Thread by @AlexMasmej

[THREAD] Last month I started a Telegram group around what I firmly believe will be the next huge #DeFi trend: undercollaterization.

Here are my takeaways and pieces from the brilliant people who took part in the discussion:

First, context. DeFi only leverages overcollaterization because it’s the easiest enforcement: you can be 100% sure one will repay risky/volatile positions (smart contract risk aside)

But the world isn’t made of rich Compound whales. Actually, undercollaterization is everywhere.

Whether it’s social credit rating, or fractional reserve banking, loans, it’s all undercollateralized.

Yes, it’s prone to mistakes, but this time we can make it *transparent* and permissionless (which means more liquid for everyone). It’s healthier

Let’s dive in: (17 tweets)

1. DeFiner.org & Transwis.cash are early solutions but are clunky: they don’t operate permisionlessly on Ethereum, hence suffer from bad or “traditional banking” UX. Loans are still financially collateralized. They’re a transition, not the next killer app.

2. Debt is such a huge market.

In fact, “the market” is debt, because it’s all loaned by central banks. Funds are then lent to commercial banks. Great @EFDevcon talk by @brewster_kahle

3. We need new blockchain enforcements other than financial:

- social (Your friends pay/lock funds for you?)
- legal? But that would need identity👇
“There’s a slippery slope between identity and decentralization.”

Censorship resistance cannot be implemented with known actors.

4. Enable.credit is a great first step to enable DAI loans, hats off to @onggunhao.

But it’s ISA based and not only it’s high trust, but it also has very long cycles in period where speculation isn’t as exciting, esp for college students. Fluctuation with grades maybe?

5. Since staking at a fixed interest guarantees income, you could lower collateralization ratio by collateralizing fixed future known yield? @maplefinance is trying that.

6. Trustlines by @pepae is dope and a great fully fledged platform and app, to be leveraged by people you *already* trust.

But the system is complex and the app isn’t mainstream yet.

7. Multisig Trust: how much money do you trust your friend/family for? A 2-person multisig quantifies that number and can enable a decentralized trust system.

Fascinating paper on it eprint.iacr.org/2017/156.pdf

8. @danfinlay shares similar points: link.medium.com/PcNkK6hF71 The problem becomes: are we going to maximize trust for socially apt people then?

It might actually be great to optimize for people with most trusted relationships, but are connectors necessarily the most trustworthy?

9. Union, a new project by @JacobShiach, is attempting to do a Union / Revenue-Generating DAO and their model is really cool, check it out here: medium.com/primeradiant/a…

9. Maybe Web2 ID like Instagram or Facebook are better for Sybil-resilient than brand new systems, because we have so much more past data.

Soon: crypto loans for 5-star AirBnb hosts?

10. World Coin is apocalyptic for most people. It’s a stealth project that uses a retina scanner by Mark Zuckerberg’s friends, in order to solve the dID problem.

11. The problem with such solutions is that identity should stay hidden at all times or censorship resistance disappears.

ZKPs could be a solution to solve undercollaterization Sybil-resilience without showing public ID data.

12. ERC-20 loans MVPs by myself and @pet3rpan_ are pretty cool. My debt was tokenized because he bet on me, I was highly speculative but I hope I come across as reassuring now ❤️

People buying the token pay Peter, who gets money before I even need to repay.

12. @tryrollhq goes the step further in terms of personal tokens by trying to extend personal time for cash. $ALEX, $DUDE and $PETER were created for that purpose. New ISAs can be done via this, since you hold a person’s asset.

3. An alternative way of undercollaterization: start with the worst credit score, level 0, and gamify the reputation with rewards for certain behaviors? A cool idea, sounds like Pepper4D for loans (imagined by @eric_rsno)

15. Credit lines to establish trustworthiness: personal.ceu.hu/staff/Adam_Sze… (read the abstract at least)

16. No more overcollateralization will be needed when less-volatile assets will be added, like RealT tokens on Uniswap for real estate or taking your house as a collateral in the future.

17. Staking reputation to get lower interests rates on loans? That could incentivize top borrowers to always repay on time.

The more reputation you put at stake, the less you pay. This again has to be supported by a great dID solution that doesn’t reveal your ID to the public.

That’s it! You liked it? Wanna voice your take? You deserve it, join: t.me/joinchat/L20Xe…