Connecting the Data Dots for a CECL-Compliant Picture


Slated to go into effect in 2020, the Financial Accounting Standards Board’s (FASB) current expected loss standard (CECL) is widely acknowledged as one of the biggest changes in accounting and one that will impact almost the entire financial industry. This four-letter acronym will require banks to calculate the expected loss over the life of each loan and book that loss at the time of origination.


Currently, financial institutions calculate potential credit losses by evaluating a portfolio’s past performance and manage their loss reserves accordingly. Credit losses are recognized once incurred. For those still trying to wrap their heads around the significance of the change, the following analogy may help. Today’s incurred loss model is similar to an individual who avoids worrying about the future, viewing potential problems with a “we’ll cross that bridge when we get there” mindset, whereas the CECL standard is akin to a doomsday prepper, planning in advance for scenarios that may or may not happen.


While this is a simple explanation, CECL is a complex change that will have a substantial effect on how financial institutions of all sizes conduct business. Since CECL requires banks to book the loss allowance for the life of loans at origination, if the credit quality of a loan changes, the bank has to adjust the allowance, which means the bank’s income could take a hit. For banks with several thousand loans on the books, there is significant potential for income volatility. To minimize this volatility and ensure CECL does not considerably disrupt a community bank’s operations, it is crucial to start planning now, especially from a data collection standpoint.


CECL Success Hinges on Data

To be compliant with the new standard, banks need a consistent and proven method for collecting and analyzing loan data. This includes how it’s stored, what types of data is stored, such as charge-offs, recovery data and accrued interest, as well as the quality of the data.


However, the data requirements are proving to be a challenge for many financial institutions. In fact, Deloitte reports that 81 percent of banks cite data management and ensuring data are of sufficient length and quality as the greatest hurdles related to CECL. To accurately predict loss allowances, FASB calls for large amounts of historical data, from three years all the way up to up to 12 years. These necessary data elements will likely include historical defaults, attrition and recovery data, delinquency data, collateral information, prepayment data and macroeconomic variables, among others depending on the bank’s portfolio makeup.


Bridge the Data Gap

It is crucial that banks conduct a gap assessment to determine the availability and quality of data and which data elements they lack. This will require involvement from various stakeholders across the institution, including accounting, finance, IT, credit and possibly even the asset/liability committee. These employees will be able to provide their unique expertise about how to locate, access and collect necessary data within their respective departments. Banks must focus on getting their data in order now because it is the foundation for CECL compliance. Adequate and accessible data helps banks understand how their portfolios perform in all types of economic markets; identify when there is a downturn in credit quality; determine what drives portfolio profitability and what hinders it. All of these factors must be considered when calculating loss allowance.


After assessing the availability of internal data, banks should evaluate strengths and weaknesses within that data. By doing so, inconsistent formatting and the range of variables captured can be corrected and standardized. For example, when inputting address information, does your bank use “St.” or “Street?” What about ZIP codes? Do you use five- or nine-digit ZIP codes? Establishing a standard protocol for such variables will make it easier to use and segment the data to quickly identify trends.


Tightly Segment Portfolios to Identify Trends

When it comes to segmenting the portfolio, loans should be organized based on the type of loan, such as auto, small business or credit card, and the term of the loan. From there, each category should be segmented even further so that each portfolio segment is comprised of loans that will perform similarly through different economic cycles. Economic cycles generally range between eight and 15 years in duration, so you can imagine how much data is needed to achieve this accurately.


Collecting the right credit quality indicators, such as FICO scores, debt service coverage and loan to value ratios, for each loan at the point of origination and throughout the life of the loan will help banks understand how different portfolios perform over time and which factors influence their performance.


Some credit quality indicators may be more predictive than others for certain types of loans. For instance, with small business loans, a bank might find that no deposit activity in the last 60 days is a highly predictive indicator of a problem. Or for an auto loan, a bank may realize that an overdrawn account is an accurate indicator.


These indicators should be collected on a regular, ongoing basis and as credit quality changes, the bank can track and quickly identify what is causing those changes. In other words, banks can understand which economic factors each portfolio is sensitive to. This requires cross-analyzing external data, like interest rates, housing values and unemployment rates, against changes in a portfolio’s credit quality. For some banks, national economic data will be useful while for others, regional values will work best – it just depends on the banks’ customer base and portfolio makeup.


Banks should start analyzing their existing data now to gain a better understanding of their loan portfolios and what influences their performance. Ramping up efforts to gather more data is also advised because banks will choose to leverage loss estimation models based on the availability and quality of their data. With greater amounts of accurate data, the bank has more flexibility in selecting which loss estimation model to use, resulting in the most predictive loss allowance calculation.


Consider Data Storage & Security Needs

With the new CECL standard quickly approaching, institutions should also plan for the massive amounts of data that will need to be stored and securely backed up. This involves assessing data architecture to ensure the bank can support the standard’s increased data requirements. Additionally, disparate sources of data that are scattered across the bank’s business lines should be integrated and stored in an accessible, centralized location. This will take time and require involvement from multiple employees, so the sooner banks get started on this, the better. Likewise, bank employees should be familiar with their institution’s data governance policies so that data ownership and quality expectations are clear as they relate to CECL.


The banks that proactively prepare for CECL and begin connecting the data dots now will not only be positioned for compliance, but also sustained growth and profitability. CECL stands to be a competitive differentiator for the institutions that approach the standard effectively and efficiently, as CECL will provide valuable insights on how credit risk can impact a bank’s profit and earnings.


In summary, connecting these data dots now will empower banks to go beyond compliance with CECL to gain a better understanding of how they can enhance their portfolios’ performance and positively impact their institution’s bottom line.



John Robertson is a Senior Business Process Architect for Baker Hill’s Advisory Services. In this role, Robertson provides guidance to Baker Hill’s clients on profitability, specifically with strategies involving risk-based pricing and relationship profitability. With 26 years of experience in the banking industry, Robertson assists banks in developing and implementing technology for commercial lending that improves the efficiency of the lending process and the productivity of the lending officers.