The conflicts of interest in the business model of credit rating agencies (CRAs) were laid bare in the 2008 financial crisis.
Investors and issuing sponsors (banks) rely on the ratings of the Big Three (Fitch, Moody’s Investor Services and Standard & Poor’s) to determine the likelihood of a bond issuer paying back the money on time.
The credit rating of a borrower (usually a company, but also a government or a multilateral agency) impacts the pricing of its debt, especially its bonds, which trade on a market. A triple-A borrower like Apple is going to pay less to borrow than a company that is at the lower end of the investment-grade scale – and way less than a company that issues junk bonds, aka high-yield debt.
Sounds good, but the problem in the model depends on who pays for the rating. In an ideal world, investors would pay a fee for an independent rating, which would ensure CRAs are free to give an honest opinion. But the asset-management industry decided it would rather save on fees in return for potentially compromised ratings (the big firms can do their own due diligence). So the CRA industry charges the issuer a fee, and issuers are incentivized to pay for a rating if it’s going to be a good one.
In theory this means only solid companies buy a rating, but the 2008 subprime-debt crisis made clear that CRAs were happy to put a triple-A rating on tranches of bundled real-estate loans that turned out to be toxic.
New models for DeFi
Because “everybody was doing it”, no one got seriously punished, and the model remains intact. It also mostly works okay, as only very large companies can afford a rating. Around 10 percent of listed companies have credit ratings, and these are very well known, well-researched companies.
The disruptive power of blockchain is slowly making its way through more corners of financial services. As more lending protocols emerge, and eventually as real-world tokenization takes root, there will be a lot of entities looking to borrow in the form of digital assets.
Fintechs are popping up to provide credit ratings within the DeFi space. A first wave of tech-based businesses have tried to produce ratings in the form of NFTs. But now a Dubai- and Hong Kong-based team is tackling the problem from the starting point of TradFi, and converting those models into ones suitable for DeFi. Along the way they are reimaging a different set of incentives to make credit ratings different and, they say, better.
“We will provide a forward-looking probability of default that will be more accurate than a traditional CRA,” said Robert Alcorn, co-founder and CEO in Dubai (pictured, left).
Synnax has just raised $1 million in a pre-seed funding round to launch a DeFi CRA that is currently in beta-testing mode, with a live launch slated for summer.
Alcorn says the idea stemmed from his previous startup, Clearpool, a DeFi protocol for uncollateralized lending. He teamed up with Dario Capodici and Alessio Quaglini of Hong Kong-based Hex Trust, a digital asset custodian, to launch Synnax. The founders have pre-crypto backgrounds in fixed-income at international banks.
Encrypted vs. transparent
Capodici, who serves as COO, says Synnax’s model takes encrypted data from ratee companies and allows a network of independent data scientists use their own machine-learning models to come up with multiple ratings. Ratee companies provide APIs to show real-time updates on data such as profitability and leverage, but the data scientists may incorporate their own sources of data into their models, such as macro information or social media sentiment.
The network of data scientists can produce dozens or even hundreds of different scores to determine a ratee’s likelihood of paying back its debt on time. Synnax amalgamates these into a single credit rating, which is made public, although the underlying ratee company’s data remains private.
“We aggregate the data and weigh the analysts based on who’s model is the most accurate,” Capodici said. “It provides a single view but based on hundreds of algorithms, with real-time updates. This is different to traditional CRAs whose rating is a the view of a single analyst or model, and which only changes when a company releases its latest financial statements.”
This not only tilts more weight in the final aggregation, but weighting also plays a part in the fee model.
Revenue streams
Synnax is building three revenue streams. First, ratee companies do pay, based on a quarterly subscription. Second, users (hedge funds, brokers) pay a subscription to get a look beneath the hood – there’s a lot of nuance in the models behind the rating itself. Third, banks, underwriters, lenders, or fintechs can buy the data to run their own models, to help them price DeFi debt for their clients.
The money also flows out, to the data scientists. They are rewarded based on their weighting, which is based on how predictive their models are, based on default rates as well as other metrics such as predicting spreads or trading volumes.
This being crypto, data scientists will be paid in a Synnax governance token, Synai, which has yet to be minted. Clients will pay for services in this format as well. Synnax will operate a treasury, with a discount window to trade Synai on virtual-asset exchanges. Synnax will also reward its users with Synai for responding to polls or participating in competitions to build new AI models.
Synnax, and the DeFi world in general, is not going to make a dent in the traditional CRA model, which is designed to serve large listed companies. The founders say it is aimed at private companies that are interested in blockchain-based financing.
But the two worlds may influence each other.
DeFi’s impact on TradFi
Alcorn says he hopes the traditional CRAs will become clients, using Synnax data to augment their own work. Synnax intends to put out ratings on about 2,000 companies, including those big public corporations. This will provide an alternative rating, even if the companies aren’t in the DeFi space.
To keep things simple, Synnax decided to stick with the TradFi rating system (triple-A and down), but it will be interesting to see how, over time, its ratings differ or conform to those issued by traditional CRAs.
DeFi ratings are also meant to change regularly. Ratee companies will see their scores improve if they become more transparent. That means providing more data, as well as ensuring data is API-enabled and not manually uploaded. This too will create a new way of looking at the creditworthiness of companies.
There is the hybrid pricing model, which could impact the final rating, if there are indeed biases hidden in the status-quo business model.
This model for a CRA will also impact the way investment banks and lenders approach issuing debt. This is from the way data is treated, and the way capital is treated.
In TradFi, underwriting banks have people on their debt-capital-market teams who work with clients to improve their credit ratings, via financial engineering and how companies report tax and accounts. That sort of thing won’t be effective in a DeFi setting, because bankers won’t (in theory) be able to influence all of these decentralized data scientists.
“Data scientists won’t take fudged data submissions at face value,” Alcorn said.
Capodici adds that TradFi ratings are designed to minimize the capital charges on banks. Under Bank of International Settlement rules, banks prefer fee-based businesses, which don’t incur capital costs. But lending does. So does equity underwriting (because the stock briefly sits on a bank’s balance sheet).
Therefore, banks only want to do underwriting for their biggest clients, from whom they expect to generate plenty of fee-based revenues in other ways (transaction banking, FX, wealth management, etc). This is one reason why only the biggest corporations get a credit rating: small and private players are locked out of traditional capital markets.
Private capital markets
It’s possible, however, that such companies could borrow in DeFi markets, and they’ll be able to get an affordable credit rating to support their treasury needs.
Alcorn said, “We expect a migration of private credit moving onchain because it’s so much more efficient, while we also thing digital-asset capital will at times want to transact offchain when there’s a yield differential.”
The pre-seed funding was led by No Limits Holdings. Other investors include Edessa Capital, Kenetic Capital, Bitscale, Ryze Capital, MH Ventures, Hex Trust, Moonvault, GameFi Ventures, Typhon Ventures, Ausvic Capital, Drops Ventures, and Everstake Ventures.
