Federal Reserve's Supervisory Stress Testing - Part 3: CCAR 2013
This is the third and last blog dedicated to the Federal Reserve’s Stress Tests 2013. While the first two blogs focused on the DFAST stress tests, this one addresses the methodology and results of the Comprehensive Capital Analysis and Review (CCAR). The main difference between the DFAST stress tests and the CCAR stress tests is that the former assumed a set of standardised capital actions, across all bank holding companies (BHC), while the latter is based on the capital actions proposed by each BHC.
The Federal Reserve announced on Thursday 14 March that it has approved the capital plans of 14 of the 18 BHCs that participated to the CCAR. Two other institutions, Goldman Sachs and JP Morgan, received conditional non-objection while the Federal Reserve objected to the capital plans of Ally Financial, Inc. and BB&T Corporation. The consequence is that these two latter BHCs are not permitted to implement their proposed capital actions, such as common stock issuance, dividend payments and share repurchases.
The assessment framework used by the Federal Reserve to review the BHCs’ capital plans is based on quantitative and qualitative factors. In its quantitative assessment, the Federal Reserve evaluated whether each BHC’s can maintain post-stress capital ratios higher than the (regulatory minimum) 5% Tier 1 common capital threshold under the severely adverse scenario. In its qualitative assessment, the Federal Reserve focused on process and governance arrangements (including the robustness of the risk measurement and management practices) to ensure that capital levels could cope with firm-specific vulnerabilities and risks.
Ally Financial Inc.’s capital plan received an objection from the Federal Reserve on both quantitative and qualitative grounds whilst the capital plan of BB&T Corporation was objected on the basis of the qualitative assessment alone. Goldman Sachs and JP Morgan exhibited weaknesses in their capital planning process but the quantitative results were assessed as reliable. These two large BHCs are required to address weaknesses and resubmit capital plans by the end of the third quarter of 2013, but they can proceed with their capital distributions plans. This decision may sound awkward to an external and independent third party…
The following chart shows the current and projected Tier 1 common ratios under the supervisory severely adverse scenario, as calculated by the six largest BHCs and by the Federal Reserve under the set of standardised capital actions (respectively DFAST-Banks projections and DFAST-Fed projections). They are compared to the Federal Reserve’s own estimates taking into account the BHCs’ proposed capital actions (CCAR-Fed projections). All of these largest BHC’s, except Citigroup, passed the quantitative assessment by only a very small margin. Surprisingly, the uncertainty in the estimates, which we would expect to be large, is not addressed at all by the Federal Reserve.
The Federal Reserve does not describe either how they have validated these post-stress capital estimates in practice. The following examples raise questions: JP Morgan estimated that its net income would remain unchanged would the severely adverse scenario materialise and the portfolio loss rate of Goldman Sachs is more than 7 times lower than the Federal Reserve’s own estimate. Also, projections made by Federal Reserve and Citigroup seem to be aligned, while differences can be substantial for the other participants. Finally, we can argue on how a BHC with a post stress Tier 1 common ratio slightly above 5% could possibly “have the ability to lend to households and businesses and to continue to meet their financial obligations, even in times of economic difficulty”.
About the Author: Olivier Salomé is a manager at Risk Dynamics and specialized in the validation of risk models for both the banking and insurance sectors in the context of Basel III and Solvency II. He is also one of the key contributors to the development of Risk Dynamics Scenario and Stress Testing validation frameworks and methodologies. Click here to email Olivier.