The uncomfortable question nobody asks before chasing stablecoin yield
Every year, tens of billions of dollars of crypto capital sit in stablecoins earning yield. Most holders never explicitly price the one risk that matters most — the probability that the stablecoin loses its dollar peg. A 15% APY looks great until you remember that Terra’s UST paid 19.5% APY on Anchor right up until it imploded and vaporized $40 billion of capital in 72 hours. Yield without a realistic view of depeg probability is not yield — it’s a covered sale of disaster insurance you forgot you were writing.
The math is simple. If you earn 10% APY on $50,000 of a stablecoin, you make $5,000 per year. If that stablecoin has a 5% annual probability of dropping 15% in a depeg event, your expected loss is $50,000 × 0.05 × 0.15 = $375. Subtract that from your yield: expected profit is $4,625. Still positive — but you’ve collapsed the entire scenario into one number and lost the tail. In reality, 95% of the time you make $5,000 and 5% of the time you lose $7,500. Same expected value, vastly different risk profile. This calculator lets you see both.
How to categorize stablecoins by depeg risk
Not all stablecoins carry the same tail risk. Broadly there are four categories.
Top-tier fiat-backed (USDC, USDP): Backed 1:1 by cash and short-term Treasuries held with regulated custodians. Transparent monthly attestations. USDC depegged briefly to ~$0.88 in March 2023 during the Silicon Valley Bank crisis, recovering within 72 hours once SVB deposits were made whole. Historical severity: 5-15% loss that usually recovers. Annual probability: likely 2-3% but trending lower as infrastructure hardens.
Tier-two centralized (USDT): Largest by market cap, backed by a mix of cash, Treasuries, and commercial paper. Operates from offshore jurisdictions. Has traded at micro-discounts ($0.95-$0.98) multiple times but always recovered. Historical severity: 5-20% in worst briefly-held moments. Annual probability: hard to estimate; roughly 3-5%.
Overcollateralized decentralized (DAI, LUSD): Backed by on-chain crypto collateral worth 150%+ of circulating supply. DAI held its peg through the 2022 and 2023 crashes but is exposed to correlated collateral risk — if ETH drops 80%, the system relies on liquidations clearing. Historical severity: small, usually <5%. Annual probability: 2-3%.
Algorithmic / undercollateralized: No backing or fractional backing. Rely on arbitrage and market confidence. Terra/UST, Iron Finance, Basis Cash — all zero now. Historical severity: 90-100% when they fail. Annual probability: high, measured not in years but in months for most algo stables.
Four real depegs and the lessons they taught
Terra/UST (May 2022): Dropped from $1.00 to $0.10 in five days. Holders who had $100,000 in UST earning 19.5% APY on Anchor ended with $10,000. Full annual yield was $19,500 — gained in months, lost in days. The lesson: unsustainable yields are always subsidized by something eventually fragile.
USDC (March 2023): Dropped to $0.88 on a Saturday following the SVB weekend, recovering to $0.99+ by Monday. Holders who panic-sold realized a 10-15% loss. Holders who held were made whole. The lesson: centralized stables with genuine reserves survive shocks, but you have to actually believe that during the shock.
Iron Finance (June 2021): Partially-collateralized TITAN token went from $64 to effectively zero in 24 hours, breaking the IRON stablecoin peg. Notable because Mark Cuban publicly disclosed a position hours before. Lesson: partial collateralization with algorithmic stabilizers is structurally fragile.
USDN (Waves, multiple times): Dropped below $0.80 repeatedly in 2022 before being delisted. Lesson: if a stablecoin repeatedly loses peg and recovers, the distribution of outcomes includes the tail where it does not recover.
How to actually use the expected-value framework
Set a realistic annual depeg probability for the stablecoin you’re using. For top-tier fiat-backed stables, 2-3% is defensible. For USDT, 3-5%. For overcollateralized decentralized, 2-4%. For algorithmic, start at 20% and negotiate up.
Set a severity — the expected loss if the tail event happens. For Tier 1, 10% is reasonable (USDC-style temporary depeg). For USDT, 15%. For algo, assume 90%+ because that’s the historical record.
Run the calculator. Compare your expected value to what you could earn in 3-month Treasury bills (currently around 5% APY, zero depeg risk). If the stablecoin expected value is lower than T-bills, you’re being paid to take uncompensated risk. If it’s meaningfully higher, you’re being compensated — though you still have to size the position against your tolerance for tail loss, not just expected value.
Sizing rules that actually protect you
Three rules experienced stablecoin yield farmers live by. First, diversify across issuers. Holding $100,000 across USDC, USDT, and DAI reduces concentration risk versus putting all $100,000 in any single one. Second, never deposit more in any single protocol than you can afford to see drop to zero overnight. Smart contract risk and peg risk compound. Third, if an APY looks obviously too good, it is — by definition — being subsidized. The subsidy has a half-life. Harvest the yield early and rotate out before the music stops.
Yield is not free. The yield on a stablecoin is always compensation for a risk — smart contract risk, reserve composition risk, regulatory risk, or peg-mechanism risk. A useful exercise: before depositing anything into a yield venue, write down in one sentence what risk you are being paid to bear. If you can’t articulate it, you don’t understand the trade.
Related calculators
- DeFi yield calculator — model gross APR returns from liquidity farming.
- Stablecoin APY comparison — compare current yield opportunities across stablecoins.
- Impermanent loss calculator — model LP exposure in stablecoin-asset pools.
- Crypto tax calculator — stablecoin yield is ordinary income in most jurisdictions.