Launch Virtual xy=k DBR Auction

Inverse Finance DAO
1

Proposal to Launch Virtual xy=k DBR Auction

Summary

This proposal outlines the launch of the Virtual xy=k DBR Auction within the Inverse Finance DAO. Inspired by Uniswap’s xy=k constant product bonding curve formula, this innovative auction aims to facilitate the purchase of DBR using DOLA, directly contributing to the reduction of DOLA bad debt on Frontier and enhancing DBR liquidity in the market.

Background

Currently, DOLA bad debt stands at 7,618,162, with 4,684,126 DOLA repaid to date (2,584,428 net DOLA reduction made in 2023). The most up-to-date view of the DAO’s bad debt can be viewed on the Transparency Portal. In order to de-risk DOLA and the DAO’s operations, it is crucial that progress in reducing DOLA bad debt continues, and is currently the top priority.

Objective

The virtual xy=k auction is an innovation being brought to market first within the Inverse Finance DAO, inspired by Uniswap’s xy=k constant product bonding curve formula. An in-depth description of the primitive can be found in this medium article. For this iteration, the virtual auction will permissionlessly allow anyone to buy DBR using DOLA (and only this trade). In the auction, the price of DBR (per DOLA) will continuously reduce every second, until a DBR purchase is made at which point the price will increase. It is expected that MEV bots will monitor the auction, executing arbitrage transactions between it and the TriDBR pool on curve as soon as it becomes profitable. A front end will be made available to make interaction with the auction easier.

The auction contract includes the sendToSaleHandler() function, that anyone can call. This moves accumulated DOLA proceeds to the saleHandler contract, which currently includes logic that uses the DOLA to repay the DOLA borrowBalance of either 0xf508c58ce37ce40a40997C715075172691F92e2D or 0xeA0c959BBb7476DDD6cD4204bDee82b790AA1562 (the 2 Frontier exploiters of 2022). The logic of the saleHandler contract can be updated by governance in the future, for example, to send the DOLA proceeds to the Treasury.

The DAO sets a yearly max budget of DBR to be used by the auction, with an operator (the Fed Chair multisig) able to set the DbrRatePerYear between 0 and this governance set max.

In addition to this becoming a powerful (automated) tool for repayment of DOLA bad debt, the auction also has the additional benefit of increasing the current depth of DBR liquidity on the market for buyers. This will improve the user experience for borrowers on FiRM, particularly ones with much larger positions who want to purchase DBR for longer-duration loans.

Impact on Alternative DBR Issuance Methods

Given that this initiative leads to the creation and sale of new DBRs in the market, it’s crucial to offset its effect against other DBR issuance methods on a one-to-one basis. Currently, the sole other functioning mechanism involves the streaming of DBRs to INV stakers on the FiRM platform. Consequently, the annual DBR streaming rate for INV stakers needs to be adjusted downwards in direct proportion to the annual rate employed in the DBR auction.

Parameter Considerations

When setting the virtual reserves for DOLA and DBR in the auction, it’s essential to strike a balance that optimizes trade execution and minimizes arbitrage profit. This decision is pivotal as it directly influences the auction’s efficiency and functionality. The following considerations are recommended for determining the optimal depth:

  1. Optimal Trade Execution: The virtual reserves should be sized to facilitate efficient trade execution, even for substantial buy orders. This approach aims to align the DAO’s realized price with the prevailing market rate in the TriDBR pool as closely as possible. When the depth is insufficient, a significant price disparity for profitable arbitrage might arise, especially after a sizable market DBR purchase that rapidly elevates the market price.
  2. Minimizing Arbitrage Profits: The depth should be calibrated to ensure arbitrage opportunities are limited primarily to covering transaction costs and a nominal profit margin. Overly generous arbitrage opportunities, resulting from excessive depth, could lead to disproportionate profits to arb bots following significant market movements, which are lossy to the DAO.
  3. Controlled Price Reduction Dynamics: It is crucial to monitor and control the rate at which DBR’s price decreases in the auction. An overly rapid price decline could lead to significant spreads between the auction and market prices during times of high network congestion (elevated gas costs).

Based on this, we recommend starting with initial virtual reserves set at:

DBR Reserves = 1,550,000
DOLA Reserves = 232,500

We recommend starting at an initial rate of 2,000,000 DBR per year, with a max rate (the maximum that Fed Chair can set the DBR rate to) of 5,000,000 per year.

(Data analysis and visualization done using Python)

These initial values are proposed as a starting point. Continuous monitoring and periodic adjustments will be necessary to ensure the auction’s objectives are consistently met, especially after changes in the depth of TriDBR. Data-driven adjustments, informed by auction performance and market conditions, should be part of the ongoing management strategy. Additionally, engaging with the community for feedback and insights can further refine these parameters, ensuring they remain aligned with the DAO’s evolving needs and market dynamics.

For 3d visualizations using plotly, use the Jupyter Notebook link here.

Deployed Contracts

DBR Auction: 0x933cBE81313d9dD523dF6dC9B899A7AF8Ba073e3
Sales Handler: 0xB4497A7351e4915182b3E577B3A2f411FA66b27f
Auction Helper: 0xC7D5E6FA4D5B4b4A82b14a256008DAfAF5232ADb

On-Chain Actions

  • Add DBR auction as a DBR minter
  • setMaxDbrRatePerYear to 5,000,000
  • setDbrRatePeryear to 2,000,000
  • overrideReserve to 1,550,000 DBR and 232,500 DOLA
3 Likes
2

Fully support. Xy=k auctions are essential for sDOLA and bad DOLA debt repayments, but also represent an important innovation that will likely be emulated throughout DeFi.

3

Awesome. Going to have to read through another couple times to fully grasp and reply but overall great work.

Would suggest to direct a higher DBR share to xyk bonds in the initial period in lieu of raised or even maintained rates to INV stakers.

Can always reduce the amount. But if this thing works very efficiently would rather start out with a big clip of ammo

1 Like
4

Thanks for the feedback @mason. The idea with the current proposed rates (2m/yr start rate with flexibility to move up to 5m/yr via Fed Chair) is to keep the proposal relatively uncontroversial to increase the likelihood of passing which is crucial for the initial launch. It’s important to note that this max rate (5m/yr), which Fed Chair can set freely, will already cut DBR yield to INV stakers in half as things stand; it’s not a small amount. Furthermore, we need to consider that sDOLA’s launch will be shared shortly, which equally will contribute to DBR issuance (and hence, taking from INV staker allocation), although in this case, the design is that it adds more lending capacity to FiRM (leading to a higher global DBR issuance rate) – more details on this later though!

In my opinion, the proposed rates for the initial launch are appropriately calibrated. They strike a balance between being impactful and maintaining broad support within the DAO community (I hope). More aggressive rates, while potentially more efficient in the short term, could be more controversial among INV voters. I believe reserving such changes for subsequent proposals would be more strategic once we have data and feedback from the initial implementation.

1 Like