Why do banks join loan syndications? The case of participant banks

Abstract

This paper examines the determinants of banks' involvement in loan syndication using the financial information of 847 participant banks. The results indicate that participant banks join loan syndications when their capital levels are sufficient enough to support the extra risk taken. Banks with lower net interest margin are found to choose syndicated lending as a way of boosting their margins. The motivation of risk diversification through participating in loan syndications is also confirmed.

Keywords:

• banks
• syndicated loans
• capital adequacy
• risk diversification
• participant banks
• arranger banks

Notes

Previous studies have also focused on the role of the arranger, their reputation, and possible asymmetric information problems (such as Lee & Mullinaeux 2001; and Panyagometh & Roberts, 2002) and the pricing of syndicated loans in relation to borrower financial characteristics and loan terms (Angbazo, Mei, &Saunders, 1998; Coleman, Esho, & Sharpe, 2006; Hubbard, Kuttner, & Palia, 2002; Yi & Mullineaux, 2005). Additionally, several studies (Billet, Flannery, & Garfinkel, 1995; Lummer & McConnel, 1989; Preece & Mullineaux, 1996) have investigated the signalling effect of borrowing through syndicated lending markets.

For example, Thomson Financial (2007) reports that lead banks earned fees worth of 12.9 billion USD in 2007.

Dennis and Mullineaux (2000) and Panyagometh and Roberts (2002) identify empirically the factors that influence an arranger bank's decision to syndicate a loan and the determinants of the proportion of the loan sold in the event of syndication. They utilise loan characteristics, arranger's financial characteristics, variables related to the opacity of the borrower and information asymmetries as explanatory variables. Altunbaş et al. (2005) employ arranger banks financial characteristics as the main determinants of arranger banks decision to syndicate loans. In contrast, Pavel and Phillis (1987), Simons (1993), and Pennacchi (1988) use share of the loans retained (or syndicated) by the arranger banks. Due to unavailability of data on participant banks share of the syndicated loan, we are unable to test our arguments on the amount of loans retained by the participant banks.

Ideally, one would like to have information from the bank's monthly internal reporting system. However, given the yearly frequency of publicly available balance sheet and profit/loss statement information, values at the end of the fiscal year preceding the loan participation decision are the most accurate measure one can use to gauge the interaction between bank characteristics and lending decisions.

Ideally, one would like to include more recent data for the analysis. Original data were collected during a funded research project, subsequently, due to financial constraints, we were unable to update the database beyond 2001.

The study restricts the analysis to institutions from industrialised countries for the sake of homogeneity and also because participation in loan syndications by banks from emerging countries represents a very small share of the total sample.

Loanware database provides shares of each participant banks only for some deals.

Income earned through means other than the core business of deposit collecting and loan making. Non-interest income is comprised of fee income from intermediary services charges such as deposit, chequing, loan arrangements, credit card, electronic funds transfer, trust and fund management, global custody services, securities brokerage, underwriting, real estate services, insurance activities, loan commitments, letter of credit and consulting.

The Hausman test is employed to select between fixed effect and random effect estimations. Results of Hausman test and random effect results are presented in Tables 3–5.

A correlation matrix of independent variables is presented in Tables 3–5.