Digitizing Stress Testing

Running the stress tests now required of banks is tedious and expensive.

But by digitizing — already common in other areas of banking — the quality of operations can be improved and cost savings introduced. Better and more efficient risk management — and easier management of the concurrent regulatory requirements now in place — is an area ripe for the digitization process. Below are some benefits of digitizing stress testing:

Quick Production of Vigorous Stress Scenarios

Building an overarching data-management system, with advanced technology and sophisticated analytics, is required. This includes the use of artificial intelligence (AI) to manage and produce vigorous stress scenarios — both quickly and routinely.

Intelligent Decision Making

In response to rapidly changing market circumstances, digitized operations allow for better data on which to make crucial decisions in the heat of battle. A greater number of company-run stress tests can be undertaken and more flexibility within scenarios can be achieved.

Stress Test Report generation for base , adverse and severely adverse scenarios

Digitized operations can also routinely produce the kind of annual stress-test reports required by regulators. Not only in the form required, but also with the kind of data that will stand up to scrutiny.

After the financial crisis of 2007–2008, stress-testing requirements became far more robust under the Federal Reserve’s Comprehensive Capital Analysis and Review (CCAR). Meeting the requirements without digitization can be labor intensive and consume a significant amount of a compliance department’s resources.

Potential Areas of Digitization in Stress Testing

A robust adaptation to digitization will create multiple layers of optimization, including the real-time visualization of outlays versus available capital, more precise sensitivity analysis, and resources better prioritized towards materiality of risk.

Data input, overlays, the creation of required reports (such as Y­14A and CCAR), and challenge-process documentation can all be well served by digitization. Output can be standardized with the use of custom-built templates. Core external data sources that are used as variables in stress scenarios can be automatically updated.

With some time and care, a transparent and easily defendable process is made a routine aspect of bank operations.

This is not to say that the digitization of risk and stress testing is simple. Test-and-learn methodologies must be profound and due diligence must be paid to preventing errors being introduced. Systems can’t go active until thorough testing and analysis have been carried out of the digitization process itself.

It has not been unheard of over the past several years for major banks — including those with significant resources — to have difficulties passing their annual stress tests. This has often been because models could not be defended with the correlated data that was used for the stress tests. Regulators would ask, “If scenario X occurs, what will happen?” and data could not be crunched quickly enough to defend the models being used. There was a lack of data agility.

Digitization makes for a better toolbox containing better tools. With a more robust core of data and AI tools on hand, revenue forecasts and capital reserve scenarios that are subject to testing can be played out far more quickly and with much more confidence. Regulators are not left with vagaries, but with robust setups that are backed up with precise data.

It should be noted that regulatory agencies such as the U.S. Securities and Exchange Commission (SEC) are increasingly using their own digitization processes to create machine-learning algorithms. These are used to detect fraud and create in-house testing scenarios.

A significant database of prior stress-test reports is also being built. This allows the SEC to analyze the stress-test reports they receive and to identify those that appear to be outside the normal range of required capitalization. It would be wise to keep up with the abilities of one’s regulators.

The digitization of stress testing also fits nicely with banking institutions’ desire for the kind of big data that is used in short- and long-term internal planning models.

Finally, the long-term cost savings that can be achieved by digitizing regulatory functions can free up resources for other purposes, including countering the never-ending threats posed in cyberspace.