Nonprofit organization fraud reporting: does governance matter?

• Pages: 409-428
• DOI: https://doi.org/10.1108/IJAIM-10-2019-0117
• Date: 03 March 2020
• Published date: 03 March 2020
• Author: Husam Abu Khadra,Dursun Delen
• Subject Matter: Accounting/accountancy,Accounting methods/systems

Nonprof‌it organization

fraud reporting: does

governance matter?

Husam Abu Khadra

Department of Accounting, Heller College of Business, Roosevelt University,

Schaumburg, Illinois, USA, and

Dursun Delen

Department of Management Science and Information Systems,

Oklahoma State University, Stillwater, Oklahoma, USA

Abstract

Purpose –This paper aims to contribute to the extant literature in this f‌ield by examining nonprof‌it

organizations’fraudreporting compliance using logisticregression and decision tree induction algorithms.

Design/methodology/approach –This study used the data from 428 nonprof‌it organizations during

2009-2015 period, andanalyzed 21 individual measures (obtained from these organizations’Internal Revenue

Service Form990 f‌ilings) using logistic regression and decision tree induction algorithms, to study the

governancecharacteristics and fraud reporting.

Findings –The study found evidence that compliance with the law, board of directors’independence,

federal audit and using independent accountants to compile and review f‌inancial statements are the most

prevailing factors affecting the odds of nonprof‌it organizations experiencing fraud reported as an asset

diversion.

Originality/value –The argument associated with using governanceto reduce the chances of fraud has

been a popular topic in industry and academia but unfortunately has limited empirical evidence in the

literature,especially when it relates to nonprof‌its.This study contributes to the literature inthis respect.

Keywords Fraud in nonprof‌its, Nonprof‌it governance, Asset diversion, IRS form 990

Paper type Research paper

1. Introduction

Due to their frequent fraud scandals, nonprof‌it organizations in the USA have been subject

to negatively portrayed media and public inquiries (McDonnell and Rutherford, 2018).

Despite these inquiries, not much attention has been devoted to analyzing and solving this

issue; as a result, more and more signif‌icant fraud cases have occurred over recent years.

Stephens and Flaherty (2013) indicate that one-sixth of all major embezzlement instances

occur in the nonprof‌it industry. Furthermore, accordingto the 2018 Association of Certif‌ied

Fraud Examiners’(ACFE’s) Report to the Nation,nonprof‌it entities made upto 9 per cent of

the reported and analyzed fraud cases and suffered a median lossof $75,000. This loss may

not seem signif‌icant; however, formany nonprof‌it entities, f‌inancial resources are extremely

limited, and a loss of $75,000 can be particularlydevastating in effectively carrying out their

missions (ACFE, 2018).

All organizations are vulnerable to fraud, but nonprof‌it organizations are subject to a

higher risk of fraud due to their greater reliance on human decency, moral missions and

ethical values and messages. An uptick in fraud activity couldbe the result of an increase in

Nonprof‌it

organization

fraud

reporting

409

Received5 October 2019

Revised27 November 2019

Accepted31 December 2019

InternationalJournal of

Accounting& Information

Management

Vol.28 No. 3, 2020

pp. 409-428

© Emerald Publishing Limited

1834-7649

DOI 10.1108/IJAIM-10-2019-0117

The current issue and full text archive of this journal is available on Emerald Insight at:

https://www.emerald.com/insight/1834-7649.htm

the availability of funds, lack of enforcing regulations or weak or nonexistent internal or

external control measures. Unfortunately, very little empirical research exists on fraud in

nonprof‌its to provide answers as to why the uptick is occurring or what could be the

recommendationson how to reduce the occurrence of fraud in the nonprof‌it sector.

Historically, the lack of research was justif‌ied by the argument that publicly available

information on nonprof‌it organizations’governance and reporting of fraudulent activities

has been very limited. However, in 2008, things had changed when the Internal Revenue

Service (IRS) began requiring nonprof‌it organizations to disclose in their 990 Form tax

f‌iling, whether they were aware of any signif‌icant asset diversions during the year. A

diversion of assets includes any unauthorizedconversion or use of the organization’s assets

(other than those of the ones for the organization’s authorized purposes) including, but not

limited to, the embezzlement or theft activities (IRS,2019). Nonprof‌it organizations are also

required to answer a comprehensivelist of governance questions (Harris et al.,2015).

The main goal of this study is to examine the governance practices of nonprof‌it

organizations and investigate the impact of these practices on reporting fraud as asset

diversions. To test our hypothesis that states “governance decreases the probability of

fraud,”we examined all of the digitized data f‌iles from annual extracts of tax-exempt

organizations’f‌inancial reporting to identify and characterize the organizations that

suffered from asset diversion.Next, as a pseudo control group, we matched thesefraudulent

cases with similar organizations from the same reporting period that did not report any

asset diversion. To conductthe study, we followed an exploratory hybrid research approach

that represents an intersection of theory- and data-driven research (Maass et al.,2018).

Initially, we followed a theory-basedapproach to develop the main study hypotheses. Then,

we followed theoretical understanding with a data-driven approachto account for all of the

relevant variables that were available in the data set. The aim was to discover any

scientif‌ically interesting patterns through the application of analytical techniques and

means of reasoning (Kitchen, 2014). This approach enabled us to def‌ine, extract and use 20

governance factors and 1 control variable, all of which collectively affect the odds of fraud

occurrence, thus expanding on the work done by Harris et al. (2017). In their study,

Harris et al. (2017) used eleven corporate governance variables within a hypothesis-driven

research approach. This paper represents a shift toward more data-driven research that

relies on large and feature-rich data set to provide a more holistic perspective related

opportunities for conducting more impactful research studies. Furthermore, unlike Harris

et al. (2017), who had used a data set created by the Washington Post for the years 2008-

2011, our data was extracted directly from the IRS Sample of Statistics of Income (SOI)

database for the years 2009-2015. Thus, it covered a longer time window and excluded the

f‌inancial crises years of 2007-2008,which was important in terms of minimizing the impacts

of any external environmental factors, such as the economic recession, on the outcomes of

the study. Due to the 2008 tax f‌iling 990 form redesign,no comparative data were considered

for the years before 2009. Our study identif‌ied compliance with the law,board of directors’

independence,federal audit and using independent accountants to compile and review

f‌inancial statements as the main factors affecting the odds of nonprof‌it organizations’

fraudulent experiences.Our results were different from those of the ones reported by Harris

et al. (2017), which leads us to believe that the environmental factors mentioned above may

have affected their studyresults.

Due to the dichotomous nature of the dependent variablein our study, we developed and

analyzed the data using a binarylogistic regression model. Furthermore, we conf‌irmed our

results by using the C5 decision tree classif‌iers, C5 is an improved version of C4.5 and ID3.

The data set that we prepared and used for the analysis contained information on 428

IJAIM

28,3

410