Gross Loan Flows, Financial Crises, and Banking Sector Reforms: Evidence from Korea

• DOI: http://doi.org/10.1111/ajfs.12149
• Author: Junghwan Hyun
• Date: 01 October 2016
• Published date: 01 October 2016

Gross Loan Flows, Financial Crises, and

Banking Sector Reforms: Evidence from

Korea*

Junghwan Hyun**

Hiroshima University

Received 27 December 2015; Accepted 25 July 2016

Abstract

This paper explores gross loan flows in Korea from 1984 to 2014 focusing on how gross loan

flows differ before and after the 1997 global financial crisis and how they respond to the cri-

sis and the associated banking sector reforms. The paper shows that sizable loan expansion

and contraction, measured as deviations from the industry loan growth trend, coexist over

time, indicating intense heterogeneity in bank lending. Also, the crisis and the reforms and

restructuring of the banking sector are largely attributable to changes in gross loan flows in

the post-crisis period. Particularly, banks’ asset portfolio reallocation occurs from the com-

mercial sector to individual customers after the crisis.

Keywords Gross loan flows; Korean banking sector; Financial crisis; Banking sector reforms

JEL Classification: E44, E51, G21

1. Introduction

Changes in aggregate loans are the result of two intrinsically different bank activi-

ties: expansion, the issuing of new loans; and cancellation, the removal of expired

loans (Dell’Ariccia and Garibaldi, 2005).

1

Therefore, decomposing aggregate net

*I am grateful to the Editor (Kee-Hong Bae) and two anonymous referees for insightful com-

ments. Other useful comments come from Hye-Won Kim, Ji-Eun Lee, Susumu Morimoto,

Ha-Chin Yi, David Vera, Eliza Wu, and several seminar participants at the 90th Western Eco-

nomics Association International (WEAI) Annual Conference and the 10th Annual Confer-

ence on Asia-Pacific Financial Markets (CAFM) of the Korean Securities Association (KSA). I

also thank Min-Je Jeon for his help in collecting data. All remaining errors are mine.

**Corresponding author: Junghwan Hyun, IDEC, Hiroshima University, Kagamiyama 1-5-1,

Higashi-Hiroshima, Japan. Tel: +81-82-424-3724, email: momos039@hiroshima-u.ac.jp.

1

When a bank extends loans, due to search frictions in loan markets, it is time-consuming

and involves costs for screening and evaluating investment projects. In addition, the lender

takes credit risks with new borrowers. When a lender contracts loans, it involves information

loss, caused by the end of the relationship with the borrowers, as well as the costs of recover-

ing some loan amounts, which may be not trivial, especially during recessions.

Asia-Pacific Journal of Financial Studies (2016) 45, 705–728 doi:10.1111/ajfs.12149

©2016 Korean Securities Association 705

change into two gross loan flows —loan expansion and loan contraction —pro-

vides a better understanding of heterogeneous bank lending at the micro level and

its interaction with the macroeconomy. Indeed, since the recent global financial cri-

sis (GFC) that has placed bank lending behaviors in the spotlight, several studies

approach banks’ lending behaviors using gross loan flows, presenting evidence that

gross loan flows capture a better picture of the bank credit market than aggregate

bank lending or the interest rate spread, because the latter cannot reflect heteroge-

neous lending behavior at the micro level (Chari et al., 2008; Cohen-Cole et al.,

2008).

2

Dell’Ariccia and Garibaldi (2005) document the stylized facts of gross loan

flows from 1979:Q2 to 1999:Q2 using United States commercial bank data. Craig

and Haubrich (2013) also explore the properties of gross loan flows in the United

States from 1960 to 2004.

This paper explores gross loan flows in South Korea from 1984 to 2014, using

the methodology proposed by Dell’Ariccia and Garibaldi (2005), to investigate how

gross loan flows differ before and after the 1997 GFC and how they respond to the

crisis and the associated banking sector reforms. The history of the South Korean

banking sector during the sample period provides a suitable environment in whi ch

to explore how gross loan flows change in response to financial crises and institu-

tional changes. To be specific, prior to the 1997 GFC, the loan market was imma-

ture; as a result, search friction might be nontrivial between lenders and borrowers,

which may result in costly loan expansion and inefficient loan reallocation (Cho,

1988; Hwang, 2012). Or given that the government heavily intervened in individual

banks’ loan market decisions to support its industrial policy, lenders’ incentive to

monitor and discipline borrowers might be marginal (Smith, 2000; Banker et al.,

2010). If so, in the pre-crisis era, it is possible that loan expansion and contraction

did not follow the usual matching process based on market-oriented lending deci-

sions and that they barely respond to the business cycle. In this case, loan realloca-

tion, whether inefficient or not, is actually driven by government policy. In either

of these two cases, the intensity of gross loan flows is not predictable. Furthermore,

during the banking sector reforms subsequent to the 1997 GFC, insolvent lenders

were bankrupt and liquidated or acquired and merged and, as a result, net loan

aggregates rapidly shrank to a substantial extent. Although prior studies (e.g. Park

and Weber, 2006; Banker et al., 2010; Hall and Simper, 2013) show that banks’ effi-

ciency improved after the Korean GFC and that the loan decision-making process

changed due to the regulatory reforms and the restructuring of the banking sectors,

they do not consider loan flows’ reaction to the reforms. The reforms might inten-

sify loan reallocation, given that banks restructured their loan portfolios during the

period. Alternatively, if banks show herd behavior in lending, the magnitude of loan

2

Chari et al. (2008), Craig and Haubrich (2013), and Peek and Rosengren (2013) point out

that aggregate data, such as the loan supply, obscure the heterogeneous lending behaviors of

banks and so provide incomplete information on loan markets. Note that the bank loan sup-

ply is decomposed into loan expansion and contraction in this paper.

J. Hyun

706 ©2016 Korean Securities Association