TY - JOUR T1 - Auto ABS Loan-Level Data: Improved Risk Analysis JF - The Journal of Structured Finance SP - 31 LP - 42 DO - 10.3905/jsf.2022.1.130 VL - 27 IS - 4 AU - Yini Yang AU - Joy Zhang AU - Jiawei “David” Zhang Y1 - 2022/01/31 UR - https://pm-research.com/content/27/4/31.abstract N2 - In 2014, the US Securities and Exchange Commission’s Regulation AB mandated loan-level data disclosure for public auto loan asset-backed securities (ABS). As a result, the loan level data from 2017 to 2021 display a rich set of loan-level variables that shed light on collateral performance patterns and improve risk analysis, especially for credit risk. Statistical predictive models that incorporate these loan-level drivers substantially improve the accuracy and granularity of default forecasts for auto ABS loans, and can provide benefits for investors and risk managers who use them. ER -