All that glitters is not gold! Independent directors' attributes and earnings quality: Beyond formal independence

• Published date: 01 November 2021
• Author: Antonio Marra
• Date: 01 November 2021
• DOI: http://doi.org/10.1111/corg.12380

ORIGINAL ARTICLE

All that glitters is not gold! Independent directors' attributes

and earnings quality: Beyond formal independence

Antonio Marra

SDA Bocconi - School of Management,

Università Bocconi, Milan, Italy

Correspondence

Antonio Marra, SDA Bocconi - School of

Management, Università Bocconi, Via

Roentgen 1 –20136, Milan, Italy.

Email: antonio.marra@unibocconi.it

[Correction added on 17 May 2022, after first

online publication: CRUI funding statement

has been added.]

Abstract

Research Question/Issue: This paper examines the effect of additional independent

directors' attributes, beyond formal independence, on earnings management

practices.

Research Findings/Insights: I find that being non-busy, having accounting expertise,

and being appointed by non-controlling shareholders are relevant directors'

attributes—beyond formal independence—in their earnings management monitoring

task, among the directors' attributes I have tested. Additionally, the paper shows that

independent directors who possess such features simultaneously outclass all other

directors—included the touted effective formally independent directors—in mitigating

earnings management activities.

Theoretical/Academic Implications: I respond to calls for dismantling common

wisdom on board independence, investigating factors leading to better monitoring,

showing that independent directors are not all-alike. With specific reference to direc-

tors' monitoring task, when controlling for additional directors' attributes, formal

independence becomes uninfluential in constraining earnings management activities. I

also show that the co-existence of attributes makes independent directors more

effective relative to board mates not sharing such attributes.

Practitioner/Policy Implications: Findings of this work might be useful for practi-

tioners in attempting to design corporate governance mechanisms better able to

monitor earnings management practices through independent directors and may

serve as a stimulus for regulators when re-thinking regulations on board composition

and structure.

KEYWORDS

corporate governance, board independence, independent directors, directors' attributes,

earnings quality

1|INTRODUCTION

The role and functioning of the board of directors are matters of con-

tinuous debate among policy makers, academics, and practitioners,

and the rules related to these issues are never stable (Davies &

Hopt, 2013). Indeed, at the European Union (EU) level, the European

Commission has implemented an Action Plan aimed at remodeling

company law to enhance corporate governance and the role of

directors. In general, the presence of board-independent representa-

tives capable of challenging management's decisions and acting as

Received: 2 October 2019 Revised: 14 April 2021 Accepted: 15 April 2021

DOI: 10.1111/corg.12380

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

provided the original work is properly cited.

© 2021 The Author. Corporate Governance: An International Review published by John Wiley & Sons Ltd.

Corp Govern Int Rev. 2021;29:567–592. wileyonlinelibrary.com/journal/corg 567

effective monitors is widely viewed as a way to protect the interests

of shareholders and, where appropriate, those of other stakeholders.

Nevertheless, there is still little consensus about what a board should

look like (Johnson et al., 2013), and findings of its impact on a

company's life are “mixed at best”(Leblanc, 2004, p. 438). Moreover,

some directors may lack the ability to work properly, thus not being

able to affect the company's performance/results/outcomes (Johnson

et al., 2013). Further, most studies worked on board composition and

structure—focusing on board formal independence—with only a

limited number of studies looking at additional individual attributes

with the potential to shape directors' behavior (Sharpe, 2013).

Board of directors has a broad range of responsibilities that—

based on theoretical work—are usually classified into two primary

functions: monitoring and advising (Adams et al., 2007; Raheja, 2005).

A specific monitoring task of the board of directors is to assure the

quality of financial reporting, proxied by Earnings Management

(EM) practices. EM hampers financial reporting quality, obscures real

performance, and lessens shareholders' ability to make informed deci-

sions, therefore generating agency costs (Xie et al., 2003). In recent

years, considerable attention has been given to reporting quality by

regulators and by the popular press (Loomis, 1999; Xie et al., 2003),

and according to the SEC chairman Arthur Levitt, falsified reports and

doctored records are a common problem, and there are “great

expanses of accounting rot, just waiting to be revealed.”The

investigation of the relation between board of directors, specifically

independent members, and EM is critical as lack in the monitoring

task generates severe effects on firms' outcomes (Trueman &

Titman, 1988).

Therefore, in this study, I aim at “understanding mechanisms that

may affect the degree to which characteristics associated with

‘potential’to influence outcomes result in the ‘actual’influence being

realized”(Johnson et al., 2013, p. 254) opening the independent

directors' “black box”

1

(Dalton & Cannella, 2003). I do so by investi-

gating whether and to what extent additional independent directors'

characteristics influence EM practices, above and beyond, formal

independence.

Specifically, I exploit the unique features of the Italian institu-

tional setting, within the time window 2005–2017, with a hand-

collected data sample of the entire board population of Italian listed

companies, for a total of 2,249 firm-year observations and 21,193

directors (of whom 8,436 are declared to be independent,

i.e., formally independent). The Italian setting is characterized by a rel-

atively small and inefficient capital market, weak legal protection for

both creditors and shareholders (La Porta et al., 1999), and highly con-

centrated ownership structure with the presence of principal–

principal (PP) conflicts, particularly between the controlling and minor-

ity shareholders. In such a context, the independent directors' role

remains central and yet unexplored as they compete in a specific labor

market (Fama, 1980; Fama & Jensen, 1983), where performing poorly

may undermine their reputation and future career (Zattoni & Cuomo,

2010). They have incentives to reinforce their monitoring tasks to

protect their human capital and avoid legal liabilities (Carcello

et al., 2002; Fama, 1980; Fama & Jensen, 1983).

In my first set of tests, I start with a validity test aiming to

confirm common wisdom related to formally independent boards'

ability to constrain EM practices. Next, I test my first hypothesis on

the effect of additional independent directors' attributes on EM

relative to the formally independent directors. This analysis heeds

the call coming from recent studies suggesting further investigation

on independent directors' effect on firms' outcomes (Dalton &

Cannella, 2003; Johnson et al., 2013). Formal independence has

been touted as panacea for board monitoring effectiveness, while

I maintain that rather than independence—per se—additional

independent directors' attributes are to be considered for proper

monitoring. Using seven directors' characteristics, namely, high

visibility,detachment from the Chief Executive Officer (CEO),tenure,

busyness,accounting expertise, being appointed by non-controlling

shareholders, and being acertified statutory auditor, I aim at capturing

additional monitoring ability of independent directors, above and

beyond, formal independence. I find that being non-busy, having

accounting expertise, and appointment by non-controlling shareholders,

are most relevant attributes for independent directors in their EM

monitoring task, among the tested directors' attributes. Findings

enlighten on the fact that—in the Italian setting—independence

per se might not be the proper solution in the EM monitoring

role directors should play, while in Italy directors' attributes such as

time to devote to the monitoring task, accounting knowledge, and

appointment by minorities, are premium factors to constrain EM

activities. On the contrary, directors' length of appointment,

detachment from the CEO, and being visible do not seem to limit EM

practices.

My second set of tests responds empirically to my second

hypothesis, where I maintain that additional independent directors'

attributes—proved to be more effective with reference to limiting

EM—should naturally be even more effective where co-existent.

To test my hypothesis, I first address the issue of independent

directors' appointment being endogenous to the firm, using the

Heckman process (Heckman, 1979; Lennox et al., 2012). Then, I

assess whether and to what extent independent directors who pool

at the same time non-busyness,accounting expertise, and are appointed

by non-controlling shareholders are better able to monitor relative to

peers. Results confirm my prediction with independent directors

embedding simultaneously the three attributes identified above,

outclassing any other independent director type. Findings show that

when those attributes co-exist, formal independent directors' ability

to limit EM practice does not differ from non-independent peers

and that only the three identified characteristics are still valid

attributes to limit EM practices.

A final set of tests aims at corroborating my inferences working

on additional analysis and sensitivities. First, to alleviate concerns

that my inferences are driven by confounding effects or a poor iden-

tification, I test a model that provides a direct assessment of each

characteristic's additional effect to reinforce inferences on the first

and second hypotheses. Second, I run a lagged first-difference (LFD)

model to tackle reverse causality (Allison, 2009; England et al., 2007;

Martin et al., 2012), on my first hypothesis. Third, I implement a

568 MARRA