In recent times, stress testing has acquired statutory status the world over and in US it has become an integral component of Comprehensive Capital Analysis and Review (CCAR) guidelines, aimed at maintaining fiscal prudence in US banking sector. As per CCAR guidelines, banks should be able to create a response mechanism in case an economic distress were to occur. Hence, it is of utmost importance that results are actionable from Capital Planning and Customer Lifecycle Management (Acquisition and Account management & Collections) perspective.
The current approaches to stress testing are applied with more focus on capital requirement. This is an outcome based approach rather than an approach which focusses on basic drivers of such an outcome. This limits the ability of stress testing results to be “actionable”.
The primary approaches to stress testing are Sensitivity Analysis and Scenario Analysis. Application of both these approaches, under capital computation focus, captures and quantifies movement of portfolio across risk grades.
There is a gap between regulatory requirements and outcomes of existing approach to stress testing. This gap relates to the ability to make stress testing output actionable for the bank from a customer life-cycle management perspective. In other words, while banks have a good grip on how much loss they are to likely face, the ability to nuance the differences in behavior is not that advanced.
At Axtria, we believe that behavior-based risk transition model can be built to overcome the above challenges. This typically will involve the following 3-step approach:
- Segment customers on the basis of their behavior
- Record the transitions in customer behavior and assign probabilities using Markov chain
- Model relationship of these transitions to capture impact of relevant macro-economic variables