Can flawed credit rating systems be rebuilt on the blockchain?
Credit scores control people's lives, and something that powerful needs to be decentralised.
On 2 October 2018, Mastercard filed a patent for connecting an individual's cryptocurrency and fiat accounts.
It's described as a "behavioural profiling method and system to authenticate a user." This verification might be an important security element in the process where credit cards or accounts will be automatically frozen when a strange transaction is detected.
Strange, in this case, is defined as a deviation from the user's typical behaviour according to their established profile, such as a transaction taking place outside their usual country of residence.
"Variations between the behavioral characteristics of the user interactions observed during the current session and a behavioral profile previously developed based on prior usage patterns of the user through the plurality of channels are identified, in real-time or near real-time," as the patent says.
Essentially, it might be a way of tracking someone's financial activity and behaviour in cryptocurrency as well as fiat currency, with applications in security as well as an extension of systems such as credit scores.
Credit scores in brief
Credit scores are an automatically generated number that controls people's financial life. Given the sheer scope of their power, it might be tempting to assume they have some kind of grand government purpose, but they're actually a business service.
Different for-profit credit rating agencies develop their own scoring systems and keep their own data, and the implicit purpose of credit scores is to maximise the profits which can be generated from the creation and maintenance of such scores.
The two best known credit rating agencies might be Experian and Equifax. You might recognise Equifax from the Equifax data breach that exposed detailed and sensitive information, such as the social security numbers, of almost 150 million people.
And you might recognise Experian from another data breach, where its partner company Alteryx left the personal information of 123 million people sitting around on the Internet for anyone to access. It was highly sensitive information too, including addresses (along with latitude and longitude) and phone numbers. Funnily enough, it still technically qualified as "anonymised" data even though it would be trivial to de-anonymise it. The real problem for the company wasn't that millions of people had sensitive personal information leaked – it was that non-paying customers were able to access the data as well. People would typically have had to pay a $33,800 licensing fee to access all that data.
The kicker is that identity theft, which can wreak havoc on a person's credit score, can be carried out with the information collected and leaked by credit rating agencies. And no major credit rating agency is especially well known for its customer service.
From one angle, the credit score system has devolved into one where credit rating agencies collect customer data, sell it for profit, then turn around and blame the customer for lax data security when fraudsters start utilising the data being publicly sold by the credit rating agencies themselves. And customers can't even opt out of this data collection because it's required if they want to live any kind of financial life. And if you suspect you might be a victim of identity theft, one go-to move is to check your credit score, which might incur extra costs on your end.
That's not even touching on the way credit scores have started morphing from a risk prevention tool into a behavioural control tool. The goal of credit rating agencies is to make money, and in many cases, that goal is best achieved at the expense of customers born into the inescapable credit score system. Some examples include making people pay for credit reports whenever they need to do certain essentials or driving more value towards partners by creating a system that penalises customers for closing unwanted accounts.
The centralisation of credit ratings under centralised profit-driven businesses has created a deeply flawed, ethically fraught and, in some ways, outright dangerous environment.
Where China's social credit system fits in
China is introducing a similar social credit system. Just like Experian- and Equifax-style credit ratings, it uses a system of rewards and punishments to shape people's behaviour in the desired ways. Credit rating scores are typically deployed for the purpose of generating business revenue, but China's seems to be gearing up as a way of silencing dissent and punishing those publicly speaking out about party corruption, through systems such as preventing "undesirables" from travelling, maintaining a voice on social media or even interacting with others – someone who associates with undesirables will in turn see their own social credit score start to suffer.
Despite the overarching similarities, China's social credit system is shaping up to be much more complex (and competently operated) than most credit rating systems we know today.
In part, this might be because it has access to much, much more data. It can tie into cameras recording on every street corner, it can see where people travel, who they associate with, what they buy, where they buy it, how their medical records look and whether the person has any exemptions from scrutiny – you can't have high-ranking party members being subject to the same scrutiny as the rabble, after all.
In this way, many of the differences between credit ratings and China's social credit system might come down to data access. If cryptocurrencies start falling under the umbrella of credit rating systems, many more data points might start getting measured.
On one hand, the very idea of a credit history is somewhat antithetical to the original idea of cryptocurrency and financial freedom. On the other hand, cryptocurrency development has seen both the intelligent design and the evolution of social and economic microcosms – some of which include similar reputation metrics for users.
One of the most apt examples of this might be Constellation, a project which aims to recreate trust on the blockchain, or DAG in this case, through a reputation system. And launching an effective reputation-based system naturally means collecting and translating user data into scores of a sort.
It has a lot of potential, but credit scores as we know them are also an example of how these systems can go wrong. With news like Mastercard's patent suggesting that cryptocurrency might start moving under the traditional financial umbrella, credit scores and all, it might be worth considering what the future looks like.
A way forwards
The problem might be as follows:
- Having better access to more detailed user data stands to make credit score systems a lot more effective and useful for their original purpose of risk mitigation
- But the sheer power that these systems accrue means they tend to follow existing incentives fairly ruthlessly. In the case of credit scores, those incentives are largely the profit motivations of credit reporting agencies, even at the cost of users. In the case of China's social credit system, that might be protecting existing power structures and silencing dissent, even at the cost of functional and well-informed national governance.
In this context, the key problem with existing credit rating systems might be that they're centralised. So even though the idea of welcoming credit score-like systems into cryptocurrency might be counter-intuitive, it might also present opportunities for re-inventing credit scores in a decentralised and incorruptible way. Data self-sovereignty in distributed ledgers might also stand to solve the ever-present data security issues plaguing the industry.
"Having access to blockchain and DLT opens up many data points that a traditional credit score system does not," explains Zac Russell, VP of marketing at Constellation, noting that bringing in more data can allow for improvements that satisfy more people.
He gives the example of small businesses, whose sporadic cash flow might algorithmically be translated into a picture of a business that's in more dire straits than it actually is. A more detailed credit rating system might more effectively differentiate between cash flow financing, expansion loans and other types of financing to give a more accurate score.
"From a business perspective, the credit score industry is notoriously hard on medium to small businesses due to how they analyze cash flow. DLT offers the ability to factor in additional variables based on an individual's behavior over a long period of time within a network. For instance, it could pull social data or transactions that are normally unaccounted for.
"There is an opportunity to build a more holistic profile of an individual, which is also more accurate. As local banks or lenders often make better decisions based on factors outside of a narrow transactional history, DLT would bring that local knowledge to a global scale."
The downside, he notes, is that no matter how data is used, a credit rating system that is trusted too implicitly and leaned on too heavily could swing back around to where it is today.
"Complete trust in data could cancel out business ideas and entrepreneurialism based on an individual's or entity's history, not valuing the idea or proposition. I think we could see a fairer, more accurate system, but I also worry that too much trust in that system could stifle entrepreneurialism," Russell says.
China's Sesame credit system (aka the Zhima credit system), which is a precursor of sorts to the full social credit system, might be a glimpse into the potential upsides offered by access to more data points.
As a centralised system, it's naturally prone to corruptibility, Russell says, but its comprehensiveness gives it new ways to prevent people from "gaming" the system by opening unnecessary credit streams or keeping accounts they don't want. Basically, with more data going into the system, people won't need to lean as heavily on "tricks" to boost their credit scores, which can also make the entire thing more useful.
"The Sesame credit system offers an interesting paradigm on how DLT credit scores systems could work in the future," Russell says. "The Zhima credit system offers a greater number of inputs (social media interactions, purchases from online retailers) outside of traditional financial history, which indeed provides a more holistic picture of an individual or business. These increased data inputs, as pointed out, help stop the questionable and sometimes harmful practice of "gaming" the credit system by opening unnecessary credit accounts, and allows for the overlooked and underserved an opportunity to enter an economy. DLT technology offers the ability to work in such a way, but without this data being owned, manipulated, and potentially stolen by a centralised entity such as Alibaba... So in a sense, we could have our cake and eat it too — a fairer, safer, more democratic system, led by secured data, without the risks of centralised ownership (like the case with Sesame)."
Decentralisation provides a powerful way of reshaping flawed existing credit systems. But like many things in cryptocurrency, the realisation of this potential might come down to the quality of a system's design.
Disclosure: At the time of writing, the author holds ETH, IOTA, ICX, VET, XLM, BTC and ADA.