Banking Market Structure and Macroeconomic Stability: Are Low‐Income Countries Special?

• Author: Franziska Bremus,Claudia M. Buch
• DOI: http://doi.org/10.1111/1468-0106.12095
• Published date: 01 February 2015
• Date: 01 February 2015

BANKING MARKET STRUCTURE AND

MACROECONOMIC STABILITY: ARE LOW-INCOME

COUNTRIES SPECIAL?

FRANZISKA BREMUS*DIW Berlin

CLAUDIA M. BUCH Deutsche Bundesbank

Abstract. Does the structure of banking markets affect macroeconomic volatility and, if yes, is this

link different in low-income countries? In this paper, we explore the channels through which the

structure of banking markets affects macroeconomic volatility. Our research has three main find-

ings. First, we study whether idiosyncratic volatility at the bank level can impact aggregate volatility.

We find weak evidence for a link between granular banking sector volatility and macroeconomic

fluctuations. Second, a higher share of domestic credit to GDP coincides with higher volatility in the

short run. Third, a higher level of cross-border asset holdings increases volatility in low-income

countries.

1. MOTIVATION

Negative effects of macroeconomic volatility on growth and welfare can be

particularly pronounced in low-income countries (Pallage and Robe, 2003;

Loayza et al., 2007; Calderon and Yeyati, 2009). In addition, low-income coun-

tries often experience excess macroeconomic volatility compared to advanced

economies. Shocks are more frequent and larger. Moreover, structural charac-

teristics of low-income economies like a low degree of diversification can amplify

the effect of shocks (Acemoglu and Zilibotti, 1997; Koren and Tenreyro, 2007,

2013). Given that real and financial cycles are closely related (Claessens et al.,

2011, 2012), an additional reason for differences between high-income and

low-income countries in terms of macroeconomic stability could be differences

in the structure of banking systems. The banking systems in low-income coun-

tries in fact differ in numerous aspects from those in higher-income economies.

Banking systems in low-income countries are typically smaller and less open

than those in developed economies. Therefore, access to finance is more limited.

In this paper, we explore the channels through which the structure of banking

markets affects macroeconomic instability as measured by the volatility of GDP

*Address for Correspondence: Franziska Bremus, German Institute for Economic Research (DIW

Berlin), Mohrenstraße 58, 10117 Berlin, Germany. E-mail: fbremus@diw.de. This paper is part of a

research project on macroeconomic policy in low-income countries supported by the UK’s Depart-

ment for International Development (DFID). Franziska Bremus acknowledges funding from the

DFID. The paper was presented at the Conference on ‘Macroeconomic Challenges Facing Low-

Income Countries: New Perspectives’ (Washington, DC, 30–31 January 2014). The views expressed

herein are those of the authors and should not be attributed to the IMF, its Executive Board, or its

management, to DFID, or to the Deutsche Bundesbank. We thank two anonymous referees, César

Calderón, Atilim Seymen and conference participants for helpful comments and suggestions. Hanna

Schwank provided very valuable research assistance. All remaining errors or inconsistencies are our

own.

bs_bs_banner

Pacific Economic Review, 20: 1 (2015) pp. 73–100

doi: 10.1111/1468-0106.12095

© 2015 Wiley Publishing Asia Pty Ltd

per capita.1We particularly focus on low-income countries. We use a linked

micro–macro panel-dataset including low-income, middle-income and high-

income countries. Bank-level data are taken from Bankscope. We investigate the

impact of three structural characteristics of banking systems. First, we link

annual GDP volatility to microeconomic risk at the bank-level by drawing on

the concept of granularity (Gabaix, 2011). Second, we evaluate how the size of

the banking sector impacts macroeconomic instability. Third, we analyse the

effect of the degree of financial openness on aggregate volatility.

Recent research shows how heterogeneous size distributions of firms can

affect macroeconomic volatility (Gabaix, 2011). If firm sizes follow a fat-tailed

power law distribution so that market concentration is high, shocks to large

firms (or banks) do not cancel out across a large number of firms as they would

under normally distributed firm sizes. In this case, macroeconomic volatility is

proportional to the product of firm-specific volatility and the Herfindahl index

of concentration: ‘granular volatility’. The link between asset growth fluctua-

tions at the bank-level and aggregate fluctuations becomes stronger as market

concentration and/or idiosyncratic bank-level volatility increase: even when

abstracting from the issue of interconnectedness between large banks. Using

matched bank–firm loan data for Japan, Amiti and Weinstein (2013) find that

idiosyncratic loan growth shocks at the bank level can explain approximately

40% of the variation in aggregate credit and investment growth. Based on

industry-level data, Carvalho and Gabaix (2013) show that part of the recent

increase in macroeconomic volatility can be attributed to the raising importance

of the financial industry.

A priori, the link between bank-specific and macroeconomic fluctuations

should be stronger if the banking sector is more concentrated. Even though

banking market concentration is high in both low-income and higher-income

countries (Fig. 1), the results in the present paper show that it is difficult to relate

macroeconomic volatility to fluctuations at the bank level. The relation between

macroeconomic volatility and ‘banking granular volatility’ (BGV), a weighted

sum of bank-specific asset growth volatility where each bank’s weight is given by

its squared market share, is mostly insignificant. Yet, as opposed to the link

between volatilities at the microeconomic and macroeconomic level, the rela-

tionship between bank-specific credit growth shocks and aggregate growth has

been shown to be positive and significant in previous work (Buch and

Neugebauer, 2011; Amiti and Weinstein, 2013; Bremus et al., 2013).

While low-income countries do not necessarily have a more concentrated

banking market structure, banking sector size as measured by domestic credit to

GDP is much smaller in low-income than in high-income countries (Fig. 1). The

expected impact of credit to GDP on macroeconomic volatility is not clear a

priori: in the literature, credit to GDP is often used as a proxy for financial

development. However, our results show a consistently positive effect of credit

to GDP on macroeconomic volatility in the short run. This hints at the

1Instability refers to higher volatility here. Other reasons for macroeconomic instability are not

analysed in this paper.

F. BREMUS AND C. M. BUCH

74

© 2015 Wiley Publishing Asia Pty Ltd