Competition and risk‐taking in investment banking

• Author: Claudia Girardone,Nemanja Radić,Franco Fiordelisi,Marta Degl'Innocenti
• Date: 01 May 2019
• DOI: http://doi.org/10.1111/fmii.12113
• Published date: 01 May 2019

DOI: 10.1111/fmii.12113

ORIGINAL ARTICLE

Competition and risk-taking in investment banking

Marta Degl'Innocenti1Franco Fiordelisi2,3 Claudia Girardone4

Nemanja Radić2,5

1University of Southampton, Highfield,

Southampton SO17 1BJ, UK

2University of Rome III, Via S. D'Amico 77,

00145 Rome, Italy

3MiddlesexUniversity, The Burroughs, London

NW4 4BT, UK

4University of Essex,Wivenhoe Park, Colchester

CO4 3SQ, UK

5CranfieldUniversity, College Rd, Cranfield,

Bedford MK43 0AL, UK

Correspondence

NemanjaRadić, Cranfield University, College Rd,

Cranfield,Bedford MK43 0AL, UK.

Email:n.radic@cranfield.ac.uk

Abstract

How does competition affect the investment banking business and

the risks individual institutions are exposed to? Using a large sample

of investment banks operating in seven developed economies over

1997–2014, we apply a panel VAR model to examine the relation-

ships between competition and risk without assuming any a priori

restrictions. Our main finding is that investment banks’ higher risk

exposure, measured as a long-term capital-at-risk and return volatil-

ity, was facilitated by greater competitive pressures for both bou-

tique investment banks and full-service investment banks. Overall,

we find some evidence that more competition leads to more fragility

before and during the recent financial crisis.

KEYWORDS

competition, investment banking, panel var, risk

JEL CLASSIFICATION

D4, G3, G24

1INTRODUCTION

Until the 2007 global financial crisis, the investment banking business enjoyed a prolonged period of prosperity and

stability. Deregulation and technological improvements havecontributed to the integration of investment and com-

mercial banking and encouraged greater competitive pressures in the financial services sector (Goddard, Molyneux,

Wilson, & Tavakoli, 2007). As the industry became more contestable, firms were increasingly driven by profit maxi-

mizing motives. Many developed as large full-service institutions and responded to the decline in commissions gained

from their traditional securities business by seeking new income sources. In particular,product and services diversifi-

cation has led to greater (and possibly excessive)risk-taking activities and exposure, including proprietary trading and

dealing with complex financial securities (Altunbas, Gambacorta, & Marques-Ibanez, 2009; Carbó-Valverde, Hannan,

& Rodriguez-Fernandez, 2011, 2012). This increase in investmentbanks’ risk exposure could have contributed signifi-

cantly to the greater fragility of the banking and financial sector.

c

2019 New YorkUniversity Salomon Center and Wiley Periodicals, Inc.

Financial Markets,Inst. &Inst. 2019;28:241–260. wileyonlinelibrary.com/journal/fmii 241

242 DEGL'INNOCENTI ET AL.

But how do investment banks compete? And how does competition affect the investmentbanking business and the

risks these banks are exposedto? So far, we could not find any answer to these questions in the existing academic liter-

ature. Previous studies on this topic mainly focus on the commercial banking industry and largely overlookinvestment

banks,although these latter played a critical role in generating and spreading the global financial crisis. Therefore, shed-

ding lights on the mechanisms through which theycan raise their risk-exposure is of great importance to policy-makers

to identify prompt and efficient interventions to make the system less fragile.

This paper covers this gap and contributes to the existing literature in severalways. First, we empirically provide

new insights on the relationship between competition and risk for a large sample of investments banks, coveringa rel-

atively long-time span for both the pre- and post-crisis period (1997-2014). Our first contribution is to construct a

unique dataset of investmentbanking institutions operating in seven developed economies (i.e. France, Germany,Italy,

Japan, Switzerland, UK and the US). Not only our dataset is larger than those analyzed in published studies on invest-

ment banks (e.g.,Beccalli, 2004; Mamatzakis & Bermpei 2014; Radić, Fiordelisi, & Girardone, 2012), but it also contains

detailed information obtained from several sources: Bankscope, DataStream, the World Bank's World Development

Indicators, and the Heritage Foundation.

Second, we measure competition in the investment banking business at the firm level using both the Efficiency-

Adjusted Lerner Indexof Monopoly Power (Koetter, Kolari, & Spierdijk, 2012) and the Excess Price-Cost Margin (Gas-

par & Massa, 2006). These measures have a number of advantages over traditional competition measures as they

enable us to better account for investment banking features and for risk originated from profit maximisation. We

also calculate several ad-hoc measures of investment banks’ risk-taking that proxy for two measures of volatility of

an investment banks’ performance (i.e., rolling volatility for both ROA and total revenues),earnings-at-risk exposure

and market risk.

Thirdly, we distinguish between boutique investment banks (BIBs) and full-service investmentbanks (FSIBs). The

former specialize in particular segments of the market; they do not offer a broad range of services and are not part of

larger financial institutions; while the latter offer clients a range of services including underwriting, merger and acqui-

sition advisory services, trading, merchant banking and prime brokerage.1Globalization,through cross border invest-

ment flows, and M&As, as well as direct and portfolio investment in emerging markets have fuelled the profitability

particularly of FSIBs while, at the same time, exposing them to foreign market risks. Buch, Koch, and Koetter (2013)

note that international diversification may reduce but also increase the risk of an international financial firm depend-

ing on the correlation between domestic and foreign returns and on the volatility of foreign markets. Therefore, it is

important to examine the relationship between competition and risk by considering the different exposures to inter-

national markets of investmentsbanks.

Finally, we formalize the relationship between risk exposure and market conditions in investment banking by

implementing a panel-data vector auto-regression (VAR) methodology.This econometric approach fits very well our

research aims since it allows us to test the impulse responses of risk exposure to changes in the market struc-

ture and competition levels, and vice versa, while considering bank- and industry-specific effects. We test for the

short- and long-run effects of a change in risk exposure on the changes in competition and vice versa. As far as

we are aware our study is the first to apply a panel VAR approach in assessing the bank competition-risk tak-

ing relationship. We also perform several robustness and sensitivity checks to assess the reliability of our baseline

results.

Our evidence shows that higher competition (low marketpower) measured by the Efficiency-Adjusted Lerner index

of Monopoly Power or the ExcessPrice-Cost Margin is associated with higher risk exposure for both BIBs and FSIBs in

terms of increase in earnings-at-risk or revenuevolatility. We therefore find some support for the competition-fragility

hypothesis in the investmentbanking industry both before and during the crisis. These results are consistent with sev-

eralprevious studies on commercial banks (e.g., Allen & Gale, 2004; Forssbæck & Shehzad, 2015; Keeley,1990; Repullo,

2004).However, we also find a positive relationship between market power and marketrisk. This result is not unusual in

the literature since market measures change more frequently than accounting measures and better takeinto account

market perceptions of the bank's soundness in the future (Zigraiova& Havranek, 2016). Finally, compared to the exist-

ing research on commercial banks, we show that business models matter for risk-exposure. We find that smaller and