Artificial intelligence is offering a widening range of tools and processes to help banks manage their credit portfolios in the face of rising economic uncertainties and market volatility. It is also helping banks to better cope with tightening regulations.
The Hong Kong Monetary Authority (HKMA), for example, has proposed that the implementation of the Basel III final reform package, a set of measures that seek to strengthen the regulation, supervision and risk management of banks, would start as soon as July 1 2024. In addition, IFRS 9, the latest International Financial Reporting Standards, require financial institutions to recognize expected credit losses (ECL) on their loan portfolios.
All this increases the complexity and rigor of credit risk management, requiring financial institutions to adopt more sophisticated risk management practices and tools.
Some Asian banks are working on improving their measurement and monitoring of credit risks through early warning systems (EWSs), which help identify potential material credit deterioration or defaults at an early stage. These allow institutions to detect potential risks, take proactive measures, and minimize losses by addressing the deterioration before it escalates, according to Moody’s Analytics.
A robust early warning system helps to build the resilience of an institution by providing warning signals in time for preventative measures to be taken.
Traditional EWSs usually require a large number of experimentally defined indicators and rely heavily on expert judgment. AI, on the other hand, excels at discovering patterns based on large volumes of high-velocity data that can be used to generate credit default signals.
With sufficient computational power, AI algorithms are capable of generating early warning signals using indicators from a wide range of sources, as well as increasing the accuracy of indicators, Deloitte says in a recent report.
AI models are engineered to retrieve, collect, and unify large datasets from diverse sources such as sentiment/news feeds, know your customer (KYC) data, forward-looking macro indicators, and cyber, supply chain, climate, and ESG scores, Moody’s Analytics explains.
Banks are starting to leverage generative AI to transition from information gathering to decision-making, and to enhance workflows in assessing credit portfolio risks under various lenses to help prevent losses from default events. For example, with natural language processing (NLP) technology, a range of output from social media posts to financial news can be captured and used in credit analysis, something that has traditionally been performed by human analysts, according to Deloitte’s report.
Although AI has many benefits in credit risk management, it also poses some challenges. "The challenges of AI-based assessment tools include a few things. First, we need a large amount of high-quality data for training. Additionally, there are interpretability and transparency issues in AI algorithms, as well as regulatory compliance concerns," says Danny Lam, industry practice lead for banking at Moody's Analytics. "High-quality means trustworthy information. It is not something obtained from the internet without validation. The source is directly from vendors or client's website, not from other internet sources."
