Bank soundness and bank lending to new firms during the global financial crisis

• Date: 01 July 2020
• DOI: http://doi.org/10.1002/rfe.1090
• Author: Eriko Naiki,Yuta Ogane
• Published date: 01 July 2020

Rev Financ Econ. 2020;38:513–541.

|

513

wileyonlinelibrary.com/journal/rfe

1

|

INTRODUCTION

The aim of this study is to investigate how the soundness of financial institutions affected bank lending to new firms during the

2008 financial crisis. Numerous previous studies have examined the effects of the soundness of financial institutions on lending

to firms and found that such soundness significantly affects lending to these firms. Specifically, several studies have shown that

sounder financial institutions are more likely to provide financing to firms (e.g., Cornett, McNutt, Strahan, & Tehranian, 2011;

Gambacorta & Mistrulli, 2004), while other studies have observed the opposite results (e.g., Diamond & Rajan, 2000; Diamond

& Rajan, 2001; Gorton & Winton, 2017). In addition, recent studies have investigated such effects on small and medium‐sized

enterprises (SMEs) while focusing on the 2008 financial crisis (e.g., Iyer, Peydró, da‐Rocha‐Lopes, & Schoar, 2013; Popov &

Udell, 2012).1

These studies have explained why financial institutions provide financing to these firms while using the risk‐tak-

ing theory and have drawn various implications for facilitating financing to these firms.

However, no study has empirically investigated the effects of the soundness of financial institutions on bank lending to firms

that have never had lending relationships with financial institutions, such as new firms. Previous studies have emphasized the

Received: 1 July 2019

|

Revised: 1 September 2019

|

Accepted: 2 October 2019

DOI: 10.1002/rfe.1090

ORIGINAL ARTICLE

Bank soundness and bank lending to new firms during the global

financial crisis

ErikoNaiki

|

YutaOgane

This is an open access article under the terms of the Creative Commons Attribution‐NonCommercial‐NoDerivs License, which permits use and distribution in any medium,

provided the original work is properly cited, the use is non‐commercial and no modifications or adaptations are made.

© 2019 The Authors. Review of Financial Economics published by Wiley Periodicals, Inc. on behalf of University of New Orleans

Faculty of Policy Studies,Aichi Gakuin

University, Nisshin, Japan

Correspondence

Yuta Ogane, Department of Economics,

Nanzan University, Nagoya, Aichi, Japan.

Email: ogane@nanzan-u.ac.jp

Funding information

Japan Society for the Promotion of

Science, Grant/Award Number: 17H07226,

19K13746, and 19K13748

Abstract

This paper examines how the soundness of financial institutions affected bank lend-

ing to new firms during the 2008 financial crisis by using a unique firm–bank match‐

level dataset of 1,467 unlisted small and medium‐sized enterprises incorporated in

Japan. We employ a within‐firm estimator that can control for unobserved firms’

demand for credit through firm ∗ time fixed effects. The major findings of this paper

are the following four points. First, sounder financial institutions may be generally

less likely to provide financing to new firms. Second, our results suggest that sounder

financial institutions were less likely to provide loans to new firms during the 2008

financial crisis. Third, financial institutions were less likely to provide financing to

new firms during such crisis as compared to those with the same soundness during

non‐crisis periods. Finally, such lending relationships to new firms that are estab-

lished during the financial crisis by sounder financial institutions are more likely to

be continued than such lending by less sound financial institutions.

KEYWORDS

bank lending, bank soundness, financial crisis, new firms

JEL CLASSIFICATION

G01; G21; L26; M13

514

|

NAIKI ANd OGANE

importance of focusing on new firms (e.g., Berger & Udell, 1998; Robb & Robinson, 2014) and other studies have suggested

that such effects on new firms differ from those on other firms (e.g., Iyer et al., 2013).2

Moreover, focusing on new firms is

likely to enhance previous studies that have examined how lending relationships between firms and financial institutions affect

sources of financing during times of economic distress (e.g., Carbó‐Valverde, Rodríguez‐Fernández, & Udell, 2016; Cotugno,

Monferrà, & Sampagnaro, 2013; Dewally & Shao, 2014; Jiangli, Unal, & Yom, 2008). These studies have focused on existing

firm–bank relationships continued before the 2008 financial crisis. Furthermore, Tsuruta (2016) investigates the effects of the

financial crisis on not only such relationship but also on small businesses without lending relationships.3

In contrast, these ef-

fects on firms without ever having lending relationships with financial institutions have not revealed in spite of its importance.

The reasons prior studies have not examined such effects are technical problems and data limitations. More specifically,

bank lending is simultaneously determined by firms’ demand for credit and financial institutions’ supply of credit, making it

difficult to disentangle demand and supply (i.e., identification problem). Although several recent studies have addressed this

problem (e.g., Jiménez, Ongena, Peydró, & Saurina, 2014; Khwaja & Mian, 2008), no study has focused on firms without

ever having lending relationships with financial institutions because these firms are mainly new firms. Even at present, data

on the lending relationship between new firms and financial institutions are difficult to obtain (i.e., data limitations). For these

reasons, it is difficult to investigate the effects of the soundness of financial institutions on bank lending to new firms while

addressing the identification problem and data limitations.

Against this background, this study is the first to examine the effects of the soundness of financial institutions on bank lend-

ing to new firms while focusing on such effects during the 2008 financial crisis. Investigating such effects should be intriguing

because several studies have suggested that lending activity among each financial institution differs greatly when the borrowers

are new firms (e.g., Furlong & Keeley, 1989) and during this financial crisis (e.g., Claessens & Van Horen, 2014; de Haas

& Lelyveld, 2014; Ongena, Peydró, & Van Horen, 2015), as compared to other firms and other periods. A unique firm–bank

match‐level dataset in Japan enables us to solve the two aforementioned problems (i.e., identification problem, data limitations)

and conduct empirical tests. This dataset includes information on 1,467 unlisted SMEs in the early stages of the entrepreneurial

process, which rely heavily on bank lending, as well as data on their correspondent financial institutions. The benefits of using

this Japanese dataset are threefold.

First, the dataset enables us to construct a three‐way panel dataset by firm, year, and lender (financial institution). To the

best of our knowledge, due to data limitations, no study has employed this panel dataset while focusing on new firms. Second,

using our Japanese dataset may make it possible to more clearly grasp the differences in the lending activities of financial in-

stitutions toward new firms, which are the most difficult borrowers to screen for creditworthiness. This is because, in Japan,

lenders other than financial institutions such as investment funds, venture capitals, and angel investors are less likely to provide

financing to these firms as compared to in other developed countries. Finally, among G7 countries, the 2008 financial crisis

had especially negative effects in Japan. In this respect, as mentioned in Tsuruta (2019), the falling rate of GDP in Japan was

much larger among comparable developed countries; thus, using this Japanese dataset may reveal our objective effects more

remarkably than data from other countries or may have the potential for revealing unknown effects that cannot be revealed with

other countries’ datasets.

The major findings of this study are the following four points. First, sounder financial institutions may be less likely to pro-

vide financing to new firms. However, this evidence is not necessarily strong. Second, our results suggest that sounder financial

institutions were less likely to provide loans to new firms during the 2008 financial crisis. Although this result is inconsistent

with the predictions of previous studies, this may be because we use new firms as a sample, whereas previous studies employ

older firms. Third, sounder financial institutions were less likely to provide financing to new firms during the 2008 financial

crisis compared to financial institutions with the same soundness during non‐crisis periods. Specifically, the following two

types of financial institutions are less likely to lend to new firms: (1) financial institutions with high capital adequacy ratios (for

those that do not have overseas business locations); (2) financial institutions with low non‐performing loan ratios (for those that

have overseas business locations). Finally, lending by sounder financial institutions to new firms that are established during

the financial crisis is more likely to be continued as compared to lending by less sound financial institutions to such firms. In

particular, lending by sounder financial institutions that do not have overseas business locations with low non‐performing loan

ratios tends to be continued.

The most significant contribution of this study is that we examine the effects of the soundness of financial institutions on

bank lending to firms while simultaneously considering the following three points: (1) focusing on new firms; (2) employing

bank lending as an indicator of firms’ credit availability; (3) controlling for unobservable firm characteristics by employing a

within‐firm estimator. The third point is used in Gan (2007) and subsequent studies such as Khwaja and Mian (2008), Jiménez

et al. (2014), and Ogura (2018). Our dataset and this powerful estimation method enable us to examine the above effects while

sufficiently controlling for all characteristics of the firms, including not only unobservable time‐invariant characteristics but