In tough times, institutions should capitalise on technological advances to shape decision making.
By Joelyn Chan
As Asian economies brace for a recession, the probability of loan default increases. Banks are planning to set aside additional provisions for loan losses, even though the risk will be short-lived. As a mitigation measure for 2020 to help borrowing affected by the global COVID-19 pandemic, some regional banks are allowing residential mortgage borrowers to defer principal payments and only pay the interest.
Singapore’s Housing & Development Board (HDB) is also exercising flexibility during the outbreak, allowing a three-month suspension of late payment charges on HDB mortgage arrears and the possible deferment of payment on loan instalments for six months.
As lenders search for ways to absorb the shock the novel coronavirus has sent through international financial systems, artificial intelligence (AI) solutions could offer a path through the trees.
AI enables more thorough borrower evaluations
According to Credit Bureau Singapore, there were 112 mortgage defaults registered in 2019. This number is more than double the 65 cases recorded in 2015. The number may further increase in year 2020 due to Covid19 and an impending global recession.
Singapore’s HDB 2018/2019 Annual Report revealed late payment charges on mortgage loans increased from S$28.3 million (US$19.7 million) in 2018 to S$31.8 million (US$22.2million) in 2019. With AI, there could be more routine checks on the lender’s ability to pay.
AI can trawl payment default rates, frequency of payment reminders sent and requests for refinancing to build an accurate picture of borrower reliability, reducing the chance of default.
These days, a borrower’s digital footprint can reveal more information on their reliability than traditional credit scores. With access to a rich data set, including social network profiles, AI solutions could build a more rounded consumer profile by curating information on social circles, spending habits and employment information. These data points could open access to credit and loans to borrowers who have been previously denied products based on narrow evaluations of their credit score.
Synergies between AI and humans should not be left untapped
Corporate institutions usually struggle to find data on first-time borrowers. At the same time, first-time home buyers are overwhelmed with the complexities of home loan application procedure from multiple providers. Therein lie untapped opportunities and room to improve the lending experience for both parties. AI can handle a large volume of queries and enhance customer experience via client personalisation. For exceptional cases, the customer could then be referred to a customer service officer.
Undoubtedly, AI has not reached a level where it can replace the human’s touch. AI algorithms can conduct extensive evaluation but require human supervision and evaluation.
When AI and humans work together, they can deliver optimal performance. AI can assist mortgage lenders to detect potentially fraudulent behaviour, while a human verifies the anomaly cases. For example, AI could run back-end checks to quickly spot inconsistencies or inaccuracies in the application information provided. For potential cases of identity fraud, a human will later intervene to physically verify the loan applicant. This lowers the cost of labour incurred, and the potential for fraud, which subsequently lowers the cost of borrowing for lenders.