Data is undeniably the foundation for driving accurate current expected credit loss models. For financial institutions that must comply with CECL and lack meaningful losses or loan history, it can be overwhelming to determine where to start. While 2023 can seem distant, community financial institutions must assess and resolve data gap challenges now. “It takes time to dig into data,” said Zach Langley, an Analyst at Abrigo, during a recent webinar, CECL in 2023: An Analyst’s Perspective. “It takes time, people, and a lot of resources to identify gaps, figure out how to bridge them, and do it. Doing that under a time crunch can make the situation more stressful.”

Why is data such a big deal? Unlike the incurred loss model, which looks at the portfolio as it exists today, CECL incorporates an expected, forward-looking component. Financial institutions must have enough historical data from a range of economic conditions to cover any point in the credit cycle, therefore greatly increasing the amount of data needed.

The accounting standard explicitly states the following:

  • Historical credit loss experience of financial assets with similar risk characteristics generally provides a basis for an entity’s assessment of expected credit losses
  • Historical loss information can be internal or external historical loss information (or a combination of both)

While CECL still allows for a qualitative component, the reality is that the institution’s reserves are built upon historical data that are more defensible and supportable to regulators and auditors.

Leveraging benchmark data

First and foremost, financial institutions should take data inventory and perform a gap analysis to get a comprehensive understanding of the institution’s data needs are for the methodologies it is planning to use for CECL. Smaller financial institutions may struggle to obtain the amount of extensive data required to estimate and justify reserves. Whether an institution has zero losses, lacks meaningful historical data, or has a new portfolio (which is often missing loss data/granularity), many community financial institutions will find that they need a benchmark for CECL allowances.

At a minimum, historical data should cover at least a full credit cycle, encompassing a steady economic environment to a recession, through recovery, and back to a stable economy. Leveraging benchmark data to bridge the data gap helps community financial institutions to quickly and effectively supplement their CECL analysis with historical loan loss data.

There are a variety of third-party data providers available for financial institutions to leverage. Government-sponsored enterprises (GSEs), such as Freddie Mac and Fannie Mae, provide publicly available loan-level data going back to the early 2000s. While there may be costs associated with importing, aggregating, and formatting data for the institution’s use, the data itself is free of charge for financial institutions.

To incorporate more robust and supportable data points, community financial institutions can also use data available from government sources. This includes the Federal Reserve Economic Data, which has an extensive collection of data on all types of asset classes, economic conditions, and a variety of different sources of data. While it may not be as granular as the institution needs, it does help institutions defend their qualitative factors and serve as a benchmark for other models.

Financial institutions can also partner with third-party vendors for data benchmarks. For example, Trepp CECL Benchmarks provides over 20 years of CRE loan data. Financial institutions are able to leverage benchmark data to establish loss rates for mean reversion, validate model output, or justify qualitative adjustments. Users can group data by property type, vintage, and region, as well as LTV and DSCR. Using a third-party data vendor helps financial institutions to identify areas where their portfolios may outperform benchmarks, as well as areas that need improvement.

Benchmark data can provide a useful starting point for an informed CECL estimate at both the input and output level. To learn more about using benchmark data for CECL, you can watch Abrigo’s complimentary webinar, “Optimizing CECL: Moving from no losses to an integrated risk framework.”


Article Tags: