On Tuesday, the Government Accountability Office (“GAO”) issued a report to the House Financial Services Committee (the “Report”) regarding stress tests conducted by the Board of Governors of the Federal Reserve System (the “Fed”).
The Report grew out of a 2014 request from Rep. Jeb Hensarling, Chairman of the House Financial Services Committee, for the GAO to review the Fed’s stress test program because he was concerned that the Fed was not adequately transparent with the banks and the public about the program’s procedures and standards. In response to the request, the GAO conducted a comprehensive review of the Fed’s stress test program that included a review of Fed rules, guidance, internal policies and procedures, and interviews with Fed staff and officials from 19 banking institutions.
In the final Report, the GAO recommends that the Fed make several changes to its stress test procedures, including increasing transparency with respect to qualitative assessments, evaluating a broader range of severe stress scenarios, and doing more to assess uncertainty associated with the Fed’s internal models for determining the test results.
Background On the Fed’s Stress Tests
All Fed-supervised bank holding companies and banks of certain specified asset sizes are subject to one or both of two stress tests: The less stringent Dodd-Frank Act Stress Test (“DFAST”) is company-run and applies to bank holding companies and state member banks with more than $10 billion in consolidated assets. The more rigorous Comprehensive Capital Analysis and Review (“CCAR”) is Fed-run and applies to bank holding companies with more than $50 billion in consolidated assets.
Both tests analyze a banking institution’s balance sheet under a baseline scenario and hypothetical stress scenarios─ an adverse scenario and a severely adverse scenario. The Fed examines the results to see if a banking institution’s capital levels remain above the minimum capital requirements. CCAR also contains a qualitative component in which the Fed evaluates the institution’s capital planning processes, including risk management procedures, internal controls, and governance practices. It is possible to satisfy the quantitative capital requirements, but still fail the test if the Fed believes there are significant issues with the qualitative portion.
The Report’s Conclusions
In the Report, the GAO concludes that the Fed’s stress test program constitutes an important advance in protecting the financial system, but that the Fed’s implementation of the program falls short in a number of areas. These shortfalls include inadequate transparency about the standards the Fed applies to the tests, and lack of consideration of a sufficiently wide range of possible severe stress scenarios and risks associated with the system of empirical models the Fed uses to evaluate the post-stress capital ratios.
In order to address these shortfalls, the Report outlines fifteen recommendations regarding how the stress test program should be modified going forward. The most significant recommendations can be divided into three broad categories:
- The Fed should be more transparent about how it makes its qualitative assessments. In the Report, the GAO states that the Fed should release to banks and the public more information about its methodology for making qualitative assessments of banking institutions’ internal capital planning and related processes. Such transparency could make the Fed more accountable and increase public and market confidence in the efficacy of the program. The GAO also recommends that the Fed increase transparency by more regularly updating its guidance to banking institutions and communicating more clearly about how quickly it will respond to questions submitted through its CCAR communications portal.
- The Fed should consider using multiple severe stress scenarios to ensure banking institutions are prepared for a variety of possible crises. In the Report, the GAO observed that the Fed only analyzes banking institutions’ balances sheets under one “severe” stress scenario, which is based on the experience of financial markets in the post-war United States. The GAO cautions that such an approach could miss dangers from relevant but unprecedented threats to the banking system. It also recommends that the Fed employ further analyses of its own scenarios to better understand the range of outcomes that might result from different scenarios.
- The Fed should take further steps to manage risks associated with its own models used to determine the results of the stress tests. In order to determine a banking institution’s capital ratios under the stress scenarios, the Fed employs a system of empirical models. This system includes component models that predict elements of the institution’s balance sheet, risk-weighted assets, or income statement for different portions of the stress test. These models are interconnected, and many component models rely on the output of others as a source of data. The GAO criticizes the Fed for inadequately considering the risk of error associated with this system of models. According to the GAO, Fed employees justify modeling choices at the component level, with inadequate attention paid to risks associated with their interaction. This distinction is important because it is the system of models that produces the post-stress capital ratio of the institution being tested. Without adequate sensitivity and uncertainty analysis of this system, the Fed cannot know how confident to be in its conclusions about the institution’s post-stress capital ratios. To remedy this issue, the GAO recommends that the Fed implement a number of new processes to evaluate risks associated with system of models, including testing the sensitivity and uncertainty of the models’ output and having the Board and senior stuff articulate tolerance levels for key risks they identify.
In response to the findings of the Report, the Fed wrote a letter (which is appended to the Report) in which it indicated broad agreement with several of the GAO’s findings, and stated that it was already taking steps that would address many of the concerns raised in the Report.
Rep. Hensarling praised the Report and criticized “the secrecy surrounding the stress tests,” which he claimed make “it almost impossible to measure the effectiveness of the Fed’s regulatory oversight or the integrity of the tests’ findings.” He also argued that the Report demonstrated the need for the next president to designate a Vice Chairman for Supervision at the Fed.
Despite these criticisms, the Report did not question the fundamental value of stress tests to the financial system, as some critics may have hoped. The Report, therefore, seems more likely to lead to relatively minor adjustments to Fed procedures than to sweeping changes to the stress test program.