Loss rate forecasting framework based on macroeconomic changes: Application to US credit card industry

Area of Science:

• Economics
• Financial Modeling
• Data Science

Background:

• Credit card portfolios are significant assets for large U.S. banks.
• Managing charge-off rates is crucial for the profitability of the credit card industry.
• Macroeconomic conditions influence consumer debt repayment behavior.

Purpose of the Study:

• To develop an expert system for forecasting credit card industry losses.
• To utilize macroeconomic indicators for improved loss prediction.
• To provide practitioners with a holistic economic view impacting future losses.

Main Methods:

• Selection of macroeconomic indicators based on literature review and expert opinions.
• Development of an expert system framework using state-of-the-art machine learning models.
• Creation of two system versions differing in indicator lag selection strategies.

Main Results:

• Two models were developed: one using optimal lags (6 indicators) and another using all lags (7 indicators).
• Selected features spanned all three economic sectors (consumer, business, government).
• Achieved Mean Squared Error (MSE) of 1.15E-03 (optimal lags) and 1.04E-03 (all lags) using bank charge-off data (1985-2019).

Conclusions:

• The expert system effectively forecasts credit card industry losses by incorporating macroeconomic factors.
• The system offers valuable insights into the relationship between economic conditions and future credit card losses.
• Practitioners can leverage this system for better risk management and strategic decision-making.