+ All Categories
Home > Documents > Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of...

Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of...

Date post: 09-Aug-2018
Category:
Upload: lamanh
View: 213 times
Download: 0 times
Share this document with a friend
60
Syndicated Loans: The Role of Covenants in Mitigating Lender Disagreements Nishant Dass, Vikram Nanda, Qinghai Wang First Draft: October 2010 This Draft: March, 2011 Abstract We study the role of covenants in syndicated bank loans. We argue that, in addition to being a device for monitoring the borrower, covenants can help mitigate conflicts of interest between the lead-arranger and participating banks in the syndicate. Such disagreements can arise when, for instance, a lead-arranger has the incentive to support a poorly performing borrower and/or offer loan modifications while other syndicate lenders may prefer to discipline the borrower by accelerating the loan or enforcing default. We develop a simple model reflecting such conflicts and find empirical support for its predictions that covenants are less likely to be present: (i) in non-syndicated versus syndicated loans; (ii) when the lead’s loan allocation is greater; and (iii) when participating bank affiliates hold substantial equity in the borrower. Consistent with this evidence, we find that lead arrangers are more likely to syndicate with banks that hold borrower’s equity through affiliated entities. Keywords: Bank Loans, Conflict of Interest, Covenants, Lending Syndicate, Monitoring JEL Codes: G20, G21, G32 We are grateful to seminar participants at GA Tech and Sudheer Chava for his comments on the paper. We thank Michael Roberts for sharing his Dealscan-Compustat link with us. College of Management, Georgia Institute of Technology, 800 West Peachtree Street NW, Atlanta, GA 30308.
Transcript
Page 1: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Syndicated Loans: The Role of Covenants in Mitigating LenderDisagreements∗

Nishant Dass, Vikram Nanda, Qinghai Wang†

First Draft: October 2010This Draft: March, 2011

Abstract

We study the role of covenants in syndicated bank loans. We argue that, in addition to beinga device for monitoring the borrower, covenants can help mitigate conflicts of interest betweenthe lead-arranger and participating banks in the syndicate. Such disagreements can arise when,for instance, a lead-arranger has the incentive to support a poorly performing borrower and/oroffer loan modifications while other syndicate lenders may prefer to discipline the borrower byaccelerating the loan or enforcing default. We develop a simple model reflecting such conflictsand find empirical support for its predictions that covenants are less likely to be present: (i)in non-syndicated versus syndicated loans; (ii) when the lead’s loan allocation is greater; and(iii) when participating bank affiliates hold substantial equity in the borrower. Consistent withthis evidence, we find that lead arrangers are more likely to syndicate with banks that holdborrower’s equity through affiliated entities.

Keywords: Bank Loans, Conflict of Interest, Covenants, Lending Syndicate, MonitoringJEL Codes: G20, G21, G32

∗We are grateful to seminar participants at GA Tech and Sudheer Chava for his comments on the paper. Wethank Michael Roberts for sharing his Dealscan-Compustat link with us.

†College of Management, Georgia Institute of Technology, 800 West Peachtree Street NW, Atlanta, GA 30308.

Page 2: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Syndicated Loans: The Role of Covenants in Mitigating LenderDisagreements

Abstract

We study the role of covenants in syndicated bank loans. We argue that, in addition to beinga device for monitoring the borrower, covenants can help mitigate conflicts of interest betweenthe lead-arranger and participating banks in the syndicate. Such disagreements can arise when,for instance, a lead-arranger has the incentive to support a poorly performing borrower and/oroffer loan modifications while other syndicate lenders may prefer to discipline the borrower byaccelerating the loan or enforcing default. We develop a simple model reflecting such conflicts andfind empirical support for its predictions that covenants are less likely to be present: (i) in non-syndicated versus syndicated loans; (ii) when the lead’s loan allocation is greater; and (iii) whenparticipating bank affiliates hold substantial equity in the borrower. Consistent with this evidence,we find that lead arrangers are more likely to syndicate with banks that hold borrower’s equitythrough affiliated entities.

Keywords: Bank Loans, Conflict of Interest, Covenants, Lending Syndicate, MonitoringJEL Classification: G20, G21, G32

Page 3: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

1 Introduction

The literature on financial intermediation over the past few decades has built on the premise that

the raison d’etre of financial intermediaries is “delegated monitoring” (Diamond, 1984). The idea

is that a lender is assigned the role of monitoring on behalf of all other investors in order to avoid

duplication of effort involved in monitoring by each individual investor. Loan covenants facilitate

such monitoring and limit moral hazard by requiring the borrower to periodically provide the

lender with accounting information that reflects its financial health. Covenants can strengthen the

incentives of the financial intermediary to monitor and collect private information as well. This

is because the (violation of) covenants give the lender an option to act on the privately-collected

information, and possibly, call the loan when the situation warrants it (Rajan and Winton (1995)).

In syndicated loans, the role of a delegated monitor is typically played by the lead arranger;

such an arrangement (potentially) saves the participants from expending resources on monitoring

the borrower. In this paper we seek to understand the role of covenants in syndicated loans: do

they play the same role in syndicated loans as they do in loans issued by a single lender? It

is noteworthy, in this context, that loans issued by a single lender are far less likely to include

covenants than syndicated loans.1 What explains this greater use of covenants? Our claim is that

covenants may be serving a somewhat different role in syndicated loans. While covenants can help

counter borrower’s moral hazard in non-syndicated loans, covenants in syndicated loans can help

the syndicate members limit lead-arranger’s moral hazard.

A syndicate is characterized by asymmetric information between the lead arranger and the par-

ticipant lenders as well as their divergent incentives. The lead arranger clearly has more information

than other syndicate participants and it may prefer to renegotiate a loan instead of enforcing the

covenant because it finds other benefits in building/maintaining its lending relationship with the

borrower (Bharath et al., 2007). The situation is further aggravated as the lead arranger retains

only a fraction of the loan while the participants are saddled with the rest. Wary of the lead

arranger’s distorted incentives, the participants would be expected to demand more control over

the loan – especially in the event of renegotiation with the borrower – thus making covenants more

likely to be included in the loan contract. We formalize these arguments in a stylized model of a

1In the full sample of Dealscan loans issued to Compustat firms, roughly half the syndicated loans have a covenantwhile only a third of the loans issued by a single lender have one.

1

Page 4: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

lending syndicate and test the model’s empirical predictions using a sample of bank loans. Our

data are drawn from a 2009 extract of the Thomson Reuters’ Dealscan database.

We start with evidence on the greater prevalence of covenants in syndicated loans, when com-

pared with single-lender loans. Using a dummy dependent variable for the presence of a covenant

in the loan, our estimates show that the odds of a covenant are 56% higher in syndicated loans. We

then measure the extent of the conflicts of interest within the syndicate by the equity stake of the

lead arranger and the participants banks in the borrower, and argue that greater equity holdings

of the participants in the borrower help align their incentives with those of the lead arranger. The

key distinction between the lead arranger’s and the participants’ incentives is the willingness to

renegotiate instead of enforcing the covenant, and with an equity stake in the borrower, the partici-

pants would be equally willing to renegotiate and avoid default because default will have an adverse

impact on the equity value. Consistent with this, we find that the odds of a covenant are 22% lower

when the participant banks have a “large” equity stake in the borrower. That equity holdings of

the lead and participant lenders proxy their conflicts of interest is further confirmed when we look

within the loans issued by a single lender. In these loans, there is no impact of the lender’s equity

holdings on the likelihood of covenants. Similar results are found in the above cases when we use

the number of covenants and the number of types of covenants as alternative dependent variables.

We argue that this negative effect of equity holdings on the presence of covenants is not simply

due to the better information that these lenders have about the borrower. Empirical evidence to this

effect is found when the negative effect of lenders’ equity holdings on the presence of covenants is

robust to controlling for any past lending relationships of the lead as well as the participant lenders

with the borrower. The impact of equity holdings on the presence of covenants is also robust to

analyzing only those loans where the lead as well as the participant lenders have a positive equity

stake. We find that the negative effect of the participants’ equity holdings on the presence of

covenants is even stronger when the conflicts of interest within the syndicate are greater, such as

when the lead arranger has had a past lending relationship with the borrower while the participants

have not.

Covenants are tighter when there is greater information asymmetry between the firm’s managers

and the lenders (Garleanu and Zwiebel, 2009); however, the effect of the intra-syndicate conflicts

of interest on the covenant-tightness is not clear. As such, we find no empirical evidence of these

2

Page 5: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

conflicts of interests affecting the tightness of covenants. Sufi (2007) has shown that when the

moral hazard in lead arranger’s monitoring effort is greater, the lead is forced to retain a greater

fraction of the loan. Thus, in some sense, retaining a greater fraction of the loan can also mitigate

the conflicts of interest illustrated above. We show that the presence of covenants is a substitute for

this – the more the covenants demanded by the participants, the lower is the fraction of loan that

the lead needs to retain. We also find that despite controlling for the fraction of loan retained by

the lead arranger, the participants’ equity stake has a negative effect on the likelihood of covenants.

The presence of covenants and retaining a greater share of the loan – both are costly frictions

from the lead-arranger’s perspective. The former is inefficient in comparison with renegotiation

with the borrower and the latter increases the credit risk on the lead arranger’s balance sheet.

Given that participants with a greater equity stake in the borrower have their incentives aligned

with those of the lead arranger, and are less likely to demand covenants even after controlling for

the lead’s share of the loan, does the lead arranger seek out such participants when forming the

syndicate? We find that given a pool of potential participant lenders, the lead arranger matches

potential participants with borrowers by their equity stake in the borrower. We form the pool of

potential participants by grouping all those lenders that are participants in at least one loan deal

arranged by that lead in the given year. Thus, the lead arranger actively tries to minimize the

inefficiencies in the loan contract by picking the participants that are most compatible for a given

deal.

The larger equity holdings of the participants in the borrower as well as fewer covenants in the

loan, both may be driven by a third factor such as the borrower’s quality – a better-quality borrower

will attract equity investment and would also need fewer covenants. To address this endogeneity,

we instrument the participants’ equity holdings in the borrower by the equity stake of the pool

of potential participants that is described above. Using this instrument, we confirm the negative

relation between the participants’ equity holdings and the presence of covenants.

While covenants have received attention in the banking literature, we know little about the use

of sweeps in bank loans. Sweeps are restrictions on the use of internally-generated cashflows or

externally-raised capital. Typically, financial intermediaries, such as banks and venture capitalists,

include sweeps so that any internally-generated cashflows or externally-raised capital are first used

to repay the intermediary. We ask whether sweeps play a role similar to those of covenants in

3

Page 6: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

mitigating the conflicts of interest within the syndicate. We find that, in fact, the odds of the

presence of at least one sweep restriction are more than doubled in a syndicated loan, and also, the

odds of a sweep are 40% lower when the participants have a “large” equity stake in the borrower.

Thus, our results suggest that sweeps may play a role similar to the covenants’ in restricting both

the borrower’s as well as the lead arranger’s moral hazard.

Our paper makes several contributions to the financial intermediation literature. First, our

paper relates to the literature on the importance of monitoring in bank loans. The very reason

that justifies the existence of banks is their expertise in monitoring. Covenants have been seen

as facilitating the lender’s monitoring of the borrower to limit the moral hazard (e.g., Berlin and

Loeys, 1988; Gorton and Kahn, 2000). Chava and Roberts (2008) see covenants as financing

frictions that can adversely affect the borrower’s investment. They show this by analyzing the

borrower’s investment in the aftermath of a covenant violation, which leads to a transfer of control

rights to the lenders. We propose another role of covenants in limiting the moral hazard, but that

of the lead arranger and not only the borrower’s. We show that the participant lenders are likely

to demand covenants when they suspect the lead arranger’s incentives to be different than their

own with respect to the particular loan.

Second, our paper is related to the recent literature that has looked at the effects of information

asymmetry within the syndicate. Sufi (2007) shows that due to the moral hazard in lead arranger’s

monitoring of the borrower, the lead is forced to hold a larger share of the loan and form a more

concentrated syndicate, especially when there is greater information asymmetry between the bor-

rower and the lenders. Ivashina (2009) has shown that the information asymmetry between the

lead arranger and the syndicate participants results in a greater cost for the borrower. Along these

lines, we show in this paper that the conflicts of interest within the syndicate make covenants more

likely as well as can lead to the use of more covenant restrictions on the borrower.

Third, our paper is related to the literature explaining the formation of lending syndicates.

Bolton and Scharfstein’s (1996) findings suggest that a syndicate that leads to inefficient renegoti-

ation of the loan upon default is useful in deterring default but is costly when default occurs due

to an exogenous shock. Syndicate size is ultimately a function of this trade-off as well as borrower

characteristics, such as tangibility and credit ratings, etc. Lee and Mullineaux (2004) examine the

size and composition of the lending syndicate, and show that syndicates are smaller when the loan

4

Page 7: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

is risky. They, like Sufi (2007), conclude that syndicates are formed so as to improve the lenders’

monitoring. We show that the lead arranger forms the syndicate such that it minimizes contractual

frictions such as covenants.

Fourth, our paper is related to the literature on the effects of dual, and possibly conflicting,

roles of financial intermediaries. Puri (1999) analyzes the underwriting and lending roles of financial

intermediaries, and shows that such banks with dual roles can be better certifiers of the firm’s

quality in the securities market. However, equity-holdings of the intermediary in the firm impedes

its certification ability. Schenone (2004) also looks at the underwriting and lending roles of financial

intermediaries and shows that prior lending relationships with a borrower serve as a positive signal

in the IPO market, resulting in lower underpricing of the IPO. Massa and Rehman (2008) study

commercial banks that hold equity through affiliated mutual funds in the borrower that the bank

lends to, and find that affiliated mutual funds increase their holdings in the borrowing firm much

more than unaffiliated mutual funds. The superior performance of the borrower’s stock relative to

other stocks in the affiliated fund’s portfolio suggests that information flows across different parts

of a financial conglomerate. Dass and Massa (2011) also analyze the lenders that hold equity of

the borrowing firm, and show that the bank’s informational advantage due to lending results in

adverse selection in the equity market that suppresses the borrower’s stock liquidity. Our paper

also uses equity holdings of the lenders in the borrower to proxy for the conflicts of interest between

the lead and the participant lenders – larger equity holdings of participants in the borrower imply

smaller conflicts within the syndicate. We show that by aligning the interests of the lenders, equity

holdings of the participants are related with fewer covenant restrictions on the borrower.

The rest of the paper is structured as follows. §2 presents a simple model to motivate the

hypotheses tested in this paper and §3 describes the data sample as well as the variables used to

empirically test these hypotheses. §4 presents the results on the use of covenants in syndicated loans

and §5 analyzes whether the tightness of covenants is different in syndicated loans. §6 compares

the role of loan allocation of the lead arranger in mitigating the lead arranger’s moral hazard and

§7 addresses the issue of endogeneity of the syndicate put together by the lead arranger. §8 studies

the role of sweeps and concluding remarks are presented in §9.

5

Page 8: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

2 Intra-syndicate Disagreements: A Simple Model & EmpiricalPredictions

We offer a simple model of loan syndication in which a borrower, under certain conditions, has

the incentive to shift to riskier, lower-value projects. Two ways in which lenders can tackle such

borrower risk-shifting are considered. The first is loan modification which can reduce the borrower’s

incentive to risk-shift. The second is through the inclusion and enforcement of debt covenants.

A special concern that can emerge in the context of syndicated loans is that the various syndicate

members may not have the same interests or the same information and, hence, may disagree as

to whether to offer loan modification to the borrower. We argue that covenants, despite some

inefficiencies, may emerge in syndicated loans in response to potential lender disagreements. We

also argue that a higher lead allocation and larger holdings of borrower’s equity by participating

banks can mitigate lender disagreements, and thus, reduce the need for covenants.

2.1 Set-up

We assume that the borrowing firm has no resources and needs a loan to finance its profitable

project. The loan can be arranged by a lending syndicate that consists of a lead bank and other

participating banks. For simplicity, we take there to be a single participating bank. The loan can

also be non-syndicated, in which the lead is allocated the entire loan. All economic agents are

risk-neutral and the risk-free discount rate is taken to be zero.

There are three relevant dates in the model, t = 0, 1, and 2. At date t = 0 the syndicate lends

I, the investment required for the firm’s project. The face value of the loan is denoted by R and

it matures at t = 2, which is also when the project’s payoff is realized. The lead’s allocation of the

loan is denoted by α ∈ (0, 1], with α = 1 indicating a non-syndicated loan. The rest of the loan, an

allocation of 1−α, is provided by the participating lender in the syndicate. It is assumed, for now,

that the lead’s allocation is determined by factors outside the model, such as capital availability,

costs of raising additional funds, and the desire to limit exposure to any single firm.

At the terminal date t = 2, the firm’s project is expected to deliver a payoff of VH (high) with

probability π and VL (low) with complementary probability 1 − π. We take VH > VL > I. As a

consequence, in the absence of any risk-shifting, the debt would be risk-free. Information about

whether the project’s payoff is going high or low is received via a public signal at date t = 1.

6

Page 9: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

This public signal, which can be interpreted as the mandatory release of accounting information

by firms, is assumed to be contractible in that covenants can be written based on such a signal.

2.1.1 Risk-shifting by borrowers

It is assumed that once the firm receives information about its prospects at date 1, it can choose

to engage in risk-shifting. The incentive to engage in risk-shifting is typically associated with

relatively high leverage and, in the model, we assume that the parameters are such that it will

tend to occur only when the payoff is expected to be poor, i.e., VL. The switch to a risky project

cannot, however, be directly verified by lenders or be contracted upon – though the lenders can

take actions by modifying the loan or enforcing covenants to head off such risk-shifting.

We assume that, while there is no information asymmetry among lenders initially at date

0, the lead, on account of its greater understanding of the borrower’s condition, obtains private

information about the nature of the risk-shifting that the borrower might resort to at date 1.

Specifically, borrowers can differ in terms of the risky project that they can shift into, and this

information is learned by the firm and by the lead bank at date 1. This introduces a layer of

information asymmetry between the lead and its syndicate participant at date 1. The firm type

determines the nature of the risky project that it can shift to. The first type of firm (and risky

project), say type h, is present with a probability σ, while the other type, say type l, is present

with probability 1− σ. The nature of these projects is that either can produce a payoff of V1 > VL

at date 2 with a probability ρ. However, at date 2, while type-h will deliver X > 0 with probability

1−ρ, type-l produces only 0. For simplicity it is assumed that R > X and, hence, as we see below,

the borrower’s incentive to engage in risk-shifting is not affected by whether it is type-h or type-l.

It is assumed that there is substantial value loss from such risk shifting i.e., VL > ρV1 + (1− ρ)X.

2.1.2 Loan modification and covenants

We analyze the possibility of loan modification and the terms that the borrowers could be offered

to induce them to refrain from shifting to risky projects. We define an R∗ such that:

ρ(V1 −R∗) = VL −R

∗.

It is easy to check that such an R∗ < R. From the definition of R∗, it follows that if the face value

of the debt is reduced from R to R∗, there will no longer be risk-shifting.

7

Page 10: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

We now examine whether the lenders will indeed have the incentive to offer such a loan mod-

ification. This decision depends on what the lenders expect to receive when there is risk-shifting.

The expected values with risk-shifting depend in the firm types h and l and can be expressed as:

Rh = ρR+(1−ρ)X and Rl = ρR. To make the problem interesting, we will assume that the types

are such that with type-l, the lender would have an incentive to offer a loan modification, but not

when the expected payoff was larger with type-h. Hence, we will assume that Rh > R∗ > Rl.

The alternative action that may be available to the lenders is to enforce covenants, if such

covenants are included in the loan contract. For simplicity, we assume that such an enforcement

of the covenant results in effectively shutting down or liquidation of the firm. We denote this

liquidation value by K. It is assumed that payoff from liquidation is better than keeping the firm

alive only when the firm is of type-l and is expected to risk-shift – i.e. VL > ρV1+(1− ρ)X > K >

ρV1. Hence, it would be value maximizing to include and enforce debt covenants when the firm is

of type-l and state VL is realized.

We now turn to the issue of the objectives of the lead and the participating lender, and ana-

lyze the possibility that may have conflicting incentives in terms of whether or not to offer loan

modification to the borrower.

2.1.3 Incentives of the lead and the participating lenders

It is assumed that the manager of the firm always wants to keep the firm alive. In terms of the

lenders, recall, that while the lead bank is aware of the lender type (whether h or l) at date 1, the

participating bank is not. The participating bank will, therefore, agree with the loan being reduced

to R∗ if it believes the interests of the lead bank are similar to its own, or if it expects that to be

the optimal action irrespective of the lead bank’s incentives.

In terms of the participating bank’s incentives, we assume that it has no interest in the borrower

or the lead bank, other than to maximize the value it derives from the specific loan (and any equity

investment it may have in the firm).

On the other hand, we allow for the lead bank to obtain some potential benefits from the

borrower that are affected by whether the borrower is liquidated or remains viable. These potential

benefits, say B, could be due to maintaining a valuable relationship and avoiding a loss of reputation

due to the borrower’s bankruptcy. It could also reflect the equity stake that the lead has in the

8

Page 11: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

borrower. As a consequence, when faced with a choice of whether or not to offer a loan modification,

the lead bank’s decision will depend on whether:

B + α(R∗) ≷ α(Rh)

We assume that there exists a level of lead’s allocation α∗ < 1 such that the above relationship

is satisfied as an equality (i.e., α∗ = B

Rh−R∗ ). We now consider various cases depending on the

lead’s allocation and the equity stake of the participating bank.

• Case-I

It follows from the discussion above that as long as the lead’s allocation is α ≥ α∗, the

preferences of the lead and the participating bank are the same for whether or not to modify

the loan. In state VL, if the firm is of type-l, the loan is reduced to R∗. On the other hand,

if the firm is of type-h, no loan reduction is offered. It is optimal for the participating bank

to not disagree with the decisions of the lead bank. Under our assumptions, covenants would

not confer any benefits since Rh > K.

Hence, we can state that:

Case-I: There exist parameter values and a lead allocation level 0 < α∗ < 1, such that for

lead allocations α ≥ α∗, no covenants are employed. Loan modification is offered with some

probability and receives the support of both the lead and the participating bank.

• Case-II

In the second case, the lead bank’s allocation is α < α∗ and it is assumed that the participating

bank is only invested in the firm’s debt (the case with equity ownership is considered below).

With this level of allocation, the lead always prefers to modify the loan terms for the borrower

in state VL. However, this may not be optimal from the perspective of the participating bank.

Suppose there are no covenants, then in state VL, the lead always offers a debt reduction

to R∗. The interesting case is when the participating bank does not agree with the loan

modification. This will occur whenever R∗ < σRh + (1 − σ)Rl, i.e., the participating bank

would rather let the firm engage in risk-shifting than agree to the loan reduction. In this

case, it will always be optimal to include a covenant in the loan contract and to enforce the

9

Page 12: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

covenant, so long as value to the lenders from liquidating is greater than if the firms always

risk-shifts in state VL (i.e., K ≥ σRh + (1− σ)Rl). We can, therefore, state that:

Case-II: There exist parameter values and a lead allocation level 0 < α∗ < 1, such that for lead

allocations α < α∗, covenants will be included in the loan contracts and enforced with some

probability. Loan modifications will be favored by the lead bank but opposed by the participating

bank.

• Case-III

We now consider the case in which α < α∗, but the participating bank owns a fraction µ of

the borrower’s equity. The presence of an equity stake will induce the participating bank to

not disagree with the preference of the lead bank to always offer loan modification in state

VL. The condition can be expressed as follows:

(1− α)R∗ + µ(VL −R∗) > ρµ(V1 −R) + σRh + (1− σ)Rl

.

Since µ(VL−R∗) > ρµ(V1−R), an increase in the participating bank’s equity ownership in the

borrower makes it more likely that it will not disagree with the lead bank’s loan modification

proposals.

Case-III: There exist parameter values, the lead bank’s allocation level 0 < α∗ < 1, and the

participating bank’s equity stake µ∗ in the borrower, such that for lead allocations α < α∗ and

the participating bank’s equity stake in the borrower µ ≥ µ∗, covenants will not be included in

loan contracts; instead, loan modification will be offered with some probability and supported

by both the lead and the participating bank.

Note that in both cases I and II, the outcomes are inefficient in the following sense: (i) In case-I,

there is risk-shifting that takes place when the firm is of type-h. Even though this is more efficient

than liquidation, there is a value loss since the value in the absence of risk-shifting is greater than

that with risk-shifting, i.e., VL is greater than Rh. (ii) In case-II, there is inefficiency in that the

liquidation resulting from the enforcement of covenants results in less value than if there is loan

modification and no liquidation.

10

Page 13: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

It is somewhat surprising that case-III, in which there is syndicated lending – but one in which

there is equity ownership by the participating banks – is actually the most efficient. In this set-up,

there is no liquidation or risk-shifting since loan modification is always offered. Here, the lead as

well as the participating bank make decisions that take account of the value loss from risk-shifting

as well as from liquidation.

As a consequence, it follows that:

1. Even if the lead bank is not resource constrained but does not own significant equity in the

borrower, there are situations where a syndicate with participating banks owning substantial

borrower equity may lead to an efficient loan arrangement. Such an arrangement includes few

covenants and a greater likelihood of loan modification when the borrower’s outlook is poor.

2. On account of the efficient outcomes that emerge when participating banks own equity in

the borrower, we expect lead banks to specifically seek out such participating banks for their

syndicates.

2.2 Empirical Predictions

We now summarize the main empirical implications that follow from the framework developed

above.

Ceteris paribus,

1. Syndicated loans are more likely to have covenants than non-syndicated loans.

2. Among syndicated loans, the likelihood of covenants is lower when participating lenders have

a significant equity stake in the borrower.

3. Among syndicated loans, a lower lead allocation will be associated with a greater use of

covenants. Hence, lead allocation and participating-lenders’ equity are substitutes in terms of

their effect on the likelihood of covenants in the loan contract.

4. The lead bank is more likely to form a syndicate with lenders that have an equity stake in the

borrower.

11

Page 14: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

3 Data and Description of Variables

3.1 Data

We draw our data from three main sources – Thomson Reuters’ Dealscan and 13F Institutional

Holdings, and S&P’s Compustat (Fundamentals Quarterly) databases – and merge them to con-

struct our final sample. We start by collecting information on bank loans from a December 2009

extract of Thomson Reuters’ Dealscan database. We only include completed loan deals that are

syndicated in the U.S. and are denominated in US$. We screen these data to ensure that we have the

necessary accounting information for the borrower. This is facilitated by the Dealscan-Compustat

link used in Chava and Roberts (2008), an updated version of which is provided to us by Michael

Roberts. We use the loan package as the unit of analysis since covenants, which are the focus of

this paper, are determined at the level of a package and not facility. For information such as the

loan spread that is unique at the facility level but may vary across facilities within the package, we

only retain the facility with the largest loan amount in the package.

To obtain equity holdings of lenders in borrowing firms, we begin by searching the 13F in-

stitutional holdings database for investment affiliates of the lending banks. Using an algorithmic

match that is further improved by manually searching for lender names in the institutional holdings

database, we are able to find a matching manager number in 13F for about four hundred lenders.

Based on the lender-affiliated investment arm thus identified, we calculate the percentage equity

ownership of the lead arranger as well as the participant banks in the borrowing firm. The sample

thus constructed – with information on loans, borrower characteristics, and lenders’ equity stake

in the borrower – forms the basis of our analysis. Our final sample period starts in 1994 due to

limited and unreliable data on covenants prior to that year. We next provide a brief discussion of

the variables used in our analysis.

3.2 Dependent Variables

The main dependent variable, Covenant Decision, is an indicator for the presence of a loan covenant.

Table 1 lists the seventeen different types of covenant restrictions that are found across all the

loans in Dealscan.2 If any one of the covenants listed in that table is found in a given loan, then

2The characteristics of our 1994-2009 sample are very similar to those of the 1994-2005 sample used in Chava andRoberts (2008).

12

Page 15: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Covenant Decision equals one for that package; it is zero otherwise. An alternative dependent

variable, Number of Covenants, is employed as well. It counts the number of covenants that are

used in a loan package, and ranges between zero (for no covenants) and eight (the maximum in the

sample). We also employ a dependent variable based on classifying covenants into four types, as

indicated in Table 1. The dependent variable Types of Covenants counts the number of different

types of covenants present in a given loan, and ranges between zero (for no covenants) and four

(the maximum number of covenant types).

In other tests, we investigate whether tightness of the covenant is influenced by the potential

for disagreements within the syndicate. Tightness is the difference between the actual value of the

accounting variable (e.g., current ratio, or net worth, etc.) from the quarter before the loan and

the initial covenant threshold, divided by the firm-specific standard deviation of that accounting

variable. Furthermore, we analyze other forms of restrictions that are frequently employed by

lenders in loan deals. These consist of either restrictions on dividends or a variety of “sweeps,”

which are contractual obligations imposed on the borrower to service the loan with cashflows

generated by the activity specified in the sweep.

As discussed, we expect the share of loan retained by the lead bank to be negatively related

to the presence of loan covenants. The lead bank’s percentage share (Lead’s Allocation) of the

loan is directly obtained from the Lender Shares file in DealScan. We also study the participant

banks comprising the lending syndicate to test the prediction that the lead bank prefers to include

participating lenders that have an equity stake in the borrower. The dependent variable in this

case, Syndicate Member, is a dummy variable that equals one when a potential participant bank is

in the actual lending syndicate for the loan under consideration; it is zero otherwise. The potential

syndicate members are all those lenders that are participants in at least one deal arranged in the

same year and by the same lead bank as the given loan.

When analyzing sweeps and other restrictions, we use the dependent variable Sweeps Decision,

which is a dummy variable that equals one whenever the loan package includes at least one of the

sweeps/restrictions listed in Table 14; it is zero otherwise. We also use the Number of Restrictions

as an alternative dependent variable, which counts the number of sweeps/restrictions found in the

given loan.

13

Page 16: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

3.3 Independent Variables of Interest

We start our analysis with a comparison of syndicated loans with loans that are issued by a single

lender. The independent variable of interest in this case is Syndicate, which is equal to one for a

syndicated loan, and zero otherwise. Lead’s Equity Holdings is the equity stake held by the lead

arranger in the borrower, measured as a percentage of the borrower’s total shares outstanding. For

syndicated loans, Participants’ Equity Holdings are measured similarly and aggregated across all

the participants in the syndicate. Both these equity holdings are measured four quarters before the

loan’s start-date.

The lenders’ equity holdings may be naturally higher in firms that have greater institutional

ownership. Therefore, we control for the overall institutional ownership in the firm’s equity with

Institutional Holdings, which is the percentage equity ownership of all the 13F institutions in the

borrower, measured four quarters before the loan’s start-date.3

We also employ alternative versions of the lead’s and participants’ equity holdings to counter

the effect of any extreme values on our inferences. Lead’s Equity Holdings are Large is a dummy

variable that equals one if the lead arranger’s equity stake in the borrower is above its mean value

of 2%; it is zero otherwise. The average number of lenders in a syndicate is six, i.e., on average

there are five participants in a syndicate. Correspondingly, we define Participants’ Equity Holdings

are Large as a dummy variable equal to one if the aggregate equity holdings of the participants

in the borrower are above 10% (i.e., 5 × 2%); it is zero otherwise. In these specifications, we

also measure the overall institutional ownership with a dummy variable, Institutional Holdings are

Large; it equals one if the overall institutional holdings in the borrower are above its mean value,

40%, and it is zero otherwise.

3.4 Loan-specific and Borrower-specific Control Variables

The regression specifications include several loan-specific control variables. Number of Lenders

counts the number of lenders in syndicated loans. Loan Size is the logarithm of the loan amount.

Loan’s Maturity is the logarithm of the loan’s maturity counted in number of months and encom-

3The three equity-holdings variables – that of lead, participants, and all institutions – are winsorized at the 99thpercentile so as to avoid the effect of outliers on our estimation. Some of the extreme values seem to indicate greaterthan 100% equity holdings, which, as is well known, arise due to data error in treating the short-sales and/or stocksplits, etc.

14

Page 17: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

passes all facilities within the package. LIBOR (Drawn) is the yield spread over LIBOR that is paid

on drawn funds.4 Secured and Senior are dummy variables that equal one if the loan is secured

and senior, respectively; they are zero otherwise.5

Additional loan-specific control variables used in some specifications include the following. Lead-

Borrower Had a Relationship is a dummy variable indicating whether the current lead arranger

was a lead arranger for another loan to the same borrower in the previous three years. Similarly,

Participants-Borrower Had a Relationship indicates whether any of the current participant banks

was a lead arranger for another loan to the same borrower in the previous three years. Lead’s

Reputation (Participants’ Reputation) is a dummy variable that equals one if the lead arranger (at

least one participating bank) was in the top decile of banks, when ranked by the aggregate size or

number of deals arranged in the previous year.

In most of the specifications, we include various borrower characteristics in addition to the

loan-specific control variables.6 Book Leverage is the sum of long term debt and debt in current

liabilities, calculated as a fraction of assets. Cash Holdings are cash holdings and short-term

investments as a fraction of lagged assets. Return on Assets is income before extraordinary items

as a percentage of lagged assets. Tobin’s Q is the ratio of the sum of assets and market equity less

book equity and deferred taxes to assets. Tangibility Ratio is the ratio of net property, plant, and

equipment to assets. Capital Expenditures is the ratio of capital expenditures to net property, plant,

and equipment. KZ Index is 3.139× (Book Leverage) +0.283× (Tobin’s Q) −1.002× (Cashflow)

−39.368× (Dividends) −1.315× (Cash Holdings). Here, Cashflow is the sum of income before

extraordinary items, depreciation and amortization, calculated as a fraction of lagged assets, and

Dividends is total dividends calculated as a fraction of lagged assets. The summary statistics for

all the above variables are reported in Table 2.

We also control for industry, ratings, and year fixed effects in the regressions. Industry fixed

effects are based on the 48 Fama and French industries, except we exclude financial firms belonging

to the four Fama-French industries covering SIC codes 6000-6999. For the fixed effects of credit

ratings, we classify all borrowers into seven groups. Borrowers without a long-term S&P credit

4Number of Lenders, Loan Size, Loan’s Maturity and LIBOR (Drawn) are also winsorized at the 99th percentile.5As mentioned above, for loan-specific information that is only available at the facility level and not the package

level, we pick the largest facility in the package. As such, LIBOR (Drawn), Secured, and Senior correspond to thelargest facility.

6Borrower Size is not included because it is multi-collinear with Loan Size.

15

Page 18: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

rating are the benchmark group. All borrowers with a long-term S&P credit rating are classified

into the following six groups: Group 1 is for ratings CCC and below. Groups 2, 3, and 4 are for

borrowers rated B, BB, and BBB, respectively. Group 5 is for borrowers rated A and Group 6 is for

those rated higher than A. All these groups also include the “–” and “+” variations of the above

ratings.

4 The Role of Covenants in Resolving the Conflicts of InterestWithin the Lending Syndicate

4.1 Syndicated Loans

We start by presenting empirical results of the test of our primary prediction – that syndicated

loans are more likely to include covenants than loans that are issued by a single lender. As described

in §2 above, the participants in the syndicate are likely to impose covenants on the borrower in

order to counter the lead arranger’s tendency to renegotiate whenever the borrower faces financial

trouble.

Our empirical test for this prediction involves estimating the odds for the presence of at least

one covenant in a syndicated loan. We estimate a Logit regression where the dependent variable

is Covenant Decision and the explanatory variable of interest is Syndicate. In addition, guided

by the recent literature on bank loans, we include several variables to control for other potential

determinants of the presence of covenants in a loan contract. Specifically, we estimate the following

Logit model:

Covenant Decision = α1 + β1Syndicate+ γ1LOAN + γ2FIRM + δi + φj + ψt + � (1)

Here, LOAN and FIRM denote the various loan- and borrower-specific control variables that

are described above, while δi, φj , and ψt represent industry, ratings, and year fixed effects, respec-

tively. Alternative versions of the model use Number of Covenants or Types of Covenants as the

dependent variable. As described in §3 above, these are ordinal variables that count the number

and types of covenants, respectively. Accordingly, these alternative versions of the model are esti-

mated as an Ordered Logit and test whether syndicated loans are likely to have a larger number

and more types of covenants.

16

Page 19: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Columns (1) and (2) of Table 3 report the odds-ratios estimated from the above Logit model.7

The results in these two columns strongly support our main prediction that syndicated loans are

more likely to include covenants. In fact, the odds of a covenant are nearly 60% higher in a

syndicated loan, and these higher odds are statistically significant at the 1% level. We also find

that the odds of a covenant are lower in larger loans (denoted by Loan Size), higher when a loan

has longer maturity (Loan’s Maturity), lower when the borrower pays a higher yield spread (LIBOR

(Drawn)), and lower when the loan is secured (Secured). All these coefficients are economically

meaningful as well as statistically significant and conform with the sign we expect. Specifically,

borrowers with better credit risk are able to borrower larger amounts and face fewer covenant

restrictions. Loans with longer maturity increase the likelihood of risk-shifting by the borrower,

and therefore require more covenants. The negative relation between yields and covenants suggests

that these are substitute tools used by lenders to monitor the borrower. Loans that are secured by,

say, a fixed asset as a collateral, do not require covenants as much because the lenders directly have

recourse to the collateral. The odds of a covenant are dramatically higher in senior loans. While

this result is statistically strong, the positive impact of Senior is not easy to interpret, especially

because a vast majority of the loans in our sample are senior. A possible interpretation is that

the covenants that protect senior claimants may be less onerous than those that protect junior

claimants and, hence, may be more acceptable to borrowers. Including borrower-specific control

variables in column (2) does not change any of the column (1) results discussed above.

For the tests in columns (3) and (4) of Table 3, the dependent variable is Number of Covenants.

We find that the odds in column (3) of having more covenants are almost doubled in syndicated

loans, when compared with non-syndicated/single-lender loans. Given the proportional slopes

assumption underlying the Ordered Logit model, this implies that the odds of a borrower having

eight (the maximum number of covenants in our sample) covenants versus less than eight covenants

as well as the odds of, say, four or more covenants versus less than four covenants are both almost

twice as large in syndicated loans. This interpretation is equally valid for the odds of a borrower

having at least one covenant versus having no covenant at all . This result is statistically significant

at the 1% level and also robust to the inclusion of borrower-specific control variables in column (4).

7Since we report odds ratios in all our tables, the coefficient estimates that are greater than one suggest that oddsof covenants are greater due to the corresponding independent variables and coefficients smaller than one suggestthat odds of covenants are smaller due to those independent variables.

17

Page 20: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

The estimated cut-offs for the eight categories of Number of Covenants are statistically signifi-

cant, thereby justifying the use of this ordinal variable instead of collapsing the categories. However,

collapsing the higher categories of four or more covenants into a single fourth category does not

alter our results in any way. These estimates, with the dependent variable Number of Covenants

now ranging between zero and four, are left unreported for brevity.

For the tests in columns (5) and (6) of Table 3, the dependent variable is Types of Covenants.

The estimates are again obtained from an Ordered Logit model and we find that the odds in column

(5) of having more types of covenants are nearly doubled in syndicated loans, when compared with

non-syndicated loans. As described above, this implies that the odds of having four types versus

less than four types of covenants as well as the odds of having, say, two or more types versus less

than two types are nearly double in syndicated loans. We find that the coefficients of the control

variables are very similar to those reported in columns (3) and (4). Also, the estimated cut-offs

for the four categories of Types of Covenants are significant, which justifies the use of this ordinal

variable instead of one where the higher categories are collapsed. However, collapsing the categories

of two or more types of covenants into a single second category (besides the base category of no

covenants and the first category of one type of covenants) does not affect our results. We leave

these estimates unreported for brevity.

Overall, the findings from this table strongly support our claim that syndicates are more likely

than single lenders to impose covenant restrictions on the borrower.

4.2 Conflict of Interests and Lenders’ Equity Holdings

As described above, the conflict of interest between the lead arranger and participant banks can be

alleviated if the latter group benefits sufficiently from the survival of the borrower and is therefore

willing to renegotiate the loan contract instead of enforcing default. We have shown in §2 above that

this can be achieved if the participant banks hold an equity stake in the borrowing firm. Specifically,

with an equity stake in the borrower, the interests of the participants are better aligned with those

of the lead bank, and thus, the participants will be more willing to renegotiate with the borrower.

Before we take this prediction to the data, it is worth discussing how the lenders hold equity

in the borrower when it is well known that banks in the US are not allowed to do so directly in

any firm. Santos and Wilson (2008) have shown that even though US banks cannot hold equity

18

Page 21: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

directly, they can hold shares of a firm in a fiduciary capacity and, thereby, exercise control rights

even without any cashflow rights. They further provide evidence that a bank holding such an equity

stake in a firm charges lower interest rates on loans to that firm. This is consistent with our claim

that an equity stake of the participants in the borrower can make them more favorable toward the

borrower. Others have also shown evidence of banks being able to indirectly hold equity stakes

in their borrowers. For instance, Massa and Rehman (2008) show that mutual funds belonging to

the same financial conglomerate as a given lender increase their equity holdings in the borrower

by much more than mutual funds unaffiliated with that lender. Thus, through its affiliated mutual

fund, the lender is able to indirectly hold an equity stake in the borrower. Dass and Massa (2011)

also proxy for the lenders’ equity stake via the affiliated mutual funds’ holdings in the borrower, and

study its impact on the borrower’s stock liquidity. We follow the same methodology and measure

the lenders’ equity stake by the holdings of their affiliated institutional investors in the borrower’s

stock.

We empirically test the above prediction by estimating the effect of lead arranger’s as well

as participant banks’ equity holdings in the borrower on the odds of there being at least one

covenant. We again estimate a Logit model with Covenant Decision as the dependent variable but

the explanatory variables of interest are Lead’s Equity Holdings and Participants’ Equity Holdings.

Since the lenders’ equity holdings may be naturally higher in firms that have greater institutional

ownership, we control for the latter with Institutional Holdings. Given our claim about greater

conflicts of interest in syndicates, it is important to control for the Number of Lenders as the conflicts

might also be greater in larger syndicates. As before, we include several loan-specific and borrower-

specific variables to control for other potential determinants of covenants in a loan contract. To

ensure that the results obtained from estimating such a model are not influenced by extreme values

of the lenders’ equity-holdings, we also estimate the model with dummy variables Lead’s Equity

Holdings are Large and Participants’ Equity Holdings are Large instead of the percentage equity

holdings. In this case, we replace the institutional ownership with its discrete version, Institutional

Holdings are Large. All these variables have been defined above in §3. The Logit model described

19

Page 22: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

here can be represented as:

Covenant Decision = α2 + β2Lead’s Equity Holdings (are Large)

+ β3Participants’ Equity Holdings (are Large)

+ β4Institutional Holdings (are Large) + β5Number of Lenders

+ γ3LOAN + γ4FIRM + δi + φj + ψt + � (2)

Besides the main variables listed above, we include industry, ratings, and year fixed effects,

denoted in (2) by δi, φj , and ψt respectively. As before, we also use Number of Covenants and

Types of Covenants as alternative dependent variables in an Ordered Logit model.

The results are reported in Table 4. We start with columns (1) and (2), which report odds ratios

estimated from the Logit model (2). These results support our prediction that those syndicated

loans, in which the conflicts of interest between the syndicate members are mitigated and the

participants’ interests are aligned with the lead arranger’s by the virtue of their equity holdings in

the borrower, are much less likely to include covenants. In fact, the odds of a covenant are 1.3%

lower in column (1) for every percentage point equity ownership of the participants in the borrower

and 22% lower in column (2) when the participants have a large equity stake in the borrower. These

results are significant at the 5% and 10% level respectively.

Results for the dependent variable Number of Covenants are reported in columns (3) and (4),

and support our prediction. Specifically, the odds of having eight covenants versus less than eight

covenants are nearly 1% lower in column (3) for every percentage point equity ownership of the

participants in the borrower and 15% lower in column (4) when the participants have a large equity

stake in the borrower. Columns (5) and (6) of Table 4 report results for the dependent variable

Types of Covenants. The estimates are again obtained from an Ordered Logit model and we find

that the odds in column (5) of having more types of covenants are 1% lower for every percentage

point equity ownership of the participants in the borrower and the odds in column (6) are 22%

lower when the participants have a large equity stake in the borrower.

Overall, the findings from this table support our claim that covenants are an important tool

for the participants to ensure that they have some control over the outcomes of the loan, especially

under situations of financial distress facing the borrower when the interests of the lead and par-

ticipants are more divergent. From the results presented in Table 4, we find that the participants

20

Page 23: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

are willing to use fewer covenants when they have a greater equity stake in the borrower and, as a

result, their interests are better aligned with those of the lead arranger.

An important corollary of this result is that when the interests of the participant banks are

aligned with those of the lead arranger, then the loan contracts are more efficient with fewer

contractual frictions such as covenant restrictions. Therefore, it is optimal from a contracting

perspective for the lead arranger to seek out participant banks that have an equity stake (albeit

indirectly through affiliated institutional investors) in the borrower. We directly address this issue

in §7 below.

4.3 The Role of Equity Holdings When There is Only One Lender

One way of checking that we are indeed capturing the intra-syndicate conflicts of interest with the

equity holdings of the lenders is to analyze the effect of the lead’s equity holdings in the borrower

when there are no participants, i.e., the loan is issued by a single lender. Given that there are no

lender conflicts of interest when there is only one lender, the equity holdings of that lender should

have no effect on the likelihood of the use of covenants. We test this by estimating the following

model only for loans issued by a single lender:

Covenant Decision = α3 + β6Lead’s Equity Holdings (are Large)

+ β7Institutional Holdings (are Large) + γ5LOAN

+ γ6FIRM + δi + φj + ψt + � (3)

All the variables in (3) are the same as those defined in (2) of the previous sub-section and

again, we also use Number of Covenants and Types of Covenants as alternative dependent variables

in an Ordered Logit version of this model. The estimated results are reported in Table 5 and the

coefficients of Lead’s Equity Holdings in columns (1), (3), and (5) or of Lead’s Equity Holdings are

Large in columns (2), (4), and (6) consistently suggest no role for the lender’s equity holdings in

the borrower when there is only one lender and thus, no conflicts of interest to be resolved with

covenants. The lead’s equity holdings neither affect the likelihood of using any covenants nor do

they affect the number and types of covenants used. Instead of relying on covenants, the lead in this

case has enough leeway with the borrower to directly renegotiate when there is financial distress.

These results further reinforce the conclusion drawn above about the role of covenants in giving

21

Page 24: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

sufficient power to the participant banks.

4.4 Robustness Checks

Given how small the equity holdings of the average lender in a borrower are, it is worth asking the

question whether our results are affected in some manner by the extremely right-skewed distribution

of the lenders’ equity holdings. Also, a plausible alternative explanation for the above results is that

the participants’ equity holdings reflect the better information of these lenders about the specific

borrower, which reduces the need for imposing covenants on the borrower. So the negative effect of

the participants’ equity holdings on covenants is not due to the lower conflicts of interest between

the lead and the participants but instead due to more information about the borrower. To address

these concerns/alternative explanations we conduct the robustness tests described below.

We first address the concern about the right-skewness of the lenders’ equity holdings by limiting

to a sample of loans where the equity holdings of both the lead arranger as well as the participant

lenders is strictly positive. We re-estimate the model (2) on this reduced sample and find that

our previous results not only remain but, in fact, become stronger, as shown in Panel A of Table

6.8 Specifically, we find that the odds of a covenant are 1.7% lower in column (1) with every

percentage point equity ownership of the participants in the borrower and 27% lower in column (2)

when the participants have a large equity stake in the borrower. These results are significant at the

1% and 5% level respectively. Equally strong evidence is found when using Number of Covenants

in columns (3)-(4) and Types of Covenants in columns (5)-(6) – the likelihood of fewer covenants

and fewer types of covenants is significantly lower with greater equity ownership of the participant

banks in the borrower.

To directly control for the effect of the participant banks’ informational advantage with respect

to the borrower on the loan covenants, we include the following two characteristics in the model

(2) shown above. First, we include Participants-Borrower Relationship, which is a dummy variable

equal to one if a participant bank has been a lead-arranger for the given borrower in the previous

three years; it is zero otherwise. If this dummy equals one then a participant bank may have

better information about the borrower and therefore may not require covenant restrictions on the

borrower. The second variable that we include is Participants’ Reputation, which is a dummy

8Although we estimate the full model shown in (2), for brevity we only report the coefficients on the main variablesof interest across both panels of Table 6. The full set of coefficients can be made available upon request.

22

Page 25: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

variable equal to one if a participant lender belonged to the top ten percentile of banks, when

sorted by either the number or dollar value of deals lead-arranged in the past year. If this dummy

variable equals one, then the covenant restrictions might be fewer either because the “well-reputed”

participant has superior monitoring technology or a better screening ability. The model including

the above two variables can be written as:

Covenant Decision = α4 + β8Lead’s Equity Holdings (are Large)

+ β9Participants’ Equity Holdings (are Large)

+ β10Participants-Borrower Relationship + β11Participants’ Reputation

+ β12Institutional Holdings (are Large) + β13Number of Lenders

+ γ7LOAN + γ8FIRM + δi + φj + ψt + � (4)

The results from estimating this model are reported in Panel B of Table 6. We find that

our conclusion remains unaltered and directly controlling for these “information” variables does

not weaken the negative effect of the participants’ equity holdings on covenant restrictions. This

suggests that the equity holdings of the lenders are not simply reflecting an informational advantage.

4.5 When Conflicts of Interests Are Greater

If there are situations when the conflict of interest between the lead-arranger and the participants is

more acute, then our predicted effect of the participants’ equity stake on the covenant restrictions

ought to be stronger. We test this by identifying two such situations, where the conflicts may be

greater – first, when the syndicate is larger, and second, when the lead arranger has had a previous

lending relationship with the given borrower while none of the participants have. Specifically, we

test the following model:

Covenant Decision = α5 + β14Lead’s Equity Holdings (are Large)

+ β15Participants’ Equity Holdings (are Large) × Conflicts are Greater

+ β16Participants’ Equity Holdings (are Large) × Conflicts are Smaller

+ β17Conflicts are Greater

+ β18Institutional Holdings (are Large) + β19Number of Lenders

+ γ9LOAN + γ10FIRM + δi + φj + ψt + � (5)

23

Page 26: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

While the rest of the variables are same as those defined above, we have interacted the par-

ticipants’ equity-holdings with two dummy variables indicating situations in which the conflicts of

interest are likely to be greater or smaller. Conflicts are Greater is first proxied by Large Syndicate

and then by Lead Has Stronger Reln. The complementary dummy variable Conflicts are Smaller

is first proxied by Small Syndicate and then by Participants Have Stronger Reln. The dummy

variable Large Syndicate equals one when the syndicate size is larger than its average of six; it is

zero otherwise. The complementary variable Small Syndicate is one when the syndicate size less

than or equal to six; it is zero otherwise. The results from this estimation are reported in Panel

A of Table 7. The dummy variable Lead Has Stronger Reln. equals one when the current lead

arranger has arranged another loan for the same borrower in the past three years while none of the

participants have; it is zero otherwise. The complementary variable Participants Have Stronger

Reln. equals one when at least one of the participants has lead-arranged another loan for the same

borrower in the past three years while the current lead-arranger has not; it is zero otherwise. These

results are reported in Panel B of Table 7.9

The estimated β15 is found to be smaller (more negative) than β16; this result is mostly sig-

nificant statistically and is always economically meaningful, thus supporting our prediction. The

inference we make from these results is that, indeed, when the conflicts of interest are greater, the

participants’ equity holdings are more effective and useful in resolving those conflicts, and thereby

reducing the need for covenant restrictions.

We end this section on the role of loan covenants by concluding that covenants do help in

resolving the conflicts of interest inherent in the groups of banks coming together in a syndicate.

There is one outstanding issue – that of endogeneity. Specifically, it is plausible that the equity

holdings of the participants as well as the lower need for covenant restrictions are both due to

the better quality of the borrower. We address this issue of endogeneity in §7, where we analyze

whether the lead arranger forms the syndicate such that a potential participant lender is matched

with the borrower on the basis of its equity ownership in that borrower.10

9Again for brevity, we only report the coefficients on the main variables of interest. The full set of coefficients canbe made available upon request.

10The instruments that we use to address the above endogeneity are based on the characteristics of the “potentialsyndicate”, which is why we deal with the issue in that section.

24

Page 27: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

5 Tightness of Covenants

Tightness of covenants is an important aspect of the loan restrictions imposed by the lenders and

bank loan covenants are typically tighter than covenants in other financial contracts (Garleanu

and Zwiebel, 2009). Tightness of loan covenants reflects the future potential agency problems

anticipated by the lenders vis-a-vis the borrower (Chava and Roberts, 2008). Selection of tight

covenants conveys important information about the future performance of the firm and about

the consequences of covenant violation (Demiroglu and James, 2010). Given the importance of

covenant tightness, in this section we ask whether the conflicts of interest highlighted above have

any influence on covenant tightness. It can be argued that the covenant boundary is set tighter

when the conflicts of interest between the lead arranger and the participants are greater. Since

the lead arranger has an informational advantage over the participants, the latter group may insist

on keeping the covenants tighter because even a small deviation in the covenant variable may lead

to a violation, thereby promptly informing the participants of the changing financial health of the

borrower. On the other hand, tightness may only vary by borrower-quality and not by the extent

of conflicting interests within the syndicate. The ambiguous effect of the conflicts of interest on

covenant tightness is therefore clarified by testing three different empirical models corresponding

to the models (1) and (2) used above, except the dependent variable now is Covenant Tightness.

Following Chava and Roberts (2008), we define the tightness for three types of covenants (current

ratio, net worth, and tangible net worth) because only the corresponding accounting variables

are standard and clearly defined while the lender has more discretion in defining other covenant-

related accounting variables, such as EBITDA or debt. For these three variables, the Covenant

Tightness is defined as the difference between the actual accounting variable (current ratio, net

worth, and tangible net worth) and the initial covenant threshold, divided by the firm-specific

standard deviation of the accounting variable. Here, the actual value of the accounting variable

is measured at the end of last fiscal quarter before the loan starts, and its standard deviation is

measured using the value over the past sixteen quarters.

Results from these tests are reported in Table 8. Columns (1)-(3) report the results using

Current Ratio covenants, columns (4)-(6) report results for Net Worth covenants, and the last

three columns (7)-(9) report those for Tangible Net Worth covenants. The first column for each

25

Page 28: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

of the covenants tests whether Covenant Tightness is different in syndicated loans than in single-

lender loans, given the conflicts of interests in the syndicate. The insignificance of the coefficient on

Syndicate dummy in columns (1), (4), and (7) clearly shows that there is no difference in between

syndicated loans and single-lender loans, when it comes to covenant tightness. Next, we look within

the syndicates and test whether the extent of the conflicts of interest, measured by Lead’s Equity

Holdings and Participants’ Equity Holdings, has an impact on the covenant tightness. Here again,

we find no impact of the lenders’ equity holdings on tightness; the coefficient on both the variables

are statistically not different from zero across columns (2), (5), and (8). This is further corroborated

by the results reported in columns (3), (6), and (9); the coefficients on Lead’s Equity Holdings are

Large and Participants’ Equity Holdings are Large are equally insignificant.

Overall, these results show that while the presence of covenants may be governed by the conflicts

of interest within the syndicate, the tightness is not. This suggests that tightness is perhaps only

determined by the borrower’s characteristics – poor-quality borrowers have tighter loan covenants

imposed on them by the lenders.

6 Loan Allocation

We have argued that a high loan allocation to the lead can obviate the need for covenants – since the

high allocation reduces the potential for moral hazard on the part of the lead bank and, thereby,

disagreement within the syndicate. We test whether a larger allocation of the loan to the lead

arranger lowers the need of covenants by estimating the following model:

Covenant Decision = α6 + β20Lead’s Allocation + β21Lead’s Equity Holdings (are Large)

+ β22Participants’ Equity Holdings (are Large)

+ β23Institutional Holdings (are Large) + β24Number of Lenders

+ γ11LOAN + γ12FIRM + δi + φj + ψt + � (6)

Here, all the variables are the same as defined earlier while Lead’s Allocation is the percentage

of loan allocated to the lead arranger. Information on individual lender’s allocation of the loan

is provided in Dealscan for some of the loans; unfortunately, this information is missing for a

majority of the loans.11 Following the previous models, we also use Number of Covenants and

11This explains the dramatically smaller number of observations in these results.

26

Page 29: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Types of Covenants as alternative dependent variables in this model (6). Results from estimating

this model for the three alternative dependent variables are reported in columns (1)-(3) of Table 9,

respectively.

However, as is clear from the discussion above, the allocation of the loan to the lead arranger is

not exogenous. It is clearly determined endogenously along with other features of the loan, such as

the presence of covenants, etc., and is a direct consequence of the extent of the conflicts of interest

within the syndicate.12 Therefore, we instrument Lead’s Allocation and re-estimate the model (6);

these results for the dependent variables Covenant Decision, Number of Covenants, and Types of

Covenants are reported in columns (4)-(6) of Table 9, respectively. The instruments used are:

(i) the loan allocation to a lead arranger averaged across all the loans to borrowers in the same

industry and year as the current loan; (ii) the reciprocal of the number of lenders averaged across

all the loans to borrowers in the same industry and year as the current loan; and (iii) the median

loan allocation of this lead arranger in all the loan deals lead-arranged in the past (since the start

of our data sample). These instruments meet the relevance and exclusion criteria, as summarized

in the Kleibergen-Paap Wald F and Hansen’s J statistics, respectively.

Note that the instrumental variables regression for a discrete dependent variable assumes that

the error terms of the structural and reduced-form equations are joint-normally distributed, due to

which the regression for Covenant Decision in column (4) is estimated using the IV-Probit model.

To make the results for Covenant Decision in column (1) comparable with those in column (4),

we estimate the model in column (1) as a Probit instead of the Logit model that we have used

throughout. For similar distributional issues, the models in columns (5) and (6) are estimated as

IV-regressions and for comparison, the models in columns (2) and (3) are estimated as simple OLS

regressions.

The Probit results in column (1) for the dependent variable Covenant Decision reveal a clear

negative correlation between the presence of covenants and the loan allocation to the lead arranger.

This is after controlling for the negative effect of the participant banks’ equity holdings documented

above. Similar evidence is found for the dependent variables Number of Covenants and Types of

Covenants in columns (2) and (3), respectively. The OLS estimates of the coefficient on Lead’s

12Indeed, in unreported results, we find that the lead’s allocation of the loan is directly proportional to the equity-holdings of the lead arranger in the borrower – i.e., exactly when the potential for moral hazard in monitoring by thelead is greater. These results can be made available upon request.

27

Page 30: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Allocation are negative and significant at the 1% level even after controlling for the negative effect

of the participant banks’ equity holdings. The results after instrumenting the Lead’s Allocation in

columns (4)-(6) yield much higher point estimates and are also much stronger statistically. Overall,

these results suggest that the need for covenants to resolve the intra-syndicate conflicts of interest

is mitigated when the lead arranger holds a bigger share of the loan. And therefore, covenants and

the lead arranger’s loan allocation are substitutes in allaying the conflicts of interest within the

syndicate.

7 Syndicate Formation by the Lead Arranger – Matching of Po-tential Participants with the Borrower

7.1 Syndicate Formation

We have shown that due to the conflicts of interest within the syndicate, the loan must either have

more covenant restrictions or the lead arranger must retain a larger share of the loan. Both of

these outcomes are difficult/inefficient for the lead arranger – covenants are costly frictions that

make renegotiation more difficult and retaining a greater share of the loan limits the lead-arranger’s

ability to diversify across more borrowers. Therefore, it is optimal for the lead arranger to mitigate

these frictions, and one way of doing so may be through selecting a certain type of participant

lender. Specifically, if the participants’ equity holdings in the borrower help with aligning their

interests with those of the lead arranger, and thereby reducing the above frictions in the loan

contracts, then it may be optimal for the lead arranger to pick participants that have an equity

stake in the borrower. This is what we address in this section.

To test this, we construct a sample of “potential” participants of the lead arranger. These

potential participants are all lenders that participate with the lead arranger on some syndicated

loan deal in the current year, and are either involved in at least ten loan deals in the given year or

do not belong to the bottom quartile of the banks, when sorted by the aggregate size of all the loan

packages that the lender is involved with during the given year. With the “potential syndicate”

thus formed for each lead arranger, we ask whether the equity stake of the potential participant

in the given borrower is a predictor of whether that potential participant ends up as an actual

28

Page 31: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

syndicate member or not. The question is answered with the following empirical model:

Syndicate Member = α7 + β25Potential Lender’s Equity Holdings (are Large)

+ β26Potential Lender-Borrower Relationship + β27Lead-Potential Interactions

+ β28Lead’s Equity Holdings (are Large) + β29Institutional Holdings (are Large)

+ γ11LOAN + γ12FIRM + δi + φj + ψt + � (7)

Syndicate Member is a dummy variable that equals one if the potential participant is on the

eventual syndicate for the given loan package; it is zero otherwise. Potential Lender’s Equity

Holdings is the equity stake of the potential lender in the borrower. It is measured as it was

earlier for all lenders – via the holdings of the affiliated institutional investor in 13F, measured

four quarters before the loan; as earlier, the equity stake is zero if the potential participant does

not have an affiliation in 13F. Potential Lender’s Equity Holdings are Large is a dummy variable

that equals one if the potential lender’s equity holdings in the borrower are greater than its mean,

which is 0.5%. Potential Lender-Borrower Relationship is a dummy variable that equals one if the

potential lender has lead-arranged a loan deal for the given borrower in the past three years; it is

zero otherwise. It controls for the potential lender’s prior lending relationship with the borrower.

Lead-Potential Interactions counts the number of syndicated loan deals that the given pair of lead

arranger and potential participant have completed together in the given year. This variable is

included to control for frequent partnering on deals by the pair of lenders. We also control for the

lead’s as well as overall institutional equity holdings in the borrower; again, these are measured

four quarters prior to the loan.13 LOAN and FIRM are the same loan- and firm-specific control

variables that have been used earlier. Descriptive statistics of these data are reported in Table 10.

We test this Logit model (7) using data from 1994 onwards on loans made to non-financial

firms. The estimated odds ratios are reported in Table 11. Results across columns (1)-(4), which

show different specifications of the same model, reveal that the odds of a potential participant being

chosen for the actual syndicate are significantly higher when the potential participant has an equity

stake in the borrower. These results are significant at the 1% level, and the coefficient estimates of

Potential Lender’s Equity Holdings are Large show that the odds are 13% higher when the potential

13All these independent variables described above, except the dummy variable Potential Lender-Borrower Rela-tionship, are winsorized at the 99th percentile to avoid the effect of outliers on our results.

29

Page 32: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

lender has at least 0.5% equity stake in the borrower. This is true even after controlling for the most

important determinant of being chosen on the actual syndicate – the prior lending relationship of

that lender with the given borrower, which increases the odds by more than ten times!

We have argued above that when the lead arranger’s equity stake in the borrower is larger, the

conflicts of interest within the syndicate might be worse. In column (5) (column (6)) of the Table,

we interact Potential Lender’s Equity Holdings (are Large) with Lead’s Equity Holdings are Large

to test the effect of the potential lender’s equity stake on being chosen is different when the conflict

of interest are worse. Specifically, when the conflicts of interest are worse, the lead’s incentives to

pick a certain type of participants to mitigate these conflicts should also be stronger. The results

in column (5) (column (6)) support this claim as we find that the odds of a potential participant

being chosen are 10% (28%) higher when the lead’s equity stake in the borrower is large.

Overall, the results in Table 11 confirm that lead arrangers do seek out a certain type of

participants to mitigate the conflicts of interest with the participants as otherwise it leads to

greater frictions in the loan contract. Anticipating more covenants or a larger loan allocation due

to the conflicts of interest within the syndicate, the lead arranger forms the syndicate in a manner

that reduces these conflicts.

7.2 Instrumented Equity Holdings

In this subsection, we address the inherent endogeneity in the negative relationship between equity

holdings and covenants that we have illustrated earlier in Table 4. It is plausible that the equity

holdings of the participants as well as the lower need for covenant restrictions are both due to the

better quality of the borrower – holding equity in this borrower is a good investment as well as the

need for covenant restrictions on this borrower is lower. We address this now by instrumenting the

Participants’ Equity Holdings in model (2) as our instrument is based on the “potential syndicate”

described above. Specifically, we pick our instruments from the following set of variables – 1) the

potential syndicate’s equity holdings in the borrower; 2) dummy for when the potential-lenders’

equity-holdings are positive; 3) dummy for when the potential-lenders’ equity-holdings large; 4) each

potential participant’s average equity-holdings in the borrower; 5) potential participant’s average

equity-holdings in the borrower times the size of the actual syndicate; and 6) the size of the potential

syndicate. The specific instruments are chosen such that they satisfy the relevance and exclusion

30

Page 33: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

conditions.

Given the assumption of jointly-normal distribution of the first- and second-stage error terms,

we must estimate the instrumental-variable regression for the dependent variable Covenant Decision

as a Probit and as an ordinary IV regression for the dependent variables Number of Covenants and

Types of Covenants. The model specification is the same as that in (2) above and the results from the

estimation are reported in Table 13. We find that the instrumented Participants’ Equity Holdings

is strongly significantly negative in all three regressions, i.e., controlling for the endogeneity, we

still find that the participants’ equity stake in the borrower lowers the likelihood of any covenants

and also reduces the number/type of covenants used, as it mitigates the conflicts of interest with

the lead arranger.

8 Sweeps – The Other Type of Loan Restrictions

Besides covenants, bank loans (and many other types of financial intermediation contracts, such as

those with private equity firms) often have another type of restrictions called sweeps. Sweeps are a

variety of restrictions imposed by the financial intermediary on the borrower that are characterized

by the limitations on the use of internally-generated cashflows or externally-raised capital. These

sweep restrictions ensure that any internally-generated cashflows or externally-raised capital is first

used to repay the lenders, before it can be used for any other purpose. Table 14 lists the types of

sweeps and other restrictions found in our sample, and also provides some summary statistics on

the loans that include these sweeps/restrictions.

As is evident from the statistics reported in Table 15, the sweeps are about as frequently used

as covenants in loans. In fact, Dividend Restriction is the most frequently used restriction of any

kind in our sample of loans. Therefore, in this section, we ask whether sweeps play a role similar to

that of covenants in resolving the intra-syndicate conflicts of interest. To answer this question, we

start with a very simple model, similar to (1) above, except now our dependent variables are the

dummy variable Sweeps Decision and the ordinal variable Number of Sweeps. As before, the model

with Sweeps Decision is estimated as a Logit and that with Number of Sweeps is estimated as an

Ordered Logit. We report the odds-ratio of the main variable of interest (Syndicate) in Panel A of

Table 15; while we include all the other variables from (1) in our estimation, we do not report the

remaining coefficients for brevity. Results in Panel A of Table 15 show that, like covenants, sweeps

31

Page 34: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

are more likely to be used in syndicated loans than loans issued by a single lender.

We next look within the sample of syndicated loans, and proxy the extent of the conflicts of

interest with the same measures as used in the model (2) above. We use Sweeps Decision and

Number of Sweeps as the dependent variables, and follow the same set of specifications as in Table

4 earlier. We report these results in Panel B of Table 15; again, the coefficients of all the variables

except those of primary interest are not reported for brevity. The odds ratios suggest that sweeps

are a lot less likely in loans when the participants’ interests are better aligned with those of the

lead-arranger, as proxied by the participants’ equity stake in the borrower. This is true in all the

columns, (1)-(4), for both dependent variables and for using both measures of the participants’

equity stake.

These results are both statistically as well as economically significant. Overall, the results in

Table 15 suggest that sweeps may also play a role similar to that of covenants in partially resolving

the conflicts of interest within the syndicate.

9 Concluding Remarks

In this paper we study the role of covenants in syndicated bank loans. Our argument is that, in

addition to being a device for monitoring the borrower, covenants can help mitigate conflicts of

interest between the lead-arranger and participating banks in the syndicate. Such disagreements

can arise when, for instance, a lead-arranger has incentives that differ from those of other members

of the lending syndicate. The lead arranger might, for instance, want to support a borrower with

which it has developed significant relationship capital. This support might take the form of the

lead-arranger being willing to offer loan modification to the borrower when it is facing financial

hardship. Such loan modification may, however, be opposed by other syndicate lenders that are

primarily interested in the value of the loan – and have little incentive to agree to a loan modification

that does not maximize the expected payoff from the loan.

We develop a simple model to show that covenants can be advantageous when there is the

potential for disagreements within the syndicate. We test and find empirical support for the

model’s predictions that covenants are less likely to be present in non-syndicated versus syndicated

loans. The model also suggests that when the lead’s loan allocation is greater, there is less potential

for conflict of interest because the incentives of the lead bank and the participating banks is more

32

Page 35: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

aligned. This is also the case when the participating banks have affiliates that own substantial

equity in the borrower. This is supported by our empirical results that covenants are less likely to

be present when the lead’s loan allocation is greater and when participating bank affiliates hold

substantial equity in the borrower. Consistent with this evidence, we find that lead arrangers are

more likely to syndicate with banks that hold borrower equity through affiliated entities.

33

Page 36: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

References

Berlin, Mitchell, and Jan Loeys (1988), “Bond Covenants and Delegated Monitoring,” Journal ofFinance, vol. 43(2), 397-412.

Bharath, Sridhar, Sandeep Dahiya, Anthony Saunders, and Anand Srinivasan (2007), “So What DoI Get? The Bank’s View of Lending Relationships,” Journal of Financial Economics, vol. 85(2),368-419.

Chava, Sudheer, and Michael Roberts (2008), “How Does Financing Impact Investment? The Roleof Debt Covenants,” Journal of Finance, vol. 63(5), 2085-2121.

Dass, Nishant, and Massimo Massa (2011), “The Impact of a Strong Bank-Firm Relationship onthe Borrowing Firm,” Review of Financial Studies, vol. 24(4), 1204-1260.

Demiroglu, Cem, and Christopher M. James (2010), “The Information Content of Bank LoanCovenants,” Review of Financial Studies, forthcoming.

Diamond, Douglas W. (1984), “Financial Intermediation and Delegated Monitoring,” Review ofEconomic Studies, vol. 51(3), 393-414.

Garleanu, Nicolae, and Jeffrey Zwiebel (2009), “Design and Renegotiation of Debt Covenants,”Review of Financial Studies, vol. 22(2), 749-781.

Gorton, Gary, and James Kahn (2000), “The Design of Bank Loan Contracts,” Review of FinancialStudies, vol. 13(2), 331-364.

Gopalan, Radhakrishnan, Vikram Nanda, and Vijay Yerramilli (2010), “Lead Arranger Reputationand the Loan Syndication Market,” Journal of Finance, forthcoming.

Ivashina, Victoria (2009), “Asymmetric Information Effects on Loan Spreads,” Journal of FinancialEconomics, vol. 92, 300-319.

Lee, Sang Whi, and Donald J. Mullineaux (2004), “Monitoring, Financial Distress, and the Struc-ture of Commercial Lending Syndicates,” Financial Management, Autumn, 107-130.

Massa, Massimo, and Zahid Rehman (2008), “Information flows within financial conglomerates:Evidence from the BanksMutual Funds Relation,” Journal of Financial Economics, vol. 89(2),288-306.

Puri, Manju (1999), “Commercial Banks as Underwriters: Implications for the Going Public Pro-cess,” Journal of Financial Economics, vol. 54, 133-163.

Rajan, Raghuram, and Andrew Winton (1995), “Covenants and Collateral as Incentives to Moni-tor,” Journal of Finance, vol. 50(4), 1113-1146.

Santos, Joao A. C., and Kristin E. Wilson (2008), “Does Banks’ Corporate Control Benefit Firms?Evidence from US Banks’ Control Over Firms’ Voting Rights,” NY Fed working paper.

Schenone, Carola (2004), “The Effect of Banking Relationships on the Firm’s IPO Underpricing,”Journal of Finance, vol. 59(6), 2903-2958.

Sufi, Amir (2007), “Information Asymmetry and Financing Arrangements: Evidence from Syndi-cated Loans”, Journal of Finance, vol. 62(2), 629-668.

34

Page 37: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Table 1: Summary of Covenant Restrictions. This table presents a list of covenant restrictions that are found in our sample of loans made to non-financialfirms belonging to the intersection of Compustat and Dealscan during the 1994-2009 period. These statistics are based on only those loans for which the deal sizeand number of lenders are not missing.

Type of Number of Average Median Aggregate SizeCovenant Covenant Packages Deal Size ($m) Deal Size ($m) of Deals ($bn)

Max. Debt to EBITDA Leverage 6806 413.11 175.00 2,811.65Min. Interest Coverage Liquidity 5040 458.90 225.00 2,312.84Min. Fixed Charge Coverage Liquidity 4948 272.36 125.00 1,347.64Tanginble Net Worth Net Worth 3166 129.68 30.00 410.55Max. Capex Capex 3086 283.54 105.00 874.99Net Worth Net Worth 2750 285.35 125.00 784.70Max. Leverage ratio Leverage 2448 507.82 250.00 1,243.13Max. Debt to Tangible Net Worth Leverage 1924 128.63 18.98 247.48Min. Current Ratio Liquidity 1735 114.04 35.00 197.86Min. Debt Service Coverage Liquidity 1369 120.41 25.00 164.85Min. EBITDA Liquidity 1303 163.00 65.00 212.39Max. Senior Debt to EBITDA Leverage 1274 405.75 180.00 516.92Min. Quick Ratio Liquidity 587 35.77 13.00 21.00Min. Cash Interest Coverage Liquidity 191 338.95 150.00 64.74Max. Debt to Equity Leverage 136 250.88 46.20 34.12Max. Senior Leverage Leverage 23 736.41 242.50 16.94Max. Loan to Value Leverage 14 217.05 112.50 3.04

35

Page 38: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Table 2: Summary Statistics. This table presents summary statistics of the main variables used in our analyses. We categorize the variables into loan-,lender-, and borrower-characteristics. Here again, we limit to only those loans for which the deal size and number of lenders are not missing. Panel A shows thestatistics for the overall sample while Panel B compares the same characteristics across syndicated and single-lender loans.

Panel A: Overall Sample

Units N Mean Median Std. Dev.

Loan Characteristics:Covenant Decision 0/1 25,052 0.549 1.000 0.498Number of Covenants integer 25,052 1.469 1.000 1.588Types of Covenants integer 25,052 1.247 1.000 1.284Syndicate 0/1 25,052 0.682 1.000 0.466Number of Lenders integer 25,052 6.262 3.000 7.150Loan Size ln(dollars) 25,052 18.517 18.660 1.701LIBOR (Drawn) basis points 18,701 166.389 150.000 115.858Loan’s Maturity ln(1 + months) 22,963 3.655 3.905 0.750Secured 0/1 16,203 0.729 1.000 0.445Senior 0/1 25,052 0.984 1.000 0.127

Lender Characteristics:Lead’s Allocation % 8,413 49.407 35.714 35.919Lead’s Equity Holdings % 19,698 1.817 0.000 4.404Participants’ Equity Holdings % 13,729 5.650 0.935 10.420Institutional Holdings % 20,924 45.441 47.790 29.514Lead’s Equity Holdings are Large 0/1 19,698 0.201 0.000 0.400Participants’ Equity Holdings are Large 0/1 13,729 0.177 0.000 0.382Institutional Holdings are Large 0/1 20,924 0.577 1.000 0.494

Borrower Characteristics:Book Leverage ratio 24,994 0.333 0.304 0.317Cash Holdings ratio 23,864 0.091 0.033 0.300Return on Assets % 23,851 -0.115 0.940 15.743Tobin’s Q ratio 25,049 2.524 1.318 128.051Tangibility Ratio ratio 24,926 0.347 0.283 0.250Capital Expenditures ratio 22,531 0.126 0.048 5.299KZ Index 22,644 1.269 1.189 1.966

36

Page 39: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Panel B: Syndicated vs. Single-Lender Loans

Syndicated Loans Single-Lender Loans

N Mean Median Std. Dev. N Mean Median Std. Dev.

Loan Characteristics:Covenant Decision 17,087 0.561 1.000 0.496 7965 0.525 1.000 0.499Number of Covenants 17,087 1.501 1.000 1.594 7965 1.400 1.000 1.571Types of Covenants 17,087 1.277 1.000 1.286 7965 1.183 1.000 1.278Number of Lenders 17,087 8.715 6.000 7.486 7965 1.000 1.000 0.000Loan Size 17,087 19.257 19.209 1.211 7965 16.930 16.811 1.504LIBOR (Drawn) 14,711 157.597 125.000 115.805 3990 198.804 200.000 110.165Loan’s Maturity 16,142 3.750 4.120 0.703 6821 3.428 3.624 0.805Secured 10,481 0.655 1.000 0.475 5722 0.863 1.000 0.344Senior 17,087 0.983 1.000 0.128 7965 0.984 1.000 0.125

Lender Characteristics:Lead’s Allocation 6,064 30.113 24.000 20.893 2349 99.217 100.000 7.178Lead’s Equity Holdings 13,729 2.289 0.198 4.839 5969 0.731 0.000 2.909Participants’ Equity Holdings 13,729 5.650 0.935 10.420 0 - - -Institutional Holdings 14,378 52.349 56.573 28.105 6546 30.269 24.812 26.730Lead’s Equity Holdings are Large 13,729 0.253 0.000 0.435 5969 0.081 0.000 0.272Participants’ Equity Holdings are Large 13,729 0.177 0.000 0.382 0 - - -Institutional Holdings are Large 14,378 0.682 1.000 0.466 6546 0.346 0.000 0.476

Borrower Characteristics:Book Leverage 17,055 0.353 0.323 0.325 7939 0.291 0.248 0.295Cash Holdings 16,478 0.068 0.029 0.120 7386 0.143 0.048 0.505Return on Assets 16,472 0.586 1.014 12.031 7379 -1.679 0.709 21.783Tobin’s Q 17,086 1.608 1.308 2.311 7963 4.489 1.348 227.085Tangibility Ratio 16,996 0.365 0.311 0.249 7930 0.309 0.230 0.247Capital Expenditures 15,608 0.076 0.045 1.047 6923 0.238 0.055 9.429KZ Index 15,654 1.338 1.227 1.572 6990 1.115 1.082 2.636

37

Page 40: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Table 3: Covenants in Syndicated Loans. This table shows that covenants are more prevalent in syndicated loans. The dependent variables are: CovenantDecision in columns (1)-(2), Number of Covenants in columns (3)-(4), and Types of Covenants in columns (5)-(6). Accordingly, the regression model in columns(1)-(2) is Logit and that in the remaining columns is Ordered-Logit; all the estimates reported here are odds-ratios. Variables are defined in §3 of the paper.

Covenant Decision Number of Covenants Types of Covenants

INDEPENDENT VARIABLES (1) (2) (3) (4) (5) (6)

Syndicate 1.579*** 1.564*** 1.949*** 1.868*** 2.019*** 1.969***[5.50] [5.03] [10.90] [9.59] [11.50] [10.28]

Loan Size 0.930** 0.921** 0.872*** 0.872*** 0.869*** 0.873***[-2.32] [-2.42] [-6.51] [-6.10] [-6.60] [-5.94]

Loan’s Maturity 1.598*** 1.580*** 1.638*** 1.578*** 1.476*** 1.432***[9.99] [9.25] [14.67] [13.13] [11.54] [10.23]

LIBOR (Drawn) 0.998*** 0.999*** 1.001*** 1.001*** 1.000 1.001**[-4.57] [-3.91] [4.00] [4.59] [0.60] [1.98]

Secured 0.816** 0.849* 1.260*** 1.271*** 1.084 1.101*[-2.36] [-1.77] [4.51] [4.50] [1.45] [1.66]

Senior 6.022*** 7.012*** 4.992** 6.573** 5.253** 6.457**[3.58] [3.76] [2.02] [2.17] [2.32] [2.29]

Book Leverage 1.301 0.910 0.768**[0.91] [-0.81] [-2.54]

Cash Holdings 0.707 0.572 0.596[-1.09] [-1.08] [-0.97]

Return on Assets 1.020*** 1.033*** 1.029***[3.46] [6.76] [6.44]

Tobin’s Q 1.093** 0.990 0.993[2.48] [-0.39] [-0.29]

Tangibility Ratio 0.935 0.790* 0.896[-0.37] [-1.71] [-0.77]

Capital Expenditures 1.119 1.278*** 1.219[1.02] [2.76] [1.28]

KZ Index 0.875 0.985 0.980[-1.61] [-0.61] [-1.18]

Observations 12,626 11,488 12,626 11,488 12,626 11,488Pseudo R2 17.9% 18.2% 8.7% 8.8% 9.3% 9.5%Industry/Ratings/Year Fixed-Effects Yes Yes Yes Yes Yes Yes

z-statistics in brackets use robust standard errors clustered at the firm level; *** p<0.01, ** p<0.05, * p<0.1

38

Page 41: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Table 4: Conflict of Interests Between the Lead Arranger and Participant Banks. This table shows evidence that the conflict of interests betweenthe lead arranger and participant banks affects prevalence of covenants in syndicated loans. The dependent variables are: Covenant Decision in columns (1)-(2),Number of Covenants in columns (3)-(4), and Types of Covenants in columns (5)-(6). Accordingly, the regression model in columns (1)-(2) is Logit and that inthe remaining columns is Ordered-Logit; all the estimates reported here are odds-ratios. Note that the number of observations in columns (1) and (2) is fewerthan the later columns because in certain industries, such as “Autos” (based on the 48 Fama-French industries classification), all loans have a covenant (i.e., theindustry dummy “completely predicts success”). As a result, those observations are dropped from the analysis.

Covenant Decision Number of Covenants Types of Covenants

INDEPENDENT VARIABLES (1) (2) (3) (4) (5) (6)

Lead’s Equity Holdings 0.997 0.991 0.993[-0.25] [-1.20] [-0.98]

Participants’ Equity Holdings 0.988** 0.992** 0.991**[-2.15] [-2.06] [-2.46]

Lead’s Equity Holdings are Large 0.873 0.874** 0.902[-1.23] [-2.08] [-1.53]

Participants’ Equity Holdings are Large 0.780* 0.852* 0.777***[-1.81] [-1.95] [-3.08]

Inst. Equity Holdings (are Large) 1.003 1.116 1.002 1.064 1.002* 1.097[1.59] [1.08] [1.59] [0.96] [1.85] [1.41]

Number of Lenders 1.070*** 1.069*** 1.028*** 1.028*** 1.027*** 1.027***[8.43] [8.35] [6.55] [6.43] [6.42] [6.33]

Loan Size 0.686*** 0.691*** 0.714*** 0.717*** 0.727*** 0.730***[-6.85] [-6.68] [-9.19] [-9.08] [-8.77] [-8.69]

LIBOR (Drawn) 0.999*** 0.999*** 1.002*** 1.002*** 1.001*** 1.001***[-2.62] [-2.73] [5.32] [5.23] [2.70] [2.61]

Loan’s Maturity 1.549*** 1.551*** 1.562*** 1.562*** 1.396*** 1.394***[6.11] [6.14] [9.31] [9.30] [7.02] [6.99]

Secured 0.814* 0.814* 1.260*** 1.260*** 1.099 1.098[-1.73] [-1.72] [3.42] [3.42] [1.31] [1.30]

Senior 8.734*** 8.854*** 19.106*** 19.127*** 14.418*** 14.409***[3.09] [3.12] [3.08] [3.13] [2.81] [2.86]

Book Leverage 15.463*** 14.046*** 0.961 0.932 0.785 0.763[2.75] [2.69] [-0.22] [-0.39] [-1.35] [-1.48]

Cash Holdings 0.055*** 0.055*** 0.229*** 0.230*** 0.236*** 0.237***[-5.47] [-5.48] [-4.80] [-4.79] [-4.73] [-4.72]

Return on Assets 1.005 1.005 1.034*** 1.034*** 1.030*** 1.029***[0.38] [0.40] [4.17] [4.14] [3.56] [3.54]

Tobin’s Q 1.490*** 1.484*** 0.993 0.993 0.988 0.987[4.21] [4.22] [-0.25] [-0.26] [-0.45] [-0.47]

39

Page 42: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Tan

gibilityRatio

0.90

40.90

80.73

2*0.73

0*0.78

80.78

4[-0.41

][-0.39

][-1.83

][-1.84

][-1.34

][-1.37

]Cap

ital

Expen

ditures

0.83

00.82

71.34

51.33

80.88

00.87

7[-0.69

][-0.70

][0.99]

[0.96]

[-0.79

][-0.80

]KZIndex

0.38

2***

0.39

1***

1.00

61.01

20.96

50.97

2[-3.11

][-3.06

][0.27]

[0.46]

[-1.18

][-0.87

]

Observations

7,37

27,37

27,38

17,38

17,38

17,38

1PseudoR

220

.8%

20.8%

10.6%

10.6%

11.3%

11.3%

Industry

Dummies

Yes

Yes

Yes

Yes

Yes

Yes

Ratings

Dummies

Yes

Yes

Yes

Yes

Yes

Yes

YearDummies

Yes

Yes

Yes

Yes

Yes

Yes

z-statistics

inbracketsuse

robust

stan

darderrors

clustered

atthefirm

level;**

*p<0.01

,**

p<0.05

,*p<0.1

40

Page 43: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Table 5: Loans Issued by a Single Lender. The table looks within loans issued by single lenders and shows that the equity holdings of the lead arrangerin the borrower have no effect on the prevalence of covenants. The dependent variables are: Covenant Decision in columns (1)-(2), Number of Covenants incolumns (3)-(4), and Types of Covenants in columns (5)-(6). Accordingly, the regression model in columns (1)-(2) is Logit and that in the remaining columns isOrdered-Logit; all the estimates reported here are odds-ratios. Note that the number of observations in columns (1) and (2) is fewer than the later columns becausein certain industries, such as “Autos” (based on the 48 Fama-French industries classification), all loans have a covenant (i.e., the industry dummy “completelypredicts success”). As a result, those observations are dropped from the analysis.

Covenant Decision Number of Covenants Types of Covenants

INDEPENDENT VARIABLES (1) (2) (3) (4) (5) (6)

Lead’s Equity Holdings 0.988 0.986 0.973[-0.43] [-0.71] [-1.48]

Lead’s Equity Holdings are Large 0.801 0.743 0.724[-0.72] [-1.47] [-1.50]

Inst. Equity Holdings (are Large) 0.991*** 0.677*** 0.993*** 0.767** 0.995** 0.799**[-3.07] [-2.65] [-3.35] [-2.49] [-2.50] [-2.04]

Loan Size 1.165** 1.142* 1.108* 1.086 1.068 1.053[1.96] [1.77] [1.92] [1.61] [1.25] [1.02]

Loan’s Maturity 1.314*** 1.317*** 1.355*** 1.366*** 1.292*** 1.298***[2.73] [2.76] [4.38] [4.50] [3.64] [3.70]

LIBOR (Drawn) 1.001 1.001 1.000 1.000 1.000 1.000[0.74] [0.81] [0.77] [0.86] [-0.51] [-0.46]

Secured 1.024 1.055 1.061 1.088 1.044 1.068[0.13] [0.28] [0.47] [0.68] [0.31] [0.49]

Senior 9.084** 9.153** 3.410 3.558 3.745 3.878[2.41] [2.36] [0.89] [0.90] [0.94] [0.96]

Book Leverage 32.347 37.096 0.416* 0.455* 0.384** 0.407**[1.46] [1.51] [-1.93] [-1.75] [-2.53] [-2.40]

Cash Holdings 0.085** 0.078** 0.372** 0.349** 0.363** 0.349**[-2.27] [-2.33] [-2.25] [-2.37] [-2.48] [-2.56]

Return on Assets 1.024* 1.023 1.045*** 1.045*** 1.044*** 1.043***[1.70] [1.62] [4.94] [4.90] [5.08] [5.04]

Tobin’s Q 1.606** 1.622** 1.008 1.014 1.009 1.014[2.17] [2.20] [0.21] [0.34] [0.26] [0.37]

Tangibility Ratio 0.777 0.787 0.761 0.763 1.039 1.044[-0.64] [-0.60] [-0.98] [-0.97] [0.13] [0.15]

Capital Expenditures 0.830 0.847 0.914 0.916 1.313 1.306[-0.31] [-0.27] [-0.21] [-0.21] [0.60] [0.58]

KZ Index 0.257* 0.248* 1.081 1.060 1.104 1.089[-1.83] [-1.87] [0.66] [0.50] [1.10] [0.97]

41

Page 44: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Observations

2,08

42,08

42,11

82,11

82,11

82,11

8PseudoR

221

.8%

21.8%

8.4%

8.4%

9.3%

9.3%

Industry

Dummies

Yes

Yes

Yes

Yes

Yes

Yes

Ratings

Dummies

Yes

Yes

Yes

Yes

Yes

Yes

YearDummies

Yes

Yes

Yes

Yes

Yes

Yes

z-statistics

inbracketsuse

robust

stan

darderrors

clustered

atthefirm

level;**

*p<0.01

,**

p<0.05

,*p<0.1

42

Page 45: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Table 6: Robustness Checks. The two panels of this table check for robustness of the results presented in Table 4 above. In Panel A of this table, we limitour sample to deals where both, the lead as well as the participant banks, have positive equity holdings in the borrower. This is to ensure that the above resultsare not being driven by the observations with a zero equity stake of the lender in the borrower. In Panel B, we repeat the analysis of Table 4 to see if it isrobust to including proxies of lenders’ superior information about the borrower. These proxies are two dummy variables that equal one if the lead and/or theparticipants have been a lead-arranger for the given borrower at any time in the past three years. If the banks have had such a lending relationship with theborrower in the past, then they may have better information about the borrower than other lenders. The dependent variables are: Covenant Decision in columns(1)-(2), Number of Covenants in columns (3)-(4), and Types of Covenants in columns (5)-(6). Accordingly, the regression model in columns (1)-(2) is Logit andthat in the remaining columns is Ordered-Logit; as before, all the estimates reported here are odds-ratios. Although the control variables used in Table 4 arealso included in all the tests reported in the two panels below, their coefficients are suppressed to keep the tables concise. Note that the number of observationsin columns (1) and (2) is fewer than the later columns because in certain industries, such as “Autos” (based on the 48 Fama-French industries classification), allloans have a covenant (i.e., the industry dummy “completely predicts success”).

Panel A: When the Equity Holdings of the Lead as well as the Participants in the Borrower are Positive

Covenant Decision Number of Covenants Types of Covenants

INDEPENDENT VARIABLES (1) (2) (3) (4) (5) (6)

Lead’s Equity Holdings 1.001 0.994 0.995[0.08] [-0.71] [-0.60]

Participants’ Equity Holdings 0.984*** 0.990** 0.991**[-2.60] [-2.19] [-2.10]

Lead’s Equity Holdings are Large 0.908 0.916 0.937[-0.78] [-1.15] [-0.87]

Participants’ Equity Holdings are Large 0.730** 0.799** 0.754***[-2.11] [-2.35] [-2.97]

Control Variables Yes Yes Yes Yes Yes Yes

Observations 3,647 3,647 3,676 3,676 3,833 3,833Pseudo R2 19.4% 19.4% 12.3% 12.3% 12.5% 12.5%Industry Dummies Yes Yes Yes Yes Yes YesRatings Dummies Yes Yes Yes Yes Yes YesYear Dummies Yes Yes Yes Yes Yes Yes

z-statistics in brackets use robust standard errors clustered at the firm level; *** p<0.01, ** p<0.05, * p<0.1

43

Page 46: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Panel B: Controlling for Lenders’ Borrower-Specific Information With Their Past Lending Relationships and forLenders’ Screening/Monitoring Ability With Their Reputation

Covenant Decision Number of Covenants Types of Covenants

INDEPENDENT VARIABLES (1) (2) (3) (4) (5) (6)

Lead’s Equity Holdings 0.997 0.991 0.993[-0.25] [-1.18] [-0.96]

Participants’ Equity Holdings 0.988** 0.992** 0.991**[-2.11] [-2.15] [-2.48]

Lead’s Equity Holdings are Large 0.871 0.879** 0.905[-1.25] [-1.99] [-1.49]

Participants’ Equity Holdings are Large 0.787* 0.845** 0.776***[-1.74] [-2.05] [-3.10]

Lead-Borrower Relationship 1.030 1.030 1.117** 1.117** 1.069 1.069[0.38] [0.38] [2.43] [2.43] [1.42] [1.42]

Lead’s Reputation 1.081 1.088 0.785*** 0.789*** 0.890* 0.893*[0.87] [0.96] [-3.77] [-3.69] [-1.76] [-1.72]

Participants-Borrower Relationship 0.931 0.930 0.991 0.989 0.949 0.947[-0.57] [-0.59] [-0.12] [-0.16] [-0.71] [-0.73]

Participants’ Reputation 1.239** 1.242** 1.156** 1.158** 1.180** 1.181**[2.13] [2.15] [1.99] [2.01] [2.21] [2.22]

Control Variables Yes Yes Yes Yes Yes Yes

Observations 7,372 7,372 7,381 7,381 7,381 7,381Pseudo R2 20.9% 20.9% 10.7% 10.7% 11.3% 11.3%Industry Dummies Yes Yes Yes Yes Yes YesRatings Dummies Yes Yes Yes Yes Yes YesYear Dummies Yes Yes Yes Yes Yes Yes

z-statistics in brackets use robust standard errors clustered at the firm level; *** p<0.01, ** p<0.05, * p<0.1

44

Page 47: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Table 7: When the Conflicts of Interest are Greater. The two panels of this table test whether the effects of the participants’ equity stake presentedabove (e.g., in Table 4) are stronger when the conflicts of interest within the syndicate are greater. In Panel A of this table, we proxy the conflicts of interestwith the syndicate size – they would be greater when the syndicate is “large”. Large Syndicate is a dummy variable that equals one if the syndicate size is largerthan its mean value (which is six), and it equals zero otherwise. Small Syndicate is a complementary dummy variable that equals one if the syndicate size issmaller than its mean value, and zero otherwise. We interact these two dummy variables with Participants’ Equity Holdings (are Large) to test our claim thatits effect on the presence of covenants should be stronger when the conflicts are greater. In Panel B of this table, we proxy for “greater” conflicts of interest withthe past lending relationships of the lead and participant lenders with the borrower – the conflicts would be greater if the lead arranger has had a past lendingrelationship with the borrower but the participants have had none. Lead Has Stronger Reln. is a dummy variable equal to one if the lead arranger has had alending relationship with the borrower in the previous three years while none of the participant lenders have; it is zero otherwise. Similarly, Participants HaveStronger Reln. is a dummy variable equal to one if at least one of the participants has had a lending relationship with the borrower in the previous three yearswhile the lead arranger has not; it is zero otherwise. Again, we interact these two dummy variables with Participants’ Equity Holdings (are Large) to test ourabove claim. The dependent variables are: Covenant Decision in columns (1)-(2), Number of Covenants in columns (3)-(4), and Types of Covenants in columns(5)-(6). Accordingly, the regression model in columns (1)-(2) is Logit and that in the remaining columns is Ordered-Logit; as before, all the estimates reportedhere are odds-ratios. Although the control variables used in Table 4 are also included in all the tests reported in the two panels below, their coefficients aresuppressed to keep the tables concise. Note that the number of observations in columns (1) and (2) is fewer than the later columns because in certain industries,such as “Autos” (based on the 48 Fama-French industries classification), all loans have a covenant (i.e., the industry dummy “completely predicts success”).

Panel A: When the Syndicate is Large

Covenant Decision Number of Covenants Types of Covenants

INDEPENDENT VARIABLES (1) (2) (3) (4) (5) (6)

Lead’s Equity Holdings 0.998 0.992 0.993[-0.16] [-1.14] [-0.91]

Participants’ Equity Holdings 0.988* 0.992** 0.990**× Large Syndicate [-1.78] [-2.20] [-2.45]

Participants’ Equity Holdings 0.988 0.998 0.996× Small Syndicate [-1.23] [-0.29] [-0.52]

Lead’s Equity Holdings are Large 0.881 0.877** 0.907[-1.14] [-2.03] [-1.46]

Participants’ Equity Holdings are Large 0.778 0.830** 0.754***× Large Syndicate [-1.61] [-2.30] [-3.36]

Participants’ Equity Holdings are Large 0.833 1.024 0.943× Small Syndicate [-0.73] [0.11] [-0.30]

Large Syndicate 1.493*** 1.512*** 1.180** 1.184** 1.214*** 1.217***[3.39] [3.54] [2.36] [2.47] [2.62] [2.69]

Control Variables Yes Yes Yes Yes Yes Yes

Observations 7,372 7,372 7,381 7,381 7,381 7,381Pseudo R2 21.0% 21.0% 10.6% 10.6% 11.3% 11.3%Industry/Ratings/Year Dummies Yes Yes Yes Yes Yes Yes

z-statistics in brackets use robust standard errors clustered at the firm level; *** p<0.01, ** p<0.05, * p<0.1

45

Page 48: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Panel B: When the Lead has a Stronger Relationship with the Borrower

Covenant Decision Number of Covenants Types of Covenants

INDEPENDENT VARIABLES (1) (2) (3) (4) (5) (6)

Lead’s Equity Holdings 0.991 0.988 0.991[-0.73] [-1.61] [-1.26]

Participants’ Equity Holdings 0.989 0.994 0.990***× Lead Has Stronger Reln. [-1.62] [-1.59] [-2.62]

Participants’ Equity Holdings 1.027 1.000 1.001× Participants Have Stronger Reln. [1.53] [0.04] [0.07]

Lead’s Equity Holdings are Large 0.842 0.861** 0.883*[-1.59] [-2.36] [-1.88]

Participants’ Equity Holdings are Large 0.727* 0.825** 0.738***× Lead Has Stronger Reln. [-1.92] [-2.18] [-3.27]

Participants’ Equity Holdings are Large 2.654** 1.060 1.014× Participants Have Stronger Reln. [2.26] [0.30] [0.07]

Lead Has Stronger Reln. 1.165* 1.164* 1.139*** 1.140*** 1.134** 1.128**[1.81] [1.80] [2.61] [2.74] [2.43] [2.42]

Control Variables Yes Yes Yes Yes Yes Yes

Observations 7,372 7,372 7,381 7,381 7,381 7,381Pseudo R2 20.8% 20.9% 10.6% 10.6% 11.3% 11.3%Industry Dummies Yes Yes Yes Yes Yes YesRatings Dummies Yes Yes Yes Yes Yes YesYear Dummies Yes Yes Yes Yes Yes Yes

z-statistics in brackets use robust standard errors clustered at the firm level; *** p<0.01, ** p<0.05, * p<0.1

46

Page 49: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Table 8: Tightness of Covenants. This table tests whether, like the presence of covenants, the tightness of the covenant boundary is also affected by theconflicts of interest within the syndicate. Due to the reasons proposed in Chava and Roberts (2008), we also study the tightness of only three covenants – theCurrent Ratio covenant, the Net Worth covenant, and the Tangible Net Worth covenant in columns (1)-(3), (4)-(6), and (7)-(9), respectively. The dependentvariable, Covenant Tightness, is defined following Chava and Roberts (2008) – it is the difference between the actual accounting variable and the initial covenantthreshold, divided by the firm-specific standard deviation of the accounting variable; as such, the coefficients reported here are OLS estimates.

Current Ratio Net Worth Tangible Net Worth

INDEPENDENT VARIABLES (1) (2) (3) (4) (5) (6) (7) (8) (9)

Syndicate 0.244 -18.108 -2.940[1.43] [-1.17] [-0.77]

Lead’s Equity Holdings 0.015 -0.930 0.607[0.47] [-0.73] [0.80]

Participants’ Equity Holdings -0.004 0.381 -0.392[-0.18] [0.47] [-1.05]

Lead’s Equity Holdings are Large 0.314 -15.350 0.974[1.03] [-1.19] [0.26]

Participants’ Equity Holdings are Large -0.205 4.065 -2.909[-0.50] [0.25] [-0.59]

Inst. Equity Holdings (are Large) 0.007* 0.243 0.521 29.616* -0.017 7.897[1.71] [1.14] [1.44] [1.66] [-0.31] [0.96]

Number of Lenders 0.026 0.026 -0.380 -0.313 0.541 0.495[0.98] [0.99] [-0.43] [-0.37] [0.90] [0.86]

Loan Size 0.044 -0.162 -0.146 3.957 3.598 4.770 -0.182 -2.417 -3.651[0.48] [-0.96] [-0.86] [0.77] [0.33] [0.45] [-0.22] [-0.86] [-0.95]

LIBOR (Drawn) -0.001 -0.002 -0.002 0.005 0.052 0.042 -0.047 -0.094 -0.087[-1.48] [-1.33] [-1.38] [0.14] [0.78] [0.71] [-1.00] [-0.97] [-0.97]

Loan’s Maturity 0.293* 0.458* 0.451* 6.684 12.862 12.257 -1.608 -6.445 -5.610[1.96] [1.85] [1.81] [1.20] [1.60] [1.58] [-0.75] [-0.92] [-0.90]

Secured -0.278 -0.253 -0.242 3.746 4.093 5.601 12.184 20.018 20.360[-1.36] [-1.02] [-0.98] [0.56] [0.53] [0.67] [0.99] [0.98] [0.99]

Senior -0.416 9.952 12.472 31.545 -8.977 -15.099 -11.719[-0.84] [1.29] [0.58] [1.38] [-0.71] [-0.75] [-0.68]

Book Leverage -1.803 4.311 3.608 11.600 15.290 12.444 -11.093 -24.056 -12.061[-1.21] [1.26] [1.06] [0.72] [0.64] [0.58] [-1.15] [-1.02] [-0.66]

Cash Holdings 4.484*** 3.388 3.628 17.988 37.359 40.092 3.673 33.232 30.234[4.37] [1.45] [1.55] [1.01] [1.21] [1.30] [0.83] [1.06] [1.05]

Return on Assets 0.053*** 0.070** 0.070** 1.265 2.784 2.724 0.043 0.125 0.102[3.13] [2.06] [2.12] [1.11] [1.18] [1.19] [0.45] [0.45] [0.37]

Tobin’s Q -0.139 0.070 0.007 4.217 3.066 3.798 -0.805 -2.149 -1.675[-0.94] [0.22] [0.02] [1.39] [1.03] [1.16] [-0.81] [-0.84] [-0.75]

47

Page 50: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Tan

gibilityRatio

-1.943

***

-2.984

***

-2.930

***

13.731

27.104

27.824

-12.67

5-14.65

4-15.28

3[-4.30

][-4.68

][-4.59

][0.82]

[1.02]

[1.06]

[-0.73

][-0.75

][-0.76

]Cap

ital

Expen

ditures

-1.912

**-1.756

*-1.889

**-1.956

-11.16

4-11.84

5-0.666

5.00

77.32

9[-2.49

][-1.92

][-1.98

][-0.26

][-0.59

][-0.63

][-0.16

][0.47]

[0.58]

KZIndex

0.41

9-1.316

-1.118

-1.379

-2.041

-1.423

3.41

76.49

83.86

3[0.93]

[-1.28

][-1.08

][-0.86

][-0.81

][-0.62

][0.87]

[0.86]

[0.66]

Con

stan

t2.37

05.09

34.64

0-284

.760

-388

.914

*-417

.516

*21

.248

88.403

96.033

[1.10]

[1.46]

[1.27]

[-1.53

][-1.68

][-1.71

][0.83]

[0.89]

[0.91]

Observations

845

485

485

1,66

41,34

21,34

21,52

188

288

2R-squared

33.6%

42.0%

41.9%

21.4%

28.7%

28.7%

4.0%

6.3%

6.3%

Industry

Dummies

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Ratings

Dummies

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

YearDummies

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

t-statistics

inbracketsuse

robust

stan

darderrors

clustered

atthefirm

level;**

*p<0.01

,**

p<0.05

,*p<0.1

48

Page 51: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Table 9: Lead Arranger’s Allocation of the Loan. This table tests whether retaining a greater fraction of the loan by the lead arranger helps resolve theconflicts of interests within the syndicate, and if so, does that make the covenants less necessary in such a situation. Lead’s Allocation is the percentage of loanretained by the lead arranger, and it is available only for a limited number of loans, which explains the smaller number of observations in this table. Note thatthe number of observations in column (1) is fewer than the later columns because in certain industries, such as “Autos” (based on the 48 Fama-French industriesclassification), all loans have a covenant (i.e., the industry dummy “completely predicts success”).

Covenant Decision Number of Covenants Types of Covenants

INDEPENDENT VARIABLES Probit IV OLS IV OLS IV

Lead’s Allocation -0.004* -0.037*** -0.004*** -0.025*** -0.003*** -0.033***[-1.94] [-8.66] [-2.73] [-4.77] [-3.18] [-6.49]

Lead’s Equity Holdings 0.009 0.017** -0.001 0.007 0.001 0.010*[1.03] [2.31] [-0.20] [1.12] [0.13] [1.83]

Participants’ Equity Holdings -0.012*** -0.011*** -0.005* -0.004 -0.006** -0.006*[-2.62] [-3.08] [-1.65] [-1.22] [-2.34] [-1.95]

Inst. Equity Holdings 0.004** 0.006*** 0.001 0.002* 0.002** 0.003***[2.36] [4.95] [1.43] [1.83] [2.56] [3.16]

Number of Lenders 0.012 -0.049*** 0.007 -0.023*** 0.003 -0.040***[1.51] [-5.55] [1.34] [-2.58] [0.84] [-4.80]

Loan Size -0.197*** 0.027 -0.221*** -0.187*** -0.159*** -0.116***[-3.39] [0.60] [-5.85] [-4.59] [-5.39] [-3.22]

LIBOR (Drawn) 0.000 0.003*** 0.002*** 0.004*** 0.001*** 0.003***[0.08] [6.74] [5.09] [7.85] [3.32] [7.76]

Loan’s Maturity 0.328*** -0.006 0.282*** 0.127*** 0.187*** 0.039[5.29] [-0.13] [6.87] [2.81] [5.89] [0.98]

Secured -0.135 -0.066 0.134** 0.161** 0.001 0.020[-1.40] [-0.88] [2.31] [2.45] [0.02] [0.35]

Senior 1.374** 0.841 1.962*** 1.734** 1.419*** 1.131[2.09] [1.11] [2.78] [2.22] [2.59] [1.56]

Book Leverage 1.228 0.382 -0.091 -0.300 -0.141 -0.434***[1.60] [0.76] [-0.57] [-1.59] [-1.18] [-2.92]

Cash Holdings -1.210*** -0.556 -0.618** -0.524* -0.409* -0.307[-2.72] [-1.63] [-2.29] [-1.74] [-1.90] [-1.24]

Return on Assets -0.000 0.001 0.025*** 0.025*** 0.013** 0.015**[-0.01] [0.15] [3.53] [3.24] [2.29] [2.16]

Tobin’s Q 0.221*** 0.155*** -0.018 0.012 -0.016 -0.003[3.04] [3.04] [-0.80] [0.45] [-0.97] [-0.15]

Tangibility Ratio -0.137 -0.196* -0.245* -0.563*** -0.125 -0.498***[-0.63] [-1.77] [-1.77] [-5.49] [-1.10] [-5.62]

Capital Expenditures 0.371 -0.162 0.261 0.071 0.141 -0.162[1.10] [-0.61] [0.95] [0.24] [0.73] [-0.71]

49

Page 52: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

KZIndex

-0.455

*-0.328

**0.00

10.01

1-0.029

***

-0.029

[-1.91

][-2.12

][0.06]

[0.36]

[-2.62

][-1.44

]Con

stan

t2.27

50.73

51.95

24.40

9***

2.17

2**

4.23

6***

[1.37]

[0.61]

[1.64]

[3.94]

[2.15]

[4.09]

Observations

3,50

83,53

63,54

63,53

63,54

63,53

6Pseudo/

Adjusted

/UncenteredR

20.34

437

.0%

77.2%

36.6%

76.9%

WaldChi-sq.

270.6*

**Han

sen’s

J(p-value)

0.67

0.88

K-P

WaldF-stat

86.57

108.7

Industry

Dummies

Yes

Yes

Yes

Ratings

Dummies

Yes

Yes

Yes

Yes

Yes

Yes

YearDummies

Yes

Yes

Yes

t-/z

-statisticsin

bracketsuse

robust

stan

darderrors

clustered

atthefirm

level;**

*p<0.01

,**

p<0.05

,*p<0.1

50

Page 53: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Table 10: Summary Statistics of the Potential Syndicate Sample. This table presents summary statistics of the main variables in this sample of potentialsyndicates. As in Table 2 above, we categorize the variables into loan-, lender-, and borrower-characteristics. Here again, we limit to only those loans for whichthe deal size and number of lenders are not missing. In addition, we limit to only those potential lenders that are identifiable in the 13F filings; this ensures thatthe Potential Lender’s Equity Holdings – the key variable in our analysis – is either zero or positive but not missing altogether.

Units N Mean Median Std. Dev.

Lender Characteristics:Syndicate Member 0/1 283,016 0.082 0.000 0.274Potential Lender’s Equity Holdings % 283,016 0.505 0.112 0.917Potential Lender-Borrower Relationship 0/1 283,016 0.009 0.000 0.096Lead-Potential Interactions integer 283,016 10.218 3.000 16.145Lead’s Equity Holdings % 283,016 5.521 2.630 7.091Institutional Holdings % 283,016 64.975 66.455 18.118Potential Lender’s Equity Holdings are Large 0/1 283,016 0.255 0.000 0.436Lead’s Equity Holdings are Large 0/1 283,016 0.559 1.000 0.497Institutional Holdings are Large 0/1 283,016 0.727 1.000 0.446

Loan Characteristics:Number of Lenders integer 283,016 12.682 11.000 8.990Loan Size logarithm(dollars) 283,016 20.097 20.030 1.116LIBOR (Drawn) basis points 262,354 96.054 62.500 88.203Loan’s Maturity logarithm(1 + months) 271,873 3.530 3.906 0.745Secured 0/1 144,726 0.388 0.000 0.487Senior 0/1 283,016 0.999 1.000 0.023

Borrower Characteristics:Book Leverage ratio 282,945 0.305 0.296 0.167Cash Holdings ratio 282,489 0.060 0.030 0.084Return on Assets % 282,354 1.350 1.259 2.446Tobin’s Q ratio 283,016 1.803 1.458 1.220Tangibility Ratio ratio 281,073 0.361 0.311 0.231Capital Expenditures ratio 269,504 0.054 0.041 0.080KZ Index 270,530 1.210 1.151 0.594

51

Page 54: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Table 11: Syndicate Formation by the Lead Arranger. In this table, we test whether, in the light of the benefits of participants with an equity stake inthe borrower, the lead arranger puts together a syndicate such that it will lead to fewer frictions such as covenant restrictions. Empirically, we test whether thepool of potential participant lenders is matched by the lead arranger to a given borrower on the basis of that potential lender’s equity stake in the borrower. Thepool of potential participants consists of all those lenders that are participants on at least one deal arranged by the lead bank in the given year. The dependentvariable, Syndicate Member, is a dummy variable that equals one if the particular potential participant is a member of the actual syndicate of the given loanpackage; it is zero otherwise. The regression is estimated using a Logit model and the estimated odds-ratios are reported in this table. Given that the conflicts ofinterest within the syndicate are higher when the lead arranger has a greater equity stake in the borrower, in columns (5) and (6) of this table, we test whether theeffect documented in columns (1)-(4) is stronger when the lead arranger’s equity stake is “large”. Lead’s Equity Holdings are Large is a dummy variable definedas above, and Lead’s Equity Holdings are Small is a complementary dummy variable that equals one if the lead arranger holds an equity stake that is equal to orsmaller than 2%, and is zero otherwise.

INDEPENDENT VARIABLES (1) (2) (3) (4) (5) (6)

Potential Lender’s Equity Holdings 1.060*** 1.062***[5.03] [5.58]

Potential Lender’s Equity Holdings are Large 1.134*** 1.136***[4.99] [5.08]

Potential Lender’s Equity Holdings 1.095***× Lead’s Equity Holdings are Large [6.83]

Potential Lender’s Equity Holdings 1.022× Lead’s Equity Holdings are Small [1.24]

Potential Lender’s Equity Holdings are Large 1.279***× Lead’s Equity Holdings are Large [7.58]

Potential Lender’s Equity Holdings are Large 0.990× Lead’s Equity Holdings are Small [-0.29]

Potential Lender–Borrower Relationship 13.379*** 13.432*** 13.346*** 13.400*** 13.260*** 13.248***[34.38] [34.54] [35.14] [35.18] [35.01] [34.96]

Lead–Potential Interactions 1.039*** 1.039*** 1.039*** 1.039*** 1.039*** 1.039***[9.42] [9.40] [9.26] [9.23] [9.24] [9.22]

Lead’s Equity Holdings 0.996 0.995 0.994[-0.78] [-0.89] [-1.28]

Lead’s Equity Holdings are Large 0.896** 0.900** 0.836***[-2.14] [-2.03] [-3.50]

Inst. Equity Holdings (are Large) 0.998* 0.971 0.999 0.998 0.999 1.000[-1.68] [-0.67] [-1.08] [-0.04] [-1.12] [0.01]

Number of Lenders 1.047*** 1.047*** 1.049*** 1.049*** 1.049*** 1.049***[28.67] [28.36] [29.19] [29.22] [29.17] [29.16]

Loan Size 1.014 1.023 1.017 1.022 1.014 1.022[0.34] [0.54] [0.42] [0.55] [0.36] [0.54]

Loan’s Maturity 1.160*** 1.156*** 1.155*** 1.153*** 1.157*** 1.153***[4.81] [4.57] [4.79] [4.64] [4.82] [4.67]

LIBOR (Drawn) 0.999*** 0.999*** 0.999*** 0.999*** 0.999*** 0.999***

52

Page 55: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

[-4.82

][-4.94

][-4.55

][-4.71

][-4.52

][-4.71

]Secured

0.85

7***

0.85

1***

0.87

7**

0.87

6**

0.87

8**

0.87

6**

[-3.01

][-3.17

][-2.53

][-2.53

][-2.53

][-2.51

]Sen

ior

0.30

7***

0.29

8***

0.30

0***

0.29

1***

0.30

2***

0.28

8***

[-3.98

][-4.04

][-3.96

][-4.07

][-3.95

][-4.10

]Book

Leverag

e1.51

01.60

81.49

81.60

5[1.26]

[1.44]

[1.23]

[1.45]

CashHoldings

1.27

01.26

71.26

91.27

1[0.64]

[0.64]

[0.63]

[0.65]

Return

onAssets

0.99

80.99

80.99

80.99

8[-0.25

][-0.31

][-0.23

][-0.30

]Tob

in’s

Q1.02

81.02

91.02

71.02

9[0.79]

[0.87]

[0.76]

[0.86]

Tan

gibilityRatio

1.01

61.01

61.01

61.01

6[0.15]

[0.15]

[0.15]

[0.14]

Cap

ital

Expen

ditures

0.69

10.67

20.69

90.67

3[-1.17

][-1.25

][-1.13

][-1.24

]KZIndex

0.86

40.85

1*0.86

50.85

1*[-1.48

][-1.67

][-1.45

][-1.67

]Observations

141,79

814

1,79

813

6,05

113

6,05

113

6,05

113

6,05

1PseudoR

214

.0%

14.0%

14.1%

14.1%

14.1%

14.1%

Industry

Dummies

Yes

Yes

Yes

Yes

Yes

Yes

Ratings

Dummies

Yes

Yes

Yes

Yes

YearDummies

Yes

Yes

Yes

Yes

Yes

Yes

z-statistics

inbracketsuse

robust

stan

darderrors

clustered

atthefirm

level;**

*p<0.01

,**

p<0.05

,*p<0.1

53

Page 56: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Table 12: Controlling for Lead-Arranger’s Fixed-Effects in Syndicate Formation. In this table, we test whether the “syndicate-formation” resultsshown in Table 11 are robust to controlling for the fixed-effects of the “important” lead banks. We identify the “important” lead banks as those that have both,above-median number of potential lenders to pick from and above-median number of deals in a given year. The dependent and independent variables as well asthe regression model specifications are the same as those in Table 11. While all the variables are included in the test, the coefficients of the control variables areleft unreported in order to keep the table concise.

INDEPENDENT VARIABLES (1) (2) (3) (4) (5) (6)

Potential Lender’s Equity Holdings 1.039*** 1.042***[4.50] [5.14]

Potential Lender’s Equity Holdings are Large 1.093*** 1.098***[3.14] [3.23]

Potential Lender’s Equity Holdings 1.089***× Lead’s Equity Holdings are Large [6.77]

Potential Lender’s Equity Holdings 0.985× Lead’s Equity Holdings are Small [-0.95]

Potential Lender’s Equity Holdings are Large 1.246***× Lead’s Equity Holdings are Large [5.11]

Potential Lender’s Equity Holdings are Large 0.945× Lead’s Equity Holdings are Small [-1.54]

Potential Lender–Borrower Relationship 8.683*** 8.693*** 8.658*** 8.668*** 8.571*** 8.559***[29.35] [29.22] [29.14] [28.95] [29.07] [28.55]

Lead–Potential Interactions 1.065*** 1.065*** 1.065*** 1.065*** 1.065*** 1.065***[16.87] [16.84] [16.94] [16.93] [16.98] [16.93]

Lead’s Equity Holdings 0.996 0.997 0.995[-1.09] [-0.84] [-1.58]

Lead’s Equity Holdings are Large 0.984 0.989 0.914***[-0.62] [-0.43] [-3.79]

Inst. Equity Holdings are Large 1.000 1.056 1.001 1.066 1.001 1.068[0.44] [1.52] [0.82] [1.53] [0.75] [1.57]

Control Variables Yes Yes Yes Yes Yes Yes

Observations 138,976 138,976 133,345 133,345 133,345 133,345Pseudo R2 23.5% 23.5% 23.6% 23.6% 23.6% 23.6%Lead-Bank Dummies Yes Yes Yes Yes Yes YesIndustry Dummies Yes Yes Yes Yes Yes YesRatings Dummies Yes Yes Yes YesYear Dummies Yes Yes Yes Yes Yes Yes

z-statistics in brackets use robust standard errors clustered at the firm level; *** p<0.01, ** p<0.05, * p<0.1

54

Page 57: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Table 13: Instrument Participants’ Equity Holdings to Address Endogeneity. In this table, we address theendogeneity of the participants’ equity holdings in the borrower – a third factor could be related with both, higherparticipants’ equity holdings as well as fewer covenants in the loan. To achieve this, we instrument the Participants’Equity Holdings with: the number of potential lenders, the average equity holdings of a potential lender in theborrower, the average equity holdings of a potential lender times the size of the actual syndicate, and a dummythat equals one if the equity holdings of the potential syndicate are high or non-zero. The dependent variablesare: Covenant Decision in column (1), Number of Covenants in column (2), and Types of Covenants in column (3).Accordingly, the regression model in column (1) is an IV-Probit and that in the remaining columns is an IV-regression.Although the control variables used in Table 4 are also included in all the tests reported in the two panels below,their coefficients are suppressed to keep the table concise. Note that the number of observations in column (1) isfewer than the later columns because in certain industries, such as “Autos” (based on the 48 Fama-French industriesclassification), all loans have a covenant (i.e., the industry dummy “completely predicts success”).

Covenant Decision Number of Covenants Types of Covenants

INDEPENDENT VARIABLES (1) (2) (3)

Lead’s Equity Holdings 0.021 0.002 0.011[1.61] [0.30] [1.35]

Instrumented Participants’ Equity Holdings -0.036** -0.015** -0.023***[-2.50] [-1.97] [-2.62]

Control Variables Yes Yes Yes

Observations 7,210 7,218 7,218Adjusted R2 27.7% 24.1%Wald Chi-sq. 4.5**Hansen’s J (p-value) 0.13 0.24K-P Wald F-stat 55.04 60.93Industry Dummies Yes Yes YesRatings Dummies Yes Yes YesYear Dummies Yes Yes Yes

z-statistics in brackets use robust standard errors clustered at the firm level; *** p<0.01, ** p<0.05, * p<0.1

55

Page 58: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Table 14: Sweeps and Other Restrictions. This table presents a list of sweeps and other restrictions that are found in our sample of loans made tonon-financial firms belonging to the intersection of Compustat and Dealscan during the 1994-2009 period. Summary statistics of loans including such sweeps andrestrictions are reported alongside.

Sweeps and Number of Average Median Aggregate SizeOther Restrictions Packages Deal Size ($m) Deal Size ($m) of Deals ($bn)

Dividend Restrictions 11,795 326.95 120.00 3,856.35Asset Sales Sweep 4,817 473.94 180.00 2,282.98Debt Issuance Sweep 3,451 571.04 220.00 1,970.66Equity Issuance Sweep 3,261 450.13 175.00 1,467.88Insurance Proceeds Sweep 2,696 497.02 200.00 1,339.97Percent of Cashflows Sweep 2,094 512.04 220.00 1,072.21Percentage of Net Income Paid as Dividends 226 293.34 148.57 66.30Percentage of Cashflows Paid as Dividends 28 518.15 300.00 14.51

56

Page 59: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Table 15: Sweeps and Other Restrictions. In this table, instead of covenants, we analyze other types of restrictions, such as sweeps, etc. We ask whether,like covenants, sweeps help resolve the conflicts of interest within the syndicate, and therefore, are more likely to be used in syndicates and less likely to benecessary when the participants have an equity stake in the borrower. This is similar to the analysis conducted for covenants in Tables 3 and 4; the results fromthe analysis of sweeps are presented in Panels A and B, respectively, of this table. Next, we instrument the Participants’ Equity Holdings, as was done in Table13 for covenants; results with Instrumented Participants’ Equity Holdings for sweeps are presented in Panel C of this table. The dependent variables are eitherdummy variables that mark the presence of a sweep or they are ordinal variables that count the number of sweeps. As shown in Table 14, we find eight differenttypes of sweeps and other restrictions, and Dividend Restrictions is the most common of them all. Therefore, we additionally split all the sweeps into two groups –dividend restrictions and all others. As such, the dummy dependent variable Sweeps Decision marks the presence of either all sweeps, or only dividend restriction,or other sweeps excluding the dividend restriction; the respective column headings (All, Dividend, and Others) indicate the sweeps used in defining the dummydependent variable. The ordinal dependent variable, Number of Sweeps, is also split accordingly – either counting all the sweeps or counting all sweeps except thedividend restriction. Accordingly, the regression model in Panels A and B is Logit for Sweeps Decision and Ordered-Logit for Number of Sweeps; as before, allthe estimates reported here are odds-ratios. In Panel C, with Instrumented Participants’ Equity Holdings, the models used are IV-Probit for Sweeps Decision andIV-regression for Number of Sweeps. Although the control variables used in Table 4 are also included in all the tests reported in the various panels below, theircoefficients are suppressed to keep the tables concise. Note that the number of observations in columns using a dummy dependent variable are fewer than thelater columns because in certain industries, such as “Autos” (based on the 48 Fama-French industries classification), all loans have a covenant (i.e., the industrydummy “completely predicts success”).

Panel A: Sweeps in Syndicated Loans

Sweeps Decision Number of Sweeps

INDEPENDENT VARIABLES All Dividend Others All Others

Syndicate 2.327*** 2.388*** 1.799*** 2.006*** 1.799***[11.42] [12.05] [7.41] [11.37] [7.96]

Control Variables Yes Yes Yes Yes Yes

Observations 11,488 11,488 11,488 11,488 11,488Pseudo R2 13.8% 13.6% 23.3% 11.7% 14.8%Industry Dummies Yes Yes Yes Yes YesRatings Dummies Yes Yes Yes Yes YesYear Dummies Yes Yes Yes Yes Yes

z-statistics in brackets use robust standard errors clustered at the firm level*** p<0.01, ** p<0.05, * p<0.1

57

Page 60: Syndicated Loans: The Role of Covenants in Mitigating ... · Syndicated Loans: The Role of Covenants in Mitigating Lender ... presence of at least one sweep restriction are more than

Panel B: Conflict of Interests Between the Lead Arranger and Participant Banks

Sweeps Decision Number of Sweeps

INDEPENDENT VARIABLES All Dividend Others All Dividend Others All Others All Others

Lead’s Equity Holdings 0.970*** 0.974*** 0.970*** 0.972*** 0.969***[-3.03] [-2.64] [-3.03] [-3.14] [-3.10]

Participants’ Equity Holdings 0.988** 0.989** 0.994 0.990** 0.996[-2.14] [-2.07] [-0.97] [-2.04] [-0.76]

Lead’s Equity Holdings are Large 0.696*** 0.719*** 0.774*** 0.745*** 0.773***[-4.10] [-3.86] [-2.83] [-4.11] [-3.03]

Participants’ Equity Holdings are Large 0.609*** 0.667*** 0.859 0.733*** 0.902[-4.33] [-3.57] [-1.18] [-2.97] [-0.86]

Control Variables Yes Yes Yes Yes Yes Yes Yes Yes Yes YesObservations 7,350 7,373 7,381 7,350 7,373 7,381 7,381 7,381 7,381 7,381Pseudo R2 15.7% 16.2% 23.5% 16.0% 16.4% 23.4% 12.1% 14.8% 12.1% 14.7%Industry Dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes YesRatings Dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes YesYear Dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

z-statistics in brackets use robust standard errors clustered at the firm level; *** p<0.01, ** p<0.05, * p<0.1

Panel C: Instrument Participants’ Equity Holdings to Address Endogeneity

Sweeps Decision Number of Sweeps

INDEPENDENT VARIABLES All Dividend Others All Others

Lead’s Equity Holdings 0.009 0.011 -0.004 -0.006 -0.000[0.73] [0.91] [-0.48] [-0.58] [-0.05]

Instrumented Participants’ Equity Holdings -0.040*** -0.040*** -0.019** -0.023** -0.021**[-3.26] [-3.36] [-2.08] [-2.19] [-2.16]

Control Variables Yes Yes Yes Yes Yes

Observations 7,188 7,210 7,381 7,218 7,218Wald Chi-sq. 7.6*** 8.5*** 3.4*Adjusted R2 33.6% 32.3%Hansen’s J (p-value) 0.65 0.42K-P Wald F-stat 55.96 55.96Industry/Ratings/Year Dummies Yes Yes Yes Yes Yes

z-statistics in brackets use robust standard errors clustered at the firm level; *** p<0.01, ** p<0.05, * p<0.1

58


Recommended