What data is needed for stress testing
For credit risk stress testing, financial institutions can choose from several different approaches, according to regulatory guidance. The most common approaches to stress testing are bottom up or top down. For many financial institutions, bottom up stress testing may be the more useful exercise, particularly for institutions with high levels of commercial real estate.
For stress testing a loan portfolio using a bottom up approach, a number of basic data fields are required, so data can be a big challenge with stress testing. Several of these fields, including basic loan information such as call code, loan type, property type, risk rating and location, are already being captured in the institution’s core accounting system. Collateral information like appraisal value and collateral type may also be held in the institution’s core. (A full checklist of data items usually used in bottom up stress testing can be found at the end of our whitepaper: Solving Data Challenges of Loan Portfolio Stress Testing.)
The following image shows how one stress testing software imports loan data fields from the core, allowing users to select different filters to build a concentration for bottom up stress testing.
In addition, banks will need to have at least a minimum amount of financial data in order to perform a meaningful stress test on sensitive concentrations like commercial real estate. Examples of financial data a bank will want to include for this type of stress test are property appraisals, rent rolls, cap rates and property net operating income (NOI). Many institutions will have this type of information in their credit files, but it isn’t typically included or updated in the core system.
Once the loan and financial data is updated and included as part of the stress testing process, the institution has the ability to run various scenarios with different levels of potential risk factors and apply them to concentrations within the loan portfolio in a bottom up approach. The FDIC suggests that common risk factors include:
• debt-service coverage
• loan-to-value ratios and capitalization rates
• property net operating income
• collateral value depreciation (regional and local)
• CRE sector performance (office, retail, multi-family, warehouse/industrial, lodging)
• interest-rate levels on variable-rate loans
• contractual terms that may introduce refinancing or repayment risk
• occupancy status
• lease rates
• unit absorption rates for real estate developments
• economic factors such as changes in local employment and house prices.
To identify stress factor changes that should be used for various scenarios (i.e. worst case, medium case, best case), financial institutions should try to tie the percentages to real data from past experience. For example, look at the time horizon when they experienced the highest losses and measure the change in certain risk factors like collateral values, vacancy rates, NOI, personal income, etc. Use this data to then develop additional scenarios that apply in better or worse economic environments.
For the results to be meaningful, the financial institution will also want to include key call report information in identifying how the scenarios could impact the bank’s balance sheet, income statement and key performance metrics like risk-based capital ratios and allowance to total loans.
Macroeconomic data for larger institutions, as well as regional and local economic data, is also helpful when performing stress testing, as it provides a means of quantifying the impact of various events on the loan portfolio and creating various scenarios with realistic assumptions. For example, in economic downturns, how much have property values declined in the bank’s region? By looking at NOI based on the vacancy and rental rates, how much has cash flow deteriorated?
To learn more about stress testing, download our recent whitepaper, “Solving Data Challenges in Loan Portfolio Stress Testing.”
Sageworks Stress Testing helps institutions perform bottom up stress tests on their riskiest concentrations. Watch a demo to find out more.