On June 27th, the Federal Reserve released the results for CCAR 2019, with none of the 18 participating banks receiving an objection to their capital plans on quantitative or qualitative grounds (although one received a conditional non-objection). The results were improved from the prior year, in which banks faced more severe economic scenarios and the effects of the 2018 tax reform; both of which had a material one-time negative impact on banks’ capital.
The day before the Fed’s release, McKinsey hosted its fifth annual CCAR Roundtable in New York, convening heads of capital planning and stress testing and other senior leaders from 22 institutions for a participant-led discussion of top-of-mind topics.
The industry perspectives from the roundtable highlighted a common theme, which was also reflected in the CCAR results: While banks have seen some reduction in stress capital requirements, they are still uncertain about evolving regulatory expectations for capital planning practices. In addition, banks are grappling with new requirements, notably the new Current Expected Credit Losses accounting standard (CECL), as well as how to streamline the CCAR process. In the following sections, we share the views of roundtable participants and our own insights on key points raised at the roundtable.
Regulatory uncertainty remains, and in some cases has increased
While the Fed has proposed several regulatory changes in recent years to simplify and clarify expectations, participants noted regulatory uncertainty remains.
For example, banks face continued uncertainty around regulatory expectations for the Global Market Shock test. For several years, the Federal Reserve has said that banks may remove the double-count between Global Market Shock losses and 9-quarter projections where they can demonstrate that the positions stressed are the same. However, only a small number of banks reported attempting to remove the double-count, in part due to unclear expectations about the right approach.
In addition, banks reported a lack of clarity around what trading hedges (so-called “macro hedges”) regulators would accept. In its results disclosure, the Fed highlighted the use of “certain large trading positions to reduce the impact of the market shock” as an area of potential concern, although did not provide guidance on what types of hedges might be considered more acceptable.
Banks should continue to seek constructive engagement with regulators with the aim to create solutions and gain clarity. This could be done by engaging directly through on-site teams or through submission of FAQs, or by working in collaboration with industry groups (such as the Clearing House or SIFMA) to voice their concerns, propose solutions, and obtain that clarity. This will be particularly important for topics such as CECL, where expectations will likely evolve in the coming years, based on banks’ approaches over the next one to two cycles.
Integration of CECL into CCAR is a top strategic priority, with a wide divergence in potential approaches
Banks expect the integration of CECL into CCAR – which will be required starting next year – to lead to material increases in forecasted provisions in CCAR, with several banks estimating increases of up to 30 percent. Accordingly, responses to a survey we took at the roundtable showed that CECL integration is a top priority. However, few have finalized their approach (Exhibit 1).
In addition, banks are considering widely different potential approaches to this integration, with varying levels of rigor. While some banks have considered developing individual forecasts for each projection quarter in the CCAR horizon, others may use a simplified “coverage ratio” approach to approximate the impact of stress on provisions.
Banks are also asking significant questions about expectations for the roles of the second line of defense and internal audit functions. And they want to know if they must treat approaches to integrating CECL and CCAR as models or as qualitative approaches, and what level of review these would face.
Banks should use the coming months before CCAR 2020 to do several things: identify the highest impact modeling decisions and assumptions (for example, the biggest drivers of losses); develop reasonable and defensible approaches to CECL/CCAR integration; and engage stakeholders from across the lines of defense and the business in an iterative process, to refine those approaches. As banks finalize CECL models for business-as-usual approaches, they should consider the need to integrate these with CCAR – including issues such as data needs, production implications, and modeling approaches – to find efficiencies and avoid unnecessary conservatism.
Streamlining and automation are top-of-mind, but banks face challenges
Many banks say streamlining and automating CCAR processes is a top priority over the coming year, potentially allowing them to decrease CCAR budget allocations (Exhibit 2).
So far, however, the scope of such work has varied widely. Many of the large banks that began filing in early CCAR cycles noted difficulties in simplifying their detailed, interconnected machines without a complete overhaul. Some of this may be due to the fact that many banks built their CCAR machine in a piecemeal fashion as new expectations and requirements arose. In contrast, banks that have more recently become subject to filing requirements generally have a newer, more cohesive infrastructure. Some have reported that they can run stress tests in hours or a few days, as opposed to six-plus weeks for others.
To address these issues, some banks suggested prioritizing processes based on feasibility and expected value, and then pursuing both quick wins and large-value, longer-term initiatives. On the quick-win side, some banks explained during the roundtable that they have reviewed their CCAR governance processes, reducing and streamlining committees, and merging some review-and-challenge cycles to reduce cycle times. In these cases, the results have been impressive, cutting cycle times by up to three weeks. Banks’ longer-term initiatives include building cloud-based workflow solutions that will enable connectivity and sequential online runs of most model components. Such enhancements may require an upfront investment, but banks can use them in multiple processes across the institution (including financial planning and analysis). These enhancements also provide ongoing value in the form of run-rate cost savings and enhanced strategic insights.
Most “Extended Stress Test Cycle” firms have made limited changes
Among the banks that are now subject to CCAR only biannually – and did not participate in CCAR 2019 – most have largely maintained the same set of processes, despite the decreased requirements. Many of these banks report running internal and supervisory scenarios. Although they reported making minor modifications to governance and timing, many continue to find value in running the annual test for capital planning purposes.
Ten years after SCAP, the Fed’s stress testing regime continues to evolve materially each year. While banks had hoped for significant relief and greater clarity over the past two years, many have been disappointed with the level of relief, and the new areas of uncertainty that have emerged. With this backdrop, banks should prioritize four key areas:
- Proactively engaging with supervisors, and proposing approaches that are backed by principles of sound risk management and clear rationales
- Responding to new requirements pragmatically, while balancing complexity and capital implications
- Exploring bold measures to simplify the CCAR process and use capabilities more broadly for strategy and financial planning
- Managing capital allocation and optimizing portfolios to right-size stress capital.