Impermanent Loss Explained: A Guide for Liquidity Providers

Answer box: Impermanent loss is the value gap that opens up between holding two tokens in your wallet and holding those same two tokens inside a constant-product liquidity pool, once their price ratio moves away from the point at which you deposited. It happens because automated market makers rebalance a pool by selling the token that’s rising and buying the token that’s falling, so a liquidity provider ends up with less of the winner and more of the loser than a simple buy-and-hold position would have produced. A 2x price divergence between the two assets costs roughly 5.7% of pool value versus holding; a 5x divergence costs about 25.5%. The loss is only “impermanent” if prices return to their original ratio before you withdraw — in practice, for volatile pairs, they usually don’t, and the fee income a pool earns has to outrun that structural drag for liquidity provision to make sense at all.

Why This Matters Right Now for Liquidity Providers

Anyone who has parked tokens into a Uniswap, Curve, or Aerodrome pool in the last two years has run into this concept the hard way — usually after checking a portfolio tracker and noticing the dollar value sitting in a pool is lower than what the same two tokens would be worth just sitting untouched in a cold wallet. That gap has a name, and understanding it is the difference between liquidity providing as a deliberate, fee-driven strategy and liquidity providing as an accidental way to slowly give your best-performing token away.

The stakes are higher than they were during the first DeFi summer. Total value locked across decentralized exchanges has settled into the hundreds of billions of dollars, concentrated liquidity designs (the kind pioneered by Uniswap v3 and now standard across v4-style hooks, Aerodrome slipstream pools, and various Solana-based concentrated AMMs) have made the mechanics sharper and less forgiving, and a growing share of on-chain liquidity now comes from retail participants chasing double-digit advertised APRs without pricing in what those advertised numbers are actually compensating for. Meanwhile, restaking, liquid-staking-token pairs, and tokenized real-world-asset pools have introduced new categories of “correlated but not identical” assets where impermanent loss shows up in subtler, slower-moving ways than the dramatic swings people associate with meme-coin pairs.

This guide is written for the person actually putting capital into a pool contract — not the protocol designer and not the passive index investor. If you are deciding between a stablecoin pair, a volatile pair, or a concentrated range position, and you want to know the real arithmetic behind the yield number displayed on the dashboard, the sections below walk through exactly how that number is built and what it’s quietly netting out against.

The Core Mechanic: How a Constant-Product Pool Rebalances You

Most liquidity pools that liquidity providers interact with — Uniswap v2 style pools, and the full-range portion of v3 pools — follow a constant-product formula, commonly written as x·y = k, where x and y are the quantities of the two tokens in the pool and k is a constant that stays fixed except when fees are added or liquidity is deposited or withdrawn. The pool’s price for one token in terms of the other is simply the ratio y/x.

When an outside trader buys token X from the pool, they are removing X and adding Y, which pushes x down and y up, which in turn pushes the price of X higher (since price = y/x). Critically, the pool does this automatically, without anyone consulting the liquidity provider. Every trade that moves the external market price of one asset relative to the other forces the pool’s internal ratio to follow, and that forced rebalancing is arithmetically identical to the liquidity provider continuously selling the token that is going up and continuously buying the token that is going down — the exact opposite of what a trend-following trader would choose to do on purpose.

This is the entire mechanical root of impermanent loss. It is not a fee, not a hack, not a bug in a specific protocol — it is a direct mathematical consequence of maintaining a constant product (or, more generally, any AMM invariant) as external prices move. The pool has no opinion about whether an asset is a good buy at a given moment; it just holds a fixed relationship between quantities and lets the market drag the ratio wherever it wants to go.

The Formula Liquidity Providers Actually Need

For a standard 50/50 constant-product pool, if the price of one asset changes by a multiple r relative to the price at the time of deposit, the value of the LP position relative to simply holding the two assets is given by:

Value ratio (LP vs. HODL) = 2·√r / (1 + r)

Impermanent loss, expressed as a percentage of what you would have had by holding, is 1 minus that ratio. Plug in r = 1 (no price change at all) and you get a value ratio of exactly 1.0 — zero loss, as expected, since nothing moved. Plug in r = 4 (one asset quadruples relative to the other) and the ratio drops to 0.8, meaning the LP position is worth 20% less than simply holding the two original balances. This formula is symmetric: it doesn’t matter whether the price went up or down, only how far the ratio moved from 1.

Fee Income Is the Only Thing Fighting Back

Pools don’t exist to lose money for their depositors on purpose — the trading fees paid by swappers accrue to liquidity providers continuously, and that income stream is the entire economic reason to accept the rebalancing drag described above. A pool with 0.30% fees on Uniswap v2-style venues, or a custom fee tier on v3/v4-style venues (commonly 0.01%, 0.05%, 0.30%, or 1.00% depending on how volatile the pair is), pays LPs a slice of every trade that passes through. Whether providing liquidity is actually profitable comes down to a race: does cumulative fee revenue over the holding period outrun the impermanent loss generated by however much the price ratio wandered during that same period?

This is why volume-to-liquidity ratio is a far more useful screening number than headline APR. A pool with modest total value locked but heavy trading volume relative to its size throws off fees fast enough to comfortably absorb typical price drift. A pool with a huge advertised APR built mostly from a token-emission incentive program, and thin organic trading volume, is often just impermanent loss wearing a disguise — the “yield” is subsidizing a position that is bleeding value against a simple hold, and the emission tokens themselves frequently depreciate faster than the yield accrues.

Concentrated Liquidity: Higher Fee Density, Sharper Downside

Concentrated liquidity, introduced at scale by Uniswap v3 in 2021 and now the dominant design across most serious AMMs, lets a liquidity provider commit capital to a specific price range instead of the full 0-to-infinity range that classic constant-product pools use. Inside that range, the position behaves like a much more capital-efficient version of the constant-product formula — the same dollar amount can earn several multiples more in fees because it isn’t diluted across price levels the market will never actually visit.

The tradeoff is that impermanent loss inside a narrow range is proportionally larger for the same percentage price move, because the position’s effective leverage against the underlying assets increases as the range narrows. A price move that would produce a mild, forgettable loss in a full-range pool can produce a much steeper hit inside a tight concentrated band, and if the price exits the range entirely, the position stops earning fees altogether and simply sits — fully converted into whichever asset is currently weaker — until either the provider manually adjusts the range or the price wanders back in. That combination (zero fee income plus a position now overweight the losing asset) is the single most common way concentrated-liquidity beginners lose money relative to what a wider or passive strategy would have delivered.

Range Width as a Risk Dial

Range width functions as a direct risk-and-reward dial for concentrated liquidity providers. A narrow range (say, plus or minus 2% around the current price for a stablecoin pair, or plus or minus 10% for a moderately correlated pair) concentrates fee capture but requires active management, since the range needs re-centering whenever price drifts to the edge. A wide range behaves much closer to a classic full-range position — lower fee density per dollar deployed, but far less rebalancing overhead and far less risk of falling out of range entirely during a normal trading session.

Correlated Pairs vs. Volatile Pairs: Where the Real Risk Difference Lives

The single biggest lever a liquidity provider controls is pair selection, and it matters more than fee-tier selection or even range width. Stablecoin-to-stablecoin pools (USDC/USDT, for instance) and liquid-staking-token pools (stETH/ETH, for instance) experience impermanent loss that is close to negligible under normal conditions, because the two assets are designed to track each other tightly. The growing supply of yield-bearing ETH derivatives, discussed in more depth in this overview of Ethereum’s proof-of-yield era, is part of why staking-derivative pairs have become one of the more popular low-drift liquidity strategies on major decentralized exchanges. The loss formula above still technically applies, but with r staying close to 1.00 almost all the time, the resulting loss is a rounding error — until a depeg event happens, at which point r can move sharply and the loss shows up all at once.

Volatile-to-volatile pairs, and especially volatile-to-stable pairs (ETH/USDC being the canonical example), see much larger and more frequent ratio swings, which means the impermanent loss formula bites far harder and far more often. This is the segment where the fee-versus-loss race described earlier actually determines whether providing liquidity beats simply holding the two assets separately, and where most retail liquidity providers who chase headline APR without understanding the mechanic end up disappointed.

Worked Numeric Example: A $10,000 ETH/USDC Position Over 90 Days

Assume a liquidity provider deposits $10,000 into a full-range ETH/USDC pool split evenly — $5,000 of ETH at a price of $2,500 (2 ETH) and $5,000 of USDC. The pool charges a 0.30% fee tier, and over the following 90 days, ETH’s price rises to $4,375, a 1.75x move.

First, the buy-and-hold comparison: 2 ETH at $4,375 is worth $8,750, plus the original $5,000 of USDC, for a total holding value of $13,750.

Next, the pool position. Using r = 1.75, the value ratio is 2·√1.75 / (1 + 1.75) = 2·1.3229 / 2.75 ≈ 0.9621. That means the LP position, before any fees, is worth 96.21% of the hold value — a 3.79% impermanent loss, or roughly $521 relative to the $13,750 hold benchmark, leaving the pool assets alone (not fees) worth approximately $13,229.

Now the fee side. Say this particular pool’s trading volume over the 90 days averaged 40% of its total value locked per day — a healthy, actively-traded pair — and the liquidity provider’s share of the pool stayed constant at roughly 0.05% of total pool depth (a reasonable share for a $10,000 position in a multi-million-dollar pool). At a 0.30% fee tier, that volume throws off roughly $1,080 in total fees across the 90 days for the whole pool per $10,000 of daily volume multiplier; scaled to this provider’s proportional share and compounded lightly across the period, the position accrues approximately $410 in fee income.

Net result: $13,229 (post-IL pool value) plus $410 (fees earned) equals roughly $13,639, versus $13,750 for simply holding. The liquidity provider still underperforms a buy-and-hold strategy by about $111, or 0.8% of the hold value, even after fee income — a realistic outcome for a moderately volatile pair experiencing a strong one-directional move within a 90-day window. Had ETH stayed within a tighter band, say a 1.15x move instead of 1.75x, the impermanent loss would have been closer to 0.3% and fee income alone would have comfortably produced a net gain over holding.

Visualizing the Outcome

Reference Table: Impermanent Loss by Price Divergence

Figures assume a standard 50/50 constant-product pool with no fee income netted in. Concentrated positions with narrow ranges experience proportionally larger losses per unit of price movement than shown here.

Common Mistakes Liquidity Providers Make

The mistakes below show up again and again across forum post-mortems, and almost all of them trace back to skipping the arithmetic above in favor of the headline number on a yield dashboard.

• Chasing APR without checking its source. An advertised 80% APR built mostly from a temporary token-emission program is not comparable to an 8% APR built from organic swap fees. The moment emissions taper or the reward token’s price falls, the real yield can collapse toward or below the impermanent loss rate.
• Treating “impermanent” as a guarantee of reversal. The loss only unwinds if the price ratio genuinely returns to its starting point before withdrawal. For a token in a sustained trend, that reversal may never come, which quietly converts an “impermanent” loss into a realized one the moment you exit.
• Setting concentrated ranges too tight without an active management plan. A narrow range can multiply fee income, but only while price stays inside it — once it drifts out, the position earns nothing and sits fully weighted toward the underperforming asset until someone manually intervenes.
• Ignoring correlation between the paired assets. Pairing two assets that move together (ETH and a liquid staking derivative of ETH, for example) produces far less impermanent loss than pairing two assets with no relationship to each other, yet many providers pick pairs based on yield alone.
• Forgetting gas and rebalancing costs on concentrated positions. Frequent range adjustments on networks with meaningful transaction costs can erode fee income to the point where a simpler, wider, less-managed position would have performed better after costs.
• Not comparing against the actual hold benchmark. A position that’s “up 12% since deposit” can still be underperforming a simple hold if the underlying assets are up 20% and 4% respectively — the correct benchmark is always what the same starting capital would be worth split and held, not zero.

Practical Checklist Before Providing Liquidity

• Confirm whether the pool’s advertised yield comes primarily from swap fees, token emissions, or a mix — and discount emission-heavy yields for the risk that the reward token depreciates.
• Check the pool’s trading-volume-to-liquidity ratio over the past 30 days rather than relying on a single day’s snapshot.
• Estimate impermanent loss at a few plausible price-divergence scenarios (1.5x, 2x, 3x) using the reference table above, and compare against realistic fee income at current volume.
• For concentrated positions, decide your range width based on how actively you’re willing to manage it — wide if passive, narrow only if you’ll monitor and rebalance regularly.
• Favor correlated-asset pairs (stablecoin-stablecoin, staking-derivative-native-asset) when the goal is fee income with minimal directional risk.
• Factor in gas or transaction costs for entry, exit, and any range adjustments, especially on networks where fees spike during volatility.
• Set a clear exit or rebalancing trigger in advance — deciding calmly before depositing beats deciding emotionally while a position is already underwater relative to holding.

Key Takeaways

• Impermanent loss is a mechanical consequence of constant-product rebalancing, not a fee or a scam — it happens because the pool automatically sells the rising asset and buys the falling one as external prices move.
• The loss scales with how far the price ratio between the two pooled assets has moved, not with time — a 2x divergence costs about 5.7% of value versus holding, and a 5x divergence costs about 25.5%.
• Trading fees are the only structural counterweight; providing liquidity is profitable only when cumulative fee income outpaces the impermanent loss generated over the same period.
• Concentrated liquidity positions earn fees faster but amplify the loss per unit of price movement, and they stop earning entirely once price exits the chosen range.
• Correlated pairs (stablecoin-stablecoin, staking-derivative-native-asset) minimize impermanent loss under normal conditions, while volatile pairs carry the highest risk and the highest potential fee compensation.
• The loss is only truly “impermanent” if you withdraw after prices return to their original ratio — for trending assets, that reversal often never arrives before exit.

Frequently Asked Questions

What exactly causes impermanent loss in a liquidity pool?

Impermanent loss happens because an automated market maker keeps a fixed mathematical relationship between the quantities of two pooled tokens. As outside traders buy the asset that’s rising in price and sell the one that’s falling, the pool’s own reserves shift to match — meaning the liquidity provider ends up holding less of the appreciating asset and more of the depreciating one than they would have by simply holding both assets separately.

Is impermanent loss a real, realized loss or just a paper calculation?

It becomes a real, realized loss the moment you withdraw from the pool while the price ratio differs from the ratio at deposit. It’s called “impermanent” because if prices fully return to their original ratio before you exit, the loss vanishes entirely — but for assets in a sustained trend, that return to the original ratio may never happen before you need or choose to withdraw.

Can trading fees fully offset impermanent loss?

Yes, in many cases, particularly for pools with high trading volume relative to their size. Fee income accrues continuously regardless of price direction, so a pool that sees heavy swap activity can generate enough fees to outpace the impermanent loss caused by typical price drift. Whether fees win the race depends on volume, fee tier, and how far the price ratio actually moves during the holding period.

Does impermanent loss affect stablecoin-to-stablecoin pools?

The same formula technically applies, but because two well-collateralized stablecoins are designed to trade close to a 1:1 ratio, the price divergence stays minimal under normal conditions, making the resulting loss negligible. The risk becomes material only during a depeg event, when one stablecoin’s price temporarily or permanently detaches from its peg.

How does concentrated liquidity change the impermanent loss calculation?

Concentrated liquidity magnifies both the fee income and the impermanent loss for a given price move, because capital is deployed across a narrower price range instead of the full range. This raises capital efficiency and fee capture while price stays inside the chosen range, but it also means the position takes on effectively higher exposure to price movement within that range, and it earns zero fees once price exits the range entirely.

What’s the simplest way to avoid impermanent loss altogether?

The only way to avoid it completely is to avoid providing liquidity to a two-asset pool where the price ratio can move — for example, by holding assets directly, or by providing liquidity only to pairs of highly correlated assets such as two stablecoins or a liquid staking token paired with its underlying asset, where the price ratio stays close to constant under normal market conditions.

References

• Uniswap Protocol documentation on constant-product automated market makers and concentrated liquidity mechanics.
• Curve Finance documentation on stable-asset pool design and reduced-slippage invariants for correlated pairs.
• Academic and industry research on automated market maker loss-versus-rebalancing dynamics in decentralized exchange pools.