Carmina Burana is widely known as one of the staples of any half decent orchestra. Twenty four poems were set to music by Carl Orff in 1936 and you would certainly know the music from the beginning if you heard it, as it appears in many movies and television shows. Benedictine Monks originally wrote the poems and they amazingly cover a variety of topics not generally associated with clergy. While only 24 were set to music, there are many more that are divided into basic groups: 55 songs of morals and mockery, 131 love songs and 40 drinking and gaming songs. These are not tame even viewed with modern sensibilities, they contain a great deal of political sarcasm and some focus on sleeping late, eating heavy food, drinking large quantities of wine and even playing dice games - weird indeed.
All of this talk brings to mind risk management of course and since we are more about banking than singing, we discuss the subject of stress testing a bank's credit portfolio. It has been some time (actually 2006) since the regulatory community created guidance surrounding stress testing and the extra due diligence required for banks with commercial real estate (CRE) concentrations.
While this guidance was dated pre credit crisis, it wasn't really rigorously enforced until people understood exposures could lead to failure and other issues. By now, most community banks with CRE concentrations have a credit test solution in place and that is good because it can help highlight issues before they surface. Whether the solution is an in-house spread sheet or an outsourced service, as with most regulatory subjects, the more complicated your bank is the more robust your solution must be.
Stress testing is important and doing just enough to check a box isn't likely to add much value. Stress testing should produce results that are robust enough to highlight potential issues so actions can be taken well in advance of any major problems. They should also be an integral part of risk management. The main purpose of a stress test is to determine if a bank is capable of withstanding a number of scenarios that could occur, particularly in terms of the effect on the bank's capital. Anything short of that is largely a waste of time, as it won't go far enough.
We do a lot of these for our customers and we see many banks approach their stress testing requirement either by using an internal model or by purchasing software that tests only CRE loans in the portfolio. Stress testing less than 100% of any portfolio adds some value, but it can also lead to unintended consequences and leaves other potential risks hidden. It is important with any model to start with the end. This is to say, ask what the model is supposed to do and then see if the one you are using is getting you there. If not, consider changing or upgrading so you are getting results you can use.
We were a pioneer in stress testing for community banks and have always approached it as a management tool that can be used to determine what could happen to the entire bank. Our methodology uses a risk-adjusted approach using probability of default and loss given default as common drivers across all the loans in the portfolio. By doing this, the drivers are the same regardless of the loan type, so the bank has a better understanding of risk across the entire portfolio. The model also looks at current and historical trends to determine the effect of declining credit quality on the portfolio. Risk management in the loan portfolio isn't the place to roll the dice, so to keep things singing along, consider a more comprehensive solution to go to the next level.