Why Yield Farming APY Hides Impermanent Loss

A liquidity pool advertises 20% APY. A trader deposits equal value of two tokens, checks back in a month, and finds the position is worth less than if they had simply held both assets. The rewards were real. The yield was real. And yet the outcome was a loss.

This is the part of yield farming that rarely makes it into the marketing. The APY number describes rewards emitted by the protocol - it says nothing about what happens to the two assets sitting inside the pool while those rewards accrue.

Key Takeaways

• Advertised APY in yield farming is a gross figure - it doesn't subtract impermanent loss
• Impermanent loss grows with price divergence between pooled assets, not with time
• High APY pools often exist because the underlying assets are volatile, not because the yield is free
• Net returns depend on price correlation between paired assets, not just the reward rate

Why Yield Farming APY Hides Impermanent Loss

Most traders read "20% APY" as a straightforward yield, similar to a savings account or a staking reward. The assumption is simple: deposit capital, collect a percentage over time, withdraw more than you put in.

That framing works for single-asset staking. It breaks down in a two-sided liquidity pool, because the pool isn't just paying you a reward - it's also rebalancing your holdings every time someone trades against it. The APY figure captures the first part. It ignores the second.

What Actually Happens

Most automated market maker (AMM) pools use a constant-product formula, where the pool holds two assets in a ratio that shifts as traders swap between them. If one asset rises in price relative to the other, arbitrageurs trade against the pool until its internal price matches the external market. That process pulls the appreciating asset out of the pool and leaves the depositor holding more of the weaker asset and less of the stronger one.

This is impermanent loss: the gap between what a liquidity position is worth and what the same two assets would be worth if simply held in a wallet. It isn't caused by a hack, a bug, or bad luck - it's a mechanical consequence of how AMMs maintain their pricing curve. The loss is called "impermanent" because it shrinks if prices converge again, but it becomes permanent the moment liquidity is withdrawn while a price gap exists.

The size of the loss scales with divergence, not with time. Two assets that move 5% apart create a small, often negligible loss. Two assets that move 50% apart - common between a volatile altcoin and a stablecoin - create a loss that can outpace months of reward emissions. This is also why the highest APY pools tend to carry the highest impermanent loss risk: the reward rate and the underlying volatility are often correlated, because protocols use high emissions to attract liquidity into pairs that are inherently harder to balance.

Slippage cost compounds this. Every swap that rebalances the pool also pays a trading fee to liquidity providers, which is part of what generates the yield. But in fast-moving markets, the same volatility that produces higher fee income is the volatility driving the price divergence that produces impermanent loss. The two effects move together, not independently - which is why farmers can watch fee income accumulate while their position's total value still declines.

Example from Crypto Markets

Consider a liquidity pool pairing ETH and a stablecoin. A trader deposits $5,000 of ETH and $5,000 of the stablecoin. If ETH rallies 40% while the stablecoin stays flat, arbitrage trades pull ETH out of the pool and push stablecoins in, rebalancing the pair back toward 50/50 by value. The result: the depositor now holds less ETH than if they had simply held it, and more stablecoin exposure they didn't necessarily want.

Modeling this scenario, a 40% price divergence produces an impermanent loss of roughly 4-5% relative to simply holding both assets. A pool advertising 20% APY sounds like it easily outpaces that. But APY is typically quoted as an annualized rate based on current conditions - the actual accrued yield over the weeks it takes ETH to move 40% may only be a fraction of that headline number, sometimes not enough to offset the divergence at all.

This dynamic played out repeatedly across DeFi pools during periods of sharp altcoin rallies and drawdowns - the same volatility that pushed emission-driven APYs higher also pushed impermanent loss higher, often in the same window.

What Traders Can Learn

The lesson isn't that yield farming is bad - it's that the APY figure is incomplete without knowing the correlation between the paired assets. A pool of two stablecoins carries minimal impermanent loss because both assets are pegged to the same value; the yield in that case is closer to what it appears to be. A pool pairing a volatile token against a stablecoin or against a different volatile token carries a structural cost that grows precisely when the market is moving the most - which is often when farmers are most tempted to chase the highest headline rate.

This mirrors a pattern seen elsewhere in DeFi risk: the way collateral ratios determine when DeFi protocols become insolvent shows that safety margins matter more than headline numbers, and the way DeFi liquidation auctions unwind positions under stress shows that mechanical processes, not sentiment, drive outcomes. Yield farming follows the same logic - the mechanics of the pool determine the real return, not the number on the dashboard.

Reading a pool's underlying asset pair, and asking how correlated those two assets are, matters more than reading the APY itself. In practice, that means working through the position in order:

1. Identify the two pooled assets and how closely their prices track each other.
2. Estimate the plausible divergence between them over your intended holding period.
3. Convert that divergence into an impermanent loss figure - 40% divergence costs roughly 4-5%.
4. Compare it against the yield actually accrued over that period, not the annualized headline rate.
5. Ask why the APY is high. Emission-driven rates usually point back to volatility in the pair.

FAQ

Does advertised APY include impermanent loss?

No. Advertised APY is a gross figure describing the rewards a protocol emits to liquidity providers. It says nothing about what happens to the two assets sitting inside the pool while those rewards accrue. Impermanent loss is subtracted separately, and in volatile pairs it can consume most or all of the headline rate. Net return is always APY minus divergence cost.

Can impermanent loss be avoided entirely?

It can be minimized by providing liquidity to pairs of correlated or pegged assets, such as two stablecoins, where price divergence is structurally limited. It cannot be fully avoided in pools pairing volatile assets, because the loss is a mechanical consequence of how automated market makers rebalance their pricing curve, not a bug or a one-off event.

Does higher APY always mean higher impermanent loss risk?

Not always, but there's often a relationship. Protocols frequently raise emissions specifically to attract liquidity into volatile or newly launched pairs, which are also the pairs most prone to price divergence. The reward rate and the underlying volatility tend to move together, so the highest headline yields usually sit on the pools hardest to balance.

How long does it take for impermanent loss to become permanent?

It becomes realized the moment liquidity is withdrawn while a price gap exists between the pooled assets. If prices converge before withdrawal, the loss can shrink or disappear entirely. Time itself is not the driver - the size of the loss scales with how far the two assets have moved apart, not with how long the position has been open.

Conclusion

Yield farming's APY figure describes only one side of the position. The other side - how the pool's asset ratio shifts as prices diverge - is where much of the real return is decided. High yield often exists precisely because the underlying pair is volatile enough to generate meaningful impermanent loss. Understanding that relationship, rather than reading APY in isolation, is what separates farming for yield from farming for a loss.

Yield is never free - it's compensation for a risk you haven't priced yet.