Stress Testing - It’s All in the Data!

Stress Testing - It’s All in the Data!

Stress testing is here to stay. No matter what your asset size, stress testing must be incorporated into your bank’s risk strategies.

Regulators have definitive stress testing requirements for banks over the $10 Billion mark. The Dodd Frank Act required banks over $10 to comply with specific methods such as DFAST Macro and Economic Stress testing plus CCAR testing for banks over $50 Billion. However, banks under $10 Billion had no regulatory requirement to stress test but are being strongly encouraged (required) by their regulators to stress test based on their concentrations. Both the OCC and FDIC have published guidelines for this under $10 Billion segment but since there is no actual “formula” for testing, there is confusion over how to actually stress test their portfolios. Methods such as “Bottom Up” and “Top Down” are the most common methods being used, but what I hear most is how do I get the data I need, where is it, and is it accurate.

Data is the key to stress testing. However, data is not easy to consolidate and use for meaningful stress testing.  There are four key components to organize your data:

  • Identify Concentrations of the portfolio - This is typically the first step in determining if the bank has specific concentrations within their loan portfolios. Running reports by loan type, NAICS codes, collateral types, etc. will determine what the concentrations leading you to determine if there may be potential for risk.
  • Identify stress testing models - Based on your concentration findings, what do you want to stress for? Scenarios such as stressing by interest rate increase, appraisal value shifts, balance sheet and income statement items and many more should be considered based on specific concentrations in the portfolio.
  • Identify the data needed to support the stress - Whatever stress model you are using, identify all data fields you will need to support your stress models.
  • Where is the source of the data - While your core accounting system can provide loan data and collateral data, there will be additional data such as financial spreads which are on another system, rent rolls that may be in the form of a spreadsheet, appraisal or valuation information that may not be kept electronically.
  • Who is the owner of the data - The owner is important as the responsibility for the accuracy of the data lies here. One of the most common mistakes is NAICS codes. There may be a description of the type of business within the loan documents but it shouldn’t be personnel guessing at the code at the time of booking. Evaluate the data elements and determine who is responsible for the accuracy of the data.
  • Efficiency - This is the last component to stress testing. As your institution grows, it will become harder to manage the data needed for your reporting. Today most community sized banks are utilizing spreadsheets which require manual input as the required data is in many different forms. Now is the time to think about databases that will consolidate the required stress testing data elements and reporting can be accomplished directly from the database.

Take advantage of the data that is currently in the databases your bank currently utilizes for loan accounting, loan origination, and any other system that captures data that can be utilized for a stress test. The Baker Hill NextGen® platform allows you to capture financial statement, loan and collateral data. In addition, Baker Hill NextGen® has a report writing tool that allows you to create your own stress testing reports and allows you to document your stress scenarios that will save you hours of time and manpower in addition to more accurate reporting.