Backers should receive rewards as interest payments are made back to the protocol, because this most closely aligns all incentives. Specifically we propose having 2% of the total GFI supply be allocated for the first $1B of interest payments, with those rewards being paid on an exponentially decelerating rate per dollar (ie. earlier interest payments are rewarded significantly higher than later interest payments).
Read on for rationale on design decisions and more specifics.
First I want to explain how the Backer Rewards contracts works, as currently written, and why it’s like that.
At the highest level, the contracts were designed with the idea that Backers should be rewarded only when their pools pay back because this most closely aligns incentives between the Backers and the protocol. For example, we don’t want Backers to YOLO into a Borrower Pool, take some token rewards early on, and then not have to care what happens to their underlying USDC. That would encourage excessive risk taking.
But why use interest payments? You might consider rewarding Backers by time (eg. tokens per second). But using time has several drawbacks. The total amount borrowed could be flat for weeks and then double over night when a big pool is announced, and paying the same reward rate through all that doesn’t make sense.
Or you could do rewards strictly by the amount of money put in. But blindly looking at contributions doesn’t factor in risk or value to the protocol. Imagine putting $10k into a Borrower Pool that pays 1% interest vs. $10k into a pool that pay 15% interest. Which Backer should get more rewards?
Hence having rewards be based on the interest that gets paid back cleanly deals with many of these drawbacks, and is the most straightforwardly aligned with the interests of the protocol. It only gives rewards when Backers have proven they chose a successful Borrower, and it encourages them to choose pools that balance being safe with generating high amounts of interest.
Once you’ve settled on that, some other key questions are “how much should the rewards be”? And “How to compensate early Backers vs. later Backers?”
As a general rule, this design assumes that on-going rewards should balance incentivizing people while not encouraging excessive risk-taking. So you want them to be just enough to incentivize participation without being enough for people to just ape in.
Further, the design follows the idea that early Backers should be rewarded more than later Backers, because the protocol is higher risk early on and still needs to prove it’s track record. Early Backers (meaning those that are contributing to earlier pools) should be compensated for that higher risk. So that means the first dollar of interest paid back is worth more than the millionth dollar which is worth more than the billionth dollar.
With that context, we propose the following specific parameters.