Exploring corporate governance research in accounting journals through latent semantic and topic analyses

• Published date: 01 October 2019
• Author: Mark Edmonds,Ferhat D. Zengul,Arline Savage,Nurettin Oner,James D. Byrd
• DOI: http://doi.org/10.1002/isaf.1461
• Date: 01 October 2019

RESEARCH ARTICLE

Exploring corporate governance research in accounting

journals through latent semantic and topic analyses

Ferhat D. Zengul

1

| James D. Byrd, Jr

2

| Nurettin Oner

1,3

| Mark Edmonds

2

|

Arline Savage

2

1

Department of Health Administration,

University of Alabama at Birmingham,

Birmingham, AL, USA

2

Department of Accounting and Finance,

University of Alabama at Birmingham,

Birmingham, AL, USA

3

Department of Healthcare Management,

Ankara University, Ankara, Turkey

Correspondence

Ferhat D. Zengul, University of Alabama at

Birmingham, Department of Health

Administration, Health Care Management

Program, 1705 University Blvd. Birmingham,

AL 35294. 1720 2nd Avenue South, School of

Health Professions Building, 540A,

Birmingham, AL 35294-3361.

Email: ferhat@uab.edu

James D. Byrd Jr.,University of Alabama at

Birmingham, Department of Accounting and

Finance, Collat School of Business –CSB316,

Birmingham, AL 35294.

Email: jimbyrd@uab.edu

Summary

The literature on corporate governance (CG) has been expanding at an unprece-

dented rate since major corporate scandals surfaced, such as Enron, WorldCom,

and HealthSouth. Corresponding with accounting's important role in CG, accounting

scholars increasingly have investigated CG in recent years, so the body of literature

is growing. Although previous attempts have been made to summarize extant litera-

ture on CG via reviews, none of these attempts has utilized recent developments

in text analyses and natural language processing. This study uses latent semantic

and topic analyses to address this research gap by analysing abstracts from 1,399

articles in all accounting journals that the Australian Business Deans Council

(ABDC) has rated A and A*. The ABDC journal list is widely recognized as a

journal-quality indicator across many universities worldwide. The analyses revealed

10 distinct research topics on CG in the ABDC's top accounting journals. The

results presented include the five most representative articles for each topic, as

distinguished by topic scores. This study carries important practice and policy impli-

cations, as it reveals major research streams and exhibits how researchers respond

to various CG problems.

KEYWORDS

corporate governance, latent semantic analysis, natural language processing, topic analysis

1|INTRODUCTION

According to the Organization for Economic Cooperation and Devel-

opment (OECD), corporate governance (CG) refers to the processes

and procedures that are used to direct and control firms (OECD,

2005). An expanded version of this definition would include the

established checks and balances that organizations use to prevent

abuses that self-interested individuals, such as executives, initiate

(Larcker & Tayan, 2015).

As big accounting scandals emerged in the past two decades,

public distrust of corporations and their managers has spiked.

Governance failures include not only fraud and theft, but also viola-

tions of fiduciary duties to shareholders by management and boards

of directors. Prominent examples of these CG scandals include using

corporate funds to pay half of a lavish $2 million tab for a birthday

party held for a chief executive officer's (CEO's) wife and attempting

to fly corporate jets to Washington, DC, to secure a public bailout

from the government (Levs, 2008; Orin, 2008; Pillmore, 2003;

Riotto, 2008).

The accounting profession plays a particularly crucial role in

preventing CG scandals. In many of these scandals (e.g. Enron,

WorldCom, HealthSouth, Olympus, and Toshiba), intentional violation

of accounting rules empowered the wrongdoers. Shareholders,

employees, creditors, other stakeholders, and the public at large bore

the very high costs from these scandals resulting from the abdication

of fiduciary duties of the board, management, and/or auditors, who

Received: 16 July 2019 Revised: 11 October 2019 Accepted: 26 November 2019

DOI: 10.1002/isaf.1461

Intell Sys Acc Fin Mgmt. 2019;26:175–192. wileyonlinelibrary.com/journal/isaf © 2020 John Wiley & Sons, Ltd. 175

are responsible for CG, as in the case of the 2008 financial bailout

(Abdel-Khalik, 2002; Arnold & De Lange, 2004; Asthana, Balsam, &

Krishnan, 2010; Canada, Kuhn, & Sutton, 2008; Carcello, Dana, &

Raghunandan, 2005; Fukase, 2015; Inagaki, 2018; Orin, 2008). Some

of these scandals eventually led to the demise of Arthur Andersen,

one of the world's largest accounting firms.

Given accounting's important role in these CG scandals, account-

ing academics have been focusing more attention on CG over the past

decade. The evidence of this increase in research can be found in a

seminal review article by Brown, Beekes, and Verhoeven (2011), in

which they detailed how they conducted a Google Scholar search in

June 2010 using the keyword ‘corporate governance’, eliciting

287,000 hits. We conducted the same search on 4 January 2019 and

yielded 810,000 hits. A large proportion of this increase in hits can be

attributed to accounting, given that adding the keyword ‘accounting’

to ‘corporate governance’with the Boolean operator ‘and’yielded

508,000 hits on 4 January 2019. Brown et al. (2011) conducted the

same search in July 2010, generating only 95,000 hits.

Our study contributes to the literature by including research con-

ducted over the past decade and applying natural language processing

(NLP) and text-mining tools to analyse the research conducted around

CG in accounting.

In this study, we apply recent NLP developments by utilizing

latent semantic analysis (LSA) on CG literature from the top

30 accounting journals, as ranked by the Australian Business Deans

Council (ABDC, 2016). We contribute to academia more specifically

by summarizing current research trends in CG.

2|LITERATURE REVIEW

CG's increasing role in business and accounting research also can

be observed in more recent literature reviews. These CG reviews

comprised varying focus areas, such as qualitative research on CG

(McNulty, Zattoni, & Douglas, 2013), CG and state-owned enterprises

(Grossi, Papenfuß, & Tremblay, 2015), cross-national governance

(Schiehll & Martins, 2016), CG and innovation (Gonzales-Bustos &

Hernández-Lara, 2016), and CG models (Ahmad & Omar, 2016). How-

ever, with the exception of Brown et al. (2011), few researchers have

attempted to summarize CG research within accounting.

This research also is relevant because of the increasing interest

in big data in accounting research (Alles, 2015; Brown-Liburd, Issa, &

Lombardi, 2015; Dagilienė& Klovienė, 2019; Drew, 2018; Gepp,

Linnenluecke, O'Neill, & Smith, 2018; Griffin & Wright, 2015;

Huerta & Jensen, 2017; Krahel & Titera, 2015; Vasarhelyi, Kogan, &

Tuttle, 2015; Warren, Moffitt, & Byrnes, 2015; Yoon, Hoogduin, &

Zhang, 2015; Zhang, Yang, & Appelbaum, 2015). Sun and Vasarhelyi

(2018) investigated the various text-mining tools available for

auditing. The methodology used in this investigation enables the min-

ing of big data related to the massive amounts of text documents

available to researchers. Furthermore, this research method can be

used to understand themes and trends in accounting topics other

than CG.

3|METHODS

This study's methodology entails three phases: (1) creating the corpus;

(2) generating and curating the term list; and (3) analysis (Delen &

Crossland, 2008; Ozaydin, Zengul, Oner, & Delen, 2017).

3.1 |Phase 1. Creating the Corpus

This phase involves the aggregation of references with keywords,

titles, and abstracts, then converting them into a data format that can

be analysed using statistical software. To achieve this, the initial data

from the top 30 accounting journals (see Figure 1) were downloaded

into Endnote by using: (1) EBSCOhost Business Source Premier and

(2) Scopus. A total of 40,923 references were gathered from the top-

ranking 30 journals listed in Figure 1. The references were down-

loaded from the first issue of each journal through 2017 for all the

journals except five, which have not been available since inception

(see Appendix A). After filtering for ‘governance’and excluding refer-

ences without abstracts, a corpus of 1,399 references was finalized,

as highlighted in the seven steps in Table 1. We used the broader

term ‘governance’instead of ‘corporate governance’for filtering

FIGURE 1 Number of articles with corporate governance focus in

30 a* and a accounting journals (n= 1399). AAAJ: Accounting Auditing

and Accountability Journal; AAJPT: Auditing: A Journal of Practice and

Theory; ABACUS: Abacus: A Journal of Accounting, Finance and

Business Studies; ABR: Accounting and Business Research; AF:

Accounting and Finance; AH: Accounting Horizons; AOS: Accounting,

Organizations and Society; BAR: British Accounting Review; BRA:

Behavioral Research in Accounting; CAR: Contemporary Accounting

Research; CPA: Critical Perspectives on Accounting; FAM: Financial

Accountability and Management; FTA: Foundations and Trends in

Accounting; IAE: Issues in Accounting Education; IJA: International

Journal of Auditing; IJAIS: International Journal of Accounting

Information Systems; JAAF: Journal of Accounting Auditing and Finance;

JAE: Journal of Accounting and Economics; JAL: Journal of Accounting

Literature; JAPP: Journal of Accounting and Public Policy; JAR: Journal

of Accounting Research; JBFA: Journal of Business Finance and

Accounting; JCAE: Journal of Contemporary Accounting and Economics;

JIAR: Journal of International Accounting Research; JMAR: Journal of

Management Accounting Research; MAR: Management Accounting

Research; RAS: Review of Accounting Studies; TAR: The Accounting

Review; TEAR: The European Accounting Review; TIJA: The International

Journal of Accounting

176 ZENGUL ET AL.