时间：2015年12月9日2:00-3:30 地点：国发院万众楼一楼小教室 主讲人：邢海鹏，(SUNY, Stony Brook)   Title: An early warning model for credit market instability using firms' rating and macroeconomic variables   Abstract: Recent studies have shown that firms' credit rating transition  process is not stationary and may have structural breaks. To study the predictability of structural breaks, we develop a predictive model for  latent structural breaks in firms rating transition dynamics, using historical records of (high-dimensional) economic and market fundamentals. As a large number of economic and market variables are sometimes involved in the study, we also introduce an inference procedure that select and estimate important economic factors at the same time from the high-dimensional factor space. Based on an empirical study using the U.S. firms' credit rating transition records and the history of economic and market variations from 1986 to 2013, we demonstrate that not all structural breaks are black-swan events and some of them can be estimated and predicted up to certain extent. The proposed model itself is also an tractable example in big-data analysis on combining large scale of micro- and macro-economic data for the prediction of macroeconomic events.