The syndicate structure of securitized corporate loans

• Author: Shage Zhang,Zhengfeng Guo
• Published date: 01 February 2020
• Date: 01 February 2020
• DOI: http://doi.org/10.1111/fire.12203

DOI: 10.1111/fire.12203

ORIGINAL ARTICLE

The syndicate structure of securitized

corporate loans

Zhengfeng Guo1Shage Zhang2

1WashingtonDC, USA

2School of Business, TrinityUniversity,

San Antonio, Texas

Correspondence

ShageZhang, School of Business, Trinity

University,One TrinityPlace, San Antonio, TX

78212.

Email:szhang@trinity.edu

Abstract

Securitized loans havelower lead bank shares, but larger shares held

by non-CLO (collateralized loan obligation) institutional investors

than nonsecuritized loans. The result can largely be explained by

their degree of information asymmetry and credit risk. We find

that lead banks increase their holdings after a nonsecuritized loan

becomes securitized, but they do not reduce financial exposure to

securitized facilities during the boom of the CLO market. Further-

more, we find that securitized loans do not perform differently from

similar nonsecuritized loans. We conclude that differences in syndi-

cate structure are likely shaped by participants’ investment prefer-

ence rather than a manifestation of adverse selection.

KEYWORDS

information asymmetry, lead bank share, securitization, syndicate

structure

JEL CLASSIFICATIONS

G21, G23, G32

1INTRODUCTION

Studies on the securitization of subprime mortgage loans document extensive evidence that shows securitized loans

perform significantly worse than nonsecuritized loans (e.g., Demyanyk & VanHemert, 2011; Keys, Mukherjee, Seru, &

Vi, 2010; Mian & Sufi, 2009; Nadauld & Sherlund, 2013; Purnanandam, 2011). The main argument is that securitiza-

tion allows originating banks to transfer credit risk to other market participants and retain minimum to zero financial

stake in the loans. Hence, their incentives for ex ante due diligence and ex post monitoring are considerably weak-

ened (Gorton & Pennacchi, 1995; Pennacchi, 1988). However,the literature on corporate loan securitization shows

mixed evidence. Wang and Xia (2014) document that securitization-active banks have weaker ex post incentives to

monitor borrowers. Bord and Santos (2015) find that securitized facilities have a higher probability of default because

of weak lead bank incentives. In contrast, Benmelech, Dlugosz, and Ivashina (2012) find no evidence that securitized

facilities systematically underperform nonsecuritized facilities using multiple performance measures. They conclude

Financial Review.2020;55:61–89. wileyonlinelibrary.com/journal/fire c

2019 The Eastern Finance Association 61

62 GUO ANDZHANG

that adverse selection is not a severe issue in corporate loan securitization. Shivdashani and Wang (2011) find that

leveraged buyout (LBO) deals financed by securitized loans do not underperform other LBO deals, suggesting that

securitization does not affect banks’ lending standards.

The key difference between the securitization of subprimemortgages and that of corporate loans is the loan struc-

ture. Subprime mortgages are financed by a single lender and are often sold in one piece to mortgage-backed security

issuers, so lenders retain no financial stake in the securitized mortgages. In contrast, securitized corporate loans are

syndicated loans jointly financed by lead banks, participatingbanks, and nonbank participants. They are often only par-

tially securitized. Benmelech et al. (2012) argue that syndicate structure is likelyan important mechanism that reduces

adverse selection in syndicated loan securitization. However,finance researchers pay little attention to the relation-

ship between securitization and syndicate structure, mainly due to a lack of detailed data. In this study, we attempt

to fill this gap and shed new light on the relationship between syndicate structure, securitization, and ex post loan

performance.

Prior studies tend to interpret lower lead bank shares in securitized loans as evidence that originating banks retain

little “skin in the game” and, hence, haveweak screening and monitoring incentives (Bord & Santos, 2015). What is lack-

ing from this argument is that securitized loans may be fundamentally different from nonsecuritized loans, and shares

retained by lead banks in syndicated loans are likely to be affected by such differences. One possibility is that securi-

tized loans mayhave certain characteristics that allow participants to have easier access to information and assess loan

credit risk in a more timely and accurate manner.Such characteristics may lower the need for intensive lead bank mon-

itoring and the shares retained by lead banks. In this circumstance, lower lead bank share is not necessarily associated

with weak incentives and subsequent poor loan performance.

Relying on a comprehensive, unbalanced panel data set with 4,199 securitized and 10,171 nonsecuritized facilities

originated from 1999 to 2010, we first document the significant differences between the two types of loans in terms

of the degree of information asymmetry and risk level. We measure information asymmetry from two aspects. The

first set of variables measure the overall information transparency of a borrower,including whether the borrower is

a public firm, whether it has a bond rating, the age of the firm, and the research and development (R&D) activities of

the borrower.The idea is that the information transparency of a borrower would reduce the overall information asym-

metry between the borrower and external lenders (Maskara &Mullineaux, 2011). The second set of variables capture

the lending history of a borrower in the syndicated loan market.Repeated syndicate lending relationships between the

borrower and the lead arranger reduce the information asymmetry between the two. Sufi (2007) indicates that a bor-

rower’s repeated appearancein a lending market reveals more information on its creditworthiness to potential lenders

in that market. Hence, we look at a borrower’s history in issuing securitized loans in the recent past. We also use the

geographic distance between the lead bank and the borrower to capture the importance of soft information needed to

be collected in a lending relationship (Petersen & Rajan, 2002). More soft information is associated with higher infor-

mation asymmetry. The credit risk of a loan is measured by the credit quality evaluation ratings in Shared National

Credit (SNC) database and the credit ratings provided byStandard & Poor’s (S&P). We also include a borrower’s stock

return volatility and Altman’sZ-score to capture its overall riskiness and the likelihood of bankruptcy.

We find that securitized loan borrowers are more likely to be public firms and less likely to be nonrated, relative

to nonsecuritized loan borrowers. Securitized loan borrowers also tend to locate further away from the lead bank,

which indicates that less soft information needs to be collected about the borrower in the loan issuance process. In

multivariate regressions, we find that these information asymmetry measures, including a borrower’s previous lending

relationships with the lead bank and its track record in the securitized loan market, are significantly correlated with

the likelihood of a loan being securitized. Conditional on a facility being securitized, the proportion of it being securi-

tized is significantly and negatively associated with the degree of information asymmetry.Such results are consistent

with the argument in DeMarzo and Deffie (1999) that more transparent loans get securitized to preventthe potential

“lemons problem” in equilibrium. In addition, we find that riskier loans are more likely to be securitized byoriginating

banks, consistent with the “risk sharing” argument: originating banks transfer part of the credit risk to other financial