4 minutes
Sometimes it’s part of a conversation about preparedness for the Financial Accounting Standards Board’s proposal for managing Current and Expected Credit Loss. Other times it’s brought up when talking through a more proactive approach to risk management under existing regulations. But the topic of accessing and analyzing “loan-level data” to better understand loan risks repeatedly comes up during our discussions with banks and credit unions alike.
Credit unions face a distinct challenge when trying to look at loan-level data in that, generally, their borrower data is stored in several different core processing or decisioning systems, making it harder to gather into one place. CUs also typically have fewer employees who can focus on data management than their bank counterparts.
Credit unions can gather loan-level data using one a variety of methods.
For CECL specifically, it’s likely that a credit union will need “life of loan” data to accommodate the forward-looking calculations. A way to capture this information is by using a “limited method,” in which the credit union uses data already stored in its core and decisioning system(s). Often these systems store data for up to 13 months. While this method provides a start, 13 months of data may be limiting if the institution is trying to distill life-of-loan insights.
The next option is similar to the first, but takes data collection a step further. With a “static method,” the credit union captures one-time archives of data from each of the different systems. These archives are stored in an accessible format so that, through spreadsheets, the risk management team at the credit union can manipulate the data for insights into the portfolio. With this model, the credit union takes responsibility for pulling the archives as well as storing the data in a query-able format. For example, PDFs with data tables probably will not help the credit union achieve its reporting goals.
In the third model, a “dynamic method,” the credit union uses a vendor to create a data bridge between the institution’s core system and a flexible data storage and reporting solution. For the purposes of the reserve, it could be an automated allowance for loan and lease loss solution, such as Sageworks ALLL. With this type of data bridge in place, the CU would not only benefit from real-time data aggregation and backup, but could also leverage that data in defensible and automated calculations for the allowance. Instead of gathering spreadsheets and linking them together through formulas, the data bridge and ALLL solution would gather data automatically and perform the calculation.
This third method certainly introduces the most change for the credit union because it requires that the institution set up the data bridge or integration between the core system(s) and the ALLL solution. So while it would bring added loan-level data benefits, it may or may not be appropriate for every institution.
Once the credit union can, using one of these methods, access historical loan-level data on its portfolio, its risk management team can leverage such added benefits as:
- more defensible, documented ALLL calculations
- easier process for balance reconciliation as part of the ALLL;
- less subjectivity in forward-looking assumptions under CECL;
- ability to enhance the ALLL with more sophisticated loss rate calculations, including migration analysis or probability of default/loss given default;
- opportunity to create different ALLL (and stress testing) scenarios to see the impact model tweaks might have on the reserve – this includes testing how CECL will impact capital;
- backtesting to validate the model’s accuracy over time; and
- better portfolio reporting to understand risk.
Loan-level detail for the portfolio enables a credit union to track changes in loans from each segment to determine how the overall portfolio is shifting. Is there substantial growth in one concentration? Should the credit union focus on another area?
Similarly, delinquency rates by concentration will give greater clarity into where portfolio risk may be increasing. If pockets of risk are uncovered, the credit union can do something (e.g., change loan pricing models, adjust risk appetites, better monitor troubled debt restructuring activity) to mitigate risk moving forward.
The credit union can also review loans by segment to see which loans provide the best chance of obtaining a recovery based on historical data. This allows the institution to potentially reallocate resources, including workout officers or special assets officers, to maximize dollars returned to the credit union.
In all, there are several ways a CU can leverage loan-level data to better understand the loan portfolio, its risk and its opportunities. And there are also several ways through which the CU can begin to gather this information. Regardless of the method chosen, credit unions would benefit from starting the data collection process early, in advance of CECL’s implementation in the next few years.
Libby Bierman is financial analyst at Sageworks, Raleigh, N.C.
