Deconstructing Stress Testing – Can We Straighten Out the Crop Circles? - Gonzobanker

 By Peter Cherpack

So why has stress testing been treated like the “crop circles” of banking since 2009? Just the mention of a stress test will induce finance and credit managers to either look the other way until the examiners force them to look at it, or cause them to shrug their shoulders in miscomprehension.

Stress Testing Classifications

There are essentially three types of stress testing regulators are looking for banks to perform today:

• Modeling interest rate stress and its impact on the bank’s balance sheet (the typical ALCO exercise);
• Capital stress testing, where losses are applied across the portfolios and remaining capital is analyzed to see if it is still sufficient and “well capitalized” (the Fed “SCAP” or “CCAR” model); and
• Sensitivity/Scenario testing for the impact of market change on segments of the portfolio (focused on CRE in community banks, as it is typically their largest and most volatile portfolio).

While these differing exercises are not specifically required by any banking regulation today, it is clear that it’s just a matter of time before they are.

If We Ignore It, Will It Just Go Away?

Having worked with banks that were attempting to make sense out of stress testing since 2007, I have had the unique opportunity to observe the continued evolution, confusion and concern over stress testing in the banking market. As a banker-turned-consultant, I have witnessed three types of approaches banks take to stress testing:

1. A large number of banks don’t do it at all. These banks are frankly waiting to be told to do it by the regulators. In most cases these are the weakest banks (with a short term view), some of the stronger community banks (regulators are more concerned about their weaker neighbors TODAY), or those banks that don’t understand the concepts or practices. My guess is this represents about a third of the banks out there. Assuming the weaker sisters survive, they will eventually be told to do something; stronger institutions not already stress testing will probably adopt the process as a risk best practice shortly anyway.  
2. A similar number of banks are doing “something” they consider stress testing using spreadsheets. They may be doing any of the different types of stress testing, but usually the process and assumptions are variable and undocumented. The process is often done without a base of firm credit logic, with little or no internal market analysis. As the FDIC and other regulators start to look at this more, these banks are likely to be given “suggestions” for improvement.
3. The rest of the banks have bought some tools to assist them in the process. It may be a standalone stress testing program or some module available from their core or spreading/credit platform. Unfortunately for many of these forward-looking bankers, due to the current lack of definition of stress testing processes and the lack of expert assistance available, they are often left to fend for themselves to leverage these tools as best they can. (Kind of like buying a car from a dealer and after the papers are signed being given a large box of tires, mechanical doohickeys and an instruction book.) These banks are typically the healthier survivor community banks that are risk management focused; they will likely expand their stress testing into all types over time.

If We Build It, Will the Regulators Come?

I spend a lot of time talking to regulators and bankers on this topic, and I’d say that all banks are now expected to do some kind of stress testing. To put a stake in the heart of the crop circle analogy, let’s look at trends in stress testing and see whether “if we build it, the regulators will come.”

From the 2006 Interagency Guidance on Commercial Real Estate Lending, this June’s Guidance on Stress Testing for Banks Over $10B in Assets, to the Dodd Frank stress testing requirements for larger institutions, we can all see the writing on the regulatory wall, right? So, for banks that haven’t been told to do it, asked to do it or even asked about it by the regulators by now, it’s just a matter of time.

At a recent industry event sponsored by the Conference of State Banking Supervisors, I got to ask an FDIC representative what his expectations are now around stress testing. The rep, while conceding that the FDIC has not been consistent enough in leading banks to apply stress testing, said that now that the most dire portion of the crisis is over, they want to see all banks set up some kind of repeatable stress testing process. He suggested that the field examiners will now be more focused on the Sensitivity/Scenario testing (CRE) for community banks, but the ideas around programs of capital stress testing for community banks are percolating in larger institutions today. The new guidance suggests that banks close to $10B will have to do it, and Dodd Frank clearly includes some kind of capital stress testing for bigger banks – though this is not totally defined.

So what’s a GonzoBanker to do? Here are a few pointers to help keep the regulators happy and manage the bank’s risk levels.

• Now is the time for all bankers to give serious attention to using stress testing and documenting a plan. The OCC, Fed and FDIC want to see some evidence of a program of stress testing for all banks, as a forward-looking risk management tool – though they are very flexible regarding what type and how it is done. This doesn’t have to be a detailed manifesto; it can be as simple as recognizing the value, discussing the resources and outlining the plan, with a definition of some key responsibilities and goals. The regulators would like to see this reviewed at the board level.
• Banks anywhere near the 300/100% CRE thresholds (and what community banks aren’t?) should be prepared to have their CRE stress testing processes reviewed. A spreadsheet process or automated program can satisfy the regulators – for now. As the regulations and “guidances” always say, “Apply this practice commensurate with the bank’s portfolio complexity and risk appetite.” Spreadsheets can be quite adequate for smaller, less diverse CRE portfolios and simpler balance sheets, though this manual process can be unexpectedly time consuming and error prone. Programs can be useful, but banks should conduct due diligence on the product’s capabilities and ease of use – many of the “add on modules” aren’t fully baked yet.
• Fully document what the assumptions are, what the process is, and what the results mean to the bank. Regulators are looking for transparency and clear risk based rules when the stress test process is defined. They do not just want to be handed a stack of reports, they want defined inputs, processes, and the assumptions and logic used in those processes. Banks should ensure their process includes sound risk management practices. The results should be provided to and reviewed by management and the board.
• Link the results to the bank’s capital, earnings and credit quality under “stressed” scenarios – and explain what bank management would do should these scenarios actually happen. This concept is key to regulator approval of the bank’s narrative report. The value of the stress testing exercise is to document what conclusions were drawn and what the bank would do about those conclusions if the stress conditions are actually realized. This is the “so what?” part of the exercise – and proof that the bank is actually using this process to proactively avoid risk.

Stress testing is clearly a growing area in bank risk management, with more definition and recipes for bankers to use coming in the near future. Yes, there is stress in the unknown, and gosh knows bankers have plenty of regulatory concerns on their plate right now that are easier for them to understand and address than stress testing. But establishing a stress testing program is something that can’t just be ignored today.

Peter Cherpack is a director of Ardmore Banking Advisors where he manages the Credit Technology Division. Contact him at pcherpack@ardmoreadvisors.com