Click to Login

Stress Testing

Tue Sep 11, 2012

We have encountered quite a bit of ambiguity and confusion around the term "stress testing" as we have spoken with bankers. Only recently has it become more clear which banks are required to implement stress testing exercises or regimens to the community as a whole. As one might guess based on our previous posts, we believe that every bank could benefit to some extent from the analytical exercise in stress tests, and our goal is to help banks make the process less tedious and as productive and efficient as possible.

Based on our discussions with regulators, the purpose of their "stress testing" requirement is to motivate a bank to determine where its correlated risks lie, and to document what measures it has taken to minimize the potential for the bank to be impacted (to the potential point of insolvency) by a single, non-obvious external event over which the bank may not have any control. For instance, that event might be an asteroid from space crashing into the city of Reno, NV or the unexpected closing of a correspondent bank. In either case, regulators want a bank to understand how the ripples from these events will affect the bank's overall position.

Our framework for thinking about these situations is based not on rooting our the correlations, but characterizing the effects. Specifically, we characterize a stress test as being the impact of the bank's default rate for specific sets of its financial instruments (presumably its assets). Here's how it works. First, when we import a bank's portfolio into our database, we expect the bank to provide us with information about how it groups its financial instruments (e.g., "Municipal Accounts", "Home Equity Lines of Credit", "Treasury Bonds", etc.). Then, to create a stress test, we ask that a bank define and name a specific scenario (e.g., "Medium Crisis", "Failure of All C&I Loans in Oklahoma", etc.).

After defining and naming a stress test scenario, the bank can identify three items:

  1. Which group(s) of assets are affected in this scenario?
  2. For each affected group, which asset-holders would be impacted to the point of failure first, and which would be affected later? In other words, what attribute of the asset holder would determine when the asset would be defaulted on (e.g., loan size, outstanding loan balance, credit score, etc.)?
  3. For each affected group of asset holders, when would their defaults first be felt by the bank relative to the beginning of a simulation, and over what period during a simulation would their defaults be seen? For example, a bank might anticipate that its asset holders have a cash reserve (in proportion to the attribute mentioned in the previous question) that would be depleted over the first two to ten months of the simulation.
The purpose of this exercise is not to place an undue burden on the bank's management team by collecting an exact understanding of the amount of cash that all C&I loan holders have at their disposal, but to motivate the planning for contingencies and the up-front mitigation of risk by defining likely risk scenarios and using a financial simulation tool to highlight how long the bank would survive if such a scenario were to occur.

Capitalytics makes stress testing easier by allowing a bank to specify the scenarios that are of interest -- assumably the most realistic potential risk scenarios, and those that may be discussed by the bank's Board of Directors on a regular basis -- and then having these scenarios run regularly and automatically. The output that we produce is intended to fuel periodic discussions about how to manage and mitigate risk so that the bank is more prepared to react when issues do arise.

Contact Capitalytics today to tell us about your bank, and how we might be able to help.