I would like to hear thoughts about:
Instead of a simple Surplus Buffer, there would be two. One would look exactly like it does now except that overflow once full would go to the second Surplus Buffer – a kind of “excess reserves” pool of DAI.
As anyone who checks the invaluable Maker Burn page can tell you, the implied yield on simply using excess DAI to buy and burn MKR has been hovering around 5%. However, the available yield on DAI in a wide variety of pools can be materially higher than that.
I think it would be useful to instead of buy/burn automatically, have excess funds deposited into a second, “Excess” Surplus Buffer. This second buffer can serve several purposes:
Lend DAI or burn MKR as a function of which yields more. This could be done either on a gradual scale of doing both with the implied yield on MKR and variable rates on DAI determining which is more utilized. When DAI returns are considerably higher, this allows for quicker compounding of the internal value of the protocol. This gives the protocol an asset that increases the internal rate of return, with some caveats below.
Function as a backstop to the Surplus Buffer that does not require minting MKR tokens. If the Surplus Buffer is drawn down in part or whole, funds from the Excess Surplus Buffer can be used to either ensure solvency or to help replenish the funds. Currently, bootstrapping our way back to full buffer from self-funding is the only way to go about this, requiring a larger buffer of non-productive assets. An Excess Surplus Buffer that can be used to refill the standard Surplus Buffer would hopefully reduce the need to raise the Surplus Buffer as quickly, as it would reduce the danger to the protocol of a second negative shock occurring a short time after an initial one. Looking at markets and protocol history, the scenario that seems most likely to bite us is not another BT, but a huge hit where the Surplus Buffer is not depleted, but then cannot refill quickly enough to respond to another large-to-medium event following the first after only a short time. An Excess Surplus Buffer would allow another layer of protection that also does not leave funds fallow.
This would be a more sophisticated approach to deploying “profits” beyond what is needed to capitalize the protocol. Currently, the only options are to add to the Surplus Buffer or to buy and burn MKR tokens. This gives an avenue for capitalizing on the success of both of the protocol’s asset tokens, which I think we can all recognize can be divergent in their demand in the marketplace. Functionally, if MKR tokens are extremely cheap, it will still burn tokens, but if that same amount of DAI can earn a higher return in the lending markets, it can still do so.
In essence, the idea is that future increases in the Surplus Buffer could partly be increases in an Excess Surplus Buffer, which should also earn some return. A few obvious risks stand out, however. The main one being that we wouldn’t want to lend out (or later withdraw) so much DAI as to disrupt markets. Not only is there the danger of moving rates, but moving them lower would probably have some impact on the demand for DAI. So this would need to be managed very conservatively in scope until markets were very deep and very liquid, perhaps requiring a cap on how quickly an Excess Surplus Buffer could act, regardless of market conditions.
Before thinking about this further, does anyone else think this is an idea worth exploring? Also, would the technical implementation and formal risk evaluation of this be a distraction from more important ongoing projects, or easy to get out the door if it’s desired?