mango-docs/litepaper.md

description
This is a rough draft paper describing the Mango objectives and tentative plans

Litepaper

The Mango Vision

Mango intends to merge the liquidity and usability of CeFi with the permissionless innovation of DeFi at a lower cost to the end user than both currently provide. Towards this goal, Mango offers margin trading and perpetual futures along with decentralized governance to decide the future evolution. In the medium-term, the goal to rival centralized exchanges in trading volume is ambitious, but we see no substantial impediments for Mango Markets. In the long run, we believe a permissionless ecosystem will produce spectacular, outlandish and unpredictable innovations which will overtake centralized finance.

Why Mango Markets Will Succeed

The key ingredients for this vision are low latency, low transaction cost and full decentralization. We believe all three ingredients are necessary for the project to be viable and all three ingredients are finally available on Solana.

Latency

The Solana blockchain provides block times of roughly one second. Although one second is still noticeable, Solana's intended 400ms latency target approaches the limit of human perception.The obvious benefit of low latency is usability—most people get anxious waiting on the status of their transaction. But while reducing user anxiety is important, there is another, oft-ignored benefit of low latency: better liquidity. Liquidity providers' quote spreads are directly proportional to the time required to change the quote. The longer it takes to change a price quote, the larger the risk of significant market movement to the market maker and the wider his quotes must be. At the current one second latency, we believe the raw bid-ask spread can be competitive with centralized exchanges.

Transaction Costs

Low transaction costs are arguably the raison d'etre of finance. We believe a financial innovation must lower transaction costs to be a full improvement. Therefore, the cost per transaction on all Mango financial tools will be comparable to or lower than the costs in CeFi. We believe this must be true. Lower costs indicate efficiency and more efficient protocols and tools tend to win in the long run. It is not possible to escape fees—service providers (e.g. liquidators, insurance fund, developers) on Mango protocols must be compensated. However, Mango will err on the side of lower fees.

Decentralization

Trying to achieve competitive latency and cost by centralizing key components (e.g. the orderbook) will fail in the long run. Centralizing any component is a security risk and severely harms composability. Ultimately, the centralizer decides how other apps may interact with the centralized component and the centralizer neither has the incentive nor the bandwidth to allow all interested parties to participate in the improvements. As a result, centralizing key components gives up the immense upsides of permissionless innovation. Mango Markets will retain the upside.

No Presales, Decentralized Governance

The Mango Token will govern the protocol. The vast majority of MANGO will be locked in the DAO treasury to be distributed according to token holder wishes. That being said, our vision is that governance ought to reward the people who provide protocol services (e.g. liquidity providers, oracles) and the people who build new protocol services (e.g. developers, or other contributors). The commitment to distribute the largest portion of the DAO’s power and wealth to future contributors will encourage the most skilled and ambitious builders to join us. Finally, in accordance with the crypto ethos of transparency and equal access there will be no presale of tokens.

Ultimately, Mango intends to win the long game in financial services. Low latency makes our tools usable. Rock bottom fees makes Mango hard to compete against. Decentralization makes Mango hard to kill through centralized incompetence or malevolence. Open governance that allocates power and wealth liberally to builders, will attract the best people to build and govern the protocol. Finally, the permissionless nature will allow the millions of tiny experiments to take place that yield the life changing innovations. We're motivated and driven to build Mango according to this vision, and we hope you'll join us.

Margin Protocol

Highlights

• No fees
• Cross margined
• Allows limit orders on margin
• Margin positions pay interest
• Pooled SRM for fee reductions

Design

The initial margin protocol adds a borrowing and lending layer on top of the Serum Dex v3. The user owns a margin account which is then associated with a serum dex open orders account for each market in the group. The user may deposit any of the tokens included in the group and its value in the quote currency (typically USDT) is calculated using an oracle. This value is then used to determine how much a user may borrow. Since positions gained from margin trading are also treated as deposits, the user may take up to 5x leverage. Mango Markets does not charge any fees.

Risks

Negative equity accounts and socialization of losses

If the price of a user's collateral falls fast enough or the price of the borrowed tokens increases fast enough, the account may fall below 100% collateral ratio without a liquidator taking the position. In this case, funds are pulled from lending pools to bring the account to 101% coll. ratio. These losses are socialized across all lenders which may trigger a liquidation cascade if most lenders are also very close to being liquidated.

Oracle error

The Solana Flux Aggregator is also brand new in Solana and may have errors in the code. The centralized exchanges feeding the price may also have errors. Since all oracle price publishers are looking at the same centralized exchanges, price errors will affect all of them. This could trigger bad liquidations.

Illiquid deposits

There is a chance the user is not able to withdraw deposits because it's borrowed. This could be a problem for someone who needs liquidity immediately. The same issue applies to positions that are lent out—the user may not be able to close a position if the utilization rate is 100%. However, the protocol guarantees a 100% APY while the user waits to be able to withdraw.

Smart contract exploit

The code has been looked over by volunteers, but there has not been a formal audit. While there are bounties offered for responsible disclosure of potential vulnerabilities, there is no guarantee that hackers will choose the bounty over a profitable exploit.

Perpetual Futures (WIP)

Highlights

• Most liquid perp on Solana
• Cross margin with Serum dex to enable easy hedging on spot
• Simple and familiar UI
• On chain CLOB
• Funding rate as a function of mark price and index price. Funding paid continuously

Token Distribution

After the initial distribution of tokens, only the DAO may distribute more tokens via governance proposals. The intent is for Mango Token to be distributed liberally to protocol builders, liquidity providers and project contributors in a fully transparent way.\

Max Supply: 10,000,000,000

Initial Circulating Supply: 1,000,000,000\

---

Creators - 5%

For the work of creating Mango Markets.\

Insurance Fund Sale - 5%

There will be a sale of Mango Token that goes directly into the DAO treasury for use as the insurance fund. The insurance fund will pay Mango Perps smart contract in the event extreme volatility causes bankrupt accounts and excess losses in the system. The sale mechanism is described in detail below.\

DAO - 90%

These tokens can only be unlocked via token holder governance. We propose that tokens are distributed roughly on a logarithmic supply model similar to Bitcoin but with a halving every two years i.e. 50% of the tokens distributed in the first two years, 75% in the first four years, 87.5% in the first 6 years and so on. The DAO is not bound to follow this proposal.

Token Sale

Mechanism

Total Sale Period: 48 hours

• Unrestricted Period: 24 hours (first half)
• Withdraw Only Period: 24 hours (second half)

The Mango Sale will happen over two days. The Sale contract will have two vaults, one with 500,000,000 Mango Token and the other vault with 0 USDC. The 2 days will be split into periods: Unrestricted and Withdraw Only. During the Unrestricted period any Solana user may deposit or withdraw their USDC from the Sale vault. The Unrestricted Period ends in 24 hours and transitions into Withdraw Only, where buyers may only withdraw their USDC. At the end of the Sale, all USDC depositors will get Mango Token in proportion to their share of the USDC vault and the USDC will be moved into the Mango DAO treasury.

Explanation

Buyers get a pro-rata price at the end and every participant gets the same price for Mango Tokens.

The sale price per token can be calculated with:

(total USDC in vault) / (500,000,000 Mango Token)

This price will fluctuate during the entire 48 hour period as USDC is deposited and withdrawn. However, during the last 24 hours, buyers may only withdraw USDC and so the Sale price can only decrease during that period. The Withdraw Only Period was added so that if a buyer deposits a large amount of USDC at the last second of the Unrestricted Period and raises the price significantly, other buyers still have an opportunity to get out of the sale.

We take the view that token sales should be simple, fully transparent and minimize randomness and luck in the distribution. Simple mechanisms are easier to build, explain, understand and are harder to exploit. A transparent mechanism increases participation because buyers are more confident there are no hidden tricks that could harm them. Elements of luck engineered into the mechanism distribute value randomly or to those who are most willing to do the arbitrary, worthless tasks to get the free value. We believe all "excess" value should be captured by token holders in the DAO. Removing the casino aspects of token sales selectively encourages participants who are focused on contributing and governing in the long term.

Other mechanisms we've looked at don't live up to this standard. For example a sale on a bonding curve will give free value to those who buy a few milliseconds before others or have the good fortune of their transaction being accepted by validators before others. A capped sale of tokens with a fixed price has similar issues. A sale of tokens to market makers who sell the tokens on the open market is not fully transparent and the agreement between the market maker and the dev team may be quite complex. There are various "fair" launch mechanisms that are transparent and minimize luck, but ultimately end up being very complex (e.g. Mesa Offering) and difficult to build.

Source Code

Mango developers, including those from the community, will build the smart contracts and the user interface. All the code used will be open source and remain free and open source for future projects to use as they wish.

Governance

The Mango Token is a governance token, first and foremost. Collectively, the token holders have the power to upgrade the protocol as they see fit, only constrained by the checks-and-balances of the DAO. This allows token holders to create incentives to reward participation and drive usage of the protocol.

We consider the COMP governance model to be a gold-standard in the industry and followed it to build out the Mango Token governance model. We mainly reworked the initial token distribution, to facilitate a DAO inception that is truly open and adapted it’s mechanics to fit the solana programming model.

Anybody with 0.1% of Mango Token staked can propose a governance action; these are simple or complex sets of actions, such as adding support for a new asset, changing an asset’s collateral factor, changing a market’s interest rate model, or changing any other parameter or variable of the protocol that the current administrator can modify.

Proposals are executable code, not suggestions for a team or foundation to implement. They are not limited to the governance protocol itself, but extend to all protocols created by the Mango Token DAO including the Margin and Perp protocols contributed by the creators at launch.

All proposals are subject to a 3 day voting period, and any Mango Token staker can vote for or against the proposal. If a majority, and at least 2% of the total Mango Token supply are cast for the proposal, it is queued in the Timelock, and can be implemented after 2 days.

Project Status & Roadmap

Mango Margin "closed alpha"

• first weeks of march
• enforces strict borrow limits for every account
• liquidator and solana program remain closed source
• 3rd party liquidators begin implementing their strategies

Mango Margin "public beta"

• begins mid-march 2021 and runs for multiple months
• borrow limits will be replaced with the introduction of partial liquidations
• margin trading interface largely improved based on feedback during alpha phase
• additional trading pairs will be released
• step-wise open source releases as independent reviews are finishing

Mango DAO Launch

• On-chain voting mechanism based on spl governance
• MNGO Token sale to create USDC governance fund

Mango Perp "closed alpha"

• Begins mid August 2021 and runs for a few weeks
• Market makers can integrate and test

Mango Perp "public beta"

• Fully cross-margined perpetual swaps and margin trading
• Largely increases the number of spot markets that can be traded
• Perpetual swaps will feature lower trading fees then spot markets

---

Edits

July 30th, 2021

It turned out that on-chain governance was earlier possible on Solana then expected, so we decided to make the move before launching Mango Perp. This made the Temporary Governance Fund unnecessary. As a result of that the float at token sale will only be 10% and not 20%.

As we progressed in the development more details of the capabilities of Mango Perp's features have become concrete and added to the roadmap.