Stress-testing CRE Loans: A Practical How-To Guide

In previous articles I mentioned that the examiners are expecting new approaches to stress-testing loans, especially commercial real estate loans (CREs). The results of that analysis have a profound impact on the loan loss reserves.

Many banks are struggling to find a different approach to stress testing that goes beyond the usual loan-to-value analysis, vacancy rates etc. Ron Piccinini of First of Nebraska's Treasury department has developed a more out-of-the-box approach to CRE stress-testing that would be helpful to you as you rethink your current practices.

Here is Ron's article:

Assumptions:

In the following, it is assumed that the bank has a simple risk scorecard for commercial real-estate loans. This scorecard is assumed to assign a risk score between 1 and 10 to each loan based on its respective Debt Service Coverage Ratio (DSCR), Loan to Value (LTV), Occupancy Rate (OR), and Market Quality (MQ). Market Quality is a subjective yet important variable that is determined jointly with the commercial banker for each and every transaction. It is also assumed that the bank has access to each deals documentation. Ideally, each risk score can be associated with a probability of default, but this is not necessary.

1st step: Identify and collect the data necessary to compute each of the risk scorecards inputs. For example, in order to compute the DSCR, you may need the following data: current loan amount, interest rate, and number of amortization periods, as well as the propertys projected income, expenses, vacancy rate, cap rate. For the LTV computation, the same inputs are needed. The Occupancy Rate is just one minus the vacancy rate, and market quality can be tied to the cap rate for example.

2nd step: Create your stress / what-if scenarios. How does the loan portfolio look if things change. How does the loan portfolio look if expenses increase by 15% across the board and vacancies rise by 500 basis points? What would likely happen if cap rates were to increase 370 basis points? What if income comes in 25% below projected levels, interest rates rise by 350 basis points and Market Quality deteriorates by 3 grades? Basically any combination of very bad news will do, but this is usually the perfect time to talk to your organizations most experienced people. It may also be useful to agree on a few official scenarios which would be presented to senior management quarterly, and create ad-hoc scenarios depending on current market conditions.

3rd step: Recalculate a new risk score for each loan given the stress scenario thought out in step 2. For example, in the first stress scenario described above, multiply projected expenses by 1.15 and add 5.00% to the vacancy rate of each property in the portfolio. Then re-score each loan using the new (stressed) values. Use the resulting scores to make a qualitative assessment of the health of the portfolio, as illustrated in Figures 2 and 3.

Figure 1: Stress-test process

What about probabilities? Some regulatory bodies insist that stress-tests should be conducted at a given confidence level, e.g., 95% or 99%. (Un)fortunately, there are no regulatory tables where the physical probability of all future events is readily available, so the idea is to model the joint probability distribution of the variables listed in step 1 above. For example, it makes sense to force cap rates and vacancy rates to move together, and in opposite direction of market quality. While the question of how to calibrate these models is beyond the scope of this article, a simple albeit pessimistic solution is to model each variable independently, and to use as stressed values the 95th worse percentile of each variable simultaneously. The resulting portfolio can be used as an estimate of the 95% level of confidence and potentially satisfy regulatory demands.