Financial Proposal: Chelo finance - Credit scoring and real-world testing proposal

Chelo Funding Request: Credit Scoring

Proposal Summary

Chelo funding request $48k USDC to fund a 16 week stakeholder driven development to create an on-chain credit rating.

Proposal description

One of the main obstacles that Web3 faces right now is the inability to capitalize on the value that on-chain reputation has for the owner of a wallet.

We propose combining a suite of already available tooling to solve real world challenges for unbanked individuals to address challenges including loan granting by DAOs.

DAOs are much more than the “new limited companies”; the community bonding, aligned purpose and ability to issue their own tokens give them a structural power that has not been seen before. We want to help them to unleash that power in a smart way.

Most DAOs have idle treasuries, why not gain financial interest with tools that help build a community?

DAOs should be a two-way street, where they have the ability to engage with products to help their community thrive. Financial products for their community are an obvious use of their resources, but in order to use them in a smart way, we need to provide them with risk assessment and rating tools, as well as plug-and-play protocols for deployment. We are solving for that.

Also many proposals do not integrate real-world stakeholders in to the process, and suffer from overly-idealistic considerations. In this proposal we specifically integrate low-income bankless or under-banked stakeholders who will be able to leverage these tools from a need-based perspective, vs a web3 enthusiast perspective. We believe that closely integrating this process - which necessarily includes legal advice as well as technical advice to support the users - will lead to both lateral adoption through boilerplate mechanisms, as well as vertical integration through the value of real-world adoption.

This proposal has the potential to significantly grow the number of active DAOs on the Aragon Network by establishing a model to serve unions and communities and supports the overall purpose of the DAO defined in the Aragon Network Charter.

Proposal

1.Rationale

This proposal attempts to address a critical issue – on-chain credit score – by focusing on the needs of real-world users for non-collateralized loans, using their on-chain activity as credit score.

The global personal loan market size was valued at $47.79 billion in 2020 and is projected to reach $719.31 by 2030, growing at a CAGR of 31.7% from 2021 to 2030. In the current space of Web3 it is impossible (or just very hard) to get non-collateralized loans. For this to be possible, it is necessary to have a credit-scoring system to allow smart contracts and DAOs to decide whether to lend or not.

  1. Objectives

We propose to provide the following solutions to:

  1. On chain credit-score: Data processing of loans given to users to get valuable information about their probability of paying loans. This information should be available on-chain through smart contracts.

  2. Secure access on chain to this credit-score: Develop ways to access this credit-score on chain through smart-contracts and/or oracles securely. The process of updating a user’s credit-score must also be secure and tamper-proof.

  3. Trustful data: Create open and reliable information about web3 users’ credit scores. This information should be accessible and verifiable at all times.

Tools Applied & Strategic Ecosystem Alignment:

  • Gitcoin Passport & BrightID: Creates anonymous, decentralized, and unique identity with strong anti-sybil security for web3 users. This will be used to differentiate user’s credit score and provides a proof-of-concept for future Decentralized ID (DID) work
  • EVM CRISPR: Tool to create transactions with Aragon DAOs
  • Kiva.org: Has public data snapshots about loans created with their product.
  • Chainlink: Their oracle system may be used to bring off chain computations on chain
    *Swollet: A learn-to-earn software to help participants earn their credit scores

Achieved to date:

  • Lending protocol to create non-collateralized loans through Aragon and Gnosis safe (Snapshot) based DAOs
  • Gitcoin passport: has been integrated with over 20 “stamp” providers and is a WC3 conformant DID solution
  • Notebooklabs.xyz has already the fractionalized KYC identity on a Demo stage
  • Chainlink oracles: currently, we have a basic example of how the off-chain/on-chain data transfer should work.
  • Swollet.com : Has app in Google and iOS store- content can be added and tokenized results are available
  1. Proposed Deliverables and duration
Deliverable Description # Weeks Cost
Gitcoin Passport integration with Aragon DID integration such that a display for users is available 4 8k
Notebooklabs.xyz Gitcoin Passport integration modified UI + smart contract Work with identity management in our contracts and interface 4 5k
Data for ML modeling de-identified data from raincard.xyz or other off-ramp tool such as nucleus Aragon dTech will ensure that any off-ramp data is available for open data science review after it is sufficiently de-identified so as not to cause any risk to users. 4 0
Chainlink (or any oracle to bring data on-chain) chainlink integration Use Chainlink (or any oracle service in case after testing chainlink isn’t suitable for our work) to allow our contracts to use user’s credit-score 3 5k
Contracts auditing audited contract Before getting into production, it’s necessary to audit our system. The cost and time depends on the audit firm and final system’s complexity. The auditor firm will be chosen on the best cost / firm reliability. 3-4 (est.) [15]k (est.)
User Testing funded through a learn-to-earn model. 5 classes with 10 users To address significant trust issues we propose rewarding users in a learn-to-earn project using Swollet.com and training provided by the Piedmont Housing Alliance to test our system and provide qualified feedback. This will underpin proof of concept, improve product and help establish trust. This user testing will also provide real-world loan data for our ml model. 4 5k
LexDAO Clinic’s “House of Bestape” review and support Min. 2 meetings LexDAO’s House of Bestape will provide email support to help users of the Chelo product in the real world as well as host 2-3 open meetings to engage with the Chelo finance users 8 5k
  1. Working Team

Gerardo Lemus | Data science TradFi & Chelo Team

Jose Valdovinos | Business management Chelo Team

Rafael Contreras | Full stack dev & Chelo Team

Grant Schneider | Upstart VP of ML modeling

Aragon dTech | Technical support, documentation, community engagement liaisons and stewards

LexDAO Clinic’s House of Bestape | Legal support, community engagement, liaisons and stewards

DhruvM#6418| Notebook Labs

Felipe Novaes Rocha#2364 | 1Hive dev UI/UX

billyjitsu#3865 | Lead Dev 1Hive NFBeez & Solidity dev

Dave Norris| Piedmont Housing Alliance

Lefteris Laliotis|Swollet CEO

  1. Limitations and Risks
  • Low user response: Users that want to stay anonymous won’t use Notebook human verification
  • Technical barriers may lower director engagement. To reduce this, we will refer to a “mid-knowledge” user.
  • Lending contracts and future modifications need to be audited
  • Single point of failure on the credit score server. We may evaluate a decentralized system (like IPFS, or storj) or qri.io for storage depending on security considerations.
  • Some users and DAOs might prefer a more decentralized credit score system
  1. Funding Request

We are asking for a total of $48k which will be used:

  • 18k for code integration and boilerplate development
  • 10k for Education & Product testing
  • 10-15k auditing
4 Likes

This is really in the trenches work here, that starting line of innovation in an industry. One thing I think token holders would want to see is an MVP of the use of this product or at least a mission to an MVP: What organization or team might trial it out? Any expressed interest from Kiva?

Additionally, not sure if it belongs in this proposal and I am not legal but I am sure there are tax implications for lending and receiving loans. Maybe reporting requirements like Nexo and blockfi have had to adhere to. Would any of this link to real world ID or be required to before it was put into practice?

Overall… this is awesome stuff man. Can’t wait to shadow you on this.

1 Like

Solving for real-world needs is something I strongly support. This is a team with specialist expertise and I believe cross-sector collaboration is a leading means to grow the web3 and DAO ecosystems. Plus, I’m a fan of Gitcoin passport, which leverages existing online presence and imo is both simple for users and effective verification.

Credit scoring is not new though and from the proposal, I can not recognise what the team identifies as their strengths, opportunities and other qualities that effectively differentiate Chelo from substitutes. Is it the offer of non-collateralized loans?

Can you provide a summary of how Chelo compares to and differentiates from the following teams?

Thank you for asking!

One specific aim of Chelo is to mimic FICO scores - these are essential for many people - and in the United States a low FICO credit score may prevent an otherwise financially stable person from being able to access affordable housing.

By focusing on an end goal that fits in to existing business models Chelo provides a value that is more specific to our needs.

Another advantage of Chelo is that they are already leveraging Aragon technology and thus are bringing more open source in to our native space = bringing features that other developers may be able to reuse such as chain.link integrations and DID boilerplate :slight_smile:

1 Like

Hi Brent - to answer part of your question = we do have an MVP goal here!

Lyrasis.org manages most of the world’s open source repository software has engaged several members of this team to present as part of their upcoming curriculum in early November

https://www.lyrasis.org/Content/Pages/Event-Details.aspx?Eid=7A771B09-873E-ED11-80FD-00155D0A2732

as part of the curriculum access to the panel discussion is not free, but we’ll be able to release recordings afterwards, and the general sense of goals and values is outlined here

1 Like

Thank you for this proposal, Anson!
Love to see the lego idea of stringing components together to make useful things. As we discussed sharing here the feedback, I provided last week. The boilerplate code, if I understand it correctly, makes a lot of sense. For example:

  • Creating a standardized way a DAO built on Aragon’s stack can interface with a Chainlink oracle that is well documented and easily used by other projects sounds dope!
  • If the gitcoin passport integration means that any DAO built on Aragon’s stack can use it for Sybil resistance (i.e., create a census relying on having an ERC 20 token and a passport) sounds very exciting as well

The above two make sense, especially if jointly funded by Aragon and Chainlink/Gitcoin, respectively.

I find it difficult to justify why Aragon should be paying for the other elements of the system. For example, the last two items will carry substantially higher benefits for Swollet and Chelo, respectively. Aragon is neither sufficiently central to the value proposition (to argue strategic relevance) nor sharing any potential upside of the success of either Swollet or Chelo.

TLDR: I think having such collaborations materialize is amazing and needed! Having Aragon pay for it all seems unjustified - split the ask across a few sponsors

2 Likes

This is an industry collaboration with real world impact, that also delivers repeatable, scalable value for Aragon. What proportion of 46k do you feel is fair representation of the benefits for Aragon. 75% 66% 50% or less?

I agree it strengthens the proposal to frame the requested funding as a shared portion of the total value of the collaboration. Could the proposal look to include the in kind and various forms of capital that partners bring to the table?

1 Like

Hello - this is Gerardo - part of the Chelo team:

We are excited to be part of this project, and to answer a few questions that have been raised:

  • I confirm that me (Gerardo in the proposal) and Jose will not be receiving fees from the work, but we still need to compensate Rafael for work to be developed - we are happy to let Anson manage the funds.

  • We have worked in various uncollaterized loan projects Chelo Finance – Medium Chelo | Twitter | Linktree Introduction to Chelo - Chelo User Guide

  • we have analysed the current set of credit scoring/DID/credit reputation and we are building an architecture that can connect to a debit credit card. Some of the current credit scoring systems (like spectral finance) analyse the onchain behaviour of wallets and interactions with DeFI - meaning some of the target population (low income) individuals who are not able to enter into DeFI transactions might not have enough data. Combining multiple sources hopefully will allow us to fine tune the probability of default to the specific portfolio of loans.

Both Jose and me are experiences in tradfi methods of credit risk, which we believe can translate to this specific project.

Thanks for the time

1 Like

image
This is a great initiative! But there are some few questions I would like to ask in the learn to earn program.
Is there a curriculum for this yet?
What aspects of web3 (finance and technology alone?) are we teaching and for how long? (Duration of a course/program usually have impact on completion rate .
Who are the target audience/intending learners e.g. age, skill level in web3 etc. (These needs to be considered before a curriculum is formed).
These are important questions that needs to be answered before a learning program is commenced

2 Likes

Hey @Blockchainlore -thank you for chiming in!

right now we are planning on working with the Piedmont Housing Alliance (PHA) here in Charlottesville because they are involved in the direct alignment with local people, and will be involved in helping us find the alpha testing candidates. They have a curriculum they’ve developed here locally

Financial Opportunity Center - Piedmont Housing Alliance and we want to keep as much local as possible for our tests


I do think we need to bring the community guild in to the process if they are interested to learn from how we work with in the process, although I’m not 100% sure what that would look like. The LexDAO legal clinic is going to be meeting with the participants 2-3 times and we could probably facilitate an introduction to DAOs, however that is not the planned focus of the training (more basics of financials)

there are some reasonable privacy considerations involved in working with personal finances that we’d need to discuss from a security and liability perspective - one more reason why the legal wrappers will benefit us :slight_smile:

As far as participants- there are no implications or requirements assumed other than a willingness to participate with PHA and the Center for Civic Innovation https://www.centerforcivic.org/web3 where we are hosting the training and community engagement projects.

On the technical side the plan is to release the courses in coordination with Swollet because they have already developed an interesting web3-centric learn-to-earn app that we can use and perhaps work with our DID solutions to simplify disbursements? not trying to over-commit here, but that’s the direction we could see things going. I believe the plan is to work of 6-8 weeks and include 2-3 meetings with the LexDAO legal clinic as well.

glad to answer more questions, however in general if you have a proposal I would be glad to include that in the budget as it does seem like a natural opportunity?

2 Likes

@fartunov
Thank you so much for taking the time to review the proposal

As per your recommendations we are working on a proposal to reach out to gitcoin - thus the delay in my response here… I don’t have a final draft yet, however it’s point well taken and @lee0007 also brings it up = duly noted

a couple things off the top -
fwiw - swollet is not planning to charge us = they’re part of the proposal because they provide a useful interface & bring some web3-ness that’s worth testing

re: chelo = we need a credit system in the space, and it is ultimately modular… for reasons mentioned we could swap it out with other crediting systems in the future, and but none of them came to the table with the same depth of knowledge in the Aragon ecosystem or the ability to deliver a product like FICO

one of my concerns is that we actually haven’t budgeted enough to support the education and general project testing. @Harry = i’m dragging you in to this too because you inspired a lot of this in the stakeholder driven discussions we had earlier this year… when we look at the opportunity to have a narrative as well as a use-case like this are there some additional considerations you’d like to see included or concerns discussed?

2 Likes

Hi @alibama,

I don’t know much about this topic, but from a product marketing perspective can only suggest the basic questions:

  • Who really needs this? (Who is it a 10X improvement for compared to the current existing solutions)
  • What problem does this solve for them?
  • What are the use cases that it will work for (and which ones won’t it work for)

Ideally you can build strong hypotheses for this (maybe amongst your team you already have these) and validate those hypotheses before actually building anything.

Can you elaborate because this is a very bold statement?

Regarding your question - I think the DAO Star proposal and the GM Bus proposal set a decent benchmark at 15~20% of the cost of worthy industry collaboration initiatives (this one fitting the bill IMO). Sometimes it makes sense to chip in more, but assuming 100% does not fit the idea of collaboration (also keeping in mind that dTech is already paying $45k monthly by Aragon and this sort of activity fits its scope, listing their effort at zero cost is not exactly true).

In a different context, I also spoke with Kyle from LexDAO Clinic. My read of the conversation is that he agrees with trying to take a more collaborative approach on the financial side of things being fair.

Here’s my reference points for the three terms

  1. Repeatable: able to be done again
  2. Scalable: able to be changed in size or scale
  3. Value: “…that capital is governance (and vice versa) leads to the source of its intrinsic value. Whoever has control over a pool of important resources also has the potential to direct some of those resources to their own benefit. So the value of a system’s capital is proportional to the value of the resources it governs…”

You identify and speak to repeatable and scaleable qualities yourself @fartunova. The following are examples of elements of this project that imo already show or could prove to be repeatable and scaleable, elements denoting growth within the Aragon network and the resources governed (value)

Believe there is work underway to quantify and communicate the total cost of the project and Aragon’s contribution as a portion of that figure. Correct @alibama et.al? However, I believe 15-20% is a very low representation of the value added by tools built

  1. on the Aragon Tech Stack or
  2. integrated with the Aragon Tech Stack

My understanding is that a proposal to expand Aragon’s tech aligns with the hyperstructure strategy and should be accounted for at a higher percentage than either paid brand advertising or DAO Standards. Aragon does not govern ecosystem policy nor industry events. It does however have direct power and control over the network infrastructure which is indicated above as driving intrinsic value.

Please correct me, as I’ve much yet to learn about the future direction but I am a bit confused because it seems to me the strategy is to build a governance hyperstructure by leveraging ecosystem-wide talent yet here we are indicating such collaboration is only low-value (15-20%) to Aragon? :confused:

Hi there - we (chelo) wrote this blog to give some initial thoughts on uncollaterized loans (like credit cards):

Excerpt:

Web 3 Credit Model

All the above elements are requirements to build a credit model of the portfolio of loans that can provide us answers such as:

  • How many loans could default ?
  • How many loans could be fraudulent ?
  • What interest rate should be charged to cover all the expected losses ?

There are many ways to reduce the quantity of defaults — limit loans only to the highest credit scores, reduce loan amounts, diversify the number of loans to individuals in different locations, professions, etc).

At Chelo we are working on the top level Web3 Credit Model that DAOs could use to be able to provide uncollateralized loans to a set of borrowers (members, outside applicants, etc). To be able to build the Credit Model we are engaging/studying/partnering with the above DID / Reputation score / Credit Score providers.

2 Likes

Appreciate your interpretation, however, I disagree with your read of 20% indicating “low value”.

More importantly, I am not saying it should be 15-20% I am providing benchmarks of what has been done in the past. One that is substantially more reasonable than 100%.

A very simple way of ballparking a more reasonable split (not to be taken literally):

  • Item 1 - 50/50 Aragon & Gitcoin
  • Item 2 - 33/33/33 Aragon & Gitcoin & Notebook
  • Item 3 - 100 Aragon (estimate needed for the proposals submitted to the other partners)
  • Item 5 - 50/50 Aragon & Chainlink
  • Item 6 - Proportionally to the above splits if those Audits pertain to the above items, not to Chelo’s credit assessment contract base
  • Items 7 & 8 - probably 10-20% Aragon. Rest split across Swallet, Chelo, and the other partners

I imagine the proposers are better positioned to work out the intricacies of how to do this (not sure how Kiva, Bright ID, and Raincard play into this). The outcome will likely be above 20% but will be quite far from 100% as well. Not to mention the massive benefit of aligning other stakeholders around the initiative’s success. It’s a way to make 1+1 > 2

Hope that helps.

1 Like

Thanks for chiming in!

  1. the people we see needing this solution presently are unbanked users who will see an almost infinite improvement over current trad-fi work. for traditionally banked users there could be very modest improvements in their ability to support unbanked people, but i don’t see it as a huge win for regular users? that said most of the traditional credit scoring groups are “pretty heinous groups” so anything that fights that seems like a “win” but not a “big win” and more of an intellectual victory at least in the short run.

  2. the way this proposal solves this issue is by separating some of the hard requirements traditional banks have = we are looking at libraries being able to issue credit cards with this solution, and while they might require some DID (and even reasonable KYC) they wouldn’t need a physical address

  3. the use case we are focusing on is low-income users specifically here in Charlottesville where we can control most of the variables… i believe for this use case we will be able to deliver a valid proof-of-concept. the places we know this won’t be successful? large industries are not likely to leverage these tools any time soon… there’s just no way to insure these numbers if the $$ amounts got large and there were major disputes, and i think insuring a credit score is essential at some point to show that it’s real

2 Likes

Interesting proposal @alibama . Given that this is a credit rating system, there’s already an interoperable attestations prototype (for https://daostar.one/attestations) being put together by several companies in the DAO space / DAOstar. The use-cases there are mainly web3 CVs and reputation, but a credit rating is clearly a sort of reputational attestation. Get in touch!

3 Likes

I’d love to hear more… i think the meat of this proposal is both the credit scoring ability, however more than that it’s the DID, chainlink and real-world integrations that bring value

that said and with a bit of a heavy heart i think we need to push this proposal back another week or two to better develop a funding split and make things work for the whole DAO

1 Like