In the seventh session of the webinar series on “Financial Transformation, Credit Markets and Household Debt in Southeast Asia”, Dr Nicolas Lainez presents on the prospects and dangers of algorithmic credit scoring in Vietnam.
Regional Economic Studies Programme Webinar
Monday, 8 March 2021 – ISEAS – The Yusof Ishak Institute hosted a webinar on “The Prospects and Dangers of Algorithmic Credit Scoring in Vietnam: Regulating a Legal Blindspot”. It was delivered by Dr Nicolas Lainez, a Visiting Fellow at the Institute and an Adjunct Assistant professor at the Department of Southeast Asian Studies at NUS in Singapore. Three discussants participated in the webinar: Dr Elliott Ash, an Assistant Professor of Law, Economics, and Data Science at ETH Zurich’s Center for Law & Economics, Switzerland; Dr Eric H.Y. Koh from the Department of Finance and Banking, Faculty of Business and Accountancy at the University of Malaya in Malaysia; and Ms Dang Linh Chi, a special counsel based in the Baker McKenzie’s Hanoi office.
Dr Nicolas started off his presentation by sharing the perimeters of his research study and the challenges he faced in carrying it out. Since the deployment of algorithmic credit scoring (ACS) in the consumer lending industry in Vietnam and Southeast Asia is fairly recent, there is little to no data and research to use as reference. Moreover, accessing digital borrowers, bankers and fintech players for the purpose of this study was challenging. Hence, his aim was to provide a glimpse into the credit scoring scene in Vietnam, weigh the promises and dangers of ACS and map areas for policy intervention and research.
Next, he shared the promises of the use of ACS. He described it as a tool to boost consumer finance in South East Asia. Traditional credit scoring utilizes few data points such as income, labour status, loan purpose and address etc. Research has shown that this data is subject to human discretion and bureaucracy, especially in developing countries, creating incorrect assumptions of consumers. Such assumptions hold limited relevance in Vietnam where precarity and informality prevails. On the other hand, ACS utilises thousands of data points as well as artificial intelligence to provide higher accuracy, efficiency, speed and customization. Thus, customers enjoy a faster and easier credit experience. the narrative goes. The data is derived from telcos, mobiles, social networks, browsers, e-commerce and financial transactions.
Then, Dr Nicolas discussed the dangers of ACS. There are conflicting public perceptions of big data and AI when it comes to fairness and privacy protection. Machine learning algorithms can make biased decisions if they are flawed or reproduce human biases incorporated to existing datasets. He noted that humans also design, program and train algorithms to reach certain outcomes, which can reflect discriminating policies. Dr Nicolas questions whether credit scoring systems and credit policies will produce discrimination in newly drawn financial landscapes made of complex risk matrices and gradations. The second and third issues he elaborated on are of ACS inhibiting individual autonomy and creating an increased need for cybersecurity.
Lastly, Dr Nicolas commented on the Vietnamese government’s framework to operate regulatory sandboxes in several key domains to support fintech development. Amongst the key domains, it is unclear if credit scoring is included. Moreover, he noted that the outcome of the program remains to be seen. From his literature reviews, he found that Vietnamese law is ill-equipped to regulate algorithmic credit scoring as they were designed before the advent of AI. In addition, the law leaves lenders unaccountable for their credit decisions and borrowers unprotected and powerless as they cannot correct data or appeal decisions. Hence, Dr Nicolas provided suggestions for regulators, fintech and financial players to address these gaps by negotiating a decree that would update credit law and find a balance between innovation and public interest. This decree could limit what data is collected, how it is used, under what circumstances it can be transferred, stored and sold, and whether and how data should be shared with the National Credit Information Centre of Vietnam. The decree could also ensure that consumers can oversee the scoring process and have the right to explanation and appeal in case of rejection, and credit data correction. Additionally, he suggested amendments to the law on personal data protection to be made as well.
In conclusion Dr Nicolas opined that it was too early to evaluate the impact of this technology in Vietnam. The discussants then shared their opinions on the issues that Dr Nicolas raised from his research. Dr Linh Chi provided her account of how ACS might be regulated by the Vietnamese government. Dr Eric commented that banks might face challenges in adopting ACS entirely due to strict regulations. Lastly, Dr Elliott discussed the various issues surrounding the use of ACS.
The webinar concluded with Dr Nicolas and the discussants answering several questions posed by the audience. Some of the questions discussed the kinds of ‘inclusion’ the models mentioned imply and whether the right to participate in the economy becomes conditional upon ‘correct’ behaviours, and if ACS can inhibit local economy’s potential for long-term development.