Menu

CECL model: Build or buy?

David Kistler
Posted by David Kistler
LinkedInTwitterGoogle+PrintFriendlyEmail

Compared to existing ALLL requirements, Accounting Standards Update 2016-13 (CECL) will require more inputs, assumptions, analysis and documentation, making the option to automate the process significantly more attractive for many institutions. Regardless of an institution’s approach, key consideration areas for a CECL-ready model include data management, contractual life, segmentation, methodologies, forecasting and adjustments, documentation, product evolution and enhancements, training, advisory services and technical support. 

The following recommendations are not intended to replace regulatory guidance as it relates to vendor management. Rather, the intent is to provide an understanding of a CECL-ready model and the critical elements that such a model should contain for tactical due diligence.

Data: Data serves as the foundation for historical loss experience calculations and represents a significant threat to a successful implementation. A solution provider should offer a supportive and effective integration process with institutions’ core providers and other data sources. The provider should help identify critical loan-level data gaps, provide remediation support for historical data inconsistences as well as future loan-level collection, all while providing adequate and accessible data storage.

Contractual Life: The expected life of each segment and/or prepayment behavior will need to be calculated within the solution.

Segmentation: Institutions must analyze assets on a collective or pooled basis unless unique risk characteristics exist. Designated pools should be relatively granular while maintaining statistical significance. Vendor models should easily illustrate pool size and offer a variety of segmentation and sub-segmentation elections based on customizable and dynamic logic. The model should allow institutions to automatically and manually identify and separate loans for individual analysis.

Methodologies: Vendors that do not provide for a host of methodology elections today can inadvertently limit your options for tomorrow as particular data issues will be made apparent via execution. Methodology options should include vintage analysis, probability of loss and loss given default (PD & LGD), migration analysis, cumulative loss rate and discounted cash flow (DCF).

Forecasting and Adjustments: Vendor models should allow for swift inclusion and exclusion of all observable analysis periods and provide forecasting intelligence, support and application.

Documentation: The new standard will require more inputs, assumptions and analysis at the pool-level; tracking, consolidating, reporting and displaying all information necessary to review, support and recalculate is a critical function of any vendor-based solution.

Product Evolution and Enhancements: Institutions should have a clear understanding of the product roadmap and the contractual obligation to remain GAAP compliant.

Training, Advisory Services and Technical Support: Understanding the availability, expertise and fee structure for training, advisory services and technical support is a must. A vendor should have internal experts that have experience working for and with financial institutions and regulatory bodies.

To learn more, download the complete CECL solution buyer’s guide: web.sageworks.com/portfolio-risk-buyer’s-guide/

tags: ALLL, allowance for loan and lease losses, CECL, CECL model, portfolio risk management