Finding the right loss rates for top down stress tests
By Emily Eustis, Sageworks
In a previous post, we looked at the benefits that come with a top-down or institution level stress test. A challenge with this analysis is finding the appropriate loss rates to use.
Regulators are not specifying the loss rate methodology that should be used for stress testing portfolio segments in a top-down stress test. The OCC states that “effective methods can range from a single spreadsheet analysis to a more sophisticated model, depending on portfolio risk and the complexity of the bank.” It is up to financial institutions to evaluate and determine the most appropriate methods for their institution. Here, we suggest some commonly used methods to determine stress period loss rates for portfolio segments.
Historical Loss Rates
Financial institutions may look at the historical performance of the portfolio segments as a starting point for what loss rates were during poor economic conditions. Looking at past trends for loss rates may give an indication of loss rate volatility and correlation with economic indicators. However, since historical losses do not predict future performance, financial institutions should adjust historical losses with reasonable expectations for future stress scenarios. For example, interest rates during stress scenarios may be more extreme than what has been observed historically. Stress period loss rates should incorporate the additional impact of extreme interest rates in addition to historical loss rates.
Bottom-up stress testing
Financial institutions may analyze a portfolio segment at the loan level and aggregate the loan-level results to obtain segment-level stress loss rates for top-down stress testing. This is a way to use bottom-up stress testing within a top-down framework. It is beneficial because the stress test is being applied at a granular level, but the results are analyzed as a whole. The aggregate impact on portfolio-level earnings and capital from individual loans may be surprising and valuable in risk management. However, many institutions find it difficult to gather the data required to perform bottom-up stress testing. Thus, it is not recommended for all portfolio segments. For instance, it may be beneficial to use bottom-up stress testing for the CRE portion of a portfolio and use other methodologies for the remaining portfolio segments.
Some institutions with large portfolios and complex financial products may choose to develop specific models for how portfolio segments would perform under various economic scenarios. The level of sophistication of these models varies greatly and depends on many factors, including the institution’s size, risk tolerance and product types. In the end, it is important to use a model that is defensible and provides timely and useful results for analyzing portfolio risks during stress scenarios. Again, it is possible to use modeling for only part of the portfolio, while employing other loss rate methods for other concentrations.
Finally, financial institutions may look to similar institutions for their portfolio segment loss rates in order to get a wide range of performance data and perhaps to get historical data that is not available at the home institution. Peer data should be used with caution – while market experience is a good benchmark, it should be adjusted for differences in financial institutions’ practices and product characteristics. In addition, historical experience should be adjusted for differences as compared to future stress scenarios. All these adjustments must be clearly described in the accompanying stress testing documentation.
Regardless of the stress testing method used, data will be the key to generating defensible test results. Unfortunately, for many financial institutions, data collection is also one of the most challenging aspects of stress testing. Depending on the data available for each portfolio segment, financial institutions may be limited by the stress testing methodologies available to them. Thus, it is important to use a top-down stress testing framework that allows the use of different stress testing methodologies in each portfolio segment and easy aggregation of the segment-level results.
Learn more about top-down stress testing by downloading a whitepaper: Benefits of Top Down Stress Testing.