How investment banks react to a decline in league table rankings. Evidence from U.S. corporate bonds

Table of contents

List of abbreviations

1. Introduction

2. Review of related literature and hypotheses
2.1 The role of reputation in debt underwriting and other investment banking diciplines
2.2 Measuring reputation
2.3 League table competition and hypotheses

3. Description of employed variables
3.1 Measuring bond quality and underwriter compensation
3.2 League table performance and investment banking revenue share
3.3 Control variables

4. Sample construction, descriptive statistics and methodology
4.1 Used data and sample construction
4.2 Descriptive statistics
4.3 Methodology

5. Empirical findings and interpretation
5.1 League table performance and quality-based reputation
5.2 Investment banking revenues
5.3 League table performance and underwriter compensation
5.4 Interpretation of results

6. Conclusions

References

Appendix

List of abbreviations

illustration not visible in this excerpt

1. Introduction

“For Investment Banks, nothing is more important than making money, except perhaps the thing that tells people just how much money they are making. Enter the league table” – Tracy Alloway, September 27, 2013^[1]

The above statement addresses one of the central aspects of the modern investment banking industry, intensely discussed by both practitioners and academics: The achieved rank in the (annual) league tables. Although criticized on a regular basis for being easy to manipulate or not being an adequate measure for the overall quality of offered services (e.g. see Lonkevich, 2004 or Berman, 2007), league tables have become a popular tool to evaluate on the reputation and the prestige and investment banks owns (Bao and Edmans, 2011, p. 2290). The majority of league tables rank investment banks according to the value of securities underwritten or mergers & acquisitions (M&A) deals advised, with few examples also being based on the absolute number of issues or deals, respectively. In modern literature, the assigned rank is widely seen as a significant driver of future income (Fang, 2005, p. 2730) or the amount of fees an investment bank is able to charge as compensation for underwriting or advisory services (Golubov et al, 2012). Moreover, the league table rank is an established measure to divide investment banks into reputable and non-reputable actors, as among others seen in Rau (2000), Livingston and Miller (2000), Kale et al. (2003), McCahery and Schwienbacher (2010) or Andres et al. (2013). For this reason, investment banks are egged to compete for the best league table ranks, causing a phenomenon named “league table competition”. In the setting of M&A advisory, Derrien and Dessaint (2013) are the first authors to report results related to this phenomenon, finding investment banks to actively manage their league table position in order to achieve the best ranking possible. Yet, to the best of our knowledge, this study is the first to examine the influence of league table competition for the investment banking discipline of debt underwriting.

As a test ground, we choose the United States (U.S.) corporate bond market. We do this, as Fang (2005, p. 2733) reports this market to be very competitive, suggesting a good league table rank to be even more important for investment banks. Besides, as seen in Andres et al. (2013) we use a data set, which comprises bonds issued post enactment of the Gramm-Leach-Bliley Act (GLBA). Abolishing the second Glass-Steagall-Act, introduced in 1933 and modified several times until 1999, the GLBA caused a significant increase of competition among debt underwriters, resulting in a sharp decrease of charged fees (Andres et al., 2013, p. 97). As a last reason, the revision of section 20 of the Glass-Steagall-Act allowed formally commercial banks, such as J.P. Morgan, to enter into the market of underwriting corporate securities (Puri, 1999, p. 134), enabling them to set up their own investment banking units and to compete with traditional investment banks like Goldman Sachs, Morgan Stanley or Lehman Brothers. Therefore, we consider the U.S. corporate bond market to be a good setting to investigate league table competition among traditional investment banks and investment banking units of commercial banks, later on uniformly referred to as “investment banks”, offering underwriter services.^[2]

In order to determine the way investment banks react to a decline in league table ranks, we examine the quality of the bonds they underwrite. Different to other studies, we do not use league table ranks to separate investment banks into reputable and non-reputable, but focus on the investigation of a rank changes’ influence on underwriter behavior. The results of this study suggest underwriters to be significantly influenced by their relative dependency on investment banking revenues, when choosing on how to behave in the event of a decline in league table ranks. In a first step, we find investment banks to focus on underwriting high-quality bonds, indicated by a lower default risk as well as by lower yield (or benchmark) spreads and total return performance. However, when we control for the fraction of total revenues generated in investment banking, we document a shift towards underwriting issues of lower quality, the higher the dependency on these revenues turns out to be. As a major conclusion of this study, we interpret this effect as a contribution to the different business models of underwriters in the U.S. corporate bond market. While underwriters with a diversified revenue base, such as commercial banks, seem to be able to go through the costly (Mathis et al., 2009) (and time-consuming) process to (re-)establish a reputation as underwriters of high-quality issues, specialized investment banks follow another strategy in order to climb up the league tables as fast as possible. Furthermore, we find that investment banks with good league table performance lower their underwriter standards, as indicated by decreasing bond quality measures. We interpret this as further evidence for the “milking” of reputation, as initially demonstrated by Chemmanur and Fulghieri (1994) and also shortly discuss the potential existence of so-called “reputation cycles” in debt underwriting, inspired by findings of Mathis et al. (2009) for market of credit ratings.

Apart from investigating on the behavior of underwriters in the event of a decline in league tables, we study the relation between league table rank and underwriter compensation. Earlier studies, as of Fang (2005), report premium fees charged by reputable underwriters, when classified according to their league table rank, which we consider to further increase league table competition among investment banks. Our results support this perspective, suggesting both a continuous slipping off in league tables and a sudden, dramatic decline to negatively influence the gross spread of underwritten bonds. Due to the fact that our results are based on changes in league table positions, our findings enable us to also explain competition among those investment banks, which earlier studies uniformly classified as either being reputable or non-reputable. Although this study concentrates on the market of debt underwriting, we expect our findings to also be of relevance for other investment banking disciplines.

The remainder of this study is structured as follows. In chapter 2, we give an overview of related literature on the role of reputation in debt underwriting and other investment banking disciplines and introduce two major approaches used to determine the reputation in investment banking. Besides, we outline the way our study contributes to the current discussion on league table competition and derive our hypotheses. Chapter three summarizes the variables we employ to test our hypotheses, while in chapter four, we describe the data we base our analysis as on as well as the used methodical approaches. Additionally, we report the descriptive statistics of the employed variables and present first univariate results. In chapter five, we turn attention to the results of the multivariate analyses we run to test our forward-looking statements towards the influence of league table competition, investigating on how the own league table performance affects the selection of bonds, certified as underwriter. Apart from reporting the results, we compare the major findings with our initial assumptions and develop first suggestions for continuing research on the topic of league table competition among investment banks. Finally, chapter six concludes this study.

2. Review of related literature and hypotheses

2.1 The role of reputation in debt underwriting and other investment banking diciplines

The importance of reputation in investment banking, and debt underwriting respectively, has been intensely discussed in literature. Chemmanur and Fulghieri (1994) show that reputation is a necessary factor for financial intermediaries to establish themselves as credible information producers. Fang (2005, p. 2730) argues that the future income of investment banks as repeated players in a financial market is directly connected to the reputation they own, which is why they prefer to be credible on the long run rather than to trick clients in order to maximize profits short term. Hence, the own reputation is a valuable asset for investment banks and other financial intermediaries to keep business going, a view also supported by Rogerson (1983), who finds that reputable companies have more customers, as less unsatisfied customers leave the company. Another topic intensely discussed is the connection between reputation and pricing of investment banking services. Shapiro (1982, 1983), Rogerson (1983), Allen (1984) or Titman and Trueman (1986) find that, in markets of ex-ante unobservable product quality, a “premium price – premium quality” equilibrium exists. Nevertheless, empirical evidence for reputation being a driver of investment banking fees is ambiguous. Fang (2005) reports a premium charged by reputable bond underwriters, whereas Rau (2000) and Golubov et al. (2012) do so for Top Tier M&A advisors and Fernando et al. (2012) for the market of equity underwriting. James (1992) as well as Livingston and Miller (2000) report contrarian results in their analysis of IPOs and non-convertible debt issuance, respectively. However, both studies do not control for self-selection in the matching of issuer and underwriter, while studies controlling for endogeneity^[3] (e.g. Fang, 2005 and Golubov et al., 2012) argue in favor of a reputation premium.

In literature, we find two major approaches to further explain the role of reputation in underwriting business: The certification hypothesis based on Klein and Leffler (1981) and formulated by Booth & Smith (1986) suggests that reputable underwriters serve a certification function and lower the issuers informational costs, as they are likely to underwrite only those bonds, which fulfill a their high quality standards. The reason for the assumed behavior lies in the importance of reputation as a valuable asset for the underwriter, who seeks to minimize the risk of facing a reputational loss. For an uninformed investor, certification by reputable underwriters will therefore be a quality signal. Fang (2005) supports the certification function for the corporate bond market, as she documents lower yields in case an issuer used a reputable underwriter. This finding underlines earlier results, as of Datta et al. (1997). On the contrary, Andres et al. (2013) show significant larger informational costs for issuers using a reputable underwriter in the high-yield segment of the corporate bond market. This is in line with results of McCahery and Schwienbacher (2010), who show higher loan spreads for syndicated loans placed by reputable underwriters. The two studies raise questions about the certification hypothesis but support the exploitation of reputation-based market power by reputable underwriters. Formulated by Chemmanur and Krishnan (2012) in their study of IPO valuation, the market power hypothesis suggests that reputable underwriters are able to price an asset higher, compared to competitors of lower reputation, neglecting a certification function based on the fear to damage the own reputation. In fact, Mathis et al. (2009) show that reputable rating agencies tend to inflate their ratings of complex products in case those account for a major part of income. These results suggest a decreasing importance of evaluation standards the more important a source of revenue is. For debt underwriting, an investment banking segment characterized by relatively low fees and high competition^[4], these effects gain even more importance in the light of Strausz (2005), who analyzes the trade off between short-term profit maximization and long-term reputation decline, also discussed by Fang (2005). He finds that the functionality of certification depends on the ability of the creditor to cover certification costs and to earn a positive rent of the offered service. Thus, low chargeable fees increase the motivation for reputable underwriters to lower evaluation standards and related costs in order to maximize profit. Chemmanur and Fulghieri (1994) underline the sensitivity to exploit the own reputation in the setting of IPOs. Via an equilibrium model they show decreasing strictness of evaluation standards but increasing investment bank fees with increasing reputation, a phenomenon they call “reputation milking”. This behavior of reputable underwriters is furthermore incented by findings of Andres et al. (2013), who show that dominant investment banks do not suffer a decline in reputation even though they underwrite bonds of lower quality^[5]. Kale et al. (2003) reports the same, documenting highly stable reputation for the top echelon of advisors.

To recap, the role of reputation in debt underwriting and other investment banking segments is a widely discussed topic, but empirical evidence is mixed and in parts inconclusive. While authers like Fang (2005), in the spirit of older work like Klein and Leffler (1981), argue in favor of price as an indicator for quality in the corporate bond market and support the certification hypothesis, Andres et. al (2013) as well as McCahery and Schwienbacher (2010) neglect the certification function of reputable underwriters in the high-yield segment and for syndicated loans, respectively. They provide empirical evidence for the application of “reputation milking” and the market power hypothesis. Nevertheless, the majority of relevant studies do not neglect the importance of reputation in debt underwriting as a source of future revenue and (or) premium prices.

2.2 Measuring reputation

Most studies either use market share based league tables or tombstone-announcement based Carter-Manaster rankings in order to determine the reputation of an investment bank. Besides, earlier studies like Logue (1973), Tinic (1988) or Johnson and Miller (1988) choose the “bracket” approach of Hayes (1971), who introduced the so-called “bulge group” of investment banks.^[6] However, this classification is also based on tombstone announcements, which is why it can be seen as a pre-stage of the Carter and Manaster approach (Fang, 2005, p. 2734). While the Carter-Manaster measure is often used in equity market based studies (Livingston and Miller, 2000, p. 24) like Carter and Manaster (1990), James (1992), Logue et al. (2002), Booth et al. (2010) or Fernando et al. (2012)^[7], the market share based approach dominates in M&A or debt underwriting related studies as of Livingston and Miller (2000), Roten and Mullineaux (2002) Hunter and Jagtiani (2003), Fang (2005), Daniels and Vijayakumar (2007), Bao and Edmans (2011), Golubov et al. (2012) or Andres et al. (2013).

Carter-Manaster (1990) use tombstone announcements in order to determine the reputation of underwriters in the IPO market. Under the implicit assumption of a reputation-equilibrium, as suggested by Shapiro (1983), and in the spirit of anecdotal evidence as of Lewis (1984) and Monroe (1986), they argue that the reputation of an underwriter is reflected by its tombstone announcement position. Based on tombstone announcements in the period from 1979 to 1983, they develop a ranking scale by assigning integer ranks from zero to nine for each listed underwriter, determined by its position in the announcement. Afterwards, by choosing the ranking based on the very first announcement as departure point, ranks are updated in case a competitor dominates an underwriter, listed in the previous stage. As a central result, the authors present a ranking of equity underwriters based on a scale from zero to nine, with the most reputable underwriters taking the value of nine.^[8] Though, this reputation measure neither takes into account the different issue volumes, nor it reflects in how many transactions an underwriter was involved.

Usually used to determine an investment banks’ market share, measures based on issue / deal volume or total number of transactions (e.g. the so-called league tables) are a popular and relatively easy way^[9] (Megginson and Weiss (1991), Daniels and Vijayakumar, 2007, p. 505) to measure reputation. In the appendix, Table I provides an example for a league table of underwriters in the U.S. corporate bond market, based on the total volume of underwritten issues. The source of the data is Bloomberg.

[Insert Table 1 about here]

Fang (2005) applies the market share, as measured by the league table ranking, to classify eight investment banks as reputable, while Golubov et al. (2012), in her spirit, names eight investment banks Top-Tier.^[10] Both studies show superior service quality in case one of these eight investment banks serves as underwriter / M&A advisor, supporting the idea of measuring reputation via market share. Andres et al. (2013) define Top 3 underwriters according to their league table position, so do McCahery and Schwienbacher (2010) and Ross (2010). Rau (2000), using the value of advised transaction, defines five investment banks as reputable advisors, while Kale et al. (2003) name seven investment banks bulge bracket. Finally, Livingston and Miller (2000) even classify ten reputable underwriters.^[11]

Although the two discussed measures use different methodologies, Megginson and Weiss (1991) point out that their market-share based measure is highly correlated with the Carter-Manaster ranking. McLaughlin (1992, p. 239) finds a high Spearman rank correlation between market share and Carter-Manaster measure, while Fang (2005, p. 2735) reports a correlation of more than 95%. In comparison to the Carter-Manaster ranking, Megginson and Weiss (1991) argue that measuring reputation by market share is favorable due to not assuming reputation to be constant over time. On the other side, reputation based measures are widely criticized for being easy to manipulate, a problem intensely discussed by financial press in the

last years.^[12] Though, Bao and Edmans (2011, p. 2290) state that market share based league tables are an established measure of expertise for academics and practitioners. Generally, the investment banks classified as reputable do not significantly vary in the studies named above (e.g. see Fang, 2005, and Golubov et al., 2012), underlining the stability of reputation levels over time and across investment banking segments. This observation finds support in Gopalan et al. (2011), who show that reputable arrangers of syndicated loans suffer few to no reputation related costs in case of borrower bankruptcies or other signs of poor performance. Additional empirical evidence is provided in Rau (2000) and Andres et al. (2013).

2.3 League table competition and hypotheses

As discussed in section 2.1, a good reputation is likely to influence the business situation of investment banks in a positive manner. Kreps and Wilson (1982, p. 275) highlight the importance of a firm’s “good name” for its products. Shapiro (1982) documents that revenue streams go up with increasing reputation, while Rogerson (1983) argues, that reputable companies tend to lose less customers due to having a smaller amount of unsatisfied customers. Rau (2000) finds the position of a M&A advisor in the league table, interpreted as reputation level, secures future business, as investment banks with a good league table ranking are associated with a higher portion of closed deals in the past. Mathis et al. (2009) as well as Bouvard and Levy (2009) present results suggesting that rating agencies with good reputation attract a larger portion of future clients, further supported by Becker and Milbourn (2010) and in line with Mariano (2008, p. 2), who states that reputation is the main asset of rating agencies. Moreover, McCahery and Schwienbacher (2010) point out that reputable arrangers of syndicated loans are not only involved in larger deals, but are also linked with borrowers of higher credit ratings, while findings of Michel et al. (1991) reinforce the idea that top tier arrangers can select deals of superior quality. These findings support the thesis that a good reputation enables investment banks to select deals of superior quality.^[13] Besides, we showed in section 2.2 that, especially in case of debt underwriting and M&A advisory, measuring reputation via market share based league tables is an established methodology, applied by both academics and practitioners (Bao and Edmans, 2011, p. 2290). Surprisingly, we find only few studies to examine the question of how investment banks react on a slipping off in league table rankings, hence a decline in reputation. Although authors like Fang (2005) highlight the stability of reputation levels, this mainly refers to the classification “reputable” or “bulge bracket”. In fact, Andres et al. (2013, p. 101) state that some variation in the TOP 10 underwriters’ league table positions exists. This finding once again turns attention to the league table rank. Usually, the league table rank is used in order to separate reputable and non-reputable investment banks. To our knowledge, Derrien and Dessaint (2013) are one of the few authors to consider the league table rank itself as a possible driver of business generation or reputation.^[14] Their study investigates on how investment banks engage in manipulating their league table position in order to improve the own competitive position. As a major result, the authors highlight the league table rank as a significant predictor of future market share and as an important driver of future business generation, incenting investment banks to actively “manage” their league table position.^[15] These observations find support in Netter et al. (2011), who underline that investment banks monitor their league table position very intensely. Additionally, Derrien and Dessaint (2013) report investment banks, performing poorly compared to their last full year rank, to be more likely to implement league table management tools than investment banks performing well. Later on, we will refer to this “chase” of investment banks for a good league table rank as “league table competition”. The pressure to retrieve a good league table ranking finds additional support in results of Plaksen (2011). He shows that investment banks, which experienced a decline in their (annual) league table position, are more likely to push on the close of deals the market actually dislikes.^[16] This is an important observation, as league tables only comprise successful transactions.

The findings described above put a new light on the importance of a high market share (and thereby a good reputation), as they underline the value investment banks assign their own position in the league tables. Besides, they explain competition inside a subgroup like the “bulge bracket” investment banks. Even tough, empirical evidence is limited, which motivates us to further investigate the reaction of investment banks to a decline in league table rankings. To be precise, we examine the influence of an underwriters’ league table performance on a selection of bond quality characteristics, discussed among other things in Andres et al. (2013). In the spirit of Mathis et al. (2009), who argue that rating agencies tend to built up a reputation based on quality, we set up our first hypothesis:

H1: Investment banks, which suffered a decline in league table ranks compared to the previous year, tend to focus on high-quality bonds in order to (re-)establish themselves as reputable underwriters.

To establish itself as a quality orientated underwriter, an investment bank needs time due to the fact, that the true quality of a bond is not observable for the market at the issuance date. However, time is quite a critical factor, as Plaksen (2011, p. 44) points out that investment banks face pressure to regain a better rank in league tables to keep their business going. The arising question is, whether investment banks can resist this pressure and keep on building up a reputation as described in H1. We assume the pressure to retrieve lost market share to increase with the relative importance of revenues derived from investment banking activities. For this reason, and once again in line with Mathis et al. (2009), we formulate the second hypothesis:

H2: When sliding in the league table, a high revenue share of investment banking fees tends to have a negative influence on certification standards, forcing investment banks to also underwrite bonds of lower quality in order to retrieve a better ranking as fast as possible.

Finally, we contribute to existing literature and examine the relation between reputation and compensation. Findings of Puri (1999), Fang (2005) and Golubov et al. (2012) suggest that reputation comes at a price, which is why we check whether a decline in league table ranks causes a drop of underwriter fees. We hypothesize the following:

H3: A decline in league table rank negatively influences the fees an underwriter is able to charge as compensation for its services, resulting in a lower gross spread of bonds underwritten.

3. Description of employed variables

In this chapter, we explain the variables used to test the influence of league table competition in the U.S. corporate bond market. As formulated in section 2.3, we believe two aspects to significantly influence an underwriters’ decision which bonds it underwrites and which it does not: The recent league table performance and the fraction of total annual revenues generated by investment banking services. To test our hypotheses, we in a first step define several variables to measure the overall risk of the bonds included in our data set. We do this as we derive a bonds quality from the amount of risk it carries, thus higher risk indicating lower quality (Andres et al., 2013, p. 112). To be able to test H1 and H2, in a next step we concentrate on the lead underwriters of all available bonds and investigate their past league table performance as well as the share of revenue they derive from investment banking and debt underwriting, respectively. In case a syndicate of underwriters is used (e.g. see Shivdasani and Song, 2011), we use the most reputable underwriter of the syndicate, as indicated by its last reported league table rank prior to the bond issuance, as representative. This procedure is in line with, among others, Fang (2005) and McCahery and Schwienbacher (2010).

In section 3.1, we explain the variables used to measure both the risk of bonds included in our sample and the fees underwriters are able to charge as compensation for their service. League table performance measures as well as indicators for the depedency of an underwriter on its investment banking revenenues are presented in section 3.2, while section 3.3 summarizes the control variables employed in all our models.

3.1 Measuring bond quality and underwriter compensation

While earlier approaches as of Livingston and Miller (2000), Fang (2005) or Daniels and Vijayakumar (2007) evaluate the quality of bonds from the issuers’ perspective, Andres et al. (2013) take an investors’ point of view. They argue, that bond investors are directly affected by the rating performance of their investment, which is determined by rating actions during the bond’s run-time^[17] as well as by the question whether a bond defaulted or survived. The use of default indicators is inspired by the work of Altman (1989) or Puri (1994), who studies the long-term default performance of investment bank-underwritten issues compared to non-investment bank-underwritten issues. Referring to the influence of rating actions, Carty and Fons (1993) show that the probability of a downgrade following a downgrade within a year significantly exceeds that of an upgrade following a downgrade. Furthermore, Lando and Skødeberg (2002, p. 424) point out that a series of downgrades may lead to higher default probability, whereas Andres et al. (2013, p. 100) highlight the importance of the first observable rating action, as they find a strong tendency towards subsequent rating changes in the direction of the initial change. To account for an initial downgrade as a source of risk for bond investors, and with respect to Andres et al. (2013), we apply the dummy variable FARDG (first-action-rating-downgrade), which is set to the value of one in case a bond’s first rating action is a downgrade, and zero otherwise. Hereby, values are not affected by the length of time since the issuance date. To control for bond defaults, we use a second dummy variable Default, which takes the value of one if a bond defaulted within our study period, ending in 2010. For this reason, bonds defaulting later than 2010 take the Default value zero.

[...]

^[1] Articel published by Finanical Times. The full text is avaiable at http://www.ft.com/intl/cms/s/0/10f8bce2-26f6-11e3-9dc0-00144feab7de.html#axzz2zomRwteL.

^[2] When we discuss the implications of a high fraction of revenues earned in the Corporate and Investment Banking Devision on underwriter behavior, we will come back to distinguish between traditional investment banks and commerical banks with a more diversified revenue base.

^[3] Both studies use the two-stage approach of Heckman (1979) to account for the endogeneity inherited in the choice of underwriter or M&A advisor, respectively. The effect of self selection has further be discussed by authors like Carter and Manaster (1990), Beatty and Welch (1996) or Andres et al. (2013).

^[4] In her paper, Fang (2005, p. 2733) describes debt underwriting as a commodity market, naming more than 6% lower charged advisory fees compared to IPO underwriting.

^[5] The authors measure reputation by using the volume based league table position the investment bank owns. Bond quality is expressed by downgrade and default risk.

^[6] The classification „bulge bracket“ still is a popular demoniation of elite investment banks on Wall Street, used regulary in financial press and related literature.

^[7] Additionally, Fernando et al. (2012) use the market share based Megginson and Weiss ranking, developed in Megginson and Weiss (1991).

^[8] The approach is explained in detail in Carter and Manaster (1990, p.1054 – 1056). Underwriters taking the value of nine dominated their competitors for the whole study period, while investment banks taking the value zero never ranked above any other firm in the tombstone announcements.

^[9] If compared with setting up a Carter-Manaster ranking, especially for longer study periods.

^[10] Accoring to Fang (2005, p. 2735), the „reputable“ investment banks are Goldman Sachs, Merrill Lynch, Morgan Stanley, Salomon Brothers, CSFB, Lehman Brothers, JP Morgan, and DLJ.
Golubov et al. (2012, p. 277) name Goldman Sachs, Merrill Lynch (now Bank of America Merrill Lynch), Morgan Stanley, JP Morgan, Credit Suisse First Boston, Citi/Salomon Smith Barney, Lehman Brothers (now Barclays Capital), and Lazard „Top-Tier“.

^[11] The used dummy variable REP is either market share based or takes the value of the Carter-Manaster ranking in the study period (see p. 26).

^[12] Articels published by Bloomberg (Lonkevich, 2004), the Wall Street Journal (WSJ) (Berman, 2007) or Handelsblatt (2007) document the manipulation of league tables by prominent investment banks, as they e.g. advise themselves in acquisitions.

^[13] McCahery and Schwienbacher (2010) point out that their results suggest the market for syndicated loans to be different from the public bond market, where they assume reputation to fulfill the certification function, already discussed in 2.1.

^[14] In line with Schenone (2004), Andres et al. (2013) use a lead underwriters’ annual market share as robustness check.

^[15] Among other thing, Derrien and Dessaint identify the selling of fairness options and the obtaining of co-advisory roles as league table management actions.

^[16] The markets refusal of a deal is indicated by negative announcement effects (Plaksen, 2011, p. 44).

^[17] As Andres et al. (2013, p. 100) point out, rating actions may, for example, directly affect institutional investors like insurance companies or pension funds, as regulation rules require them to adjust capital buffers to an investment’s credit rating. The costs of rating changes out of a firm’s perspective and their influence on the choosen capital structre are further described in Kisgen (2006).