Saving on a Rainy Day? Income Smoothing and Procyclicality of Loan‐Loss Provisions in Central European Banks

• Date: 01 March 2015
• Author: Dorota Skała
• Published date: 01 March 2015
• DOI: http://doi.org/10.1111/1468-2362.12058

Saving on a Rainy Day? Income

Smoothing and Procyclicality of

Loan-Loss Provisions in Central

European Banks

Dorota Skała

WNEiZ, University of Szczecin, Szczecin, Poland.

Abstract

This paper analyses income smoothing and cyclicality of loan-loss

provisions (LLP) in Central European banks. I provide strong empirical

evidence that banks in the region use loan-loss provisions to smooth

their income streams, and that these provisions are procyclical with

respect to national business cycles. In addition, I find that income

smoothing may only partly be explained through the concept of ‘saving

for a rainy day’. Banks do use periods of high earnings to smooth

income, but they also elect to build further reserves during periods of

heavy losses—that is, on the ‘rai ny days’themselves. This behaviour

deepens existing losses and may obscure the bank’sunderlying

I would like to thank two anonymous referees and the e ditor, Benn Steil, for comments and

suggestions. Andrew Hende rson has provided excellent editin g support. In addition, I am g rateful to

Régis Blazy, Christophe Godlewski , Oskar Kowalewski, DobromiłSerwa, Laurent Weill and Marek

Zwolankowski, as well as participants of th e 10th INFINITI Conferen ce in Dublin, the 29th GdR E

Annual International Symposiu m on Money, Banking and Finance in Nantes, and se minar audiences

at the National Bank of Poland and University of Szczeci n for their valuable comm ents on previous

versions of the paper. I am in debted to Teresa Lubi

nska and Waldemar Tarczy

nski for continuous

encouragement and suppor t. All remaining errors are mine.

International Finance 18:1, 2015: pp. 25–46

DOI: 10.1111/1468-2362.12058

© 2015 John Wiley & Sons Ltd

profitability. Introducing regulatory measures in line with the Bank of

Spain’s dynamic provisioning system would make income smoothing in

Central European banks more transparent and could limit the scope for

discretionary provisioning during per iods of low profitability.

I. Introduction

Banks across the world have been found to engage in income smoothing by

adjusting the ir annual loan-l oss provisions (LLP) not only to their underlying credit

risk, but also to the level of the ir pre-provisioning profit (Bikke r and Metzemakers

2005; Fonseca and González 2008; Pérez et al. 2008; Bouvatier et al. 2014).

Supporters of income smoothi ng claim that it shoul d be viewed as an elem ent of

prudent credit-risk p olicy, in which forward-lo oking reserves a re created to accom-

modate expected p ortfolio deteri orations that emerg e during downturns (Wall and

Koch 2000; Laeven and Majnoni 2003; Fonseca and González 2008). This approach is

frequently referred to as ‘saving for a ra iny day’behaviour (Greenawalt and Sinkey

1988). Basel III recommendations in line with this reasoning call for a change in

accounting standards to in corporate the expec ted-loss perspec tive, and a similar

rationale has led th e Bank of Spain to introduce regulator y measures that ob lige

Spanish banks to pe rform income smooth ing.

1

Opponents of income smooth ing,

however, argue that it distorts bank financial results such that they do not

adequately convey underly ing profitability at a given point in ti me (Wall and Koch

2000; Bushman and Williams 2012).

This paper addresses two important questions. First, I verify whether banks in

Central Europe apply the income -smoothing approach , as the major ity of previous

analyses have overlooked thi s region. If c onfirmed, in come smooth ing carr ies

important policy implications for regulators an d supervisors in th is part of the

European Union. Second, this paper explores the ‘saving for a rainy day’hypothesi s.

I investigate whether income s moothing necessa rily implies that ban ks create higher

reserves during t he boom periods a nd benefit from them during downturns. Unlike

the existing literature, however, I differentiate between the individual bank’s

earnings cycle and th e business cycle in ana lysing changes in rese rve levels. Thus,

I am able to disentangle two see mingly opposing results found in the li terature that

suggest that reserves a re countercyclical with respect to b ank earnings (‘income

smoothing’) but procyclical with respect to t he business cycl e.

The structure of th e paper is as follows: Sect ion II presents a brief literature

review and previous empirical findings. Section III describes the methodology and

1

See, for example, Pérez et al. (2 008, 2011), Balla and McKenna (2009), Saurina (2 009) and de Lis and

García-Herrero (2010).

26 Dorota Skała

© 2015 John Wiley & Sons Ltd