Back-Testing Your First CECL Year

Is your institution's CECL allowance defensible? This vital paper, authored by ARCSys President and CEO Michael Umscheid, gives you the roadmap to confidently answer that question. Back-testing is the critical process of validating your Current Expected Credit Losses (CECL) model by rigorously comparing your forecasted loss and reversion rates against actual outcomes. The paper provides a clear framework for both short-term and long-term back-testing. You’ll learn how to analyze net charge-offs, prepayments, and the impact of Qualitative Factors (Q-factors) that could signal problems with your model's underlying data.

Don't wait for an audit to uncover model weaknesses. This guide addresses the common hurdles, warning that excessive Q-factors create a significant hurdle in back-testing and may indicate that historical data is not representative of expected results. The paper outlines potential model adjustments, such as reviewing model inputs and evaluating covariates, if actual losses are higher than expected. ARCSys is committed to helping your institution navigate these complexities. We offer expert back-testing services—along with solutions for benchmarking and sensitivity analysis—to strengthen your CECL model, ensure accuracy, and achieve full regulatory compliance. Read the full paper to empower your institution with the knowledge needed for a robust and supportable CECL process.