The availability of big data and AI-powered consumer credit scoring and decisioning has not only helped China establish a consumer credit risk scoring industry from scratch but also drove the surge of consumer finance.
There weren’t FICO-like credit scorers or privately-owned credit bureaus in China until 2015 when China's central bank allowed several selected private companies to roll out pilot systems for personal credit risk assessment. Since then, a variety of Chinese companies including existing commercial credit reporting agencies and major internet companies, have entered this field.
They have adopted big data and AI techniques, feeding algorithms with various types of data sourced from telecom operators, e-commerce platforms, online payment and banking services, social networks, and other online and mobile services.
Technological availability aside, the big data-based approach was an obvious choice for newly emerging Chinese credit scorers.
Before 2015, only China’s central bank had maintained a national database for personal credit information and mainly served established banking institutions, mostly state-owned. Banks were the main channel for Chinese consumers to get a mortgage, auto loans or credit cards, and most of the alternative lenders didn’t have access to the central bank’s credit information database. That means hundreds of millions of Chinese consumers who aren’t covered by the central bank would face difficulty accessing either traditional or alternative financing.
When it comes to AI, China is known for its advantage in consumer data, not only in quantity but also in quality.
Due to Chinese consumers being heavy users of mobile services and China having relatively lax regulations on personal data collection and usage in the last decade, Chinese internet companies and telecom operators were able to amass a wide range of consumer data, many types considered being of high predictive power for credit scoring, such as the history of payments and other financial transactions. Since Chinese consumers are increasingly more dependent on their smartphones in their daily life, their digital traces are believed to be increasingly more relevant for creditworthiness assessment and credit monitoring.
More recently, Chinese credit scorers are now able to add conventional credit information by working with commercial banks, China’s central bank and other established financial institutions, and public sector information, since consumer finance has been supported and promoted in recent years by the Chinese government as a new economic growth driver after the export and infrastructure-driven growth flattened. As credit scoring services are increasingly adopted by consumer lenders, Chinese consumers are also becoming more willing to submit personal information like ownership of properties or automobiles, to credit scorers to boost credit.
Data sharing among credit scorers and data vendors has been common in China, though personal data breaches are becoming an increasing concern.
Not only do scorers let other scorers or lenders access their own data platform, some online lenders also share data such as repayments and other credit attributes with their partner scorers. Baihang Credit, the personal credit scoring company co-established in 2018 by the National Internet Finance Association and several private credit scoring service developers, introduced some 15 online consumer lenders and conventional consumer finance companies as its first data sources.
Boosting consumer finance: Chinese consumers can check their credit scores or take out loans on credit scoring apps. (Source: Wecash)
Freemium has been a well-perceived business model in China’s commercial internet market. So it’s not surprising to see Chinese credit scorers to offer credit scores or personal credit analysis reports for free. And they wasted no time adding online lending services, either developed in-house or run by third parties.
Taking advantage of the existing consumer data on the Chinese web, Chinese credit scorers scored hundreds of millions of internet users and pushed credit products directly to users through existing online services without waiting for users to ask. E-commerce giant JD, one of the first to offer online purchase credit product, had reportedly scored about 150 million shoppers of its e-retail platform as of the end of 2015. Tencent, the leading social networking service provider, and Ant Financial, the leading digital payment service provider affiliated to e-commerce company Alibaba, each scored tens of millions of users.
Starting with e-commerce platforms like JD and Alibaba, a wide variety of online services, ranging from ride-hailing apps to food delivery apps, would begin providing credit options to their users in the last couple of years, with risk assessed and credit limits decided by in-house developers or third-party credit scoring systems. As of September 2017, 59% of the online and physical businesses and services that accepted mobile payments began offering credit products to their customers, according to Chinese internet research firm iResearch.
Independent credit scorers like Wecash, developed by Chinese fintech company 9F Group, and CCX Credit, whose parent company is a major commercial credit rating service provider, have become online platforms connecting investors and borrowers.
Independent solution providers are enabling not only online services but also traditional physical businesses like retailers to make instant or near-instant decisions for credit offerings like point-of-sales financing.
Applying machine learning, credit scorers have increased approval rates for many online and traditional lenders.
The approval rates of JD’s credit product increased by 150% after it revamped its credit assessment and underwriting models with additional online data with help from ZestFinance, a US-based machine learning underwriting solution developer with which JD has established a joint venture. Zhima Credit, the credit scoring service developed by fintech giant Ant Financial, claims that it has increased the approval rating by 7% for credit card business of a local bank.
Variable pricing enabled by credit scoring systems has also boosted consumer loans by offering lowered interest rates or higher credit limits to more creditworthy borrowers. Risk-based pricing is new in China. Local commercial banks didn’t implement it largely due to lack of credit data and high cost associated with manual loan approval procedures, stated China Rapid Finance, which has been a long-time consultant to major Chinese banks before building its own online consumer lending business.
Driven largely by online consumer credit offerings supported by credit scorers, the total amount of consumer credit (excluding mortgage) originated by online lenders as a percentage of China’s total, increased from 0.5% in 2014 to 5.5% in 2016 and is expected to reach 9% in 2020, according to a report by Chinese investment banking firm CICC.
Apart from boosting consumer finance, big data and AI-powered credit scoring is increasingly as important as risk management tools.
Before the advent of credit scoring, thousands of online lending platforms had emerged in China, many extending loans to consumers or individually owned businesses. Due to poor risk management, frauds and scams were rampant in China’s online lending sector.
Leading peer-to-peer lending sites like Yirendai and PPDai have developed their own big data-based credit scoring systems to help investors on their platforms screen prospective borrowers and make investment decisions. PPDai, one of China’s earliest peer-to-peer lending sites, rolled out in 2014 its own big data and AI-based credit scoring system. It would later introduce many other credit scoring services to its platform, such as Zhima Credit and the credit information platform by the National Internet Finance Association. PPDai now hosts an annual machine learning contest, letting outside developers use their data warehouse to improve their credit assessment models.
Some Chinese credit scorers have expanded further to provide a variety of anti-fraud and risk management services powered by big data and machine learning and, more recently, vision AI for identity risk management or blockchain for secure data sharing, risk management as well as addressing personal data safety concerns.
Tracey Xiang is a tech writer, specializing in ChinaTech, Digital Economy, FinTech and AI.