16. The Interplay Between Banks and Markets: Supervisory Stress Test Results and Investor Reactions

Committee of European Banking Supervisors (2009, 2010), European Banking Authority (2011, 2014).

Federal Reserve (2009b, 2011, 2012, 2013a, 2014a, 2015a). As mentioned in the section “Institutional Framework”, a stress test exercise was carried out by the Federal Reserve (Fed) also in 2011, but results were not publicly disclosed.

To quote the Federal Reserve report whereby the 2009 stress test results were released, “the decision to depart from the standard practice of keeping examination information confidential stemmed from the belief that greater clarity […] will make the exercise more effective at reducing uncertainty and restoring confidence in our financial institutions” (Federal Reserve 2009a, page 1).

By regulatory capture (Laffont and Tirole 1991) we mean the risk that the special interest of a regulated entity (here, the bank) might influence the regulation or the supervision in any of its form.

Additionally, as argued by (D’Cruz and Crippa 2012), releasing bad results helps focus industry attention on the need for better stress testing tools and principles, thereby improving the quality of future bank management.

Stress testing exercises have also been carried out, with different scope and depth, by most national authorities in the European Union (EU). See Quagliariello (2009) for a complete picture of macro and micro stress testing approaches in many European countries.

As most banks did not have an exposure to all countries and duration bands, many data points in the 2011 European stress test exercise were filled with zeroes (about 73% of the total); note, however, that those were meaningful zeroes, informing that a given bank was not exposed to that specific combination of duration and nationality. Another 5.3% of the data points included missing values; most of them were, again, “meaningful missings”, for example, to indicate that banks did not expect any state aid in the following months.

The 1999 Basel I capital requirements were regulated by the EU through the so-called Capital Adequacy Directive (Directive 93/6/EEC) issued in 1993. The first single Banking Directive (2000/12/EC), unifying previous regulatory texts, was passed in 2000 to improve the clarity and consistency of the EU banking legislation. The 2004 “Basel II” Accord was implemented in 2006 by two directives recasting the Capital Adequacy Directive (Directive 2006/49/EC) and the Banking Directive (Directive 2006/48/EC). Following the global financial crisis, prudential rules on banks were further tightened through Directive 2009/111/EC (CRD2) and Directive 2010/76/EU (CRD3) (focusing on capital requirements for the trading book and remuneration policies). The 2010 Basel III reform was transposed into European law in 2013 through the so-called CRD4 package (including the Capital Requirements Regulation 575/2013 EU (CRR) and the Capital Requirements Directive 2013/36/EU (CRD4)).

See BBC News (2010), reporting statements by the Bank of Spain, Bafin, the Greek Finance Ministry and others.

A variable number of banks were included for each country, starting from the largest one and stopping when at least 50% of the total assets were covered. National supervisors could add further banks to the sample (e.g., Spain included several other minor banks).

The trading portfolio, or trading book, consists of financial assets held at fair value and includes the bank’s own position in financial instruments for trading purposes (i.e., aimed at generating profits in the short run by buying and selling securities) as well as derivatives held to hedge that position. The banking book includes all other exposures (often long term) which are not held in the trading book.

To some extent, their critique also applies to the 2011 exercise, as again no proper haircut was imposed on sovereign bonds outside the trading book; however, the credibility of the results was improved by the fact that losses were imposed also on sovereign exposures held in the banking book, although indirectly (through a floor on the “expected loss rate” applied to exposures towards banks and sovereigns located in different countries).

The paper was written prior to the release of the 2011 results. Accordingly, it only focuses on the market reaction occurring when the stress test was announced and on days when clarifications were issued by EBA.

This included the decision to focus on a “core Tier 1” target of 5% which was considerably more conservative than the one used in the past. Core Tier 1 refers to high-quality capital, such as common equity and government-sponsored hybrid securities, net of a number of deductions to make the final result robust and comparable across different banks and countries.

The section of the Fed website devoted to “Stress Tests and Capital Planning” has two different subsections, one for the CCAR (http://​www.​federalreserve.​gov/​bankinforeg/​ccar.​htm) and a second for the DFAST programme run by the Fed (http://​www.​federalreserve.​gov/​bankinforeg/​dfa-stress-tests.​htm).

The material presented in this section is based on Federal Reserve (2009b, c, 2011, 2013a, b, 2014a, b, 2015a, b).

Under the Dodd-Frank Act, all financial companies with more than US$10 billion in total assets are required to conduct a company-run stress test on an annual basis. Additionally, covered companies are subject to a second company-run stress test each year, and to the supervisory stress test.

For bank holding companies with more than US$10 billion but less than US$50 billion in total consolidated assets, a new rule adopted by the Fed on 25 November 2015—effective on 1 January 1 2016—eliminates the assumption of fixed dividends in company-run stress tests and requires to incorporate reasonable assumptions regarding payments of dividends consistent with internal capital needs and projections. This change does not apply to large banks as they are subject to the capital plan rule, and are required to incorporate their planned capital actions in their post-stress capital analysis. Thus, large banks can already incorporate more realistic dividend assumptions into their capital plans (see the section “The Comprehensive Capital Analysis and Review (CCAR) Program”).

Each bank participating to the CCAR is also required to use at least one stress scenario and a baseline scenario developed by itself.

For bank holding companies that have total consolidated assets of US$50 billion or more, a new rule adopted by the Fed on 25 November 2015—effective on 1 January 1 2016—removes the Tier 1 common capital ratio requirement. According to this requirement, a large bank should demonstrate its ability to maintain a Tier 1 common capital ratio of 5% under expected and stressed scenarios. The Fed introduced the Tier 1 common capital ratio requirement in 2009 as part of the SCAP. At that time, the it had not yet adopted a minimum common equity capital requirement. In 2013, it revised its regulatory capital rules and introduced—effective on 1 January 2015—a minimum common equity Tier 1 capital requirement of 4.5% of risk-weighted assets (RWAs).

This means that the CCAR post-stress capital analysis assumes that the bank continues to pay dividends according to the original capital plans, as in a baseline scenario, even if market conditions severely deteriorate. Hirtle and Lehnert (2014) refer to this feature as a “stringent test of capital actions”.

However, Hirtle, Schuermann, and Stiroh also note that “whether the reception would have been positive if […] there had been a negative “surprise” about a firm or a group of firms, remains open to debate” (page 11).

Consistently with the evidence reported by Flannery et al. (2015), Candelon and Sy (2015) find that the 2009 US stress test had a large positive outcome, whereas subsequent US stress test exercises had lower impact on the market valuation of stressed banks as time went by.

Starting in 2014 mid-sized banks (with US$10–50 billion in assets) were also required to conduct DFA stress tests. See also the section “The Dodd-Frank Act Stress Testing (DFAST) Program”.

European Central Bank (2014).

On a purely national level, most European financial systems would be more concentrated than the US. Stress tests, however, are a pan-European phenomenon and require a Europe-wide perspective.

“Superior information” here means information that persuades investors toward a socially optimal trade-off between individual risk-bearing and the provision of liquidity to the economy. See Gick and Pausch (2012) for details.