Extracting U.S. Credit Cycle Using Dynamic Linear Model

• Activity Number: 240
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Jie Chen*+ and Agus Sudjianto
• Companies: Bank of America and Bank of America
• Keywords: credit cycle ; leading indicator ; multivariate dynamic linear model

Abstract

The recent economic crisis driven by consumer credit problem motivated us to study the dynamic of credit cycle. Of particular interest is the ability to obtain a leading indicator to credit default that can be used as early warning and effective risk mitigation for commercial banks. In this paper, we study the common dynamics of consumer revolving balance, personal consumption expenditures, personal deposit and credit card net charge-offs. We employed a new multivariate dynamic linear model to obtain a leading indicator of the US credit cycle. The model consists of local trend and common cycle components, and is estimated by maximum likelihood method. The cycle component with high order has been proved to have band-pass filter property. Cycle shifts for individual time series are also incorporated in the model, allowing us to obtain leading and lagging relationship automatically.