CCAR benchmark 2018
- 18 January 2019
CCAR (Comprehensive Capital Analysis and Review) is a key US supervisory requirement, and one of the most complex stress tests. It continues to take up considerable time and resources of financial institutions as the requirements for reporting relevant exposures evolve. You can now download the CCAR benchmark 2018 summary report below.
About this study
Following on from the success and popularity of the CCAR benchmark 2017, we re-ran the study for 2018. The goal of this project is to provide participants with a benchmark of operational risk modelling methodologies, and model outputs for their 2018 CCAR submissions.
A key feature of the report was the provision of benchmark ratios of different supervisory scenarios (e.g. the ratio of adverse to baseline scenarios), which gave participants a simple method to compare their results against those of their peers, while maintaining a high level of confidentiality.
The report also collected information on forecasting inputs used, modelling approaches, risk identification methods, validation processes, and common challenges, inter alia. The full study results are only available to participating institutions.
SME inputs continue to increase for stressed forecasting
The use of expert input continues to increase in the case of forecasting under stressed scenarios, although baseline approaches have remained mostly unchanged. This is a continuation of the trend we observed last year – subject matter expert (SME) input and business judgement continue to take center stage over more traditional quantitative modelling aspects, both in usage and focus of challenge. Firms continue to use SME input and business intuition to guide, gauge and challenge the results of traditional models.
Scenario analysis and risk identification programs are still in focus
Although not all bank holding companies (BHCs) are required to include quantified results from scenario analysis, all participants in this study did. The main reason given for incorporating scenario analysis and idiosyncratic events was to capture losses that are not covered by other modelling or as an overlay. This is driven by the need to identify and quantify risks which are material to the BHC. Formal risk identification programs have been put into place to identify the BHC’s top material risks and ensure that they are considered in the various stress scenarios.
Some links to the external environment
Most banks identify a significant relationship between losses in execution, delivery and process management, and the macroeconomic environment. Specific factors include market volatility, foreign exchange volatility, credit spread and term premiums.
The only other event types where the majority of banks have found a significant link to the external environment are clients, products and business practices, and external fraud.
Risk taxonomies are on the rise
Increasingly, firms are linking their top risks to internally developed taxonomies. In a recent ORX study, it was clear that a growing number of firms are establishing operational risk taxonomies which much more accurately capture their material risks than Basel event-based taxonomies do. As these taxonomies are still in development, some firms find linking to internal or external factors a challenge.
The forecasting of legal losses remains a challenge for many. Typically, a range of information is combined with subject matter expertise to provide estimates. Participants reported that although they now have good relationships with legal teams for this purpose, validating the outcome of the process remains a challenge.
Data availability and quality are a challenge
There is commonality between the top challenges identified by both SR 15-18 and SR 15-19 firms, with most difficulties relating to data issues. Reliable data is the key to reliable models, and many firms expressed that they either lack suitable data or they cannot rely on it. This is an area where firms will need to take steps to improve in the future.
You can now download the CCAR benchmark 2018 summary report below.