Impermanent Loss Explained: The Hidden Cost of Providing Liquidity

Liquidity providers get promised a cut of every trading fee that flows through a pool, and the yields advertised on DeFi dashboards can look irresistible. But there's a quieter number that never makes it into the marketing copy: impermanent loss. It's the reason a liquidity position can underperform simply holding the same two tokens in a wallet, even while collecting fees the entire time.

Automated market makers (AMMs) like Uniswap or Raydium hold reserves of two assets — say ETH and USDC — in a pool, priced against each other by a formula (commonly x * y = k, where the product of the two reserves stays constant). When you deposit into the pool, you're not choosing a ratio; the contract sets it based on the pool's current price, and you receive LP tokens representing your share.

As traders swap through the pool, the ratio of the two assets shifts. If ETH's price rises elsewhere, arbitrageurs buy ETH from the pool until the pool's price matches the broader market — which means the pool ends up holding less ETH and more USDC than when you deposited. When you withdraw, you get back that new mix: less of the asset that went up, more of the asset that didn't. You still made money if ETH rose, just less than if you'd simply held it.

The loss is calculated relative to what you'd have if you'd just held the two assets instead of pooling them. It's "impermanent" because if prices move back to where they were when you deposited, the loss disappears — the pool's ratio returns to your original split. It only becomes a permanent, realized loss the moment you withdraw while the price ratio has diverged from your entry point.

That framing is a little generous, though. In practice, prices rarely snap back to the exact moment you deposited, and the term has drawn criticism for making the effect sound less serious than it is. A more honest name, and one that's gained traction, is "divergence loss" — it's a loss driven by how far the two assets' prices have diverged from each other since you entered the pool, in either direction.

For a standard two-asset AMM, impermanent loss scales with the price ratio change between the two assets, not just how much either one moved individually. A price ratio change of 1.25x (one asset up 25% relative to the other) produces roughly 0.6% impermanent loss. A 2x ratio change produces about 5.7%. A 5x ratio change — not unusual for volatile altcoin pairs — produces roughly 25.5%. The loss grows faster than the price divergence, which is why volatile pairs are the riskiest place to provide liquidity, even though they often advertise the highest fee yields.

Pairing two correlated assets — like two stablecoins, or ETH with a liquid-staked ETH derivative — minimizes this risk because their price ratio barely moves. That's why stable-stable and correlated-asset pools consistently show far lower impermanent loss than volatile pairs, even with thinner fee income.

Impermanent loss isn't automatically a dealbreaker — the question is whether accumulated trading fees (and any additional incentive rewards) outpace it over your holding period. High-volume pools with moderate volatility, like major stablecoin pairs or blue-chip pairs during active trading periods, often generate enough fee revenue to comfortably offset the divergence loss. Low-volume pools paired with volatile, low-cap tokens are the classic trap: the advertised APY looks huge because it includes token emissions, but the underlying divergence loss can quietly erase most of the gain once you account for both sides of the position.

Before providing liquidity, check the pair's historical volatility relative to each other, not just each token's price chart in isolation. Favor correlated or stable pairs if you want to minimize impermanent loss, and treat headline APYs on volatile pairs with skepticism until you've estimated what divergence loss would do to that yield if the pair moved 2x or 5x apart. Tools that show a pool's realized fees versus its impermanent loss over time are worth checking before committing capital, not after.