To strengthen the micro-analysis of household debt, the Bank of Korea has established a household debt database (Household Debt DB), benchmarking the Federal Reserve Bank of New York Consumer Credit Panel. The Household Debt DB is a longitudinal panel of individuals that tracks their credit information from credit bureaus at a quarterly frequency and is compiled in a statistically usable form. The sample population represents about 2.4%(approximately 1 million persons) of the total population engaged in credit activities in Korea. This DB is very timely in that the data at the end of each quarter is recorded after the passing of approximately two months. Its underlying data is also highly reliable, because it is based on actual financial transaction-related data, not survey results.
The Household Debt DB gives details on information related to the unique characteristics of individuals, such as age, address, credit rating, and income level; on financial transaction information, including records of loans and credit and check card usage; and on information on credit standing, such as loan and tax delinquencies and loan defaults. This DB allows us to make various micro-level analyses that could not be performed based on existing macro-data, since it includes individual-level information such as age, credit rating, income level and delinquency rate. Also, it allows analysis of debt structures, including debt type, repayment method and term structure, based on a wide range of credit information from an extensive sample exceeding 1 million persons.
Looking briefly into recent changes in the household loan structure in Korea using the Household Debt DB, we see that the share of mortgage loans has been on the rise, while the loan structure has been improving with the share of amortization mortgage loans increasing and their terms lengthening. Looking at the characteristics of household borrowers, we find that most loans are taken out by those in their 30s to 50s who are actively engaged in economic activities, and borrowing has increased particularly among those with high incomes and high credit ratings. Meanwhile, the delinquency rate of household loans has shown a low and stable trend.
The Household Debt DB is expected to be used more widely, since it has significant room for expansion by adding important variables and merging with other external data.