ou need to put yourself as a reviewer in an official journal and provide a detailed peer review feedback (2-3 pages) on a paper. Please see the attachment for the paper and some PowerPoints regarding how to provide good detailed feedback.

What do you think about the paper?

  • »  What is good about it? What is bad?
  • »  How clear is the theoretical contribution?
  • »  How sound are empirics?

Agenda

1. Editorial policy and research fields
2. Providing feedback

1. Guidance on reviewer reports
2. Preparing and presenting a reviewer report

3. Comparing feedback
4. Responding to feedback
5. Revisions and further rounds of reviews
6. Presenting, providing feedback to, and discussing research (I)
7. Presenting, providing feedback to, and discussing research (II)
8. Presenting, providing feedback to, and discussing research (III)

Page 27UIBK | Thomas Lindner

Agenda

1. Editorial policy and research fields
2. Providing feedback

1. Guidance on reviewer reports
2. Preparing and presenting a reviewer report

3. Comparing feedback
4. Responding to feedback
5. Revisions and further rounds of reviews
6. Presenting, providing feedback to, and discussing research (I)
7. Presenting, providing feedback to, and discussing research (II)
8. Presenting, providing feedback to, and discussing research (III)

Page 29UIBK | Thomas Lindner

Guidance on how to prepare a review report1 (I)

» General guidance

» The job of the referee is to provide expert and unambiguous advice to the editor about whether or not a paper is
publishable. The referee advises, the editor decides. In the case of a recommendation to invite resubmission, the
referee should advise the editor about any changes that the reviewer believes are needed to make the paper
publishable. In contrast, the referee should not advise the editor to require any change that does not affect the
paper’s publishability. Referees are free to make suggestions for improving a paper, but it is important to make
clear in their reports that these comments are suggestions to the authors for improvement or extension, not
advice to the editor on requirements for publication.

Page 30UIBK | Thomas Lindner

1Note that this is a particular finance (in fact, Journal of Finance) perspective

Guidance on how to prepare a review report1 (II)
» Cover Letter
» The ideal cover letter provides three types of information:

1. A summary of the paper’s core contribution. The editor may not be an expert in this field, and it is often hard to figure out the paper’s main
point or line of reasoning.

2. What are the strengths and weaknesses of the research in its current state; and
3. A frank assessment: is the core contribution of the paper, as it stands, a publishable result, and is it likely to be publishable with one round of

revision?

» Advice to the editor should be decisive. You are being asked to make a recommendation: accept, revise or reject.
Reasons for uncertainty can be included in the explanation for the recommendation. If the paper is somewhat outside
your area, you might suggest that a second opinion be sought and you should provide names of candidate referees, and
if possible, what specific issues the alternative referee can address that you felt were outside your area of expertise.

» A revise recommendation is a serious commitment given that A-level publications are rare. You should not make a
revise recommendation to simply defer judgment. It is not helpful for Editors to hear that ‘this paper seems ok, but I am
not sure, let’s see what the authors can do.’ Most of the work you put into the refereeing process should be at the
initial stage. That said, it is crucial that you stick with the paper and make sure the paper attains the journal’s high
standard when it is resubmitted.

Page 31UIBK | Thomas Lindner

1Note that this is a particular finance (in fact, Journal of Finance) perspective

Guidance on how to prepare a review report1 (II)

» Referee report

» The first section of the report (often one paragraph) should contain a concise summary of the paper’s claimed results,
contributions, and general line of reasoning. The editor is typically not an expert in the paper’s subfield, so it is important
for this summary to be clear. Only after this, turn to substantive issues about the importance and validity of the claimed
results.

» The main job of the referee is not:
1. To help write the paper as a quasi-coauthor
2. Make an unpublishable paper publishable by directing the research
3. To ensure that the paper cites the referee’s work.

» Take a scientific stance in your report. Do not insult the authors, or use overly emotional or accusatory language. Avoid
ascribing bad intent to authors (“The authors were trying for a cheap publication,” “The authors were trying to brush
past literature/conflicting findings under the rug…”) and focus on the substance of the paper. If there are indications of
intellectual dishonesty, state the facts rather than speculating on intent. If an accusation is made, leave it for the cover
letter to the Editor.

Page 32UIBK | Thomas Lindner

1Note that this is a particular finance (in fact, Journal of Finance) perspective

Guidance on how to prepare a review report1 (III)

» Content of the report

1. The importance of the paper. This is the most subjective part of the report. Space is limited in A-level journals; in
many economics-related fields, they routinely reject more than 90% of submissions. There are plenty of “correct”
papers that do not make a significant enough marginal contribution to existing knowledge. The editor needs to
assess the importance of the contribution aided by your report. The report should contain an argument that
supports your assessment of the importance of the work and detail the considerations that bear upon your
judgment.

2. Problems with the paper that render it unpublishable. In this case provide a scientifically convincing argument
for why these problems render the paper unpublishable (i.e. you believe the problems are serious enough that
they are unlikely to be fixable in the next round). The argument needs to be clear and understandable to the
editor (and authors). You should not merely list your objections to the paper. You need to explain why those
objections are serious enough to render the paper unpublishable.

Page 33UIBK | Thomas Lindner

1Note that this is a particular finance (in fact, Journal of Finance) perspective

Guidance on how to prepare a review report1 (IV)

» Content of the report
3. Problems with the paper that currently make it unpublishable, but which you believe could be corrected. In this

case be very clear why the paper is unpublishable (see above) and what a correction to the problem would look
like. If you suspect that there are problems with the empirical work, this is where you put those concerns and
specify what additional work the authors would need to do to satisfy you. Whatever you suggest is going to cost
the author significant time. If the author satisfactorily addresses the issues, you should recommend publication.

4. Problems with the paper that do not render the paper unpublishable. Here you do not need to provide reasons
for your opinion, but you cannot hold up publication if the authors do not address these problems. In many cases
people disagree about what should and should not go into the paper. Ultimately, the author’s name goes on the
paper, not yours. It is the author’s decision on how best to write the paper, not yours. Authors also have a
responsibility not to waste good or important suggestions, subject to the fact that there are differences of
opinion and that suggestions are costly to implement. It is not appropriate for referees to try to enforce author
responsibility by taking a paper hostage if that paper is already publishable.

Page 34UIBK | Thomas Lindner

1Note that this is a particular finance (in fact, Journal of Finance) perspective

Agenda

1. Editorial policy and research fields
2. Providing feedback

1. Guidance on reviewer reports
2. Preparing and presenting a reviewer report

3. Comparing feedback
4. Responding to feedback
5. Revisions and further rounds of reviews
6. Presenting, providing feedback to, and discussing research (I)
7. Presenting, providing feedback to, and discussing research (II)
8. Presenting, providing feedback to, and discussing research (III)

Page 35UIBK | Thomas Lindner

Agenda

1. Editorial policy and research fields
2. Providing feedback

1. Guidance on reviewer reports
2. Preparing and presenting a reviewer report

3. Comparing feedback
4. Responding to feedback
5. Revisions and further rounds of reviews
6. Presenting, providing feedback to, and discussing research (I)
7. Presenting, providing feedback to, and discussing research (II)
8. Presenting, providing feedback to, and discussing research (III)

Page 39UIBK | Thomas Lindner

What do you think about the paper?

» Interorganizational Diversity, Institutional Risk, and the Formation of Multipartner Syndicates

» What is good about it? What is bad?

» How clear is the theoretical contribution?

» How sound are empirics?

Page 40UIBK | Thomas Lindner

What do you think about the paper?

» Interorganizational Diversity, Institutional Risk, and the Formation of Multipartner Syndicates

» What is good about it? What is bad?

» How clear is the theoretical contribution?

» How sound are empirics?

» Prepare a review presentation with your group (90 minutes)

» Leave 45 minutes for presentation (5 minutes max) and discussion

Page 41UIBK | Thomas Lindner

1

Interorganizational Diversity, Institutional Risk,

and the Formation of Multipartner Syndicates

Abstract: We examine how interorganizational diversity affects the formation of multipartner

banking syndicates that come together to finance some of the world’s largest infrastructure

projects. Previous studies have shown that diversity negatively affects partnership formation

because it increases the costs of cooperation among the partners and the probability of conflict

between them. We extend this research and argue that the negative relationship between

interorganizational diversity and partnership formation is moderated by the institutional context

in the country where the infrastructure project will be located. We compare 1,100 realized with

over 25,000 unrealized multipartner syndicates and examine the likelihood of their formation in

countries with different levels of institutional (specifically, political and social) risk. We show

that diverse syndicates are more likely to form when investing in high-risk countries than when

investing in low-risk countries. Our findings suggest that the country-level institutional

environment is an important consideration in the formation of interorganizational partnerships.

2

INTRODUCTION

Diverse organizations often come together to accomplish common goals. Corporate

organizations collaborate across industry lines and national borders. Increasingly, they work with

nongovernmental organizations and form public–private partnerships with government entities.

Large firms work together with small firms, experienced firms with less experienced ones, and

publicly traded firms with privately or state-owned firms. The vast literature on

interorganizational partnerships, which has analyzed cooperative agreements under a multitude

of settings, shows that firms value resource complementarity but also prefer to collaborate with

similar firms (Chung, Singh, & Lee, 2000; e.g., Gulati, 1995; Rothaermel & Boeker, 2008).

Differences between firms and, more broadly, interorganizational diversity—that is, “the

distribution of differences” (Harrison & Klein, 2007: 1200; emphasis added) between

organizations when more than two form a partnership—increase the costs of cooperation and

monitoring, and the probability of conflict between partners (Goerzen & Beamish, 2005; Gulati

& Singh, 1998; Parkhe, 1993). Given such costs, the formation of diverse interorganizational

partnerships, especially those involving a large number of partners, is somewhat surprising.

Yet, diversity also enhances access to complementary resources (Rothaermel & Boeker,

2008; Sampson, 2007) that are particularly valuable for managing environmental uncertainty

(Gulati & Gargiulo, 1999; Pfeffer & Salancik, 1978; Thompson, 1967). We argue in this paper

that diverse partnerships are more likely to form when firms invest in a high-risk environment.

The local institutional context of an investment—which we define as country-level political and

social risk—affects the extent to which complementary resources are needed and should

therefore influence partnership composition. When an investment in a start-up venture or a large

infrastructure project is made in a risky institutional context, complementary resources are more

3

likely to be required to manage the complexities of operating in that context. Moreover, in a

high-risk institutional environment, it is more difficult to know ex ante what specific resources

will be needed over the course of a project. Like members of an expedition who know little about

the terrain ahead, firms investing in high-risk institutional environments might want to equip

their partnership with diverse resources that can be used to manage different types of challenging

situations. Thus, when investing in high-risk locations, organizations are more likely to form

diverse partnerships to better position themselves to develop and protect their investment.

To understand how diversity affects the formation of interorganizational partnerships, we

examine the composition of multipartner banking syndicates that finance large infrastructure

projects in 72 countries around the world. More precisely, we compare the composition of 1,110

banking syndicates that formed (i.e., realized syndicates) with the composition of over 26,000

syndicates of similar size that were sampled randomly from the full population of syndicates that

could have been formed by the same banks (unrealized syndicates). We study variations in the

syndicates’ composition using a new measure of partnership-level diversity that considers the

distribution of differences among all participating organizations across four dimensions: firm

size, country of origin, previous experience, and ownership type. While on average the two

samples (realized and unrealized syndicates) have similar interorganizational diversity, we find

that highly diverse partnerships are more likely to form when the project that brings them

together is located in countries with high levels of institutional (i.e., political and social) risk.

Our study advances research on interorganizational partnerships in several ways. First, we

highlight that the institutional context influences the composition of interorganizational

partnerships. Prior research on interorganizational partnerships has emphasized that market

uncertainty impedes the formation of distant ties (Beckman, Haunschild, & Phillips, 2004;

4

Podolny, 1994; Sorenson & Stuart, 2008). This research has focused on market uncertainty

without considering the political and social environment of the investment. By contrast,

international business studies have highlighted that institutional risk in the host country strongly

affects the formation of joint ventures with local partners (Brouthers, Brouthers, & Werner,

2003; Delios & Henisz, 2000; Liu & Maula, 2016; Meyer, Estrin, Bhaumik, & Peng, 2009), but

these studies have focused largely on partnerships between multinationals and local firms.

Across these fields of research, few studies (e.g., Vasudeva, Spencer, & Teegen, 2013) have

analyzed the effects of the institutional context on the formation and composition of

interorganizational partnerships. We suggest that firms are more likely to seek diverse partners

(foreign or local) when they operate in high-risk institutional contexts. In such environments, it is

very difficult to know ex ante which resources are required over the duration of the project, and

therefore which partner is best suited for it. Hence, we argue that diverse partnerships are more

likely to be formed to complete projects in risky institutional environments.

Second, we evaluate how diversity affects the formation of multipartner syndicates (of

three or more banks) by focusing on the partnership as the unit of analysis. Prior research has

examined the formation of interorganizational partnerships as dyadic ties (Ahuja, Polidoro, &

Mitchell, 2009; e.g., Gulati, 1995). By contrast, studies that evaluate the formation of

interorganizational partnerships by focusing on the entire partnership as the level of analysis are

still rare (Heidl, Steensma, & Phelps, 2014). Yet, when considering partnership formation, an

important distinction needs to be highlighted between the effects of a difference between two

firms and the effects of diversity (i.e., the distribution of differences) that describe a partnership

involving more than two firms. This distinction is important both at a theoretical level—to

separate the concepts of difference (or distance) and diversity—and in terms of concept

5

measurement. At a theoretical level, both the costs and the benefits of diversity are nonlinear

functions of the number of partners, as differences between partners are not additive. As the

number of partners increases, information asymmetries, incentives to free-ride, and the costs of

monitoring increase considerably. Moreover, processes of coordination and exchange work

differently in multipartner settings, and the formation of coalitions within the partnership can

affect its performance and survival (Heidl et al., 2014). Consequently, interorganizational

diversity is likely to affect the formation of partnerships differently between multipartner

coalitions and two-party ones. While differences between two partners may be addressed ex-ante

with relational contracts that complement formal ones (Poppo & Zenger, 2002), such governance

structures become exponentially more difficult to devise and implement when multiple partners

with differing interests work together (García-Canal, Valdés-Llaneza, & Ariño, 2003).

Considering these complexities, we expect interorganizational diversity to have a more

pronounced negative effect on the formation of large, multipartner syndicates. Second, the

measurement of diversity as the distribution of differences between three or more organizations

is not a simple extension of the measurement of difference in interorganizational dyads. We

propose in this paper a new approach to measuring interorganizational diversity that (1) focuses

on the partnership as the level of analysis, (2) reflects the distribution of differences across all

the organizations involved, and (3) captures it along a number of relevant dimensions.

Finally, we extend the study of interorganizational partnerships to a new setting: the

formation of large, multipartner banking syndicates that come together to finance some of the

largest infrastructure investments around the world. This setting is known in the industry as

project finance (hereafter, PF) and serves as a favorable empirical context for testing the

influence of the institutional environment on the formation of diverse interorganizational

6

partnerships. The success of the large infrastructure projects in our data (including pipelines,

airport expansions, and large hydroelectric power plants) requires the continuous support of

many local stakeholders. Different national government agencies, local communities, their

mayors, and local councils are involved, as are suppliers of equipment, materials, and workforce.

Opportunities for opportunistic behavior abound, and project success depends on the successful

management of such risks. In PF, banks assume most of these risks by offering the project

company a nonrecourse loan that can only be repaid from the cash flows generated by the

infrastructure project, once successfully completed. Thus, to recover their loans, banks must

work together to mitigate the risks in the institutional environment, much like firms joined by an

R&D alliance must work together to ensure the success of their joint project.

INTERORGANIZATIONAL DIVERSITY AND PARTNERSHIP FORMATION

Prior literature defines interorganizational partnerships as cooperative strategies that cover “any

type of agreement between two or more firms, contractual or otherwise, involving mutual

forbearance towards one or more (typically not identical) goals by providing capital, knowledge,

technology, managerial talent, and/or other valuable assets under the purview of said firms”

(Beamish & Lupton, 2016: 163 ). Research on the formation of interorganizational partnerships

has focused extensively on the structural and relational factors affecting the collaborating parties

(and Ahuja et al., 2009; e.g. Gulati & Gargiulo, 1999). But scholarship in this area has paid less

attention to the attributes of the partners involved (Lavie & Miller, 2008: 623 ) or to how the

selection of partners might be influenced by “the moderating effects of country-level institutional

characteristics” (Vasudeva et al., 2013: 319; see also Sorenson and Stuart, 2008).

We suggest here not only that both the composition of a partnership and the context within

which it operates are important determinants of its formation, but also that the two might

7

interact. Specifically, the broader institutional context and the challenges therein might affect the

selection of partners with specific attributes. In the discussion that follows, we build, first, on

existing literature to highlight that in the context of multipartner syndicates or alliances, more

diverse partnerships are less likely to form than less diverse partnerships. We then examine

whether this is true in both low-risk and high-risk institutional environments and argue that the

relationship between interorganizational diversity and partnership formation is moderated by the

degree of risk in the institutional environment where the investment is located. In the context of

syndicates bringing together three or more banks to finance some of the largest infrastructure

projects around the world (the empirical setting in our study), our argument implies two

observable implications. First, we are more likely to observe less diverse syndicates forming

(compared with counterfactual syndicates that could have formed but did not), and, second, we

are likely to observe more diverse syndicates forming when their projects are located in high-risk

locations than when the projects are located in low-risk locations.

Prior research on interorganizational partnerships and alliances has studied them either

from the perspective of one firm seeking new partners or as dyadic relationships between two

firms seeking a “match” (Mindruta, Moeen, & Agarwal, 2016). In both conceptualizations, the

discussion has largely focused on differences (or lack thereof) between firms (Chung et al., 2000;

Gulati & Gargiulo, 1999; Harrison, Hitt, Hoskisson, & Ireland, 2001; Mitsuhashi & Greve,

2009). Scholars highlighted the tension between the benefits of combining complementary

resources from different firms (e.g., combining complementary technologies or markets) and the

costs of managing the differences between them (Colombo, Grilli, & Piva, 2006; Jiang, Tao, &

Santoro, 2010; White & Siu-Yun Lui, 2005). Some studies suggested that complementarities

increase the likelihood of a partnership (Chung et al., 2000; Colombo et al., 2006). For instance,

8

Gulati and Gargiulo (1999) analyze alliances between industrial firms and find that

complementarities in industry segments and geographical markets increase the likelihood of

alliance formation.1 Chung et al. (2000) analyze underwriting syndicates and find that

complementarities in industry, status, clientele, and specialization increase the formation of

dyadic ties. Similarly, Mitsuhashi and Greve (2009) find that resource and market

complementarity increase the likelihood of alliance formation in liner shipping; and in the

biopharma industry, Rothaermel and Boeker (2008) and Mindruta et al. (2016) find that

complementarities in market niches, size, and research and development capabilities increase the

probability of an alliance between two firms. Other studies, however, show that differences

between firms lower the probability of their forming a partnership. Some scholars highlight that

differences in status impede partnership formation (Chung et al., 2000; Gulati, 1995; Hallen,

2008; Podolny, 1994). Ahuja et al. (2009) find that technical, product, and market dissimilarity

decrease the likelihood of alliance formation.2

We argue that this tension is exacerbated in partnerships involving more than two firms. A

more diverse partnership has access to a wider range of resources, but it is also more difficult to

manage. A large number of partners makes it more difficult to devise formal structures to govern

the resource exchange (Li, Eden, Hitt, Ireland, & Garrett, 2012). Coordinating activities is also

more difficult when the partnership involves a large number of firms (Heidl et al., 2014), and

monitoring the efforts of different partners is cumbersome because not every firm can easily

observe the others’ activities and must instead rely on indirect reports and assessments (Lavie,

1 Authors conceptualize complementarity as “absence of overlap” in geographic regions and industry subsegments

(Gulati & Gargiulo, 1999: 1466).
2 Parkhe (1991) highlighted this tension by distinguishing between “Type I diversity,” which refers to “reciprocal

strengths and complementary resources furnished by the alliance partners,” and “Type II diversity,” which refers to

costs that threaten the stability and efficiency of cooperation. Parkhe’s terminology, however, employs the term

“diversity” as a synonym for differences between two firms. We use “diversity” to refer to the distribution of

differences between three or more firms.

9

Lechner, & Singh, 2007). Moreover, the partnership cannot rely on direct reciprocity or on the

relational contracts that often complement formal ones in interorganizational partnerships. As a

result, free-riding and conflict between partners are more likely (Fonti, Maoret, & Whitbred,

2016; García-Canal et al., 2003), and the partnership must devise more complex (and therefore

more expensive) rules of exchange.

Research analyzing partnerships between more than two organizations has examined their

governance (Li et al., 2012) and their stability and performance (Beamish & Kachra, 2004;

Dussauge, Garrette, & Mitchell, 2000; Heidl et al., 2014), whereas studies analyzing their

formation have been more tentative. Dussauge et al. (2000) examined the stability of

manufacturing alliances and showed that those involving more than two firms are more likely to

dissolve. Heidl, Steensma, and Phelps (2014) found that unplanned dissolutions are more likely

in multipartner alliances because “divisive faultlines” can split the alliance partners into

competing subgroups. Fonti, Maoret, and Whitbred (2016) further highlighted the challenges of

cooperation in multipartner alliances, showing that perceptions of alliance effectiveness translate

into free-riding behavior when the alliance includes more than two partners. As a result, the

governance of multipartner alliances differs considerably from governance in dyadic relations, as

shown by Li et al. (2012).

To the best of our knowledge, only Li (2013) and Heidl et al. (2014) have examined the

formation of multipartner alliances, and they have done so as first-stage models in the analysis of

performance and stability, respectively. Li (2013) argued that market competition between

partners and market uncertainty have an inverse U-shaped effect on the formation of multipartner

R&D alliances, and that top managers’ social capital and firms’ technological capabilities

increase the probability of forming such alliances. Heidl et al. (2014) showed that geographic

10

diversity (measured as the number of nations in which the partners are headquartered) decreases

the likelihood of forming a multipartner alliance, whereas the partners’ positional embeddedness

slightly increases the probability of forming the alliance. We build on these two studies and

argue, first, that interorganizational diversity (measured at the partnership level) impedes alliance

formation and, second, that this effect is moderated by the broader institutional environment. We

build our argument below as we highlight how diversity and the institutional environment both

affect the functioning of large, multipartner syndicates who come together to finance large

infrastructure projects around the world.

Diversity in the Context of Multipartner Loan Syndicates

Multipartner banking syndicates lend themselves in a unique way to studying the effects of

diversity on partnership formation. They involve a large number of partners that commit to a

joint loan. Syndicated PF loans, in particular, have become the most important means of

financing large-scale and often controversial projects in developing countries (Esty, Chavich, &

Sesia, 2014). PF syndicates typically provide more than 80 percent of capital to a project

(Brealey, Cooper, & Habib, 1996). One or two lead arrangers assemble a syndicate with

sufficient capital, knowledge, and influence to realize the project (Gatti, Kleimeier, Megginson,

& Steffanoni, 2013). PF loans are nonrecourse—that is, without the possibility to recover the

investment from the sponsors (Byoun, Kim, & Yoo, 2013; Esty, 2004a). As a result, the

participating banks actively structure and manage the project and the risk of default. The banks

also bind themselves together through a set of agreements that give each bank a veto over the

restructuring of a loan, should such restructuring (or refinancing) be required in case of default

(Esty & Megginson, 2001; François & Missonier-Piera, 2007). Thus, similar to other alliances

11

decisions, the banks must a-priori weigh each partner’s potential for opportunistic behavior and

the complementary resources that partner contributes to the syndicate.

Diverse lending syndicate offer several advantages. Banks differ in their experience, financial

strength, host-country legitimacy, political networks, and host-country influence. Syndicating with

a diverse set of banks increases the availability of complementary resources that can be leveraged

to complete a large-scale project. The popularity of Islamic finance in PF, for example, can be

explained by such unique complementary knowledge, legitimacy, and socio-political networks

(Esty, 2000; McMillen, 2001). In a different example, when Chase Manhattan Bank evaluated

syndication strategies for the financing of the $3.3 billion Hong Kong Disneyland, it decided on a

large, global syndicate involving local Chinese and state-owned banks because of their knowledge

of the context and the political connections they provided (Esty, 2001).

At the same time, diverse syndicates are more vulnerable to free-riding, moral hazard, and

opportunistic holdup, especially when a multipartner loan requires refinancing (Pichler &

Wilhelm, 2001; Simons, 1993). Contractual agreements signed when forming the syndicate

require all banks to agree to alterations of the loan terms, should alterations be necessary (Dennis

& Mullineaux, 2000: 408). When predetermined performance indicators are not met, the loan

covenants stipulate that the lending syndicate assumes control of the project (Byoun et al., 2013;

Subramanian & Tung, 2016). The costs of diversity are most evident in such cases: renegotiation

of agreements are time-consuming (Nevitt & Fabozzi, 2000; Roberts & Sufi, 2009), especially

when partners differ in their risk aversion, governance, language, and institutional context

(Bolton & Scharfstein, 1996; Giannetti & Yafeh, 2012); and veto rights allow single partners to

engage in holdup through “forced lending” (Krugman, 1985: 84 ). James (1990: 329)

documented such episodes of “exploitation of the large by the small” in the U.S. syndicated loan

12

market in the 1980s, and Lipson explained that “banks with less at stake are less willing to

provide time or money. For them, rescue operations may require undue managerial attention, or,

in the case of smaller banks, an excessive share of capital” (1981: 616 ). The following quote

from a banking official describes this renegotiation problem: “a project is only as good as its

weakest link. And that will be the ultimate problem. What will they do when the project goes

wrong?” (Nevitt & Fabozzi, 2000: 54).

The Eurotunnel megaproject provides a clear example of the potential cost involved in

diverse banking syndicates. To raise the necessary capital, 50 lead banks arranged a syndicate of

200 banks. Within two years of starting construction, technical difficulties forced the project to

refinance. It took 18 months for the syndicate to reach a new agreement (Vilanova, 2005: 22).3

“When there are more than 200 banks amongst whom agreement is required, the difficulties can

only be imagined” (Grant, 1997: 52). In 1994, the banking syndicate had to refinance for a

second time, resulting in “the most intense and prolonged arguments of the project’s history—

mostly between the banks themselves” (Grant, 1997: 51 ). The three-year negotiations divided

banks along political lines. In contrast to British and French banks, Japanese banks and other

foreign lenders were largely “indifferent to the political pressures from English and French

governments” (Vilanova, 2005: 33) and requested (and were ultimately allowed) to sell their loan

shares. European banks ended up with larger shares of the loan and thus with higher exposure to

the risks of this project than they wanted.

Despite the benefits and costs of diversity described earlier, a number of arguments speak

for a negative net effect of diversity in the context of international loan syndicates. First, while

banks themselves are heterogeneous, their primary financial contributions are highly

3 A second syndicated loan over 2 billion GBP struggled to attract international capital and required excessive

commitments by four regional agent banks (Midland, Natwest, Credit Lyonnais, and Banque Nationale de Paris).

13

substitutable. If diverse syndicates form, they form to pool secondary resources related to

experience, specialized knowledge, access to networks, and political capital. Second, in contrast

to equity providers, lenders’ gains from loans are limited to the interest on the loan. Default risk,

on the other hand, is substantial—especially when nonrecourse provisions prevent banks from

seeking to recover their loans from the project sponsors and the assets financed are physically

tied to the foreign location. Therefore, lenders are generally more risk-averse. More importantly,

the default risk is closely linked to the syndicate’s ability to reorganize in the event of financial

problems (Esty & Megginson, 2003; Paligorova & Santos, 2015; Sang Whi & Mullineaux, 2004)

and, as a result, to the predictability of partners’ behavior (Esty & Megginson, 2000, 2001). A PF

practitioner described the problems related to reorganization in diverse syndicates: “A

reorganization is likely the best alternative in many troubled projects. Yet, [. . .] many diverse

interests in a project financing—export banks, political risk insurers, commercial banks, [. . .]—

may make a structured reorganization impossible” (Hoffman, 2007: 391). As a result, lead

arrangers must anticipate reorganization costs ex-ante by creating a reliable, homogenous

syndicate that is ideally tied together by interests, norms, and mutual dependence (Esty, 2001;

Esty & Megginson, 2000, 2001; Paligorova & Santos, 2015). We therefore hypothesize:

Hypothesis 1: The greater the interorganizational diversity among partners, the lower the

likelihood that a syndicate will form.

THE INSTITUTIONAL ENVIRONMENT AND THE FORMATION OF DIVERSE

PARTNERSHIPS

The formation of diverse interorganizational partnerships requires some form of inducement

(Ahuja, 2000), which can come from within the firm or from the external environment. Among

the studies that concentrate on external inducements, Beckman et al. (2004) distinguish between

14

market uncertainty and firm-specific uncertainty and find that under market uncertainty firms

reinforce existing networks, whereas under firm-specific uncertainty they seek distant or diverse

ties that can provide new information. Also, industry-level competition has been shown to

encourage alliance formation (Eisenhardt & Schoonhoven, 1996; Sahaym, Steensma, &

Schilling, 2007; Stuart, 1998; Yu, Subramaniam, & Cannella, 2013), and some studies relate

alliance formation to the firms’ home-country institutional environment (Hitt, Ahlstrom, Dacin,

Levitas, & Svobodina, 2004; Vasudeva et al., 2013).

Two studies within the alliance formation literature explicitly address host-country-level

inducements.4 Yu et al. (2013) analyze the global automotive industry and find that host-

government restrictions moderate the effect of competitive intensity on the likelihood of alliance

formation. They show that political hazards negatively affect alliance formation but do not

examine differences between partners. Most closely related to our work—albeit from an equity

investor’s perspective—Sorenson and Stuart (2008: 266) analyze how venture capital investors

form distant ties to other investors, depending on “the places and times” in which they invest.5

In contrast to this prior work, our study focuses on how institutional risks, or, more

specifically, political risk in the host country, encourage the formation of more diverse

syndicates. Host-country political risks, we argue, can be an equally important inducement.

Mitigation of political risk often requires firms to reach out to (diverse) partners with

complementary resources, including local knowledge, pertinent experience, risk-management

4 International business literature has examined the formation of joint ventures and has shown that the host-country

environment affects the choices for local partners (e.g., Delios & Henisz, 2000; Liu & Maula, 2016). Separately,

Vasudeva et al. (2013) relate alliance-partner selection to the partners’ home-country level of corporatism. Similarly,

in a comparison of Russian and Chinese companies, Li et al. (2004) show that partner selection varies with the

home-country institutional context.
5 Sorenson and Stuart (2008) hypothesize that risk reduces the likelihood of distant ties forming. However, their

conceptualization of risk relates not to the host-country institutional context but to information asymmetries between

the partners using measures such as stage of investment, syndicate size, and prior cooperation between partners.

15

capabilities, and social capital (Guler & Guillén, 2010; Koka & Prescott, 2002). International

business research has studied the role of local partners (Brouthers et al., 2003; Delios & Henisz,

2003; Liu & Maula, 2016) and partnerships with the host-country government (Inoue, Lazzarini,

& Musacchio, 2013; Vaaler & Schrage, 2009). We build on this research and consider country of

origin and ownership type (i.e., commercial vs. state-owned bank) as two of the dimensions

along which we evaluate interorganizational diversity.

The Moderating Role of Political Risk in the Context of Multipartner Loan Syndicates

Political risk is a key consideration in loan syndication, especially in PF loans, because these

finance large, location-specific infrastructure projects, and lenders can rely only on project cash

flows for repayment of the loan (Brealey et al., 1996; Esty, 1999). Moreover, many PF investments

are located in countries with high political risk, where limited political constraints on government

actors allow them to unilaterally change regulations that directly affect the infrastructure project

(Henisz, 2000; Henisz & Williamson, 1999). As Esty and Megginson highlight from a Moody’s

report, “in many countries, investors simply do not know if these [PF] contracts will be upheld as

legal, binding, or enforceable. [. . .] These contracts are worth little more than the paper on which

they are written if the host country’s legal and political system cannot guarantee that they will be

consistently enforced” (2003: 41 ). As a result, “banks endogenously structure syndicates to

minimize the costs associated with political risk” (Esty & Megginson, 2000: 8).

We argue that banks strategically adjust the diversity of a syndicate according to the

political risk of a given project. First, involving partners from a multitude of countries allows the

syndicate to secure the support of more governments having different sources of political

leverage (Lipson, 1981; Rangan & Sengul, 2009). A PF practitioner highlighted the benefits of

geographic diversity in high-risk countries when explaining that

16

[s]ometimes the lenders are strategically selected from a range of countries. The purpose of

this syndicate diversity is to discourage the host-government from expropriatory acts or

other discriminatory action. If the host-government elected to do so, it could thereby

endanger economic relations with the home country of each lender. (Hoffman, 2007: 72).

Second, in high-risk environments, it is unclear ex-ante what problems will occur during the

completion of the project and therefore which resources will be most valuable. As a result, lead

arrangers have an incentive to invite different types of banks to the syndicate. Large banks are

more likely to extend loans in the event of default because they can more easily diversify credit

risk (Esty & Megginson, 2003). State-owned banks have direct contacts to home-country

decision makers (Musacchio, Lazzarini, & Aguilera, 2015). Banks specializing in PF may be

able to provide complementary expertise developed through prior experience (Cai, Saunders, &

Steffen, 2010; François & Missonier-Piera, 2007). Export credit agencies can provide political-

risk guarantees which are inaccessible to commercial banks (Esty & Megginson, 2003). Finally,

development banks are critical in guaranteeing a host country’s access to concessionary loans

and therefore wield considerable political and economic clout (Hainz & Kleimeier, 2012). Lead

arrangers maximize such complementary resources by creating a more diverse syndicate. We

argue that, in balancing the costs and benefits associated with diverse syndicates, lead arranges

are more likely to form diverse syndicates when investing in a country with high political risk.

For example, banks anticipated that political risk would affect the development of the $4

billion Chad–Cameroon oil pipeline project, and formed a syndicate with 17 banks from 10

countries (Esty, 2004a). Similarly, when financing the $3.5 billion development of the Hamaca

oil field in Venezuela, ConocoPhillips did not rely on U.S. funding but instead assembled a

global syndicate of nine banks from six countries, two of which were state-owned export credit

agencies. The syndicate covered a vast economic area between Japan (Bank of Tokyo-

Mitsubishi) and Canada (Export Development Canada), five languages, and 17 time zones and

17

involved considerable setup costs. The syndicate, however, also amassed substantial economic

leverage over the host country, as banks originated in countries accounting for over 70 percent of

foreign aid going to Venezuela every year.

Building upon these arguments, we hypothesize:

Hypothesis 2: In a more uncertain institutional environment, the negative effect of

interorganizational diversity on the probability of forming a syndicate is reduced.

EMPIRICAL SETTING AND DATA

The empirical context of this paper is the PF industry. In 2014, the total volume of projects

amounted to $257.5 billion (Reuters, 2015). This is equivalent to the total IPO volume (Pinelli,

Kelley, Steinbach, & Suzuki, 2014) or to almost 60 percent of total annual R&D spending in the

United States, a context often used for the study of interorganizational partnerships and alliances.

The study of PF offers a number of empirical benefits. First, PF projects are financially and

organizationally separate from the companies that propose to build them and the banks that

finance them. Thus, unlike traditional corporate investments, which are part of a portfolio, PF

investments are independent units that can be observed as “strategic research sites” (Esty, 2004b:

214). Second, because PF projects are separate organizations, their risks are assessed ex-ante,

and partnerships form as a function of these risks for a given project, clearly separating project

creation from syndicate formation. Third, PF banking syndicates involve large numbers of banks,

allowing for better measurement and conceptualization of interorganizational diversity. Finally,

PF syndicates often form between banks that are foreign to the market in which the project is

realized. This allows researchers to distinguish the effects of the local institutional context from

the effects of interorganizational diversity.

18

We obtained data on syndicated PF loans from the Dealogic Projectware and the Thomson

SDC Platinum databases. The two databases are widely used and are considered the most

complete sources of information on syndicated lending in PF (Buscaino, Caselli, Corielli, &

Gatti, 2012; Byoun et al., 2013; Esty & Sesia, 2011). We exclude loans provided by only one or

two banks because of our focus on multipartner syndicates. The loans in our sample funded

1,110 project investments in 72 countries between 2000 and 2012. The average project value is

$267 million; the total value of all projects in our sample is $519 billion.

Table 1 gives an overview of the frequency of different types of infrastructure projects in

our sample. The largest project is the Pluto LNG oilfield project in Western Australia ($11.5

billion), followed by the Jirau Hydroelectric Plant in Brazil ($5.4 billion) and the Qingdao Metro

System in China ($5.1 billion). Geographically, projects are spread around the world, with the

largest number of projects in India, the US, and Australia.

——————————-

Insert Table 1 about here

——————————-

More than 1,500 individual banks from 100 countries participate in the syndicates in our

sample. We obtain bank-level data from the Bankscope database maintained by Bureau Van Dijk.

Our sample includes 222 banks that are majority-owned by a state, 735 owned by other banking

institutions, and the remainder distributed across public, family, and other types of ownership. The

largest financial markets in our sample are the US (141 banks), Japan (119), Spain (88), Italy (77),

Germany (68), India (62), and the UK (55). The highest market share in terms of syndicated

volume over the period of observation was held by Deutsche Bank, with 18.7 percent in 2000

(Bloomberg, 2016). Other very active banks in the PF market were HSBC (with an annual average

of $24 billion), Royal Bank of Scotland ($23.5 billion), Citi Group ($22.6 billion), Société

Générale ($19.1 billion), and BNP Paribas ($18.2 billion) (Bloomberg, 2016).

nasseralajmi
Highlight

19

METHODS

Comparing Realized and Unrealized (Counterfactual) Syndicates

To analyze the likelihood that a certain syndicate forms, we compare the syndicates that

formed—that is, the realized syndicates—with hypothetical (unrealized) syndicates that could

have formed to finance the same project. Naturally, unrealized syndicates cannot be observed in

the data. Alliance research has addressed this issue in different ways (for an overview of

approaches and empirical studies see Mindruta et al., 2016). Studies analyzing focal companies

commonly assign a dummy to the formation of an alliance with any other company. Research on

the dyadic level generates unrealized dyads (Ahuja et al., 2009; Chung et al., 2000; Garcia-Pont

& Nohria, 2002; Gimeno, 2004; Gulati, 1995). Including a full list of possible alliances may,

however, introduce some bias of its own because of very different propensities to form alliances

between firms—especially in multipartner settings with a very large number of possible partners

(Gulati & Gargiulo, 1999; Mindruta et al., 2016). Thus, when extending the analysis beyond

dyads, a simulation-based approach is necessary because the number of possible unrealized

groups grows exponentially (Sorenson & Stuart, 2008).

Following Heidl et al. (2014), we simulate hypothetical or unrealized syndicates of similar

sets of banks to compare them with the realized ones. For every project that received financing,

we randomly select a set of unrealized syndicates from the population of syndicates that could

have formed. We simulate unrealized syndicates in six different ways: by (1) replacing, (2)

adding, or (3) deleting one randomly chosen bank from the realized syndicate, and by creating

syndicates of randomly selected banks that are (4) of the same size, (5) larger by one bank, and

(6) smaller by one bank (see Table 2 for the distribution and description of each category). Our

method adds 26,552 unrealized syndicates to the 1,110 syndicates financing the observed

20

projects, for a total sample of 27,662 syndicates.6 For a given project, we compare the realized

syndicate (for which the dependent variable is equal to 1) with the unrealized syndicates (for

which the dependent variable is equal to 0). Figure 1 shows the distribution of the

interorganizational diversity across the realized and unrealized syndicates.

———————————————–

Insert Table 2 and Figure 1 about here

———————————————–

Only a few studies look at multipartner cooperation in general (Heidl et al., 2014; Li, 2013;

Sorenson & Stuart, 2008), and we are not aware of any study that investigates the formation of

PF syndicates. One reason for this may be difficulties with defining counterfactuals for the

syndicates that actually formed. Our method of comparing realized and unrealized syndicates

that are similar in composition (i.e., interorganizational diversity) follows Heidl et al.’s (2014)

work, but we take their approach further by creating different sets of alternatives rather than

finding a matched alternative. Our method requires fewer assumptions about potential partners

and considers different kinds of alternatives to the realized syndicate. To our knowledge, our

study is one of the first to also consider syndicates of different numbers of partners as

counterfactuals to the realized syndicates.

Measurement of Interorganizational Diversity in Syndicates

We use several demographic characteristics for our computation. First, and most importantly,

banks originate from different home-country contexts. As such, they differ in their understanding

of the institutional environment and in their ability to operate in environments with weak

institutions. Banks from countries with weak institutional environments have a better

understanding of the challenges that arise when investing in other countries with weak

6 Due to data availability, the number of randomly generated consortia varies across types. The larger the number of

randomly chosen banks, the lower the probability that data are available for all the randomly chosen banks.

21

institutional environments (Mian, 2006). This knowledge might be very valuable when the

project is located in a high-risk host country, creating incentives for the syndicate to include

banks from countries with different institutional environments despite the costs of managing a

diverse syndicate. To capture the diversity of institutional environments familiar to banks in the

syndicate, we represent each bank by the degree of political competition in its home country,

measured using POLITY IV data (Marshall & Gurr, 2012). The measure ranges from −10, for

countries with very weak institutions, to +10, for countries with very strong institutions. This

measure allows us to differentiate between banks from different institutional backgrounds,

capturing more information than we would if we simply counted the countries of origin.

Second, banks differ in size and, as a result, in ability to diversify the risk of a certain project. To

create a strong syndicate, one that is capable of refinancing in the event of project failure, smaller

lead arrangers may invite larger commercial banks, despite their lower expertise in PF. Bank size is

measured as total bank assets as reported in the Bankscope database. Third, banks vary in PF

experience. Such experience, especially within the host-country context, can be an additional

reason to include a bank in the syndicate, despite possible costs from interorganizational diversity.

We compute a PF experience variable as the cumulative number of projects a bank has participated

in during the years before the focal syndication was agreed upon. Fourth, we further distinguish

between banks by ownership type. The Bankscope database differentiates bank-owned, state-

owned, or “other” ownership types, where “other” includes insurance companies, financial

investors, trusts, and industrial companies. State-owned banks, for example, have both superior

capacity to assume risks (García-Canal & Guillén, 2008), due to their implicit or explicit

government protection, and superior access to political decision makers in their home countries,

22

which they can use to protect the project. Hence, commercial banks may decide to syndicate with

state-owned banks, despite the costs of cooperating with such banks.

Measuring diversity is a challenging proposition (Li & Hambrick, 2005). We argue that

syndicates benefit from a broad distribution of characteristics, especially when the challenges

that might arise are unclear ex ante. Consequently, we base our measure of interorganizational

diversity on what Harrison and Klein (2007) call “disparity,” and apply this concept at the

organizational level. Just as people are characterized by a number of attributes, firm resources

are better represented as bundles that result from a variety of factors (Parkhe, 1991) rather than

as individual metrics (i.e., by separate variables). Rather than trying to unpack the individual

effects of various characteristics on which banks vary, we measure interorganizational diversity

at the partnership level across several dimensions (home-country institutional environment, size,

experience, and ownership type). To aggregate these dimensions into one variable of

interorganizational diversity, we calculate the sum of each syndicate member i’s deviation from

the syndicate mean for each dimension j. Our interorganizational diversity score (div) is therefore

a partnership-level construct representing a sum across firms and across dimensions. We

normalize the underlying metric variables (size, experience, and home-country institutions) to

range from 0 to 1 and weight types of ownership with 1 √2⁄ , following Thatcher, Jehn, and

Zanutto (2003). The diversity score is calculated as

= ∑∑( − �̅�)²

=1

=1

.

where represents the value for bank i on dimension j, and �̅� is the mean of dimension j.

23

Contextual Variables

Our hypothesis 2, presented above, relates to the context in which an investment project takes

place. In particular, we hypothesize that partnership formation differs as a function of the

political risk in the project country. Our assumption is that local risk in the project location

affects task complexity and hence both the necessity to pool complementary resources and the

cooperation costs associated with managing the project. Hence, the degree to which banking

syndicates maximize or minimize interorganizational diversity varies according to the local risk

of the project.

We include two measures of local risk. First, we use the POLCON measure (Henisz, 2000)

constructed to capture the degree of constraints on policy change using data on

the number of government branches with veto power over policy change (executive, lower

and upper legislative chambers, judicial and subfederal institutions) and the distribution of

party preferences across and within these branches. (Henisz & Zelner, 2001: 133)

The POLCON measure, however, was intended to capture political constraints or political

risk rather than a broader set of political and social risks in a country. POLITY, by contrast, is a

combined measure of democracy and autocracy (Marshall & Gurr, 2012) which captures a wider

range of political and social risks, including the risk of social unrest, that might affect an

infrastructure project. We use both measures in our analysis and we invert the POLCON measure

to capture more directly local political risk. We run additional robustness checks using the ICRG

political-risk measure published by the PRS Group (2015).

Control Variables

We include in the analysis project-level fixed effects and several control variables obtained from

the Dealogic database. Foremost, we control for previous ties between potential syndicate

members as proposed by Heidl et al. (2014). On the project level, we control for the number of

project sponsors, for the size of the project, and the number of tranches in the loan. We also

24

indicate whether the loans were given to refinance an existing project, and we include the debt-

to-equity ratio to account for the relative importance of banks and sponsors. We also control for

financial market volatility and income in the target country. We introduce fixed effects on the

project level in our analyses below to account for project-level variance that remains unexplained

by the control variables outlined here. Descriptive statistics on the variables in our model and

pairwise correlations are presented in Table 3 and Table 4.

———————————————

Insert Table 3 and Table 4 about here

———————————————

Potential Endogeneity

Drawing a sample of unobserved syndicates from the universe of possible syndicates may lead to

selection of a certain kind of syndicate and hence bias the regression results toward this type of

selection. We pursue three strategies to avoid this problem. First, we generate six different types

of counterfactual syndicates, by randomly replacing, adding, or deleting one bank from realized

syndicates, and by generating random unrealized syndicates of the same size, larger by one bank,

and smaller by one bank. Second, we draw two completely independent samples of unobserved

syndicates from the population of possible syndicates. Doing so, we not only rely on the law of

large numbers, but actually resample to see whether our original draw from the population was

biased. Third, we generate counterfactuals using both uniform distributions and frequency-

weighted distributions,7 and compare whether the sampling strategies lead to different results.

7 For the frequency-weighted sampling we approximate the probability of a bank being part of a specific syndicate

with the relative frequency it shows up in our data until the day of signing the syndication agreement of the project

in question. We then use these probabilities to draw banks (without replacement) for counterfactual syndicates.

25

Estimation Strategy

The dependent variable in our analysis is a dummy indicating whether a syndicate has formed in

reality (= 1) or whether an observation has been randomly generated by the process outlined

above (= 0). We use hierarchical binomial regression to estimate our models.8 We conduct probit

estimation with a logistic link function. Our focus is on syndicate-level factors (rather than bank

characteristics or project characteristics) that might explain why a syndicate is formed, so we use

fixed effects on the project level (cj). Our baseline estimation takes the form


,

= + ′ + ∙ , , + .

In this specification, y* is the estimated latent utility of a syndicate i financing a project j (with

the observed binary outcome indicating whether y > y*), α is the grand-mean-intercept, β is the

vector of coefficients of our control variables, X is the matrix of observations of control

variables, and γ is the coefficient of diversity.

Later in our analysis, we introduce political risk as a factor conditioning the impact of

interorganizational diversity on the likelihood that a syndicate is formed. We argue that the

degree to which interorganizational diversity affects the likelihood that a syndicate is formed

differs across projects. Following Alcácer et al. (2013), we hence add a random element to the

diversity coefficient and move the analysis technique to a random coefficient model (RCM). As

we argue above, we believe it is project context that drives these differences between

coefficients. Therefore, we introduce local risk as a moderator on the coefficient γ into our RCM.

More formally, we decompose the coefficient γ as shown below. In the language of multilevel or

hierarchical modelling (Hox, Moerbeek, & van de Schoot, 2010; Steenbergen, 2012), this means

8 We use the hglm package (Alam, Ronnegard, & Shen, 2014) in an R 3.2.3 distribution “Wooden Christmas-Tree”

(R Core Team, 2015) for estimation.

26

that we introduce random slopes dj on top of the project fixed effects (or random intercepts) cj.

We therefore estimate different (random) coefficients for each project:

= 0 + 1 ∙ , + ,

where 0 and 1 are the coefficients for the mean direct effect of diversity and for the mean

interaction effect with political risk, respectively.

RESULTS

Table 5 shows our main regression results, displaying only variables that change across the

syndicates that did, or could have, come together to finance a project. To capture variation across

projects, we introduce random intercepts and random slopes of the diversity coefficient on the

project level. We begin with a baseline model that builds on the argumentation of Heidl et al.

(2014). In model 0 in Table 5, we confirm their results: previous ties increase the likelihood of

participants’ forming a syndicate. Building on this finding, we test our hypotheses against the

data explained above. In model 1 we add the interorganizational diversity variable introduced

above. Both without (model 1) and with (model 2) random slopes we find very strong negative

effects of interorganizational diversity on the likelihood of a syndicate to form. As argued above,

the more diverse potential partners are, the less likely they will come together to finance a

project. Figure 2 illustrates the dependency.

———————————————–

Insert Table 5 and Figure 2 about here

———————————————–

Before moving to the analysis of cross-level interactions, we look more closely for

evidence of variation in the random slopes of interorganizational diversity in models 1 and 2,

following Steenbergen (2012). The left panel of Figure 3 shows a histogram of random

coefficients of interorganizational diversity on the project level, and reveals that the distribution

is skewed on the positive side. The right panel of Figure 3 shows that random slopes of

27

interorganizational diversity are increasingly positive as local risk increases. The correlation

between the two, as shown as a dashed (blue) line, is 0.14.

——————————–

Insert Figure 3 about here

———————————

Models 3 and 4 show results of tests of local context influencing the willingness of

syndicate members to partner with more diverse banks to finance projects. Indeed, there is

economically and statistically significant support for our reasoning in the data. In model 3 we

add an interaction between interorganizational diversity and local risk, as captured by an inverse

POLCON measure, and we find that the negative effect of diversity on the likelihood to form a

syndicate is moderated by local risk. Figure 4 shows the difference in effects. If we replace the

POLCON measure with an inverted POLITY measure in model 4, we find the same effect to an

economically and statistically very similar extent (note that POLITY scales from −10 to 10,

while POLCON scales from 0 to 1). In the last two columns of Table 5, we show results for a

sample split between projects in OECD countries (model 6) and in non-OECD countries (model

7). As we expect, the effect of local risk on the connection between interorganizational diversity

and the likelihood to form a syndicate is substantially stronger in non-OECD countries.

——————————–

Insert Figure 4 about here

———————————

To assess model fit, we provide descriptive statistics on the fixed and random parts of our

models. The log-likelihood figures in Table 5 are computed for the fixed-effects part of the

model and hence do not vary across models with the same fixed-effects specification (such as

model 1 and model 2). Unexplained variance decreases significantly (in terms of deviance) from

model 0 to model 1 and from model 3 to model 5. The improvement in the model fit (change in

deviance) of the fixed-effects specification when adding local risk (in terms of POLCON risk) is

28

2.482, which is significant at the 0.1 level. The key change in model fit, however, occurs when

introducing random slopes. Benchmarked against a model with fixed effects on the project level

(model 1), the level 2 R-squared (Steenbergen, 2012) of model 2 is 61 percent, and it jumps to 66

percent in model 3 when adding first random slopes and then the cross-level interaction. The

values in the OECD and non-OECD sample split are benchmarked against fixed-effects models

with the corresponding sample split (not shown in Table 5). Comparing those two subsamples,

we see that the non-OECD observations have substantially more unobserved variation in the

random part, which is in line with our argumentation that the effect of interorganizational

diversity on the likelihood to choose a syndicate is less linear in this case.

Robustness Checks

We employ four strategies to ensure that our findings are robust. We evaluate robustness with

regard to estimation technique, measures, control variables, and sampling. First, we provide

results of estimations of generalized linear models (glm) and hierarchical generalized linear

models (hglm). Changing our specification to exclude random effects or to include only random

intercepts or only random slopes for the interorganizational diversity variable does not change

our results significantly. Second, replacing the POLCON measure we use for political risk with

the POLITY measure leads to similar results. Third, adding project-level control variables to our

models also does not change the results qualitatively.

Finally, we resample the unobserved syndicates. Doing so replaces more than 95 percent of

our data with another set of randomly generated syndicates.9 The results estimated with the

“new” data of unrealized syndicates are the same (see Table 6 in comparison to Table 5). In

9 This approach is unique to settings where the observations are drawn from a population of hypothetically possible

observations. It is somewhat similar to bootstrapping, but with the additional characteristic that we do not use our

original sample as the population to draw from, but rather the “true” population of unobserved consortia.

29

addition, we also replace our sampling strategy for counterfactual syndicates. Instead of

assuming that all banks have the same probability of being sampled into counterfactual

syndicates, we weigh that probability according to the frequency of their participation in the

realized syndicates. This approach yields a set of unrealized syndicates that differ both

qualitatively and quantitatively from our initial set. These counterfactuals are “closer” to the

realized syndicates because the banks that are more active in PF (i.e., more likely to be in the

realized syndicates) are also more likely to be part of the counterfactuals. In the models

estimated with the frequency-weighted counterfactuals (Table 7), the effects remain the same,

and their statistical significance increases compared with the results in Tables 5 and 6.

———————————————-

Insert Table 6 and Table 7 about here

———————————————-

As a final check, we follow Sorenson and Stuart (2008) and investigate whether the ratio of

realized to unrealized syndicates affects the statistical significance of our results. As we use

project-level fixed effects and random coefficients on the project level, we believe our approach

to be more robust against variance deflation because of increased counterfactuals. We investigate

ratios between 1:1 and 1:115 and find that our results are robust to the amount of counterfactuals.

The standard errors of the estimated coefficients increase with decreasing counterfactual

samples, but the estimates remain statistically equivalent and significant.

DISCUSSION AND CONCLUSION

Our study analyzes the probability of syndicate formation as a function of interorganizational

diversity in different institutional contexts. We use a novel empirical setting—PF loans—which

allows us to observe alliance formation between diverse partners and, more importantly, to

examine the effect of institutional risk in the project location on the likelihood of forming more or

less diverse syndicates. We find support for our baseline hypothesis that interorganizational

30

diversity decreases the likelihood of syndicate formation as banks face higher costs of cooperation

and potential opportunistic behavior by partners. On average, the effect of diversity on the

probability of forming a partnership is negative. However, when investing in projects located in

high-risk countries with weak institutional environments, banks seem more likely to accept the

higher costs of working in a diverse syndicate in order to pool the resources needed to address local

risk. Thus, while interorganizational diversity negatively affects the probability of forming a

syndicate, the effect is less pronounced when the investment is in a high-risk location.

Our study makes several theoretical and empirical contributions. At a theoretical level, we

contribute to research on partnership formation by highlighting that the local host-country affects

partnering decisions. We show that diverse syndicates are more likely to form when investing in

high-risk locations, as syndicates seek to mitigate institutional risk by bringing together diverse

banks with a wide range of political resources, experience, and risk-management capabilities.

Second, we build on a new set of studies examining the formation of multipartner alliances

(Heidl et al., 2014; Lavie et al., 2007), and we highlight that diversity on a multipartner level is

not additive (Liu & Maula, 2016). We show that interorganizational diversity hinders the

formation of multipartner alliances because the complexity of multipartner collaboration makes

it exponentially more difficult to devise contractual or relational structures that reduce

transaction costs (Poppo & Zenger, 2002). On the empirical side, we extend methodologies used

in prior research that compare realized partnerships with unrealized ones to account for selection.

By generating unrealized syndicates in six different ways, we make fewer assumptions about

partner selection and improve the generalizability of results. We also design and implement a

new approach to measuring interorganizational diversity at the partnership level that captures the

distribution of differences across multiple dimensions and across all the firms in the syndicate.

31

REFERENCES

Ahuja, G. 2000. The duality of collaboration: Inducements and opportunities in the formation of interfirm

linkages. Strategic Management Journal, 21(3): 317–343.

Ahuja, G., Polidoro, F., Jr., & Mitchell, W. 2009. Structural homophily or social asymmetry? The

formation of alliances by poorly embedded firms. Strategic Management Journal, 30(9): 941.

Alam, M., Ronnegard, L., & Shen, X. 2014. Fitting Conditional and Simultaneous Autoregressive

Spatial Models in hglm.

Alcácer, J., Chung, W., Hawk, A., & Pacheco-de-Almeida, G. 2013. Applying random coefficient models

to strategy research: testing for firm heterogeneity, predicting firm-specific coefficients, and

estimating strategy trade-offs. Working Paper. http://ssrn.com/abstract=2323578.

Beamish, P. W., & Kachra, A. 2004. Number of partners and JV performance. Journal of World

Business, 39(2): 107–120.

Beamish, P. W., & Lupton, N. C. 2016. Cooperative strategies in international business and management:

Reflections on the past 50 years and future directions. Journal of World Business, 51(1): 163–175.

Beckman, C. M., Haunschild, P. R., & Phillips, D. J. 2004. Friends or Strangers? Firm-Specific

Uncertainty, Market Uncertainty, and Network Partner Selection. Organization Science, 15(3):

259–275.

Bloomberg. 2016. Bloomberg Data Services.

Bolton, P., & Scharfstein, D. S. 1996. Optimal debt structure and the number of creditors. Journal of

Political Economy, 104(1): 1.

Brealey, R. A., Cooper, I. A., & Habib, M. A. 1996. Using project finance to fund infrastructure

investments. Journal of Applied Corporate Finance, 9(3): 25–39.

Brouthers, K. D., Brouthers, L. E., & Werner, S. 2003. Transaction cost-enhanced entry mode choices and

firm performance. Strategic Management Journal, 24(12): 1239–1248.

Buscaino, V., Caselli, S., Corielli, F., & Gatti, S. 2012. Project Finance Collateralised Debt Obligations: an

Empirical Analysis of Spread Determinants. European Financial Management, 18(5): 950–969.

Byoun, S., Kim, J., & Yoo, S. S. 2013. Risk Management with Leverage: Evidence from Project Finance.

Journal of Financial and Quantitative Analysis, 48(2): 549–577.

Cai, J., Saunders, A., & Steffen, S. 2010. Diversification or specialization? An analysis of distance and

collaboration in loan syndication networks. Unpublished Working Paper, Fordham University.

Chung, S., Singh, H., & Lee, K. 2000. Complementarity, status similarity and social capital as drivers of

alliance formation. Strategic Management Journal, 21(1): 1–1.

Colombo, M. G., Grilli, L., & Piva, E. 2006. In search of complementary assets: The determinants of

alliance formation of high-tech start-ups. Research Policy, 35(8): 1166–1199.

Delios, A., & Henisz, W. 2000. Japanese firms’ investment strategies in emerging economies. Academy

of Management Journal, 43(3): 305–323.

Delios, A., & Henisz, W. J. 2003. Policy uncertainty and the sequence of entry by Japanese firms, 1980-

1998. Journal of International Business Studies, 34(3): 227–241.

Dennis, S. A., & Mullineaux, D. J. 2000. Syndicated loans. Journal of Financial Intermediation, 9(4):

404–426.

Dussauge, P., Garrette, B., & Mitchell, W. 2000. Learning from competing partners: outcomes and

durations of scale and link alliances in Europe, North America and Asia. Strategic Management

Journal, 21(2): 99–126.

Eisenhardt, K. M., & Schoonhoven, C. B. 1996. Resource-based view of strategic alliance formation:

Strategic and social effects in entrepreneurial firms. Organization Science, 7(2): 136–150.

Esty, B. C. 1999. Petrozuata: a case study of effective use of project finance. Journal of Applied

Corporate Finance, 12(3): 26–42.

Esty, B. C. 2000. The equate project: An introduction to Islamic project finance. The Journal of

Structured Finance, 5(4): 7–20.

32

Esty, B. C. 2001. Structuring loan syndicates: a case study of the Hong Kong Disneyland project loan.

Journal of Applied Corporate Finance, 14(3): 80–95.

Esty, B. C. 2004a. Modern project finance: a casebook (Wiley internat. ed.). New York, NY: Wiley.

Esty, B. C. 2004b. Why study large projects? An introduction to research on project finance. European

Financial Management, 10(2): 213–224.

Esty, B. C., Chavich, C., & Sesia, A. 2014. An Overview of Project Finance and Infrastructure Finance—

2014 Update. Working Paper, (214-083).

Esty, B. C., & Megginson, W. L. 2000. Syndicate structure as a response to political risk in the project

finance loan market. Working Paper Harvard Business School.

Esty, B. C., & Megginson, W. L. 2001. Legal risk as a determinant of syndicate structure in the project

finance loan market.

Esty, B. C., & Megginson, W. L. . 2003. Creditor rights, enforcement, and debt ownership structure:

Evidence from the global syndicated loan market. Journal of Financial and Quantitative

Analysis, 38(1): 37–59.

Esty, B. C., & Sesia, A. 2011. An overview of project finance and infrastructure finance – 2009 update.

Working Paper.

Fonti, F., Maoret, M., & Whitbred, R. 2016. Free‐riding in multi‐party alliances: The role of perceived

alliance effectiveness and peers’ collaboration in a research consortium. Strategic Management

Journal.

François, P., & Missonier-Piera, F. 2007. The Agency Structure of Loan Syndicates. Financial Review,

42(2): 227–245.

García-Canal, E., & Guillén, M. F. 2008. Risk and the strategy of foreign location choice in regulated

industries. Strategic Management Journal, 29(10): 1097–1115.

García-Canal, E., Valdés-Llaneza, A., & Ariño, A. 2003. Effectiveness of Dyadic and Multi-Party Joint

Ventures. Organization Studies, 24(5): 743–770.

Garcia-Pont, C., & Nohria, N. 2002. Local versus global mimetism: The dynamics of alliance formation

in the automobile industry. Strategic Management Journal, 23(4): 307–321.

Gatti, S., Kleimeier, S., Megginson, W. L., & Steffanoni, A. 2013. Arranger certification in project

finance. Financial Management, (Spring): 1–40.

Giannetti, M., & Yafeh, Y. 2012. Do Cultural Differences Between Contracting Parties Matter? Evidence

from Syndicated Bank Loans. Management Science, 58(2): 365–383.

Gimeno, J. 2004. Competition within and between networks: The contingent effect of competitive

embeddedness on alliance formation. Academy of Management Journal, 47(6): 820–842.

Goerzen, A., & Beamish, P. W. 2005. The effect of alliance network diversity on multinational enterprise

performance. Strategic Management Journal, 26(4): 333–354.

Grant, M. 1997. Financing Eurotunnel. Japan Railway & Transport Review, April: 46–53.

Gulati, R. 1995. Social Structure and Alliance Formation Patterns: A Longitudinal Analysis.

Administrative Science Quarterly, 40(4): 619–652.

Gulati, R., & Gargiulo, M. 1999. Where Do Interorganizational Networks Come From? American

Journal of Sociology, 104(5): 177–231.

Gulati, R., & Singh, H. 1998. The architecture of cooperation: Managing coordination costs and

appropriation concerns in strategic alliances. Administrative Science Quarterly, 43(4): 781–814.

Guler, I., & Guillén, M. 2010. Home Country Networks And Foreign Expansion: Evidence From The

Venture Capital Industry. Academy of Management Journal, 53(2): 390.

Hainz, C., & Kleimeier, S. 2012. Political risk, project finance, and the participation of development

banks in syndicated lending. Journal of Financial Intermediation, 21(2): 287–314.

Hallen, B. L. 2008. The Causes and Consequences of the Initial Network Positions of New Organizations:

From Whom Do Entrepreneurs Receive Investments? Administrative Science Quarterly, 53(4):

685–718.

Harrison, D. A., & Klein, K. J. 2007. What’s the Difference? Diversity Constructs as Separation, Variety,

or Disparity in Organizations. The Academy of Management Review, 32(4): 1199–1228.

33

Harrison, J. S., Hitt, M. A., Hoskisson, R. E., & Ireland, R. D. 2001. Resource complementarity in

business combinations: Extending the logic to organizational alliances. Journal of Management,

27(6): 679–690.

Heidl, R. A., Steensma, H. K., & Phelps, C. 2014. Divisive Faultlines and the Unplanned Dissolutions of

Multipartner Alliances. Organization Science, 25(5): 1351–1371.

Henisz, W. J. 2000. The institutional environment for multinational investment. Journal of Law

Economics & Organization, 16(2): 334–364.

Henisz, W. J., & Williamson, O. E. 1999. Comparative economic organization—within and between

countries. Business and Politics, 1(3): 261–278.

Henisz, W. J., & Zelner, B. A. 2001. The institutional environment for telecommunications investment.

Journal of Economics & Management Strategy, 10(1): 123–147.

Hitt, M. A., Ahlstrom, D., Dacin, M. T., Levitas, E., & Svobodina, L. 2004. The Institutional Effects on

Strategic Alliance Partner Selection in Transition Economies: China vs. Russia. Organization

Science, 15(2): 173–185.

Hoffman, S. L. 2007. The Law and Business of International Project Finance: A Resource for

Governments, Sponsors, Lawyers, and Project Participants. Cambridge University Press.

Hox, J. J., Moerbeek, M., & van de Schoot, R. 2010. Multilevel analysis: Techniques and applications.

Routledge.

Inoue, C. F. K. V., Lazzarini, S. G., & Musacchio, A. 2013. Leviathan as a Minority Shareholder: Firm-

Level Implications of State Equity Purchases. Academy of Management Journal, 56(6): 1775–

1801.

James, C. 1990. Heterogeneous creditors and the market value of bank LDC loan portfolios. Journal of

Monetary Economics, 25(3): 325–346.

Jiang, R. J., Tao, Q. T., & Santoro, M. D. 2010. Alliance portfolio diversity and firm performance.

Strategic Management Journal, 31(10): 1136.

Koka, B. R., & Prescott, J. E. 2002. Strategic alliances as social capital: A multidimensional view.

Strategic Management Journal, 23(9): 795–816.

Krugman, P. 1985. International debt strategies in an uncertain world. In G. W. Smith & J. D. Cuddington

(Eds.), International debt and the developing countries – A World Bank Symposium: 79–100.

Washington.

Lavie, D., Lechner, C., & Singh, H. 2007. The Performance Implications of Timing of Entry and

Involvement in Multipartner Alliances. Academy of Management Journal, 50(3): 578–604.

Lavie, D., & Miller, S. R. 2008. Alliance Portfolio Internationalization and Firm Performance.

Organization Science, 19(4): 623–646.

Li, D. 2013. Multilateral R&D alliances by new ventures. Journal of Business Venturing, 28(2): 241–260.

Li, D., Eden, L., Hitt, M. A., Ireland, R. D., & Garrett, R. P. 2012. Governance in Multilateral R&D

Alliances. Organization Science, 23(4): 1191–1210.

Li, J. T., & Hambrick, D. C. 2005. Factional groups: A new vantage on demographic faultlines, conflict,

and disintegration in work teams. Academy of Management Journal, 48(5): 794–813.

Lipson, C. 1981. The international organization of Third World debt. International Organization, 35(4):

603–631.

Liu, Y., & Maula, M. 2016. Local Partnering in Foreign Ventures: Uncertainty, Experiential Learning,

and Syndication in Cross-Border Venture Capital Investments. Academy of Management

Journal, 59(4): 1407–1429.

Marshall, M. G., & Gurr, T. R. 2012. Polity IV Project: Political regime characteristics and transitions,

1800-2013. Maryland: Political Instability Task Force.

McMillen, M. J. 2001. Islamic Shari’ah-Compliant Project Finance: Collateral Security and Financing

Structure Case Studies.

Meyer, K. F., Estrin, S., Bhaumik, S. K., & Peng, M. W. 2009. Institutions, resources, and entry strategies

in emerging economies. Strategic Management Journal, 30(1): 61.

34

Mian, A. 2006. Distance Constraints: The Limits of Foreign Lending in Poor Economies. The Journal of

Finance, 61(3): 1465–1505.

Mindruta, D., Moeen, M., & Agarwal, R. 2016. A two-sided matching approach for partner selection and

assessing complementarities in partners’ attributes in inter-firm alliances. Strategic Management

Journal, 37(1): 206–231.

Mitsuhashi, H., & Greve, H. R. 2009. A matching theory of alliance formation and organizational success:

Complementarity and compatibility. Academy of Management Journal, 52(5): 975–995.

Musacchio, A., Lazzarini, S. G., & Aguilera, R. V. 2015. New Varieties of State Capitalism: Strategic and

Governance Implications. The Academy of Management Perspectives, 29(1): 115–131.

Nevitt, P. K., & Fabozzi, F. J. 2000. Project financing (Seventh Edition). London: Euromoney Books.

Paligorova, T., & Santos, J. A. C. 2015. Non-bank Loan Investors and Borrowers’ Renegotiation

Prospects. http://cff.handels.gu.se/digitalAssets/1561/1561459_paligorova.pdf.

Parkhe, A. 1991. Interfirm diversity, organizational learnings, and longevity in global strategic alliances.

Journal of International Business Studies, 22(4): 579–601.

Parkhe, A. 1993. Partner nationality and the structure-performance relationship in strategic alliances.

Organization Science, 4(2): 301–324.

Pfeffer, J., & Salancik, G. R. 1978. The external control of organizations: a resource dependence

perspective. New York, NY: Harper & Row.

Pichler, P., & Wilhelm, W. 2001. A Theory of the Syndicate: Form Follows Function. Journal of

Finance, 56(6): 2237–2264.

Pinelli, M., Kelley, J., Steinbach, M., & Suzuki, S. 2014. EY Global IPO Trends (2015 4Q). London:

Ernst & Young.

Podolny, J. M. 1994. Market uncertainty and the social character of economic exchange. Administrative

Science Quarterly, 39(3): 458–484.

Poppo, L., & Zenger, T. 2002. Do formal contracts and relational governance function as substitutes or

complements? Strategic Management Journal, 23(8): 707–725.

R Core Team. 2015. R: A Language and Environment for Statistical Computing. Vienna, Austria: R

Foundation for Statistical Computing.

Rangan, S., & Sengul, M. 2009. The Influence of Macro Structure on the Foreign Market Performance of

Transnational Firms: The Value of IGO Connections, Export Dependence, and Immigration

Links. Administrative Science Quarterly, 54(2): 229–267.

Reuters, T. 2015. Global Project Finance Review – Managing underwriters – Full Year 2014. London.

https://www.thomsonone.com.

Roberts, M. R., & Sufi, A. 2009. Control Rights and Capital Structure: An Empirical Investigation.

Journal of Finance, 64(4): 1657–1695.

Rothaermel, F. T., & Boeker, W. 2008. Old Technology Meets New Technology: Complementarities,

Similarities, and Alliance Formation. Strategic Management Journal, 29(1): 47–77.

Sahaym, A., Steensma, H. K., & Schilling, M. A. 2007. The influence of information technology on the

use of loosely coupled organizational forms: An industry-level analysis. Organization Science,

18(5): 865–880.

Sampson, R. C. 2007. R&D alliances and firm performance: The impact of technological diversity and

alliance organization on innovation. Academy of Management Journal, 50(2): 364–386.

Sang Whi, L., & Mullineaux, D. J. 2004. Monitoring, financial distress, and the structure of commercial

lending syndicates. Financial Management (Wiley-Blackwell), 33(3): 107–130.

Simons, K. V. 1993. Why Do Banks Syndicate Loans? New England Economic Review,

(January/February): 45–52.

Sorenson, O., & Stuart, T. E. 2008. Bringing the context back in: settings and the search for syndicate

partners in venture capital investment networks. Administrative Science Quarterly, 53(2): 266–294.

Steenbergen, M. R. 2012. Hierarchical Linear Models for Electoral Research: A Worked Example in Stat.

http://www.elecdem.eu/media/universityofexeter/elecdem/pdfs/istanbulwkspjan2012/Hierarchical

_Linear_Models_for_Electoral_Research_A_worked_example_in_Stata.pdf.

35

Stuart, T. E. 1998. Network positions and propensities to collaborate: An investigation of strategic alliance

formation in a high-technology industry. Administrative Science Quarterly, 43(3): 668–698.

Subramanian, K. V., & Tung, F. 2016. Law and Project Finance. Journal of Financial Intermediation,

25: 154–177.

Thatcher, S. M., Jehn, K. A., & Zanutto, E. 2003. Cracks in diversity research: The effects of diversity

faultlines on conflict and performance. Group Decision and Negotiation, 12(3): 217–241.

The PRS Group. 2015. International Country Risk Guide Methodology. Syracuse, NY.

http://www.prsgroup.com/wp-content/uploads/2012/11/icrgmethodology.pdf.

Thompson, J. D. 1967. Organizations in Action Social Science Bases of Administrative Theory (1st

edition assumed edition). New York: Mcgraw-Hill College.

Vaaler, P. M., & Schrage, B. N. 2009. Residual state ownership, policy stability and financial

performance following strategic decisions by privatizing telecoms. Journal of International

Business Studies, 40(4): 621–641.

Vasudeva, G., Spencer, J. W., & Teegen, H. J. 2013. Bringing the Institutional Context Back In: A Cross-

National Comparison of Alliance Partner Selection and Knowledge Acquisition. Organization

Science, 24(2): 319–338.

Vilanova, L. 2005. Financial distress, lender passivity and project finance : the case of Eurotunnel.

Working Paper.

White, S., & Siu-Yun Lui, S. 2005. Distinguishing costs of cooperation and control in alliances. Strategic

Management Journal, 26(10): 913–932.

Yu, T. Y., Subramaniam, M., & Cannella, A. A. 2013. Competing globally, allying locally: Alliances

between global rivals and host-country factors. Journal of International Business Studies, 44(2):

117–137.

FIGURE 1:
Diversity in Realized and Hypothetical Syndicatesb

bDistribution of diversity in realized syndicates (left) and hypothetical counterfactuals following the above logic of six types of
counterfactuals (right). All types of counterfactuals are included in the right panel.

FIGURE 2:
The Effect of Diversity on Syndicate Formatione

Baseline effect

Diversity

P
ro

ba
bi

lit
y

to
fo

rm
c

on
so

rt
iu

m

0.01

0.02

0.03

0.04

0.05

0 1 2 3 4 5 6

ePartial effect of diversity on the probability to form a syndicate for projects. The confidence interval is computed for a one-
level model and is more conservative than the more precise estimates shown in Table 5. Black ticks in the plot area in the bottom
indicate the incidence of observations.

FIGURE 3:
Random Slope Coefficients of Diversityf

Distribution of diversity random coefficients

Random coefficients

F
re

qu
en

cy

−0.004 −0.002 0.000 0.002 0.004 0.006

0
20

40
60

80
10

0

0.2 0.4 0.6 0.8 1.0


0.

00
4


0.

00
2

0.
00

0
0.

00
2

0.
00

4
0.

00
6

Variation in diversity slopes

Local risk

D
iv

er
si

ty
r

an
do

m
s

lo
pe

fHistogram (left) and scatterplot (right) of diversity slopes relative to the coefficient shown in model 1 in Table 5 in dependence
of local risk (POLCON risk). The dashed line indicates the correlation between the random slope of diversity and local risk.

FIGURE 4:
The Effect of Diversity on Syndicate Formation (continued)g

Low Polcon risk (0.2)

Diversity

P
ro

ba
bi

lit
y

to
fo

rm
c

on
so

rt
iu

m

0.01

0.02

0.03

0.04

0.05
0.06
0.07

0 1 2 3 4 5 6

High Polcon risk (0.9)

Diversity

P
ro

ba
bi

lit
y

to
fo

rm
c

on
so

rt
iu

m

0.01

0.02

0.03

0.04

0.05
0.06
0.07

0 1 2 3 4 5 6

gPartial effect of diversity on the probability to form a syndicate for projects in countries (non-OECD) with low (left) and high
(right) POLCON risk. The confidence interval is computed for a one-level model and is more conservative than the more precise
estimates shown in Table 5.

TABLE 1:
Projects by Industry Sector

Industry Frequency Share
Power 194 17.5 %
Wind farm 144 13.0 %
Renewable fuel 101 9.0 %
Road 100 9.0 %
Oilfield exploration 58 5.2 %
Mining 57 5.1 %
Petrochemical plant 46 4.1 %
Oil fefinery / LNG 33 3.0 %
Telecom 32 2.9 %
Ports 32 2.9 %

TABLE 2:
Diversity in Realized and Counterfactual Syndicatesa

Realized Replacement (1) Addition (2) Deletion (3) Same size (4) Larger (5) Smaller (6)
Frequency 1, 110 9, 155 9, 057 5, 631 715 300 1, 694
Mean diversity 1.03 1.03 1.50 0.81 1.08 1.22 0.47

aFrequencies and mean diversity scores for realized and the six types of counterfactual syndicates. The counterfactual syn-
dicates are created by (1) replacing a randomly chosen bank from the realized syndicate with a random bank in the sample, (2)
adding a random bank from the sample to the realized syndicate, (3) deleting a randomly chosen bank from the realized syndicate,
(4) creating a syndicate of the same size as the realized one from randomly chosen banks in the sample, (5) creating a syndicate of
one more bank than in the realized one from randomly chosen banks in the sample, and (6) creating a syndicate of one bank fewer
than in the realized one from randomly chosen banks in the sample. All randomly chosen banks are chosen without replacement
from the sample without the banks in the respective realized syndicate.

TABLE 3:
Descriptive Statisticsc

Statistic N Mean St. Dev. Min Max
Realized syndicate 27,662 0.04 0.20 0.00 1.00
Average previous tie strength 27,662 4.46 6.31 0.00 89.00
Consortium diversity 27,662 1.21 0.93 0.00 6.94
POLCON risk in project country 27,662 0.33 0.25 0.11 1.00
POLITY in project country 27,662 7.81 4.62 −10.00 10.00
Number of tranches 27,662 9.44 5.15 0.00 33.00
Number of sponsors 27,662 2.08 2.33 0.00 29.00
Refinancing 27,662 0.18 0.38 0.00 1.00
Project debt-to-equity ratio 27,662 0.83 0.21 0.00 1.30
Project size (million USD) 27,662 473.09 701.39 2.54 11,500.00
Financial market volatility 27,662 0.18 0.07 0.08 0.43
GDP per capita in project country 27,662 25,669.00 22,031.00 432.07 101,563.70

cData for all (realized and hypothetical) syndicates. Project-, bank-, and syndicate-level data from Dealogic, SDC Platinum,
and Bankscope.

TABLE 4:
Pairwise Correlations

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)

Realized syndicate (1) 1 0.09 -0.04 -0.00 0.00 -0.00 -0.00 0.00 -0.00 -0.00 0.00 0.00
Average previous tie strength (2) 0.09 1 -0.05 -0.09 0.05 0.06 0.00 0.08 0.06 0.08 0.04 0.18
Consortium diversity (3) -0.04 -0.05 1 0.19 -0.20 0.52 0.10 0.03 0.01 0.24 -0.06 0.04
POLCON risk in project country (4) -0.00 -0.09 0.19 1 -0.81 0.01 0.14 -0.05 -0.00 0.11 -0.04 -0.45
POLITY in project country (5) 0.00 0.05 -0.20 -0.81 1 -0.03 -0.17 0.03 -0.02 -0.15 0.04 0.26
Number of tranches (6) -0.00 0.06 0.52 0.01 -0.03 1 0.03 0.07 0.10 0.17 -0.00 0.18
Number of sponsors (7) -0.00 0.00 0.10 0.14 -0.17 0.03 1 -0.06 -0.04 0.15 -0.00 0.04
Refinancing (8) 0.00 0.08 0.03 -0.05 0.03 0.07 -0.06 1 0.12 0.02 -0.02 0.13
Project debt-to-equity ratio (9) -0.00 0.06 0.01 -0.00 -0.02 0.10 -0.04 0.12 1 -0.21 -0.04 0.08
Project size (10) -0.00 0.08 0.24 0.11 -0.15 0.17 0.15 0.02 -0.21 1 0.03 0.01
Financial market volatility (11) 0.00 0.04 -0.06 -0.04 0.04 -0.00 -0.00 -0.02 -0.04 0.03 1 -0.06
GDP per capita in project country (12) 0.00 0.18 0.04 -0.45 0.26 0.18 0.04 0.13 0.08 0.01 -0.06 1

TABLE 5:
Diversity and Syndicate Formationd

Full sample Sub samples
0 1 2 3 4 5 6 7

Intercept −3.42∗∗∗ −3.17∗∗∗ −3.17∗∗∗ −3.16∗∗∗ −3.16∗∗∗ −2.94∗∗∗ −3.23∗∗∗ −3.09∗∗∗

(0.02) (0.03) (0.03) (0.03) (0.03) (0.11) (0.04) (0.05)
Tie strength 0.04∗∗∗ 0.04∗∗∗ 0.04∗∗∗ 0.04∗∗∗ 0.04∗∗∗ 0.04∗∗∗ 0.04∗∗∗ 0.05∗∗∗

(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Diversity −0.20∗∗∗ −0.20∗∗∗ −0.26∗∗∗ −0.17∗∗∗ −0.30∗∗∗ −0.20∗∗∗ −0.45∗∗∗

(0.02) (0.02) (0.03) (0.02) (0.03) (0.05) (0.07)
(1-POLCON)*Diversity 0.12∗∗ 0.10∗ 0.24 0.28∗∗∗

(0.04) (0.05) (0.18) (0.06)
-Polity*Diversity 0.01∗∗

(0.00)
Log Likelihood -4,579.88 -4,564.34 -4,564.34 -4,563.10 -4,563.19 -4,555.49 -2,451.52 -2,105.18
Num. obs. 27,662 27,662 27,662 27,662 27,662 27,662 14,777 12,89
Random variance 0.05 0.087 0.03 0.03 0.03 0.03 0.01 0.10
Level 2 R2 0.00 0.000 0.61 0.66 0.64 0.62 0.86 0.24
Fixed effects TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE
Random slope FALSE FALSE TRUE TRUE TRUE TRUE TRUE TRUE
Project controls FALSE FALSE FALSE FALSE FALSE TRUE FALSE FALSE

dRegression results for main model. DV is a dummy indicating realized or hypothetical syndicate. Estimation method is
hierarchical glm with project-level fixed effects. Project-level control variables are employed in model 5, but output not shown.
The subsamples in model 6 (7) are low risk (high risk) project countries. POLCON and Polity are inverted such that higher numbers
indicate higher risk.
∗p < 0.05
∗∗p < 0.01
∗∗∗p < 0.001

TABLE 6:
Diversity and Syndicate Formation (resampled)h

Full sample Sub samples
0 1 2 3 4 5 6 7

Intercept −3.42∗∗∗ −3.16∗∗∗ −3.16∗∗∗ −3.14∗∗∗ −3.14∗∗∗ −2.91∗∗∗ −3.22∗∗∗ −3.06∗∗∗

(0.02) (0.03) (0.03) (0.03) (0.03) (0.11) (0.04) (0.05)
Tie strength 0.04∗∗∗ 0.04∗∗∗ 0.04∗∗∗ 0.04∗∗∗ 0.04∗∗∗ 0.043∗∗∗ 0.04∗∗∗ 0.05∗∗∗

(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Diversity −0.22∗∗∗ −0.22∗∗∗ −0.278∗∗∗ −0.18∗∗∗ −0.32∗∗∗ −0.22∗∗∗ −0.50∗∗∗

(0.02) (0.02) (0.03) (0.02) (0.033) (0.05) (0.06)
(1-POLCON)*Diversity 0.14∗∗ 0.11∗ 0.28 0.32∗∗∗

(0.04) (0.045) (0.18) (0.06)
-Polity*Diversity 0.01∗∗

(0.00)
Log Likelihood -4,583.78 -4,566.39 -4,566.39 -4,564.80 -4,564.88 -4,557.15 -2,453.41 -2,104.01
Num. obs. 27,737 27,737 27,737 27,737 27,737 27,737 14,827 12,910
Random variance 0.064 0.13 0.03 0.05 0.05 0.03 0.01 0.12
Level 2 R2 0.000 0.000 0.78 0.63 0.62 0.77 0.90 -0.66
Fixed effects TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE
Random slope FALSE FALSE TRUE TRUE TRUE TRUE TRUE TRUE
Project controls FALSE FALSE FALSE FALSE FALSE TRUE FALSE FALSE

hRegression results for resampled main model. Hypothetical syndicates were resampled from the universe of potential syn-
dicates following the same rules as before. DV is a dummy indicating realized or hypothetical syndicate. Estimation method is
hierarchical glm with project-level fixed effects. Project-level control variables are employed in model 5, but output not shown. The
subsamples in model 6 (7) are low risk (high risk) project countries. POLCON and Polity are inverted such that higher numbers
indicate higher risk.
∗p < 0.05
∗∗p < 0.01
∗∗∗p < 0.001

TABLE 7:
Diversity and Syndicate Formation (frequency weighted)i

Full sample Sub samples
0 1 2 3 4 5 6 7

Intercept −4.94∗∗∗ −4.12∗∗∗ −3.93∗∗∗ −3.93∗∗∗ −3.93∗∗∗ −3.53∗∗∗ −4.15∗∗∗ −3.71∗∗∗

(0.01) (0.03) (0.02) (0.02) (0.02) (0.09) (0.03) (0.03)
Tie strength 0.06∗∗∗ 0.06∗∗∗ 0.05∗∗∗ 0.05∗∗∗ 0.05∗∗∗ 0.05∗∗∗ 0.05∗∗∗ 0.04∗∗∗

(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Diversity −0.61∗∗∗ −0.99∗∗∗ −1.12∗∗∗ −0.80∗∗∗ −1.10∗∗∗ −0.92∗∗∗ −1.60∗∗∗

(0.01) (0.03) (0.04) (0.04) (0.04) (0.06) (0.08)
(1 – POLCON) * Diversity 0.41∗∗∗ 0.27∗∗ 0.82∗∗∗ 0.76∗∗∗

(0.08) (0.08) (0.25) (0.13)
-Polity * Diversity 0.02∗∗∗

(0.00)
Log Likelihood -6,158.53 -6081.58 -6,081.58 -6,074.55 -6,075.13 -6,035.81 -3,268.93 -2,796.34
Num. obs. 117,741 117,741 117,741 117,741 117,741 117,741 63,060 54,681
Random variance component 0.00 0.00 37.98 36.51 36.42 33.07 20.70 57.80
Fixed effects TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE
Random slope FALSE FALSE TRUE TRUE TRUE TRUE TRUE TRUE
Project-level controls FALSE FALSE FALSE FALSE FALSE TRUE FALSE FALSE

iRegression results for resampled main model with weights for the frequency in data. DV is a dummy indicating realized or
hypothetical syndicate. Estimation method is hierarchical glm with project-level fixed effects. Project-level control variables are
employed in model 5, but output not shown. The subsamples in model 6 (7) are low risk (high risk) project countries. POLCON
and Polity are inverted such that higher numbers indicate higher risk.
∗p < 0.05
∗∗p < 0.01
∗∗∗p < 0.001

  • Text
  • Tables