Regulatory stress testing and bank performance

• Published date: 01 November 2020
• Author: Lukas Ahnert,Pascal Vogt,Volker Vonhoff,Florian Weigert
• Date: 01 November 2020
• DOI: http://doi.org/10.1111/eufm.12267

Eur Financ Manag. 2020;26:1449–1488. wileyonlinelibrary.com/journal/eufm © 2020 John Wiley & Sons Ltd.

|

1449

DOI: 10.1111/eufm.12267

ORIGINAL ARTICLE

Regulatory stress testing and bank

performance

Lukas Ahnert

1

|Pascal Vogt

2

|Volker Vonhoff

3

|

Florian Weigert

4

1

School of Finance, The University of

St. Gallen, Switzerland

2

The Boston Consulting Group, Cologne,

Germany

3

The Boston Consulting Group, Stuttgart,

Germany

4

Institute of Financial Analysis, The

University of Neuchâtel, Neuchâtel,

Switzerland

Correspondence

Florian Weigert, Institute of Financial

Analysis, University of Neuchâtel, Rue

A.‐L. Breguet 2, 2000 Neuchâtel,

Switzerland.

Email: florian.weigert@unine.ch

Abstract

This paper investigates the impact of stress testing

results on banks’equity and CDS performance using

a large sample of 12 tests from the US CCAR and the

European EBA regimes in the time period from 2010

to 2018. Passing banks experience positive abnormal

equity returns and tighter CDS spreads, while failing

banks show strong drops in equity prices and

widening CDS spreads. We also document strong

market reactions at the announcement date of the

stress tests. We complement existing studies by

investigating the predictability of stress test outcomes

and evaluating strategic options for affected banks

and investors.

KEYWORDS

banks, CDS performance, equity performance, stress testing

JEL CLASSIFICATION

G00; G21; G28

1|INTRODUCTION

Over the course of little more than a decade, bank stress tests have developed into a key

supervisory tool that shapes headlines and fundamentally affects the main business operations of

banks worldwide. The financial crisis of 2008–2009 revealed that banks’capital was insufficient

EUROPEAN

FINANCIAL MANAGEMENT

The authors thank three anonymous referees and seminar participants at the 8th Financial Engineering and Banking

Society 2018 in Rome and the University of St Gallen. All errors are our own. Florian Weigert is affiliated with the

Centre of Financial Research (CFR) Cologne and is thankful for their financial support.

—both in quantity and quality —to withstand large adverse shocks. Hence, following the crisis,

regulators pushed for tighter rules and a systemic overhaul of the precrisis Basel II framework

with a key focus on the refinement of banks’risk‐weighted asset calculation and capital

requirements. The postcrisis adjustments of these capital standards continue to present a

challenge for financial institutions, especially in times of overall low bank profitability and crisis

legacy costs. To assess if banks are keeping pace with regulatory demands, supervisors around the

globe increasingly rely on stress testing.

Bank stress tests assess the capitalization of banks on a forward‐looking basis under

simulated unfavorable economic scenarios. They focus on several key risk types, such as credit

risk, market risk, and liquidity risk, to determine the banks’financial health in crisis situations.

Stress tests break with traditional supervisory approaches in their far‐reaching disclosures and

provide market participants with unprecedented insights into bank balance sheets and their

reaction to macro variable changes. Moreover, at the same time, stress tests place follow‐up

requirements on banks that fall short of supervisory expectations.

The aim of this paper is to develop our understanding of the consequences of stress tests for

affected banks and investors along two dimensions. On the one hand, we study the impact of

stress testing results and stress testing announcements on banks’equity and CDS performance.

On the other hand, we examine whether stress testing results are predictable by different bank

characteristics as well as the announcement return and whether banks and investors can

strategically profit from such predictability. Our analyses are performed on a sample of 12 stress

tests; seven tests from the US Comprehensive Capital Analysis and Review (CCAR) and five

tests from the European Banking Authority (EBA) regimes in the time period from 2010

to 2018.

We provide five contributions that advance our understanding of the capital market effects

of stress tests. First —using data from 12 US and European stress tests (seven stress tests from

the United States and five from Europe) and an event study design in the period from 2010 to

2018 —we show that stress test results reveal new information to market participants as they

significantly affect banks’equity and CDS performance. Passing banks, on average, experi-

ence significantly positive abnormal equity returns of 50 basis points and significantly tighter

abnormal CDS spreads of −52 basis points on the stress test release day. To the contrary,

failing banks earn significantly negative abnormal equity returns of −174 basis points and

widening abnormal CDS spreads of 83 basis points on that day. Taken together, we observe

that the overall effect of stress tests is performance positive: On average, tested banks ex-

perience significantly positive abnormal equity returns of 33 basis points and significantly

abnormal tighter CDS spreads of −51basispointsontheresultreleaseday.

1

Relating these

results to the effectiveness of stress tests as a policy instrument, we can observe (without

taking account of the effect of stress test announcements) that past stress test result releases

were, on average, successful in improving investor sentiment towards banks’equity and

credit risk.

To the best of our knowledge, this paper is the first to analyze the market impact of stress

test results based on a panel of 12 US and European stress tests. Up to now, existing studies

1

Note that this result is derived from retrieved data of 392 equity return and 277 CDS spread observations of tested

banks in our sample. Out of the 392 equity observations, we count 361 passes and 31 fails. Out of the 277 CDS spread

observations, we count 256 passes and 21 fails. Hence, one has to be cautious to not overinterpret this result due to the

small number of stress test fails. Nevertheless, we observe that, on average, the positive impact of the majority of stress

test passes exceeds the negative impact of few stress test fails (which can be quite significant on an individual basis).

1450

|

EUROPEAN

FINANCIAL MANAGEMENT

AHNERT ET AL.

mainly focus either on (a) investigating a time series of stress tests in a jurisdiction (for the

United States or Europe) or (b) comparing two individual stress tests between the United States

and Europe. In the time‐series setting, for US banks, Fernandes, Igan, and Pinheiro (2020),

Flannery, Hirtle, and Kovner (2017), and Morgan, Peristiani, and Savino (2014) show that bank

equity and CDS performance are significantly affected by stress test releases, whereas Neretina,

Sahin, and De Haan (2014) do not observe significant capital market effects on the release day.

For European banks, Alves, Mendes, and Pereira da Silva (2015), Ellahie (2012), and Georgescu,

Gross, Kapp, and Kok (2017) document that stress test releases trigger significant capital market

consequences, while Petrella and Resti (2013) only observe a minor impact on bank equity

performance. Candelon and Sy (2015) compare the CCAR 2009 and the EBA 2011 stress tests

and conclude that the overall capital market impact is different across jurisdictions. Due to the

contrasting empirical findings retrieved in the literature, we recognize an important merit in

synthesizing the results for the up‐to‐date largest panel of stress tests from both the United

States and Europe. This dataset also includes the four most recent stress test results (i.e., the

CCAR stress tests of 2016, 2017, and 2018 as well as the EBA stress test of 2018) which have

never been assessed in the literature before.

2

,

3

Our second contribution concerns the market reaction on stress test announcement events.

On these days, we reveal market effects that point towards the opposite direction of the result

release dates, that is, banks that are announced to be stress tested earn significantly negative

abnormal equity returns of −21 basis points and significantly abnormal wider CDS spreads of 52

basis points on the eventday. We find that these results are driven by banks thatare tested for the

first time (abnormal equity return of −34 basis points and abnormal CDS spreads of 72 basis

points) and banks which are on the size threshold of being tested (abnormal equity return of −42

basis points and abnormal CDS spreads of 71 basis points). Moreover, when combining the result

release and the announcement event,we document that the overall market reaction is statistically

not distinguishable from zero.

Previous studies, such as Candelon and Sy (2015), Fernandes et al. (2020), Morgan et al.

(2014), and Petrella and Resti (2013), examine the market impact of stress test announcements

with mixed results. While Petrella and Resti (2013) do not find evidence of a market reaction of

stress test announcements, Candelon and Sy (2015), Fernandes et al. (2020), and Morgan et al.

(2014) find empirical support for a significant market impact for some stress test announce-

ments (e.g., for the EBA 2011 stress test and the CCAR 2013 stress test). Again, as in the case of

the result release effect, our study is the first to apply a panel dataset of 12 US and European

stress tests to document that the market impact of stress test announcements is, on average,

performance negative. Moreover, we contribute to show that the positive effect of stress tests on

banks’equity and CDS spreads on the result release day is subsumed by a negative effect

realized on the stress test announcement day.

4

2

The research papers with the most recent stress test samples, that we are aware of, are Fernandes et al. (2020) and

Neretina, Sahin, and de Haan (2015), who study the US stress tests between 2012 and 2015, as well as Georgescu et al.

(2017) who study the 2014 and 2016 European stress tests. For a recent overview on employed data samples and results

of the existing stress testing literature, see supporting information appendices A1 and A2.

3

Note that this large dataset also enables us to compare market reactions of stress test releases and announcements

(section 5.3), predictability of stress test outcomes (section 5.4), and trading strategies to exploit stress test outcome

predictability (section 5.5) between the United States and Europe, which are all new contributions to the literature.

4

Note that the finding of a significant negative market impact of stress test announcements only holds when we pool the

CCAR and EBA stress tests. We do not find a significant market effect when examining the US and European stress

tests individually.

AHNERT ET AL.EUROPEAN

FINANCIAL MANAGEMENT

|

1451