Credit and Crypto: A Brief History and the Future
Creating access to credit with fair interest rates has been a target for many traditional FinTech companies for well over a decade. With the boom of Bitcoin and Ethereum, the scope and effectiveness of these services has given opportunity for peer-to-peer (P2P) lending and credit scoring to become the killer apps for crypto. In this article, I will cover the progression of the crypto lending and credit scoring space from its early days of more centralized lending solutions denominated in crypto to the present decentralized lending and credit scoring protocols. There is still so much work to be done before P2P crypto lending fully takes off, but the building blocks are starting to form.
In the beginning, there was BitcoinTalk and BTCjam
All roads lead to Rome, and all crypto product ideas can be traced back to BitcoinTalk. Founded by Satoshi themselves, BitcoinTalk became the hub for everything Bitcoin: discussion, announcements, trading, and even lending bitcoin to other users. With an innate trust system on this potent forum, marking a user’s trustworthiness after trading with them and/or lending was a much used feature. (side note: with an account, you can see Satoshi’s trust page here where people like to leave nice messages for them to this day.) Additionally, you could easily create your own trust network on the website. By adding the usernames of those you trust and don’t trust, would contribute to a user’s trust rating. This considerably sophisticated forum feature paved the way for many different products in the space, the first one being BTCjam.
In 2012, BTCjam was announced on BitcoinTalk. Founded by Celso Cardoso, a Brazilian cryptographer and patented AI technologist, BTCjam was the first P2P lending platform to use bitcoin (and only bitcoin). I joined in the beginning of 2014 as the first employee, helped build the company until its dissolution.
Without FICO, banks, or even institutional money to fund loans, BTCjam developed a product that was servicing US$1million in loans per month by June 2014, operating 100+ countries, with an investor platform average of 13% ROI, and a robust community. As said by Dr. Wade Rousse in the textbook Banking Beyond Banks and Money,
“The transnational scope of Bitcoin is what is impressive with BTCjam being the first P2P lending platform to cross national borders successfully.”.
The soft credit check mentioned in the ANN thread was the beginning of the credit scoring algorithm. With his AI background, Cardoso created a machine learning neural network that utilized thousands of data points from a user’s digital footprint (social networks, Paypal, search), in addition to historical lending data, and the user’s own trust network. This was essentially the disruptive feature of its time: BTCjam successfully scored thousands of borrowers from 100+ countries using their digital footprint alone.
Growing BTCjam was more difficult than your average Silicon Valley startup, but the pains were shared with all bitcoin companies at the time. A month after raising the 2014 seed round, the Mt.Gox tragedy began with the ensued Crypto Winter. We were faced with regulation from multiple government agencies in dozens of countries with regards to using cryptocurrency, AML, lending, investing, and credit scoring.
Services provided by BTCjam:
• Credit scoring with P2P attestation
• KYC/AML: ID, address, and income/accreditation
• Bitcoin loan servicing
• Custodial bitcoin wallet
• Investing with reporting and AutoInvest
• Debt marketplace
• Arbitration
In retrospect, we were doing a lot at once and there were two competitors: Bitlendingclub and Bitbond. As the crypto lending space was small, having two aggressive competitors who did the same exact thing did not promote growth. It was Peter Thiel’s theory around perfect competition exemplified. Today’s crypto lending landscape is very different, and much more harmonious due to web3 and ICOs. Explanation in the next sections.
Like this and like dApp
The creation of turing complete blockchains, such as Ethereum and the subsequent development of layered protocol technology like 0x protocol will foster a new internet in many ways, but most notably for this post, by being a disruptor for any type of marketplace based product.
Instead of creating multiple different components to provide a single service, web3’s loose coupling structure allows businesses to focus on providing the best experience by building one component within a network of others that work together to provide this single service.
A general idea of how the borrower and investor user flow could look:
In this, niche or market-specific dApps are needed. For example, a lending dApp interface that targets borrowers in a certain category or region would create a product according to these targeted borrower’s needs.
A few use cases where niche markets would require an interface specific to their needs in order to be successful are:
• Miners (zCash, LTC, others) due to the typically collateralized nature of lending
• Unbanked communities in developing countries due to language barriers, technology proficiency, and cultural peculiarities that globalized loan dApps won’t know how to address
• Privacy advocates due to specific user expectations in how sensitive financial information is handled, the dApp may use protocol services differently with zk-SNARKs
Fundraising
In the Crypto Winter (2014–2016), investment interest was mutually exclusive to price and many companies folded under the pressure or pivoted to “blockchain not bitcoin” narratives. Also during this time, publicly traded and P2P fiat lending pioneer LendingClub (LC) share prices fell 50% in June 2015 just six months after its NYSE debut. It was a chilly winter for bitcoin companies and holders, but it was even colder for P2P bitcoin lending companies trying to fundraise.
Fast forward to August 2018 and the P2P crypto lending space has a plethora of players. This is in part due to Initial Coin Offerings. ICOs have changed the way in which projects raise funds and the benefits and harms they bring have been explored in many other articles. As it pertains to this one, borderless fundraising where an investment signals a project’s merit and potential has created an opportunity for these projects to raise funds that weren’t possible before.
With all the momentum, BTCjam raised a total of $9.2m over 3 equity rounds. The average raise of current P2P lending platforms, dApps, and protocols is at $25.1m. I do not consider the larger raises that new P2P lending companies have as superfluous and greedy. Innovations in lending technology and mechanisms require a budget towards R&D. Additionally, regulatory waters are still murky and even when they are clear, all projects will have to dedicate a considerable cost to compliance, the amount of which is dependent upon the number of jurisdictions where one intends to serve among other variables. Though it is a brave new world where pre-product projects are raising 2x what BTCjam did, it is an experiment that I am willing to see through, and one that I believe is necessary.
Price Volatility
When the market became aware of BTCjam’s “bitcoin denominated loans”, the top question we would always receive would be how we circumvent price volatility. When a loan was initiated (deposited into the borrower’s account), the repayment terms would be fixed to the bitcoin price in the borrower’s local fiat currency based on the timestamp of the loan. Investors would sometimes either be at a small loss or gain in terms of the bitcoin they lended, but obviously always at a gain in terms of fiat. Though returns between loans denominated in fiat vs. pure bitcoin loans had the same performance in ROI, accounting for the price fluctuation, many investors and borrowers found it confusing. In this time period, Bitcoin wasn’t having large rallies or losses compared to the current market and most of the loans were short term, fixing loans to be denominated in fiat was not problematic.
As the market has changed much since 2016 due to an influx of retail investors in crypto, designing solutions that take advantage of recent advancements in stablecoins such as DAI will help mitigate the increased risks that exist for both borrowers and lenders alike.
Identity
There are multiple identity projects in the space, and this topic has been covered in many, many different articles. Instead of regurgitating here as I find this topic well covered, see these articles below:
• Identity use cases for developing countries
• BrightID’s take on stamping uniqueness through network analysis
• Ryan Shea’s video on identity
Arbitration
One of the more difficult components around P2P bitcoin lending was arbitration. This was exclusively due to regulatory concerns. For BTCjam, though the success of the credit score was great, life happens and sometimes people default on their loans. We ended up developing a note marketplace where investors could sell their investments as a way to mitigate loss for their portfolio and to offload defaults (this is something investing dApps and protocols can now adopt). When a default process was initiated, the investor would be supplied with borrower's contact information in order to pursue collections. This worked to an extent, but it was a heavy burden on the investor side and investors would tend to focus on individual loss rather than their overall portfolio performance.
Considering that P2P lending investors will typically be investing small amounts in one person, it is not the best use of an investor’s time (nor expertise) to be left with the burden of arbitration. Arbitration dApps like Kleros and decentralized court systems on Aragon can save an investor’s time by creating new, incentivized markets for distributed dispute resolution.
Credit Scoring
Lending isn’t necessarily sexy. It is of my most upmost contrarian opinion that there are many P2P lending projects in the blockchain space since trust, the key element in lending, can be effectively measured through this structure. Money acts as an extension of ourselves as our use of it dictates the outcome of our livelihood. When someone risks their money with an expectation for it to be returned, this illicits an innate contract of trust. Trust metrics are a fascinating field and there are few projects in the space exploring this.
A high potential product in credit scoring is the Bloom protocol. BloomScore, their credit scoring engine, maps a borrower’s positioning within the Bloom network and evaluates them based on the scores of their peers in addition to any historical financial data. “Peer-to peer staking” is Bloom’s vouching mechanism that is their foundational layer designed to establish both a creditworthiness indicator and a Sybil-resistant identity. Bloom’s overall product trajectory should not be mistaken as a FICO replacement, but rather a privacy-forward scoring engine that can synthesize all ID and financial data on a person as long (and only if) the person’s financial actions are performed on providers within the Bloom network.
Another credit scoring protocol is Colendi. Using big-data and a borrower’s digital footprint, Colendi provides scores to borrowers with incentive mechanisms for data providers. To compare to Bloom would be unfair as their credit scoring mechanics are completely different, and hence Colendi can act as an organizational staker in the Bloom network. However, the advantage over Bloom is that a borrower could theoretically procure a credit score without any personal financial data. A limitation to this is that in order to provide this service Colendi needs historical data on potentially thousands of similarly positioned borrowers in order to properly train their algorithm. It’s not impossible at all, as this has similarities to BTCjam’s model, although it will be interesting to see how they plan on doing this.
In Conclusion
Credit scoring engines and digital footprint algorithms need historical data. This is going to take time to mature, and scoring on a global scale is difficult. The social media and digital footprint of your typical borrower living in the Lower East Side in Manhattan will be exponentially different than one in Manaus, Brazil. The same goes for establishing scores for borrowers within attestation networks and decentralized credit scoring methods. I see these credit scorer companies needing to work with locals within certain regions in order be an international scoring algorithm.
We considered 2013 as the “early days” for P2P crypto lending, when really it was the wild west precursor for the early days of today. There is plenty of R&D left, and credit scoring algorithms need time to prove themselves and a lot of historical data to train on. Hence, if underlying protocol components promoted interfaces in geographically local, market-specific, and niche categories, we can optimize at more accelerated timelines. We can get closer to the original goal of crypto which is to bank the unbanked and ultimately build the bridge between our nascent, small crypto-community and the billions of others in the world.
Found this informative? Please share!
@AlexisAiono on Twitter
Big thanks to Yalda Mousavinia for helping review this article. Follow her, she’s great!
