As we discussed in our most recent post
, the FDIC recently published an article outlining some stress testing principles for community banks in an article published in its supervisory newsletter, Supervisory Insights.
This article has major ramifications for community banks of all sizes, and in order to help community bankers understand key themes and the potential impact on their banks, Ardmore Banking Advisors will be holding a free webinar on Tuesday, June 26th at 2:00 PM EST.
If you are interested in attending this free event, please register here
: “FDIC’s New Directive on Community Bank Stress Testing: What Does it Mean For You?”
: June 26, 2012Time
: 2:00-3:00 PM ESTSpeaker
: Peter L. Cherpack, Senior Vice President & Director of Credit Technology, Ardmore Banking AdvisorsREGISTER
Last week, the FDIC published their Supervisory newsletter “FDIC Supervisory Insights Vol. 9, Issue 1 Summer 2012”. In that newsletter was an article by Bill Baxter and Tom Lyons of the FDIC outlining some stress testing principles for community banks “Stress Testing Credit Risk at Community Banks”.
This is, to our knowledge, the first publication from any of the primary bank regulators directed specifically to community banks below $10 billion in assets, and intended to help clarify regulatory expectations around stress testing. Based on our analysis, this article clarified and discussed a number of key themes:• CRE is still the most important portfolio for community banks to stress test and is what is expected. For growing banks who want to display better risk control, the process can be expanded to other portfolios.
• Stress testing is an industry best practice and will be expected in some way by regulators for all community banks that want to survive and grow. The key value of the process is in anticipating potential losses based on local market events.
• Follow key the basic principles, as regulatory expectations around the internal process for smaller institutions are not especially complex, or prescriptive as to specific methods or processes. The larger the community bank, the more useful more complex processes and software applications can be.
• Tactical Stress Testing Advice: Select the largest and/or most risky segment to stress, pick the largest credits in them. Focus on stress testing pass and pass watch credits
• The board needs to review the results of the stress testing process and may need to direct management to take mitigating action.
• Governance of the process is important. This supports validating a formal process that is auditable, repeatable and controlled.
• Stress testing is not just for difficult times, but also for when things get better. This underlines that the need for stress testing is not “going away” and can bring the bank value in many ways including helping banks grow with less risk.
In the article five types of stress testing approaches were described:• Transactional Sensitivity Analysis
– Stress testing individual loans to determine whether to make a loan and if so, formulate a more appropriate loan structure to mitigate credit risk.• Stressed Portfolio Loss Rates
- Use portfolio loss rates from a previous economic recession and apply those to their current portfolio can help identify the extent to which capital might be at risk given the bank’s balance-sheet structure and loan mix.• Scenario Analysis
– Evaluate how a certain portfolio or portfolio segment may respond to different levels in a key performance metric (e.g., cash flow or interest rates).• Loan Migration Analysis
– Evaluate how a downward migration in internal loan ratings, consistent with migrations that might be expected during adverse financial conditions, would impact asset quality and capital
• Reverse Stress Testing
– Assume a known adverse outcome, such as severe credit losses that reduce regulatory capital ratios to below satisfactory levels, and determine the loss event and associated circumstances that could lead to that outcome
It is our opinion that this article will be very helpful to community banks when approaching credit stress testing, and also reaffirms Ardmore Banking Advisors’ “Low Stress Stress Testing” approach for community banks. It makes clear that while stress testing is not a regulatory requirement for banks with assets below $10B, it is a simple to manage but valuable credit risk control within the reach of banks of all sizes.
Q: The OCC just published it's proposed rule making on stress testing for banks over $10B in Assets - per Dodd Frank. How is it different from the "Proposed Guidance" issues in June? Do you think that this will have any direct impact on Community banks as well?
A: This new OCC proposed rule for Stress Testing banks over $10B is similar to the "Guidance" offered in June, but it much more specifically defines how the OCC plans to make Banks over $10B conduct stress testing. I think there are some things we can gather from the OCC's proposed approach to the implementation of the stress testing process for banks over $10B - as it outlines how the OCC will apply stress testing as a part of the examination process for banks. While this rule is primarily centered on Capital Management, it is still instructive for banks of all sizes. Salient points they are proposing include:
- The OCC will supply a set of "Base line", "Adverse" and "Severely Adverse" scenarios to the banks to use for input to the test
- They suggest a repeating annual test over a nine-quarter planning horizon
- There will be a two month time period from when the scenario's are provided by the OCC to the institution and the time Banks have to respond with their stress testing results
- The OCC plans to coordinate the development of the annual stress test scenario's for all agencies
- These scenario's will be identical for all institutions
- The institutions will be required to calculate for each quarter end for the nine quarter period estimates of the impact of the scenario's on:
- Potential losses
- Loss provisions
- Regulatory capital
- Capital ratio's (referring to some of those established in Basel I and Basel)
- "Any other capital ratio's specified by the OCC"
- The allowance for loan losses
- Finally, institutions will need to provide reports and documentation, including policies and procedures to ensure that the stress testing is being done in accordance with the requirements of the OCC
- And it is stated that the results of these tests should be made public by the institution (on their website for example).
Questions the OCC poses for comment include:
- Should institutions be allowed to develop their own scenario's for stress testing?
- What are the costs/benefits of the process of stress testing?
- What level of detail should be in the scenario's? For example "macroeconomic factors" or more specific shocks to market and rate scenarios?
While these ideas are suggested for banks over $10B only at this point - it is clear that these are considered "best practices" in setting up the stress testing process and the types of things to test. Of course, most all of this is specifically oriented to capital adequacy and capital planning - and appears to be based largely on macroeconomic correlations to actual loan performance.
As is often the case in stress test modeling, community banks find local market conditions and local borrower financial issues much more relevant to their portfolio risk. So, all of this would have to be modified to be truly useful to smaller institutions - but many of the aspects of what they put out there are likely to become part of stress testing future guidance's for all financial institutions eventually.
Q: The regulators put out for comment in June 2011 a new Guidance on Stress Testing for Banks Over $10 Billion is Assets. Now they say that they are creating a new Guidance for community banks. In your opinion are the same best practices outlined in the first Guidance going to be applied now to Community Banks as well?
A: Based on regulator comments at the RMA National Conference in October 2011
I believe that there is an understanding amongst the Agencies that there is a fundamental difference between how larger banks and community banks can effectively approach stress testing.
In fact, representatives from all of the agencies commented at the conference that they intend to release a new Stress Testing Guidance specifically for banks under $10B - addressing needs more relevant to community banks.
It is understood that smaller community banks (those under $500M in particular) have issues with a number of the concepts outlined in the Guidance for banks over $10B including:
- The lack of diversity of products, market locations, portfolios
- Typically less capabilities in data management and credit systems
- Less resources including expertise in using credit models and technologies
- More reliance on local market influences than macro economic indicators and analysis
A number of individuals well placed in the agencies have told me and the industry that they don't expect community banks to run complex stress test modeling scenarios or systems, and are more concerned with CRE and other concentrations and how they create vunerabilities in community bank's portfolios. They also suggest that the new Guidance on Stress Testing for banks under $10B will focus more on the tactical needs of having forward looking risk analysis on CRE concentration and less on "frameworks" as outlined in the other Guidance.
Clearly stress testing is here to stay, but it appears that cooler heads will prevail to keep the "one size fits all" view of CRE stress testing implementations from harshly impacting community banks in the near term.
Q: I get asked frequently to dispel confusion around what the Regulators want community bankers to do in relation to stress testing their CRE Portfolio. There are some bankers that believe that they need highly complex involved models and processes to satisfy their examiners, while others believe that if they can "just make it through the exam", they can avoid any significant efforts.
I had a dialog recently with a highly placed regulatory official on this topic. What follows are some excerpts from that dialog. (They were sure to tell me that this information is NOT an official policy statement from any agency, and simply one regulator's view of what is going on in this area today).
A: Stress testing need not be a highly complex and mathematically intense practice and meaningful risk management information may be obtained from simple models and evaluations based on credit fundamentals.
All banks should recognize this (importance of stress testing) and really strive to embrace the process the better. This will most likely require serious evaluation of MIS capabilities and perhaps improving systems and processes to capture the key drivers(data points) for stress testing at origination. One of the other risk management processes that SR Letter 07-01 requires for high CRE concentrated banks is market analysis. Market analysis generally to be lacking in the community bank area and, its lack of utilization being part of the reason for such a large overhang of residential lots in many markets
Regulators expectations are that a firm's stress testing efforts will be consistent with the size and complexity of a firm's business. Not all firms will have to employ extensive econometric models.
A couple of other key points:
- Regulators do expect firms to start capturing basic risk information at origination, and to be able to use that information in performing sensitivity/scenario analysis not just at the loan level but also at the portfolio level. (For example a firm with a sizable CRE portfolio was unable to tell an examiner what their exposure was to Hotels in Florida).
- Regulators expect a firm to be able to support the level/type of stress testing they choose to do, and to justify all assumptions they use in the process, including in the development of specific stress scenarios.
- Ultimately, it is the goal of the agencies to have firms to produce stress results across all areas of the firm's operations, and to flow those results through the financial statements to assess the impact on capital. A Bank can no longer do interest rate stresses in a silo separate from credit stresses, etc.
- Stress scenarios and the results are expected to be shared with the Board of Directors on a regular basis. Board members should discuss the process, the assumptions, the results and their implications. Board minutes should reflect these discussions as well as any actions (or inactions) taken as a result of the stress tests.
Q: I understand that there is a new Interagency Guidance on Stress Testing out for comment, that will force community banks to use stress testing as a regulatory requirement. How will smaller community banks (under $10B) be impacted by this?
A: The Guidance is called "Proposed Guidance on Stress Testing for Banking Organizations with more than $10 Billion in Total Consolidated Assets" and explains how a larger institution should use the best practice of stress testing throughout their bank. It is specifically geared towards larger banks, and is more of an overview and suggestion of best practices than a regulatory requirement.
That understood, this new guidance offers some good information about the value of stress testing for all banks -- as well as outlining concepts such as "reverse stress testing" and "scenario testing". It also speaks to capital stressing.
My converations with regulators suggest that this should not be seen as an addtional burden on smaller institutions (as the very name of the Guidance would indicate) and in fact not as punitive for any instituion.
Follow this link to an article published in American Banker by me about the impact of the Guidance and Stress testing in general on community banks - first published June 17th.http://www.americanbanker.com/bankthink/stess-test-Cherpack-1039109-1.html
Q: I was thinking it would make sense to perform a 5% (Stress 1) and 10% (Stress 2) , and see what would the effects be to the bank. The stress level 3 would probably be around 15%.Does this make sense?
A: Actually, the answer to your question about whether a 5/10/15% stress makes sense is not simple. In our "CREAssist" service, we help the bank decide on what is appropriate and why. We don’t usually recommend a 5/10/15 step up stress, as it appears arbitrary and is hard to justify from a credit perspective. What we recommend is to look at a number of factors to create the stress amounts including:
- Actual market performance of the segements to be stressed (as local as possible)
- Stress modles to be applied to the type of properties/locations to be stressed
- Age and completeness of the financial data to be stressed (appraisals, cash flow, cap rates)
- Performance history of similar loans in the bank's portfolio
Of course, this is not a truly "scientific" process, and we use approximations and include subjective evaluation within the selection of the amount to stress by.
Q: I understand that CRE in general is referring to those types of loans where cashflow is coming from an unrelated third party or from sale of the property.
My confusion lies here - we do have lines of credit that are included in my CRE report, because they are secured by commercial real estate collateral, and used by investors, for real estate investment purposes. Am I supposed to exclude these from stress testing?
A: Lines of credit extended to developers are often secured by liens on real estate collateral. The lines typically revolve, that is, advances are made and repaid frequently during their term and the percentage advance is predicated on the value of the collateral.
Collateral can be single or multiple properties or even partnership interests if the latter can be valued. These lines are typically excluded from our CRE stress testing for two reasons:
(1) they are based solely upon the value of the collateral at the time of approval without regard to the operating performance of the collateral
(2) since the outstanding balances fluctuate, they do not amortize like traditional CRE term loans and you must focus on the line commitment, not present outstanding balances.
As a result, these lines do not lend themselves to CRE stress testing where we focus on possible changes in value to similar property types brought about by NOI and cap rate fluctuations. These lines are often not significant in number or in aggregate exposure to be statistically meaningful.
Q: The OCC was here and they audited our portfolio stress testing program. The only item that I need to improve on is “the stress test results need to be quantified by showing how the results impact earnings and capital.” How would we use our CRE Portfolio Stress Testing results to show that impact? I see in your “Outline of CRE Stress Testing Narrative Report” you say we need to comment on this. Please explain.
A. The following comments come from a recent conversation with an ex-OCC regulator, and is based on the assumption that due to the “what if” scenarios of the stress testing the results create potential loans that have significantly increased LTV due to lower valuation and/or reduced DSCR due to higher debt service:
After stress testing, loans that become exceptions based on increased LTV’s would migrate in the bank’s portfolio from pass to substandard or other downgrade categories – and would therefore require additional reserves. So the ALLL amount associated with these stressed loans would increase based on the increased ALLL loss factors.
Loans that become exceptions based on lower debt service coverage ratios after stressing would create a loss of bank income based on a reduction in interest income (increase in NPA’s and Non Accruals) – so that would impact bank earnings negatively that way.
Other increased costs/reduced earnings to the bank due to the creation of problem loans or exceptions as fallout from stress testing would include those associated with increased workout and collection costs.
The impact on capital is a little less precise to try and define as a direct correlation – as you would have to know how loss of interest income and increased reserves impact the capital ratios that the regulators want a bank to maintain (or stay within). Depending on the institution’s overall level of these ratios, a change in interest income and/or reserves may or may not push a bank over the capital ratio threshold – requiring additional capital allocations.
Q: I have a question about mixed use properties and their classification as CRE. My examiner told me to look at the square footage and use that as a way to determine if the property is CRE (based on non-owner occupancy) or considered as owner occupied real estate. This appears to be different from your statement that I should look at the source of the cash flow. What do you think of this ?
A: This question has come up before, as you noted above. I double checked with my sources at the regulatory agencies and they told me “The type of property is always determined by source of cash flow. This approach is provided in the Call Report Instructions and is consistent with GAAP”. It may be that the examiner was in this case looking at a multi tenanted property and equating square footage with cash flow income, but the bottom line is still the source of cash flow. Per the Guidance, if 51% of the cash flow from the property is from the non-owners, it’s CRE.