Published July 2, 2026 · by ILCalc
Stablecoin pools — USDC/USDT, DAI/USDC, USDC/USDe and their cousins — are marketed as the "safe corner" of liquidity provision: two assets that both target $1, so the price ratio barely moves and impermanent loss (IL) is negligible. That reputation is mostly deserved. In normal conditions, IL on a well-behaved stable pair rounds to zero, and even a small fee stream leaves you comfortably ahead of just holding the coins.
But "negligible" is not "zero," and the entire assumption rests on both tokens actually staying near $1. When one of them depegs — as USDC briefly did in March 2023, and as UST did far more violently in 2022 — the price divergence spikes and the IL that was rounding error suddenly becomes a real, measurable haircut on top of whatever is happening to the coin itself. This article shows exactly how small IL is at the peg, gives you a table of the small-move numbers, and walks through what a depeg does to the math so you can size the risk before you deposit.
For a standard 50/50 constant-product pool (Uniswap V2 and its forks), impermanent loss depends only on how much the price ratio of the two assets changes. Using k as the relative price factor (the new ratio divided by the entry ratio), the closed form is:
IL = 2·√k / (1 + k) − 1
The key property for stablecoins: this function is flat at the bottom. When k = 1 (no divergence) IL is exactly 0, and near k = 1 the curve is quadratic — the loss grows with the square of the divergence, not linearly. That is why a small wobble around the peg is almost free. If USDC trades at $0.99 against a $1.00 partner, that is a 1% ratio move (k = 0.99), and plugging in gives IL ≈ −0.0012% — roughly one hundredth of one basis point of your position value. You can see the formula behind these near-zero figures worked out step by step in our companion guide.
Here is what constant-product IL looks like across the range of divergences a stablecoin pair realistically experiences while both coins are still "pegged-ish." These are computed directly from the formula above.
| Price divergence between the two stables | Relative factor k | Impermanent loss | On a $10,000 position |
|---|---|---|---|
| 1% (e.g. $1.00 vs $0.99) | 0.99 | ≈ 0.0012% | ≈ $0.12 |
| 2% ($1.00 vs $0.98) | 0.98 | ≈ 0.005% | ≈ $0.50 |
| 5% ($1.00 vs $0.95) | 0.95 | ≈ 0.033% | ≈ $3.30 |
| 10% ($1.00 vs $0.90) | 0.90 | ≈ 0.14% | ≈ $13.86 |
Read that carefully: even a 10% gap — which would be considered a serious stablecoin stress event — costs only about 0.14% in IL. For comparison, a 2× move in a volatile pair like ETH/USDC produces about 5.7% IL. Stablecoins live in the extreme-left, ultra-flat part of the curve, and that is the whole reason the pools are attractive.
When Silicon Valley Bank failed, a portion of USDC's cash reserves was briefly stranded, and USDC traded down to roughly $0.88 over a weekend before fully recovering. Against a stable partner still at $1.00, that is a ~12% divergence (k ≈ 0.88), which the constant-product formula turns into about 0.20% IL — around $20 on a $10,000 position. Notice that the IL itself was still small. The real damage for LPs came from a different channel: as USDC fell, arbitrageurs drained the good coin and left the pool holding a larger share of the discounted USDC, so an LP who panicked and withdrew at the bottom effectively locked in the depeg. The IL number and the "what am I now holding" question are separate risks, and the second one is usually bigger.
TerraUSD (UST) was an algorithmic stablecoin that lost its peg and collapsed toward zero within days. As a price move this is off the end of the table: a stable going from $1.00 to $0.10 is a 90% divergence, and to a few cents is worse. At that point "impermanent loss" is the wrong lens entirely — you are simply holding a failing asset, and the pool mechanics guarantee you accumulate more of the loser on the way down. The lesson is that IL math describes the friction of divergence; it does not protect you from a coin that is fundamentally breaking.
The figures above use the plain constant-product invariant, which is what Uniswap-style pools use. Purpose-built stable venues like Curve's StableSwap use a flatter invariant that behaves almost like a constant-sum pool near the peg and only curves toward constant-product as the imbalance grows. In practice that means a Curve-style stable pool has even less IL than the numbers here for small moves. So treat the constant-product estimates as a conservative upper bound for those venues: if the math says 0.03% on Uniswap, the equivalent StableSwap pool will typically be less. During a real depeg the two converge, because both eventually end up heavily weighted in the discounted coin.
Because IL is so tiny, the break-even question in a stablecoin pool is lopsided in the LP's favor. Break-even is simply fees − IL$: as long as accrued trading fees exceed the dollar IL, you come out ahead of holding. With IL measured in fractions of a basis point during normal weeks, even a modest fee APR clears that bar almost immediately. A pool earning 4% APR in fees generates roughly 0.077% per week — already more than 25× the ~0.003% weekly IL from typical peg wobble. This is exactly the dynamic explained in our guide on why fees beat IL in stablecoin pools: the loss is real but small, and the income is small but steady and larger. The one scenario that flips it is a genuine depeg, where IL and inventory risk spike faster than fees can compensate.
You do not have to trust a table. A depeg is just a price move, so you can model it directly: open the calculator, treat one stable as the "price" asset, and enter the depegged level (say $0.90 or $0.85) as the new price. The tool returns the exact IL and, once you add your fee APR and holding period, the net-of-fees outcome — the number that tells you whether the position still beats holding through the stress.
Try a scenario now. Enter a depeg price and your fee assumptions to see IL and net return side by side.
Model a stablecoin depeg with the IL calculatorNo. The impermanent loss is negligible as long as both coins hold their peg — well under a basis point in a normal week. But you are still exposed to the depeg risk of both tokens simultaneously, plus smart-contract and venue risk. "Very low IL" is not the same as "risk-free"; the risk lives in the peg, not the IL formula.
That is a different animal entirely. A stable/ETH pair is a volatile pair, not a stable pair — its IL is governed by ETH's price swings, so a 2× ETH move produces about 5.7% IL, not 0.005%. Everything in this article applies only to stable/stable pools where both assets target the same value.
The IL formula is identical — it only sees the price ratio. What changes is the probability and depth of a depeg. Algorithmic or under-collateralized stables have historically depegged harder and faster (UST being the extreme), so the same formula can produce catastrophic inputs. The math is honest; the risk is in whether the peg holds.