Hey @eddieb. An insurance fund and safety module have different purposes. The insurance fund acts as a backstop for liquidations and is constantly in use. The safety module acts as a protocol-wide backstop; in the case of an unexpected loss of user funds (e.g. a hack), the DAO can vote to slash the safety module and make users whole.
More generally, it could be possible to implement an insurance fund (not a safety module) where users could deposit USDC and earn some yield for doing so (either paid by profitable liquidations or by some rewards from the DAO).
However, I think outsourcing risk-capital is appropriate when the protocol needs leverage. Let’s say it needs $300M in insurance, but it doesn’t have $300M to spend right now. Instead, it pays a ~10% APR ($30M) in exchange for the option to slash a $300M notional.
In this case, dYdX certainly has $1M to fund the insurance fund itself. By doing so, it internalizes the risk of losing that $1M if the insurance fund is depleted, but it also avoids:
- Having to pay an APR that reflects this risk to the depositors.
- Dealing with the uncertainty about the insurance fund’s TVL, as depositors must be allowed to withdraw their funds in some reasonable timeframe. Given how important the insurance fund is to ensuring liquidations happen smoothly, I’d say avoiding this uncertainty is pretty important.
In a nutshell, the protocol seeding the insurance fund itself (instead of outsourcing the deposits) seems to be a more predictable, and in the long-run potentially cheaper, solution. Curious if you have other suggestions for how the DAO can efficiently purchase the necessary USDC.
On Safety Modules Generally
~For those new to the dYdX community~
dYdX has considered variations of safety modules in the past and implemented a fork of Aave’s Safety Module on dYdX v3. Following some research from Xenophon Labs and lots of discussion, the community voted to wind down the safety module last year.
One reason the module was wound-down was because it was entirely denominated in DYDX token. Intuitively, this poses some “wrong-way risk” in that, if the protocol suffered a shortfall, DYDX token price and liquidity would also decline.
However, as dYdX Chain matures, the community might choose to re-surface the safety module discussion and consider a USDC-based safety module in a separate thread. We discussed this in detail in our paper. Thanks!