Banking Performance Insights

Stress-Testing Software: Choosing the Right Solution

Oct. 26, 2016


By Oleg A. Blokhin, Jack A. Gregory, and David W. Keever

Does Your Bank Have the Stress Testing Data You Need
Choosing the right software to manage mandated stress-testing requirements can be a difficult challenge for financial services organizations.

In addition to the complexities that are inherent with stress-testing requirements in general, the software selection process itself can add further complications. Despite the challenges, choosing an effective technology solution to help manage the complex data acquisition, analysis, and reporting processes is one of the most critical components of a successful response to mandated stress testing.

 

At a minimum, the solution that is chosen should be capable of producing a submission file that is compliant with Dodd-Frank Act stress-testing (DFAST) requirements. This includes DFAST-formatted reporting, along with executive-level views of the required nine-quarter stress-testing submission for baseline, adverse, and severely adverse scenarios.

Beyond meeting that basic requirement, management teams should be looking for specific capabilities as they choose a stress-testing software solution. The necessary and desirable features that should be at the top of any organization’s list can be organized into four broad categories.

1. General Characteristics

Banks should expect their chosen stress-testing solution to provide support for all aspects of the stress-testing process, including the initial assessment and road map planning, model and process validation, model development, internal audit supervision, data aggregation, and, ultimately, the production of an auditable Federal Reserve submission file.

Other general characteristics of a successful solution include flexible scenario management and analysis capabilities, as well as forecast production and supporting analytic reporting facilities that produce confirmable confidence at every step of the forecast preparation, review, and approval processes.

2. Independent Review Features

The system’s ability to produce integrated management overlays and make adjustments to forecasts with audit trails are critical concerns. Specific features to look for include:

  • Support for internal audit and examiner reviews, including detailed support for the management overrides and audit trails
  • Comprehensive role-based user access security and secured hosting capabilities that meet recognized information security standards
  • A detailed audit trail of user activities and actions
  • Automatic, systemwide identification and tracking of management overrides and calculation tracing steps
  • A document management library to support the various documentation needs

3. Model Development and Maintenance Features

Because stress-testing regimens are highly dependent on the use of sophisticated models, a stress-testing solution should feature a comprehensive model deployment and production platform, along with extensive and adaptable model library management capabilities. The model library should include:

  • Loan and deposit balance forecasts
  • Charge-off forecasts
  • Asset quality forecasts
  • Noninterest income and expense forecasts
  • Capital calculations
  • Other specific vendor-developed and bank-defined models

Models should be developed using statistical languages that can be easily configured for production and should be capable of linking all data inputs for the model, including nonaccrual for a broad range of models and calculations.

4. Systems and Data Integration Features

Data is the fuel the drives DFAST forecast production, so the stress-testing software must support robust loan and credit data integration from both internal systems and third-party sources. For most organizations, the scope of required data is substantial – typically 10,000 or more specific data elements are involved.

Internal data sources include:

  • Asset liability management (ALM) systems, which provide information needed for calculating stress-related net interest income
  • Allowance for loan and lease losses (ALLL) systems, which are used to estimate asset quality-related forecasts
  • General ledger accounts to provide updated balances and product-level details
  • Securities portfolio, credit portfolio, and credit data warehouse systems, which provide crucial asset quality information
  • Facility-level information that can affect variations in risk exposure by market or geography

External or public data sources also must be integrated into the system. Among the public sources or data systems that contribute to the stress-testing process are:

  • Federal Reserve macroeconomic data forecasts
  • Forecasts and analysis from other public agencies such as the FDIC and the Federal Reserve Economic Data (FRED) database maintained by the Federal Reserve Bank of St. Louis
  • Public DFAST disclosures and other DFAST-specific sources
  • Public call report and merger and acquisition data
  • Industry-specific data from private sources such as Compustat and SNL Financial

In addition, because the data is imported from many sources, both internal and external, it often requires transformation, reformatting, or translation to be usable. The stress-testing solution should be capable of performing such functions with minimal effort.

An Organized, Systematic Approach

Choosing the right software can help banks manage the complexities of DFAST compliance more efficiently and effectively, while contributing significantly to the organization’s overall risk management efforts. Although the many variables can make the choice seem difficult at first, an organized checklist of features can help banks prioritize the features and benefits that are most applicable to their particular situation.

Authors
Oleg Blokhin
Jack Gregory
keever-dave-150
Dave Keever
Principal