As Bank of Korea Governor Shin Hyun-song said in his inaugural address that the central bank would "make more active use of market price indicators to strengthen early warning functions," the Bank of Korea has begun work to enhance the features of its Machine Learning-based "financial and foreign exchange early warning model" developed two years ago. The central bank is considering reviving an artificial intelligence (AI) model that was developed two years ago and effectively left unattended, and using it for early warning operations.
The Bank of Korea reviews the overall financial system and publishes the results in the Financial Stability Report, which comes out every six months. In this process, it analyzes various market indicators such as stock and exchange rate fluctuations and the spread between long- and short-term bond yields, but it has faced criticism that this is insufficient to determine whether the changes signal a future financial crisis.
The Bank of Korea developed the financial and foreign exchange early warning model to improve predictive power. The model analyzes vulnerabilities in three institutional sectors—bonds and stocks, banks, and foreign exchange—on a monthly basis. When monthly indicators are entered, it shows the probability of a financial crisis six months later as a number (an alert index). According to simulations conducted at the time of development, the model detected the 2008 financial crisis and the 2022 Gangwon Jungdo Development Corporation rehabilitation filing (Legoland) incident early. The index rose starting three months before the incident, signaling the crisis.
The Bank of Korea plans to verify whether the model can still predict financial crises by adding new data. The current model signals stress by having the Composite Financial Pressure Index (CFPI), which it developed in-house, rise as risks increase, and the bank is also considering strengthening the model by adding other variables.
A Bank of Korea official said, "Assessments of vulnerabilities in the financial system focus on the asset soundness of financial institutions," and noted, "It is understood as an effort to strengthen a framework that monitors financial stability conditions more quickly by using market indicators."