Introduction

In order to access credit, currently you have to deposit between 133% — 150% of the value you intend to borrow. This works fine, if you are already crypto rich, or for speculative use cases like leverage or shorting an asset, but it’s not very capital efficient.

So I propose a design for a credit union on the blockchain. In which members stake funds. Receive token shares on a bonding curve, the Union Contract receives Interest, and that interest is then lent out unsecured or under-collateralized.

Major Components

It consists of:

1) A DAO, let’s call it Union, that is responsible for creating and managing credit vehicles, include setting rates and membership conditions.

2) A Staking Mechanism that takes in DAI or similar asset and then for a period of time deposits it into a yield generating protocol like meta’s maxDAI and deposits the U. In return emits a UnionToken.

3)The Lending pool holds the DAO’s funds to be lent out. It also tracks funds as they are paid back to ensure that the DAO only ever pays a dividend when it’s profitable.

4) Credit vehicles are created by a factory. Each vehicle owns a percent of the Lending Pool, has a max loan size, collateral requirement, collateral type. And is responsible for minting and burning loan NFT’s.

Examples of potential Vehicles:
* [50%]: Fully backed: deposit 100 dai => earn union & borrow 100 dai @ 10% fixed a year.
* [20 %]: Half backed: deposit 100 dai => earn union / borrow 200 dai => 20% a year.
* [10%]: Unbacked: borrow {max loan size} dai @ 30%
* [20%]: Untouched: Ensures the lending pool is always growing.

Membership Process

In order to stake dai or take out credit you must first be a member of the dao. Membership is linked to a single address.

[Note: If an account is compromised it can cause problems and so it will be necessary to explore how the system clears a compromised account, perhaps a `delete_membership` method]

Membership application

1. You submit an application by buying 1 share via uniswap.
2. each union has the option for implementing pre-filter limits such as for a kyc filter or only holders of an Amex Black Card Kudos.
3. You must be staked by at least 2(set by dao) members in a bid/reveal mechanism ie if 5 bid then 5 bid you. [bid/reveal = soft hide, it’s good enough if most people aren’t crawling the chain.]
4. Staking adds someone to your web of trust and so should only be done if you trust them. You can’t unbid someone for some number of epochs (3?)
5. If < 2 bids then your share is burnt and you must re-submit.

Once you become a member you get a number of Membership Benefits:

• Allowed to stake and borrow
• Voting on things like rates,
• Quarterly Dividends: If the DAO is profitable that epoch paid back > lent. Dividend allowed up to some % of profit to ensure the pool grows
• Propose a new vehicle.
• Governance only takes place each epoch (month or quarter, slow banking).

Staking DAI for UnionShares & to fund the lending pool

Once you become a member, you are then able to stake dai or an equivalent token in order to gain membership shares. In return for your stake you receive a share of the Union, distributed on a continuous basis.

Each epoch UnionShares are minted proportional to the quantity of Dai staked. The number of Shares minted will be determined on a bonding curve similar to Livepeer. The purpose of the bonding curve is to incentivize a minimum viable amount of stake e.g. Each Epoch with less than say $1,000,000 on stake the # of UnionShares emitted increase.

Your Staked Dai also serves as your available collateral for borrowing.

As long as you are staking, your stake is sent to a money market (e.g. MetaMoneyMarket) to earn interest. This interest is then deposited into the Lending Pool.

[SideNote: this is not ‘risk free investing’ as there is unknown smart contract risk, I am exploring how to hedge this risk with ‘defi insurance’, but there will likely always be some risk]

Borrowing from the Union

As a member of the Union you also have access to the credit vehicles. Your stake serves as your initial collateral base. And then you may also stake dai or metaDai if you want to increase your collateral without forgoing the variable rate.

Example: If you chose to borrow from the 50% collateral vehicle described above and had 1000DAI staked you would be able to borrow up to 2000 DAI at the union fixed rate. After taking out the loan your stake gets locked in and you then start paying it back on an epoch basis (or at the terms set by the dao).

If the loan was fully unsecured then you would be able to borrow up to the max with 0 dai staked. And if you were a rascal you could simply run off with the funds.

Web of Trust

The big question with under-collateralized loans is how do you stop the 1 not nice person from exploiting the system and running off with all of the funds. The first method is simply charging for that risk. If you have 10 people borrowing $1,000 and you charge 33%. As long as less than 3.3 people default you come out ahead.

In crypto where identity is less known, those 10 people could really just be one person. So we propose an additional step, securing it with a web of trust.

1. Requires 2 (3?) people to vouch for you to enter the DAO.
2. When in the DAO you have access to collateralized debt equal to your stake, and unsecured debt equal to the max the dao allows.
3. If you miss a payment you become a member in bad standing:
   — you don’t get shares for staking in any epoch you’re in bad standing.
   — If you miss 2, your Web of Trust gets withdrawals locked and stops earning shares until it’s paid back.
4. You don’t get kicked for bad debt. The goal is to incentivize your web of trust (aka friends & family) to help keep you in good standing.
5. You can get kicked for bad behavior though.

Note: It might also be necessary to give a reward for staking others so that people are not just disincentivized to stake bad people but also incentivized to stake and refer good people.

Interest Pool build up

The size of the credit pool each epoch is equal to the interest earned last epoch. This means that if no one is borrowing the Unions principal balance will grow.

Banking Dividends

If the members of the Union do a good job and the amount paid back is greater than the amount lent out, they may vote to issue a dividend up to some percent of the profit. This scenario is likely not going to take place until the union has been operational for a while and there’s a steady stream of payments or if there’s a number of loans that are being taken out and paid back within the timeframe of the epoch.

The mechanism in which the dividend is paid out could be by deposit into one side of an AMM, a Share Buy & Burn, or make the funds claimable as an actual dividend. On top of the gas efficiency and regulatory considerations, you’d also want to model out which mechanism most incentivizes staking behavior.

Some initial thoughts:

• Adding to one side of the AMM seems like it would really only benefit arbitrageurs who could sandwhich the deposit with an ‘Add Liquidity’ then ‘Withdraw Liquidity’ transaction.
• Buy & Burn helps create a base taker volume for the shares market, incentivizes liquidity providers and distributes value equally with a single transaction. But would leak value to Dai:ETH liquidity providers as the contract uses Dai to buy UNION (on uniswap).
• Traditional dividend is a lot more inefficient from a gas perspective but members would be able to realize gains without selling shares. And they could choose how to spend their Dai ie staking, buying more UNION, or something else entirely.

I prefer the dividend method, but gas prices being what they are the buy and burn method would probably be best. What do you think?

Does the math check out?

Kinda.

If you have 100 people staking $1,000 at 10% apr that would only yield ~$2,500/quarter in potential loans from the interest. This might not seem like a good start but it would allow for the Union to establish itself without risking principal to start, and as long as it can keep default rates low and loan volumes up it should earn a rate higher than the underlying money market. Not to mention there are a lot of people for whom, just a $100 loan is all they need to get their head back above water.

It’s important to couch expectations of a sustainable and fair implementation of under collateralized lending. But if all we end up doing with this is creating a sustainable charity that undercuts payday loan centers, I’ll be happy.

Initial Pool

All that said because it is essential to have a decent amount of principal in order to be able to offer loans worth taking out, it will likely be necessary to sell some initial UnionShares for a price rather than stake, in order to prime the Lending Pool and ensure governance has some training wheels (if you’ll allow a mixed metaphor).

If able to raise a small seed round of $500,000, at an 8% apr that would mean $10,000 in loans / quarter / perpetuity.

This could also create initial liquidity for an AMM. Which could be used to incentivize staking. If the initial price was set correctly, the first years worth of stakers could earn more in value staking Dai for Union Shares than staking directly with a money market. Of course, this should be considered “marketing spend” not a sustainable system cause if the Union doesn’t ever turn profitable then this would just turn into a donation of ETH to those initial stakers.

Possible, not perfect.

If you have any thoughts as to how this system could be exploited or improved, or if you’d like to help make it happen @ me on twitter.

[Update: We ended up building Union!]

Further Reading:

• An introduction to bonding curves, shapes and use cases
• ERC-2212: Interest Earning Stakes
• ROSCA’s: https://en.wikipedia.org/wiki/Rotating_savings_and_credit_association
• Chit Funds: https://en.wikipedia.org/wiki/Chit_fund
• Saving Strategies in the absence of banking: https://www.cgdev.org/blog/mindful-saving-tanzania-strategies-women-use