CCAR benchmark 2017

  • 13 March 2018

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.

Building on our work in previous CCAR benchmarks, in 2017 we ran this exercise for the third time to provide interested institutions with a benchmark of methodology and model outputs for their submissions. The summary report is available to download below.

We're running a 2018 CCAR benchmark, which will begin in July. If you’re interested in participating, please contact us.

Key findings

SME inputs based on scenarios and historical data increasingly chosen over modelling input

The inputs used for forecasting show a move away from models and towards the use of subject matter expert (SME) input, especially based on scenarios but also using historical data. In fact, SME input and business judgement are increasingly relied upon throughout the entire CCAR process, year on year. 

The importance of scenario analysis and risk identification programs are growing

Although it is not required of all bank holding companies (BHCs), firms are usually incorporating scenario analysis and idiosyncratic events into their submissions to capture losses that are not covered by other modelling. This is driven by the need to identify and quantify risks which are material to the BHCs. From this comes risk identification programs and linkage to risk taxonomies – two practices which are being employed by the majority of firms.

Firms increasingly establish their own risk taxonomies

Most of the firms surveyed stated that they link their risks to taxonomies. Furthermore, an increasing number of firms are establishing risk taxonomies to map their material risks to. This was reflected in the findings of this report, although mapping material risks identified for CCAR to a newly established risk taxonomy was also found to be a challenging process by at least one firm.

The modelling and forecasting process remains the key challenge

The top challenges experienced by the BHCs undergoing CCAR show an appreciable amount of commonality between SR 15-18 and SR 15-19 firms, with the majority of difficulties relating to the modelling and forecasting process.

About the project

Twenty firms (members and non-members of ORX) took part in the CCAR benchmark 2017. This project provided participants with:

  • Insight into common problems and solutions created by their peers
  • The opportunity to participate in a webinar to present the key results
  • A detailed benchmark report comparing CCAR model outputs in different scenarios
  • A bespoke fact sheet comparing their key results to those of their peers

Download the summary report

CCAR benchmark 2017 – summary report

Find out more

If you want to learn more about this study, or find out how you could get involved in similar projects in the future, contact us.

Contact See our other projects