OCC Issues New Guidance on Credit Concentrations and Stress Testing

OCC Issues New Guidance on Credit Concentrations and Stress Testing

Wednesday, December 14, 2011

Yesterday the OCC issued a new version of their “Concentrations of Credit” booklet. The booklet directs examiners to include a page in each Report of Examination that lists concentrations posing a challenge to management or presenting unusual or significant risk to banks, collectively. Following are some highlights from the sections of the booklet along with my commentary but I encourage institutions supervised by the OCC to read the entire booklet (which for regulatory material is fairly short at 29 pages!).

From the Introduction:

“Credit risk management does not conclude with the supervision of individual transactions. It also encompasses the management of concentrations, or pools of exposures, whose collective performance has the potential to affect a bank negatively even if each individual transaction within a pool is soundly underwritten. When exposures in a pool are sensitive to the same economic, financial, or business development, that sensitivity, if triggered, may cause the sum of the transactions to perform as if it were a single, large exposure.”

I think this is a key opening statement. As an examiner and even now in my present role as a consultant I have always heard statements from bankers such as “we know our credits,” “we underwrite conservatively,” and “we know our market,” among others. This opening statement makes it clear that concentration management of pools is always necessary because there are risks that can affect a bank’s viability even when individual credits are structured and underwritten in a prudent manner.

From the Definition:

“Other credit concentrations, such as loans secured by first liens on residential real estate, have historically posed few problems. However, during the recession of 2007-2009, the banking industry experienced significant losses in these exposures when the national housing market suffered broad declines in home values. This experience indicates that although a concentration has not proven problematic in the past does not mean that it is precluded from becoming a problem in the future."

Hopefully we have all learned this lesson from this banking crisis. Historically home mortgages had excellent repayment histories. A review of the FDIC’s Quarterly Banking Profile reveals that the Net Charge-Off rate for closed end 1-4 family residential mortgages was as low as 0.04% during the first quarter of 2005 and was at similar levels from 2002 through the first half of 2007. This rate rose to as high as 1.81% as of the fourth quarter of 2009 and is currently at 0.97%. But there are still elevated levels of loan amounts in the 30-89 past due category indicating that there may be more to come. The lesson is here is that when analyzing the bank’s loan loss reserve analysis and a particular loan category does not have any historical loan loss don’t be fooled into believing that future experience will mirror the past.

From Pools of Transactions With Similar Characteristics:

The booklet lists the historical pools of transactions that may perform similarly but also goes on to state that industry practitioners and supervisors have further refined the original framework to include other sources of concentrations which augments and focuses the framework based on the results of stress testing. This can be very revealing of otherwise unidentified concentrations. The booklet lists several but I’d like to discuss commercial real estate which includes construction and development:

“CRE merits explicit mention because of its historical volatility and its role in a disproportionate number of bank failures. Banks may view CRE as a product, which would include all transactions secured by CRE. Alternatively, banks may also take an “industry” view, which would include only those transactions where the primary source of repayment is sale or refinancing of CRE or collection of lease/rental payments. A CRE pool may be further segmented by:
     * property type;
     * geography;
     * tenant concentrations (listed by name of tenant or by industry);
     * risk rating;
     * credit structure (e.g. fixed versus variable interest rate);
     *and debt service coverage."

As a consultant, the above list is a starting point for my analysis with clients. We also go on to include LTV, the Business Type of the borrower, relationship(s) as well as others. Also the list above is considered to be high level segments that can be refined into much greater detail. For example, property type historically has been limited to groups such as Retail, Office, and Industrial but the analysis doesn’t have to stop there. With the proper MIS or analysis tool, the property type of Retail can be broken down into more granular segments, such as Gas Station – individually owned and operated, Gas Station – individual owner of several stations, Hotel – Long Term/Extended Stay, Hotel – Motel, Hotel – Resort, Hotel – Unbranded, and so on. Risk factors for the higher level groups vary widely and grouping them together does not always provide the level of detail that institutions can benefit from.

From Identifying Concentrations:

“For very broadly defined pools such as CRE, the concentration limits would necessarily be higher than for more narrowly defined sub-segments such as acquisition, development, and construction loans. When banks set higher concentration limits for broadly defined pools-especially where those limits are more than 100 percent of capital-the OCC expects appropriate sub-limits for material groups of segmented exposures."

As we know many community banks especially are heavily concentrated in the broader category of CRE. However, the commentary above requires this group to be broken down into more refined segments, such as:

*Lots – held for long term speculation and lots for immediate development by type, residential or commercial, and then also by geographic area. If necessary the groups can also be further segmented by developer. A community bank shouldn’t want to hold a significant amount of investment in this category for the individual/company, same purpose, and in the same area.

From Stress Testing: 

“Stress testing is an effective tool for identifying correlated pools of loans. Stress testing can be used to quantify the potential impact from different scenarios on those pools of credits…It is critical to ask the ‘what if’ questions and incorporate the answers into the risk management process. Stress tests can reveal the kinds of events that might present problems…As the bank’s knowledge of stress testing grows, it should strive to make the analysis more robust by simultaneously stressing a number of related variables. Banks of all sizes will benefit by supplementing stress testing of significant individual loans with portfolio and firm-wide stress testing. The overall goal is to quantify loss potential and the impact on earnings and capital adequacy."

The footnote to this section states: Bank management may want to consider modeling software as it becomes more refined and readily available.

The 2006 Interagency CRE Guidance has been public information for quite some time now. The Stress Testing section in the new booklet not only reaffirms the recommendation for bankers to perform this exercise at regular intervals on a portfolio level basis using multiple variables simultaneously but makes it a requirement because examiners are instructed to “obtain and review the bank’s capital planning and stress testing policies, procedures, and results” when assessing the adequacy of a bank’s credit concentration management (Refer to Examination Procedures page 19).

Although this exercise can sound like a difficult task as a novice, it really is not. I think the availability of tools in this industry has progressed to a point that it is fair for regulators to expect this analysis even from the smallest community bank. Bankers no longer have to rely on the simple excel spreadsheet, which works for a single variable stress test on an individual loan, but falls short when trying to stress test multiple variables on a portfolio-wide basis.

From working with our customers continuously I have found that examiners are very pleased with the results when the banker can demonstrate that:

  1. this analysis can be performed at least quarterly because what value does this really have if the process takes six months to do and the report is using balances that are already stale-dated;
  2. the results are being quantified as to the impact on the institution’s capital and earnings under at least two stress scenarios;
  3. the credits displaying risk (or failing the stress test) are isolated with immediate action being taken by management to shore up the credit as best as possible; such as ordering a new appraisal; requesting updated financial statements; and performing a site analysis;
  4. groups of credits displaying the most risk are being evaluated in terms of:
    • the overall concentration levels as permitted by bank policy with the goal of reassessing whether or not this risk is acceptable to the institution or if the level should be reduced with action plans to be taken to diversify; and
    • assessing whether or not the loan loss reserve is adequately funding this group. Many times, especially in a young institution, the group of loans that display the most risk as determined by the stress testing results is a group that has not even exhibited any loss history to date. In this case is a qualitative factor adjustment being provided in the ALLL?

From the Conclusion:

“The size of a concentration, however, does not necessarily determine the risk. Different pools of the same size may represent very different levels of risk. Although 25 percent of capital remains the threshold for capturing concentrations for regulatory purposes, the OCC expects that institutions will build their concentration management process based on the risk that a pool of loans represents.”

There are also examination procedures contained in this document that should be reviewed. For the full text click here.

Add new comment