Find answers to questions such as: What makes an ERM program effective? How should it be implemented?
RMA Survey: Stress Testing and Scenario Analysis
Editor’s Note: The following executive summary of a study released in 2008 by The Risk Management Association is reprinted here with permission.
On Dec. 6, 2006, the three bank regulatory agencies — the Office of the Comptroller of the Currency, the Federal Reserve System, and the FDIC — released their final interagency guidance on commercial real estate concentration risk, called the “Guidance on Concentrations in Commercial Real Estate Lending: Sound Risk Management Practices.” One of the components of that guidance is that if CRE concentrations exist, banks will have to maintain more stringent risk management practices. One of the practices mentioned is to perform stress testing. The guidance indicated that the stress testing may be as basic as analyzing the potential effects of stressed loss rates on the CRE portfolio, capital and earnings.
Stress testing and scenario analysis of the loan portfolio is unquestionably a good risk management practice, but the debate continues as to what it is, how it should be done and what the regulators require of it. Lacking hard-and-fast answers, risk management practitioners typically find the debate quite frustrating.
To shed light on the subject, The Risk Management Association’s Community Bank Council, working closely with RMA staff, developed a survey to catalog stress testing and scenario analysis practices among smaller regional banks and community banks. The intent of the survey was to identify portfolio-level stress testing that is occurring within community banks. There were 316 completed surveys received and more than 50 percent of respondents represented institutions with less than $500 million in assets.
According to the study, stress testing is conducted in 42 percent of the surveyed institutions. The commercial loan as well as the non-owner-occupied commercial real estate portfolios are stress tested by the highest percentage of participants. The consumer portfolios, including residential mortgage loans and home equity, are stressed by 10 percent of participants.
Common stress tests conducted on the commercial and CRE portfolios include interest rate shocks, collateral value or loan-to-value shocks, and tests to determine debt service capacity and sales/margin declines. Unique tests to the CRE portfolio include rental rate and vacancy shocks, as well as cap rate changes.
The majority of participants perform worst-case assumptions during the underwriting or renewal phase of individual credits, with no portfolio-level testing. Staffing, knowledge of the local market, insufficient data and lack of an off-the-shelf stress-testing package were commonly cited barriers to implementing portfolio-level stress testing.
Stress testing can be used for many purposes, including risk management, capital planning, contingency planning, and communication with regulators and rating agencies along with other external parties. Many community banks struggle with how stress testing should be incorporated into the overall risk management framework and, equally important, how management should respond to the test results. In 78 percent of the organizations, senior management is either substantially or moderately involved in the process. Findings are shared with the board (54 percent) and shared with C-level management (36 percent). Tangible outcomes from the stress-testing results include additional oversight of the credit and changes in underwriting.
Institutional benefits of stress testing include identification of threats to the bank, development of contingency plans, and identification of downside risks that need to be reflected in pricing. For risk management, benefits include quantifying downside impacts of adverse events, identifying the most severe outcomes and possible mitigations, and determining the key risk driver of downside scenarios. Additional benefits include proactively monitoring the commercial loan and CRE portfolios, understanding the impact on the allowance for loan and lease loss, followed by understanding the impact on capital.
Three of the necessary components of a stress-testing infrastructure include people, data and systems. Nobody has perfect data, and the data that does exist has not been collected with stress testing in mind. As much as possible, the data needs to be complete, accurate and representative. Institutions should always look to improve their data by fixing data in the source system and by having loan officers, or the lines of business, maintain a log of significant events that might impact the data. In order for the data to be utilized in stress testing, it needs to be able to be segmented into portfolios — for example, owner- and non-owner-occupied real estate rather than simply real estate.
Any stress-testing process should start with the institution’s own assessment of possible specific vulnerabilities. In general, institutions should conduct adequate and proportionate stress tests on all the risk they have identified as material. At the strategic level within community banks, stress-testing results are most often incorporated into the individual and portfolio lending decisions, provisioning for the ALLL, and concentrations and limits setting.
When asked who is responsible for deciding the shocks or scenarios, 58 percent of the institutions cited the chief credit officer. The data used are internal data as well as the local economic data provided by the government and universities. Regardless of its source, the data should reflect the bank’s portfolio and be consistent from period to period.
Because many banks perform stress testing at origination or renewal, there is no set schedule for testing. Quarterly testing, though, is conducted in 24 percent of the institutions.
When asked what regulators have implied they are looking for when conducting stress testing, institutions said that the top reason provided was to see if adequate processes are in place.
Regulators are also looking to find out if the tests are applied to specific portfolios and to understand the impacts on the allowance. When asked if regulators have indicated a specific test they would like to see conducted, 87 percent of respondents said no. The 13 percent that received regulatory guidance mentioned stress tests for declining collateral value, interest rate shocks and loan-to-value changes. These are the tests that community banks currently employ.
The top two barriers cited for not using stress testing are inadequate MIS systems and failure to understand the process.
To overcome these barriers, community bankers need training and education on the topic as well as resources to put the process in place. Given that community banks have a limited number of staff to perform many tasks that are siloed in larger institutions, the effort to research the topic adequately and implement it is difficult.
A simple next step for banks performing stress testing at the individual loan level would be to aggregate the results of the individual loan stresses into a portfolio view. These results then should be shared with the board of directors or a management committee so that the information can be incorporated into the strategic or business planning process.
© 2008 The Risk Management Association. All Rights Reserved. Confidential & Proprietary.
All of the information contained herein is obtained from sources believed to be accurate and reliable. All representations contained herein are believed by RMA to be as accurate as the data and methodologies will allow. However, because of the possibilities of human and mechanical error, as well as unforeseen factors beyond RMA’s control, the information herein is provided “as is” without warranty of any kind, and RMA makes no representations or warranties express or implied to a subscriber or any other person or entity as to the accuracy, timeliness, completeness, merchantability, or fitness for any particular purpose of any of the information contained herein. Furthermore, RMA disclaims any responsibility to continue to update the information. Moreover, information is supplied without warranty on the understanding that any person who acts upon it or otherwise changes position in reliance thereon does so entirely at such person’s own risk.
Founded in 1914, The Risk Management Association is a not-for-profit, member-driven professional association whose sole purpose is to advance the use of sound risk principles in the financial services industry. RMA promotes an enterprise-wide approach to risk management that focuses on credit risk, market risk, and operational risk. Headquartered in Philadelphia, RMA has 3,000 institutional members that include banks of all sizes as well as non-bank financial institutions. They are represented in the association by more than 18,000 risk management professionals who are chapter members in financial centers throughout North America, Europe and Asia/Pacific. Visit RMA on the Web at www.rmahq.org.
Acknowledgments: Staff members contributing to the study were Suzanne Wharton, associate director of Risk Analytics, RMA; Mark Zmiewski, director of Strategic Learning and Research, RMA; and Pamela Martin, director of Regulatory Relations, RMA.
Copyright © April 2011 BankNews Media