There ain't no such thing as a free lunch, however.
Bancor is updating its protocol once more to defeat the insidious issue of impermanent loss, which it earlier called “DeFi’s dirty little secret.”
Impermanent loss, also called divergence loss, affects exchanges based on automated market makers like Bancor or Uniswap. It happens when the prices of two assets in a liquidity pool diverge significantly, with one side going strongly up or down in value.
The effect translates to a loss of value compared to a benchmark “buy and hold” portfolio. Liquidity providers (LP) may take out less money than they would have had if they just held the tokens separately — despite the fact that they earn trading fees from the protocols.
The issue occurs because of arbitrage traders, who are necessary for AMMs to bring their prices in line with the other markets. Their activity nevertheless extracts value from LPs who facilitate the exchange.
The loss was initially called "impermanent" because if the prices return to their initial state, the loss is reverted. However, even in the optimistic scenario, divergence loss cuts into extra profits LPs would have otherwise obtained from the price swings.
Bancor has made the elimination of impermanent loss one of the key features for its second iteration. Its initial approach relied on oracles, which would read the true prices of each token and make arbitrage largely unnecessary.
However, Nate Hindman, Bancor’s head of growth, told Cointelegraph that this approach was eventually revealed to be too risky. Oracles update slowly and can be exploited by fast traders, he argued.
There has also been a growing realization in the industry that impermanent loss is impossible to truly solve. Each solution presents certain drawbacks or merely shifts the loss to somebody else.
The latter approach is what Bancor is going for with V2.1. It’s introducing the concept of impermanent loss insurance, which guarantees that liquidity providers will receive up to 100% of their initial capital, plus fees accrued. The exact percentage is subject to a vesting schedule based on how long the user is providing liquidity, Hindman explained. Total coverage is reached after 100 days, but there is an initial 30 day cliff during which no payments will be disbursed.
The insurance claim itself is paid indirectly by the protocol and the BNT holders through on-demand minting of new tokens, if necessary. The pools covered by this insurance and the exact vesting parameters are decided by community governance. The solution is in some ways similar to how protocols like Uniswap are currently subsidizing some liquidity providers with new UNI tokens.
But in Bancor’s case, there is also a deflationary mechanism. Since all pools have BNT as the second token, the protocol is able to offer single-token liquidity provision — it simply mints the corresponding amount of BNT required. The protocol then receives fees as a co-investor in the pool. When someone decides to only supply BNT, the previously minted supply and the fees accrued are burned, resulting in a net supply restriction.
The expectation is that with enough usage and in periods of low volatility, deflation through fees will prevail and accrue value for token holders.
“We view this as liquidity mining 2.0 - instead of arbitrarily paying LPs to provide liquidity on our protocol, we are compensating based exactly on their individual impermanent loss incurred.”
As it is becoming increasingly clear that divergence loss is unavoidable, future solutions may offer a spectrum of ways to spread risk between different market participants. Mooniswap, the DEX launched by the 1inch Exchange team, works under a similar principle by limiting the profit arbitrage traders can make.