How the system currently handles a default in a bear market
Imagine we have a 100M pool with 26M in idle USDC. A bear market hits. Under our current system, ~25M rapidly gets withdrawn from the pool, netting the remaining holders roughly half a percent in fees. The remaining holders are unable to leave until USDC enters the pool or loans are repaid(at which point, people are competing to exit on repayment.) Then a 5M loan defaults, causing those who didn’t exit to suffer a 6.7% loss. Most of the remaining liquidity gets drained quickly, earning another 0.05% in fees for the pool.
Problems
The losses of people in the pool are amplified by 34% due to the liquid exits.
The remaining lenders are stuck in the pool for months and are fighting to exit the moment USDC enters the pool.
The fees to compensate for this are small.
The system is not intuitive. It has prompted a fair number of questions and complaints on Discord.
A Path Forward
The two broad options I see are to either rework liquid exiting or to move rapid selling to a strong secondary market such as an LP pool. Separately I am working on a proposal together to incentivize a secondary market. While I believe liquid exiting could be improved, I suspect it will always be inferior to secondary markets and would take considerable effort to rework.
A liquid secondary market would ensure any issues occur outside the pool itself. Lenders wouldn’t get amplified losses in a bear market. Instead, the market would determine a fair price based on liquidity and default risks.
I am not enough of an expert on tokenomics to add anything useful here but for the sake of discussion wanted to share my thoughts more generally on the issue of liquidity.
Recent well-publicised events have put a spotlight on the issue of capital and risk management and the role/trust put into CeFi institutions to manage depositor’s capital. The role that these institutions have played are more similar to a Tradfi bank, albeit lacking any regulatory, capital or risk management oversight.
The benefit of Defi is that it should all be transparent and the depositors, or lenders/investors in this case should be very aware of the use of these funds, who they are being loaned to and when they can expect repayment. The issue with these CeFi institutions has been the lack of transparency and centralised nature which allows them to halt withdrawals of depositor’s funds, which leads to destruction of trust and a classic bank run ensues.
I think a long term solution here would be to educate and make it clear to lenders that they are in fact lenders, or investors and that the duration of their investment will be locked up for a period of time to match the use of funds.
As a decentralised protocol, we wouldn’t/shouldn’t have the ability to prevent withdrawal of funds after the fact and so if it is made clear up front that funds are locked for a period of time we would not breach trust with users in that regard.
I do acknowledge perhaps the current defi market is not mature enough to attract more patient capital and short-term ability to create liquidity is required to grow the protocol.
To look at it another way, allowing lenders to withdraw funds prior to a default would circumvent their very role as investors who have committed capital and are on-risk. Lending is a risky exercise and losses can/will happen. We will do everything to manage risk and avoid loss but it is risk/reward and front running to withdraw funds in anticipation of default isn’t fair to those that remain committed as they should.
I think liquid exit is good because it penalizes early exitors and benefits those who stay in the pool for the long term. Liquid exit combined with a good secondary market for tfUSDC/USDC seems like a good path, as each form of exiting has its own benefits.
The difference between liquid exit vs. a secondary market is who is providing liquidity and how much they earn in fees. For a protocol like curve.fi or uniswap v3, the LPs in those pools earn fees when users exit through the secondary market. One of the advantages of liquid exit is that 100% of fees go to LPs, while using curve would have a portion of fees go to the curve protocol.
Only if there is no default. If there is a default, then liquid exit amplifies the losses of long term holders.
This is especially concerning because, sometimes, lenders will know about the default before we can act. We will eventually have someone pulls 10 million out right before a default is announced and it will look a lot like insider trading at the expense of long term holders.
Secondary markets are taking on this default risk in return for fees and they have better tools to manage this risk.