What factors determine dividend smoothing by US and EU banks?

• Author: Nicos Koussis,Michalis Makrominas
• Date: 01 July 2019
• Published date: 01 July 2019
• DOI: http://doi.org/10.1111/jbfa.12399

DOI: 10.1111/jbfa.12399

What factors determine dividend smoothing by US

and EU banks?

Nicos Koussis1Michalis Makrominas2

1Department of Accounting, Finance and

Economics, FrederickUniversity Cyprus, Cyprus

2Department of Maritime Studies, Frederick

University Cyprus, Cyprus

Correspondence

MichalisMakrominas, Department of

MaritimeStudies, Frederick University Cyprus,

7,Y. FrederickouStr. Pallouriotisa, Nicosia 1036,

Cyprus.

Email:bus.kn@fit.ac.cy

Abstract

Using a relatively large sample of European and US banks for

the period 1998–2016, we investigate the determinants of bank

dividend smoothing based on agency, asymmetric information

and risk-shifting theories. We show that dividend payout ratio

smoothing practices were implemented on both continents before

and after the crisis of 2007 and were more strongly pronounced for

EU banks. Our findings mostly support agency-based explanations

of bank dividend behavior as evidenced by higher payout ratio

smoothing for banks with higher (initial) dividend payouts, lower

ownership concentration,public banks, and banks with lower growth

opportunities and weaker investor protection. Evidence in favor of

asymmetric information explanations is stronger for EU countries,

where smaller (more opaque) banks appear to smooth more. In

both continents, banks that rely more heavily on equity issuances

are found to smooth dividend payout ratios more, suggesting that

banks aim at improving access to equity markets. We also provide

evidence in support of risk-shifting, as evidenced by the persistence

of dividend payout ratio smoothing in the crisis years and higher

dividend smoothing for banks under greater regulatory pressure.

Additional analysis using a time series partial adjustment model

for dividend levels provides evidence supporting the prevalence of

dividend smoothing and the suggested theoretical explanations.

KEYWORDS

agency, asymmetric information, bank dividend policy, bank

regulation, legal origin, risk-shifting

JEL CLASSIFICATION

G20, G21, G31, G35.e

1INTRODUCTION

Bank dividend policy attracted attention when banks maintained high dividend payouts long into the financial crisis

despite having depleted liquidity and solvency positions. A large body of empirical studies supports a theory of

1030 c

2019 John Wiley & Sons Ltd wileyonlinelibrary.com/journal/jbfa JBus Fin Acc. 2019;46:1030–1059.

KOUSSIS ANDMAKROMINAS 1031

risk-shifting that suggests a transfer of cash from depositors and taxpayers to shareholders (Kanas, 2013, Onali,

2014). The notion of excessive risk-taking at the expense of depositors and in anticipation of possible bailouts has

prompted the Basel Committee on Banking Supervision to call for greater scrutiny of bank dividend policy,including

provisions for a “dividend stopper” (BCBS, 2011). Nevertheless, designing an effective policy response to observed

practices requires a better understanding of bank dividend behavior and its determinants.

This paper revisits the question of bank dividend policy by focusing on the determinants of bank dividend smooth-

ing for both EU and US banks. While bank behavior during the crisis has typically been addressed in terms of exuberant

payoutsamid financial distress, it can be argued that banks have sustained rather than otherwise altered their dividend

payouts.Acharya, Le, and Shin (2017) and Floyd, Li, and Skinner (2015) find that banks maintained a “relatively smooth

dividend policy” into the crisis period. From that perspective,the question of bank dividends during the crisis becomes

an issue of dividend adjustment or,rather, a lack of it: Why did banks fail to cut dividends despite economic and regula-

tory pressures? What explains bank dividend persistence?

Toaddress these questions, we estimate a partial adjustment model that characterizes banks’ realized dividend pay-

out ratio as a function of the last period’s dividend payout ratioand a target dividend payout ratio determined by fun-

damentals. A lack of adjustment of dividend payoutratios in response to changes in economic factors (e.g., profitability

or in-place regulatory requirements) signifies dividend smoothing. This approach of focusing on the determinants of

the dividend payoutratio (instead of levels) and its persistence (smoothness) follows the capital structure literature on

modeling leverageratios, as analyzed in Flannery and Rangan (2006). By utilizing a relatively large sample of European

and US banks overthe 1998–2016 period, we measure bank dividend payout ratio smoothing and seek to characterize

its determinants. We provide a comparison of bank dividend payoutratios and smoothing before and during the crisis,

between European and US banks and across different legal origins (common law vs. civil law). Importantly,we identify

the determinants of dividend smoothing based on prevalent economic explanationsrelated to theories on asymmetric

information, agency costs and risk-shifting incentives. Following Lintner’s tradition,as analyzed in Leary and Michaely

(2011), our analysis is supplemented with additional results from estimating a partial adjustment model for dividend

levels and adjustments towards a long-term target payoutratio. Despite the stricter data requirements for time-series

analysis, our analysis produces similar insights. Our findings are summarized as follows.

First, bydocumenting dividend cuts and omissions, we corroborate evidence that bank dividend smoothing remains,

albeit at a lower level, during and following the crisis of 2007. Second, using estimates from a partial adjustment model

for dividend payout ratios, we show that dividend smoothing is more significant for European banks than for their US

counterparts and more significant in civil law compared to common law countries. Evidence of higher dividend payout

ratio smoothing in civil law countries may suggest a “substitute”agency model explanation, where banks with poorer

legal protection smooth more to retain access to equity markets (La Porta, Lopez-de-Silanes,Shleifer, & Vishny, 2000).

Third, we find that dividend payout ratiosmoothing in both continents is higher for banks with a higher initial dividend

payoutand a lower concentration of ownership, in line with manager-shareholder agency theory-related explanations.

Consistent with an agency-based explanation,Khan (2006) finds a negative relation between dividends and ownership

concentration in a panel of large UK quoted firms. Our results further support an agency-based explanation because

we show stronger dividend smoothing for public firms that generally have lower ownership concentration than pri-

vate firms (supporting the evidence provided in Michaely and Roberts (2012) for non-financial firms). Wealso provide

evidence of dividend payout ratio smoothing linkedto higher stock issuances, showing that banks aim to attract new

investorsand improve access to equity markets. Fourth, we find evidence suggesting that risk-shifting is related to both

the level of dividend payoutand dividend smoothing before and after the crisis and across legal origins. Risk-shifting is

also evidenced in the persistence of dividend payout ratio smoothing in the crisis period and the higher dividend pay-

out ratiosmoothing for banks under regulatory pressure. Fifth, we find that EU banks’ dividend payout ratio smoothing

is strongly associated with asymmetric information problems, as evidenced by the significantly higher smoothing of

smaller banks relative to larger banks. In contrast, larger US banks exhibitmore dividend payout ratio smoothing than

smaller banks in the same region. Our time-series analysis of a partial adjustment model for dividend levels, following

Leary and Michaely (2011), mostly corroboratesthe above findings.