Publications & Resources

January/February 2010
Managing Troubled Assets

 

Identifying Tomorrow’s Problem Loans

By Neal Brauner

Banks spend a great deal of time analyzing their loan portfolios to develop defensible levels for their Allowance for Loan & Lease Losses (ALLL). The objective here is to establish a reserve for losses inherent in the loan portfolio as of the reporting date. But as we’ve witnessed over the past couple of years, these collective analyses have consistently failed to capture the industry’s true loss exposure as economic conditions continued to deteriorate. Each quarter we’ve seen banks make significant additional provision expense to cover credit losses that were not anticipated in the previous quarter’s ALLL analysis.

The inability of banks to get a good handle on potential losses has created a number of problems, including continuing earnings “surprises”, the inability to accurately assess capital needs, and the loss of investor and customer confidence. And let’s not forget about all the C&Ds that have been triggered by the ongoing deterioration in credit quality and the associated effects on capital adequacy.

A major part of this problem is due to the fact that we have been experiencing the banking equivalent of a “100-year flood”. Our existing quantitative models are based, for the most part, on extrapolating trends from recent history. Unfortunately in this situation, history did not give us much guidance on the potential impact of a recession of this magnitude.

Stress Testing Tips for Community Banks

You don’t need sophisticated econometric models (or rocket scientists to run them) in order to stress test your loan portfolio. The following are some of the key considerations in stress testing, whether at the borrower or portfolio levels:

Factors to stress. Keep in mind that different portfolios are sensitive to different variables. As an example, expected losses on income property loans will mostly be affected by changes in occupancy rates, interest rates, and capitalization rates, while consumer loan portfolios are going to be more sensitive to factors such as unemployment rates, personal income, and housing values. You may want to keep it simple by selecting 3 to 5 economic variables to analyze, based on the characteristics of your loan portfolio.

Size of the “shock”. Modeling too small a change in the key stress variables may not be worthwhile; while too large a change may stretch the credibility of the exercise. A good starting point would be to look at the biggest swing in each variable that occurred over the last business cycle or two.

Other considerations. You can perform single factor analyses (just shock one variable at a time to see the impact), or construct various economic scenarios to test. Another high-level approach is to just stress loss rates directly (for example, analyze the impact on capital if loss rates on C&I loans rose to twice the current industry average over the next 12 months).

Even the most basic of stress tests will likely yield some valuable insight into your bank’s loan portfolio, and will be time well spent. The regulators will look favorably upon it too!

The events of the past two years should serve as a reminder to banks of the importance of periodically stress testing their loan portfolios to better understand the potential risks. Stress testing is a different concept from the “forecasting” that occurs in most banks. Forecasting is mainly focused on anticipating a “most likely” outcome over a fairly short time horizon. With stress testing, we subject the loan portfolio to “exceptional but plausible” changes in key economic variables. The objective here is to identify those scenarios which, if they did actually occur, could result in severe losses and jeopardize the financial stability of the bank.

These stress tests can be performed at the individual borrower (“micro”) level, and at the total portfolio (“macro”) level. Stress testing individual loans can serve as a valuable early-warning system to identify those customers most likely to experience financial stress under adverse economic conditions. This can help flag areas where the bank’s exposure should be reduced if possible, and help focus management attention on those customers most in need of monitoring and assistance.

At the portfolio or macro level, stress testing can give valuable insights into potential future losses, identify key areas of risk exposure within the portfolio, and assess the adequacy of the bank’s current and prospective capital position. Stress testing can also be very useful as a portfolio management tool to help optimize portfolio composition, particularly when careful attention is given to correlations between industries and between loan portfolios.

The actual mechanics of stress testing can be fairly simple or extremely complex, depending on the time and resources a bank wishes to devote to the task (please see the accompanying article Stress Testing Tips for Community Banks). There are many factors that can come into play, including assessing the impact of key economic variables on loan performance, creation of appropriate stressed economic scenarios to test, and correlation of risks and risk measures across portfolios and industries.

We are likely to see a much greater emphasis placed on loan portfolio stress testing in the future. One potential outcome of this would be for banks to begin to develop credit risk exposure limits in response to pre-defined economic shocks, much like the current interest rate risk exposure limits are defined in terms of specific interest rate shocks.

Regardless of whether or not loan portfolio stress testing becomes a regulatory mandate, it is a tool that bankers should embrace to improve risk management processes and management decision making. It is likely that you will be hearing a lot more on this topic in the near future.

Neal Brauner is a consultant with RLR Management Consulting, Inc. He can be reached at (760) 771-5036 or neal.brauner@rlrmgmt.com.

Unauthorized reproduction of all or part of this material without the express written consent of the author is strictly prohibited. All rights reserved.