Liquidity Pools Are Not Savings Accounts

A savings account promises two things: your principal back, and a stated interest rate on top of it. A liquidity pool promises neither. When you deposit funds into a liquidity pool, you’re not lending money and waiting for interest. You’re becoming a market maker, supplying a pair of assets that traders swap against, and your share of that pool constantly rebalances as prices move. That rebalancing can leave you with less value than you deposited even if both assets you put in went up in price, a structural feature called impermanent loss, not a bug or a rare edge case. An academic study of Uniswap, the largest decentralized exchange, found that during a five-month period in 2021, roughly half of liquidity providers ended up with negative returns once impermanent loss was weighed against the fees they earned.

The confusion this article addresses isn’t a minor technicality. It’s the reason Anchor Protocol could market itself as offering “20% yield” on a stablecoin right up until the underlying system collapsed and wiped out tens of billions of dollars in a matter of days.

Key Facts

TL;DR

• A liquidity pool position is a market-making stake in a pair of assets, not a loan earning fixed interest, and its value moves with the relative prices of both assets, independent of the fees it earns.
• Impermanent loss happens whenever the price ratio between two pooled assets changes, and it can leave a liquidity provider with less value than simply holding the same assets, even during a period when both assets rose in price.
• An academic study of Uniswap v3 found that during the period studied, total impermanent loss across the platform exceeded total fees earned, leaving close to half of liquidity providers with a net loss.
• Terra’s Anchor Protocol advertised roughly 20% annual yield on a stablecoin, a number that sounded like a savings account rate but was actually subsidized by reserves that ran out, triggering a collapse that erased around $50 billion in days.
• The word “yield” in DeFi describes several very different things, fee income, subsidized token emissions, or interest from actual lending, and conflating them with a bank’s interest rate is the single most consequential misunderstanding a new DeFi user can make.

What a Liquidity Pool Actually Is

Most decentralized exchanges rely on automated market makers, smart contracts that hold a pool of two assets, say ETH and USDC, and let traders swap one for the other based on a pricing formula built into the contract. Liquidity providers deposit both assets into the pool in order to enable that trading, and in exchange, they earn a share of the trading fees generated. This is fundamentally a market-making role, not a lending role. A bank takes your deposit and lends it out at a rate it controls, promising your principal back regardless of what happens to the loans it makes. A liquidity pool has no such promise built in. Your position’s value is determined entirely by the pool’s current holdings and the current market prices of the two assets, which shift automatically as trades occur.

Impermanent Loss, Explained Without the Math

Impermanent loss happens because an automated market maker’s pricing formula forces the pool to sell the asset that’s rising in price and buy the asset that’s falling, in order to keep the pool balanced according to its formula. If you deposit equal values of ETH and USDC and ETH’s price then rises significantly, the pool will have automatically sold some of your ETH into USDC along the way to stay balanced, which means when you withdraw, you’ll have less ETH and more USDC than if you had simply held both assets separately. You still made money if ETH went up, but you made less than you would have by just holding it, and that gap is impermanent loss. It’s called “impermanent” because it only becomes a fully realized loss when you actually withdraw from the pool at a moment when that price gap hasn’t been offset by fees earned, though for many providers, that’s exactly when they end up withdrawing.

This is precisely what the Loesch, Hindman, Richardson, and Welch study of Uniswap v3 measured directly. Over the roughly five months examined, the pools studied generated $199.3 million in trading fees, which sounds like meaningful income for liquidity providers. But those same pools incurred more than $260.1 million in impermanent loss over the same period, meaning the fees didn’t come close to covering it in aggregate, and just under half of individual liquidity providers ended the period with negative returns compared to simply holding the underlying assets.

When “Yield” Isn’t Interest: The Anchor Protocol Collapse

Impermanent loss explains why liquidity pools can quietly underperform. Terra’s Anchor Protocol showed a different, more dramatic version of the same underlying confusion: DeFi “yield” isn’t automatically comparable to a bank’s interest rate, even when it’s advertised on a supposedly stable asset.

Anchor offered depositors of UST, an algorithmic stablecoin, approximately 20% annual yield, a rate that dwarfed anything available at a bank and was marketed heavily as a reason to hold UST. According to research published by the National Bureau of Economic Research, that yield was not primarily generated by organic borrowing demand within the protocol. It was substantially subsidized by a reserve fund that Terraform Labs and an affiliated foundation had to repeatedly top up to keep the advertised rate alive. By mid-April 2022, according to Chainalysis’s reporting, more than 72% of all circulating UST was sitting in Anchor specifically to capture that yield, meaning a huge share of the ecosystem’s supposed “stability” depended on a subsidy program that was, by its own internal projections, running out of money.

When large withdrawals began on May 7, 2022, UST slipped below its dollar peg, and the mechanism designed to restore that peg instead accelerated its collapse as holders rushed to redeem UST for LUNA, flooding the market with newly minted LUNA and crashing its price. Within days, according to the NBER paper’s analysis, roughly $50 billion in value had been wiped out, and both UST and LUNA were left worth a small fraction of a cent.

Comparison: Savings Account vs. Liquidity Pool

What Users Actually Lose, and Who Benefits

When impermanent loss erodes a liquidity position, the loss is largely invisible until withdrawal, and many users never realize it happened at all if they don’t compare their position against simply holding the underlying assets. When a yield program collapses outright, as Anchor’s did, the loss is immediate and total for anyone still holding UST or LUNA when the peg broke. In both cases, the protocol’s designers and early participants who withdrew before a collapse can come out ahead, while later depositors, often drawn in by an advertised rate that sounded comparable to traditional savings, absorb the loss. Marketing language describing DeFi yield in terms borrowed from traditional banking, “interest,” “APY,” “savings,” makes this confusion worse by implying a guarantee that was never actually part of the underlying mechanism.

Practical Guidance

• Before providing liquidity to any pool, understand that your return depends on both the fees earned and the price movement between the two assets, not fees alone, and that the second factor can outweigh the first.
• Treat any advertised “yield” above what a normal savings account or money market fund offers as a signal to investigate where that yield actually comes from, organic trading activity, borrowing demand, or a subsidy that could run out.
• Avoid concentrating funds in single-asset-pair pools where the two assets are likely to diverge significantly in price, since that divergence is the direct driver of impermanent loss.
• Recognize that “stablecoin yield” is not automatically safe simply because the stablecoin’s price doesn’t move; the yield program built around it can still collapse independently, as Anchor Protocol demonstrated.
• Use available impermanent loss calculators before entering a position to model outcomes across a range of price scenarios, rather than focusing only on the advertised fee rate.

What Happens Next

Expect continued innovation aimed at reducing, though not eliminating, impermanent loss, including concentrated liquidity designs, single-sided liquidity provision, and dynamic fee structures, several of which existed before Anchor’s collapse and are still being refined. None of these designs remove the core structural fact that a liquidity pool position’s value depends on relative price movement between its underlying assets, not a fixed promised rate. Expect regulators and industry commentators to keep scrutinizing DeFi “yield” marketing specifically in the wake of Anchor’s collapse, though no comprehensive rule currently requires DeFi protocols to disclose whether an advertised rate is organic or subsidized the way securities law requires similar disclosures for regulated investment products.

FAQs

Can I lose money in a liquidity pool even if I never sell or withdraw at a loss?

Yes, in the sense that impermanent loss reduces the value of your position relative to simply holding the assets, even before you withdraw. It only becomes a realized loss compared to holding once you actually exit the position.

Why did Anchor Protocol’s 20% yield collapse?

The yield was substantially subsidized by a reserve fund rather than generated purely by organic borrowing demand. When that reserve couldn’t keep pace and large withdrawals began, the mechanism meant to maintain UST’s dollar peg failed, triggering a collapse that erased roughly $50 billion in days.

Is impermanent loss a rare risk or a common one?

It’s a structural feature of automated market makers, not a rare event. A study of Uniswap v3 found that during the period examined, total impermanent loss across the platform exceeded total fees earned, and about half of individual liquidity providers had negative returns.

Does a stablecoin’s price stability mean a yield program built around it is safe?

No. Anchor Protocol’s UST maintained its price stability for an extended period while the yield subsidy behind it was quietly running out, showing that price stability and program solvency are separate questions.

Sources

• Loesch, Hindman, Richardson, and Welch, “Impermanent Loss in Uniswap v3,” arXiv, 2021
• National Bureau of Economic Research, “Anatomy of a Run: The Terra Luna Crash”
• Chainalysis, “How TerraUSD Collapsed”

This article is for educational purposes and does not constitute financial or legal advice. Providing liquidity to decentralized exchanges and participating in DeFi yield programs carries risk of partial or total loss of principal, independent of any hack or fraud. If you are making decisions about participating in DeFi protocols, consider consulting a licensed financial advisor.