Bancor’s Death Spiral.
Bancor is a decentralized exchange (DEX), and DEXs have impermanent loss (IL). One of the reasons why some people shy away from providing liquidity is IL, but Bancor promised protection.
The protection mechanism had risks, which brought extensive losses to liquidity providers (LPs) after a sharp market downturn. People who provided lifeblood to the protocol suffered the most when the team attempted to save the Bancor token ($BNT) and its holders. Is this how it’s supposed to work?
More importantly, do holders of various protocols’ tokens fully understand that they might act as a backstop when underlying platforms are experiencing issues? In crypto, it’s not all fun and games, and yields reflect risks. But sometimes, since some people don’t want to bear responsibility, others may suffer.
Bancor’s Death Spiral
On exchanges like Uniswap, LPs need to deposit 50/50 in two tokens, e.g., ETH/USDC. Bancor is more convenient because LPs provide one side of liquidity in a token like $ETH, and the protocol provides an equal amount of $BNT from its reserves.
Let’s say the price of $ETH in Bancor’s pool increases by 10%. This means someone put USDC into the pool and got $ETH back. The ratio between the two changes from 50/50 to, say, 45 for $ETH and 55 for $BNT. This event triggers IL, which you can read more about here.
Both Bancor’s and Uniswap’s LPs are affected by changes in pool ratios and IL, but differently. On Uniswap, it means a LP may withdraw different amounts of assets than what they deposited. On Bancor, it means that some of the LP’s position will be returned in $BNT token, which the team called IL protection.
In technical jargon, LPs are “short volatility.” This means that the best time to be an LP is when the market is calm. When prices jump or dump, IL ramps up.
Remember, the ratio between two tokens in a given pool constantly changes. Bancor rebalances ratios. If the paired asset is growing and the market is overall doing well, it’s not a big deal: the protocol sells a portion of the paired asset, say $ETH, for $BNT. When an LP withdraws, they receive some of their position in $BNT, which they can then sell onto the market and get their money back.
The problem occurs when the market isn’t doing well. When people actively sell and withdraw liquidity, $BNT in pools starts growing fast while the number of valuable tokens begins to shrink.
Exchanges like Uniswap and Bancor don’t know about what the price is on, say, Binance. They calculate prices based on the ratios in the pools. If there’s a significant inflow of $BNT into pools, it starts declining in price. Other LPs and holders see $BNT declining and rush to sell. This triggers something called a “death spiral.”
$BNT is still alive mainly because the team decided to stop IL protection and not pay LPs their $BNT compensation. Given that ratios of valuable tokens in pools are low, LPs should be receiving large amounts of $BNT on withdrawals, which would kill the token and the project. What happened technically meant a default on Bancor’s debt, which caused outrage on social media.
The discussion, however, focused on the centralization aspect. People are furious that a team can unilaterally make such serious decisions. But what would happen if Bancor didn’t do it?
The Implications of Risk
If you watched “The Big Short,” you probably would remember a scene where Jared Vennett explains how tranches work using Jenga. Do yourself a favor, and watch it again.
When Jared speaks about BBB bonds, he says they’re a little risky, but you can also make more money off them. The rest of the world works the same way: in a system with a risk, it’s most likely divided into tranches, and those who risk more get higher rewards.
Now take a look at the tweet below. Which of the two options looks more profitable? Staking for a single pool or staking for all pools?
If your yield is higher than that of your peers, you are the backstop when things go wrong. The same happened to $LUNA holders who ensured liquidity for some people running from $UST and $MKR holders who had to suffer significant dilution after the platform was on the brink of insolvency in 2020.
The above doesn’t mean that single pool stakers should not bear losses when the platform goes down. But, the most logical is that they suffer losses only after $BNT holders as enjoyers of higher returns during good times would suffer from an extensive sell-off.
The “protection” Bancor advertised was a senior tranche of risk. They generated some low-to-mid-sized yields, and the platform slowly reimbursed them by using junior tranche owners’ money.
When the team stepped in, it probably had no other choice. It was either seeing your project die or this. Should LPs think ahead about who backstopping their risk, and would they be prioritized if things went south? Did words like “protection” play their role in masking the risk? These are open questions that I’d like you to think about.
We operate in a largely unregulated environment, and teams care about themselves and their projects in the first place. In this kind of environment, it’s important to at least look for potential risks and emergency mechanisms before putting your money on the line.