn a floating rate, borrowers pay a floating rate, and a utilization-based formula sets both — spiking sharply as the pool nears fully lent.
• All loans are overcollateralized; your health factor (collateral value × liquidation threshold ÷ debt) tells you how far prices can move before liquidation.
• Liquidation is automatic, executed by bots, and costs you a penalty bonus of roughly 5–10% of the liquidated amount — keep health factors comfortably high and set alerts.
• Depositors face smart contract risk, bad-debt scenarios, and possible withdrawal delays at extreme utilization; yield quotes are snapshots, not promises.
• Stick to large, battle-tested money markets and mainstream collateral while learning; exotic markets are where depositor losses concentrate.