Scant coverage of fraud in nonprofit organizations and smaller firms exists despite surveys and headlines indicating fraud prevalence is at greater levels among these firms than at for-profit and larger firms. Applying a combination of stakeholder salience theory and whistleblowing theory, this paper establishes a background for such comparisons. Results of a survey of 153 fulltime employees uncover differences in whistleblowing between for-profit and nonprofit organization types. We find those who work at for-profit organizations are more likely to whistleblow than those who work for nonprofit organizations. The results and supplemental analysis suggest whistleblowing intention for fraud is significantly more likely for larger organizations (measured by number of employees) regardless of organization type. Further, as geographic reach increases, for-profit employees are increasingly likely to report.
A Washington Post investigation found that over 1,000 nonprofits reported a significant diversion of assets between 2008 and 2012 on Form 990 (Stephens and Flaherty 2013). Form 990 only requires that such diversions be reported if they are in excess of $250,000 or 5 percent of gross receipts or total assets (IRS 2016), an amount that is significant for many nonprofits. Recent research finds that nonprofits misreport fundraising expenses (Krishnan, M. Yetman, and R. Yetman 2006), are more aggressive than for-profits in managing pension assumptions (Vermeer, Edmonds, and Asthana 2014), and report financial statement error rates that are 60 percent higher than those reported by public for-profit organizations (Burks 2015). Further, only 34 percent of nonprofit employees believe their organization understands its fraud risks, and only 38 percent believe management is sufficiently educated to be able to avoid fraud (Zietlow, Hankin, Seidner, and O'Brien 2018). The 2018 Global Fraud Survey published by the Association of Certified Fraud Examiners (ACFE), which polls certified fraud examiners (CFEs), finds that 9 percent of total fraud cases occurred in nonprofits and that the median loss was $75,000 (ACFE 2018).1 Further, the number of cases is highly volatile. While fairly constant overall, the case share spiked more than 20 percent between 2010 and 2016 before returning to previous levels (ACFE 2018, 2016, 2010). Given the higher error rates at nonprofits, the higher number of self-reported cases at nonprofits in proportion to their contribution to the economy (Urban Institute 2013), and the relative abundance of research in for-profit fraud compared to an insufficiency of coverage in nonprofit fraud, such statistics bring to light a need to further investigate whether whistleblowing intentions in for-profits differ from those in nonprofits so fraud can be detected and deterred.
There are about 1.58 million tax-exempt organizations in the U.S. (Urban Institute 2013). These organizations contribute nearly $837 billion to the economy and make up 5.6 percent of the gross domestic product (Urban Institute 2013). The nonprofit sector provides over 11 million jobs and accounts for about 9 percent of all wages and salaries in the U.S. (BLS 2016a; NCCS 2017). Even during the recession, when most sectors were losing jobs, employment in the nonprofit sector continued to rise (BLS 2016a). About 62.6 million people, or approximately 25 percent of individuals in the U.S., served as nonprofit volunteers in 2015 (BLS 2016b). Furthermore, individuals gave more than $410 billion to nonprofits in 2017 (Giving USA 2018). Nonprofits not only are an important sector of our economy, but also are important in the lives of many individuals.
If one of these nonprofit organizations is the victim of asset misappropriation or is the perpetrator of financial statement fraud, volunteers, donors, recipients of the services, and the public will want to know. Empowered and engaged employees play a vital role in fraud prevention and deterrence in nonprofit organizations (Bradley 2015). Thus, the majority—about 40 percent—of fraud is detected through tips from whistleblowers (ACFE 2018). There is a paucity of research on intention to report fraud in nonprofit organizations. Current literature has not provided evidence that we can generalize findings from the for-profit sector to the nonprofit sector. SEC reporting is not an option for nonprofits due to the inapplicability of several Sarbanes-Oxley and Dodd-Frank statutes to the nonprofit community. Further, nonprofit researchers may experience a community with less auditing, largely nonregulated compliance, and greater difficulty in accessing participant samples that are diverse. This study seeks to add to both the nonprofit and whistleblowing literature by examining intention to report fraud in both for-profits and nonprofits and in organizations of varying magnitudes (as measured both by the size based on number of employees and the geographic reach of the organization).
We conduct a survey of individuals currently employed fulltime in either a for-profit or nonprofit organization. We collect information about the organization in which the individual works and intentions to report fraud, in two different scenarios. Our final sample includes a total of 153 responses. Our results indicate that those at for-profit organizations were more likely to whistleblow than those at nonprofit organizations. This fraud reporting intention may be attributable to for-profits having regulated reporting and increased audit requirements. The results indicate reporting intentions for fraud are significantly more likely for larger organizations than for smaller organizations, regardless of organization type. As geographic reach increases, for-profit employees are increasingly likely to whistleblow. The opposite is also true: nonprofit employees in international organizations are less likely to report than nonprofit employees in local organizations.
This study contributes to literature on nonprofit organizations and on whistleblowing. There are a limited number of studies that compare employee behavior at for-profits and nonprofits. Few, if any, combine the number of employees or geographic reach of the organization with the type of organization to measure whistleblowing. Likewise, literature has little to offer in regard to whistleblowing in nonprofits. Finally, differences in effects of fraud type on nonprofit whistleblowing are mostly unexplored. This paper begins to address these gaps and is useful to researchers seeking to expand knowledge of differences in fraud between nonprofit and for-profit organizations. Further, our results help illuminate how magnitude (as measured by number of employees and geographic reach) provides context to fraud reporting decisions. Our findings are also useful to practitioners seeking to design or update whistleblowing policies and procedures.
Section II briefly discusses the background literature on nonprofit stakeholders and whistleblowing, leading to the hypotheses investigated in this study. Our survey method, a description of our sample, and variable definitions are offered in Section III. We provide descriptive statistics, our empirical results, and additional robustness tests in Section IV. Section V concludes this paper with a discussion of contributions, limitations, and avenues for future research.
II. LITERATURE REVIEW AND HYPOTHESES DEVELOPMENT
Whistleblowing intentions theory (Keil, Tiwana, Sainsbury, and Sneha 2010) suggests that a potential whistleblower weighs the cost of reporting against the benefits that might be obtained. However, identifying such costs and benefits is challenging (Dyck, Morse, and Zingales 2010). For example, literature has looked at fear of retaliation (Mesmer-Magnus and Viswesvaran 2005; Liu, Liao, and Wei 2015) but barely broaches external costs to the organization. Similarly, monetary rewards from the SEC's Office of the Whistleblower (Rose, Brink, and Norman 2016) and internal rewards (Brink, Lowe, and Victoravich 2013) have been studied, while the benefits of not reporting have been largely ignored.
Personality traits (Brink, Cereola, and Menk 2015a), ethical position (Brink et al. 2015a), locus of control (Taylor and Curtis 2010), ethical style (Taylor and Curtis 2010), monetary attitude (Brink, Lowe, and Victoravich 2017), and Machiavellianism (Dalton and Radtke 2013) all influence how whistleblowers view the costs and benefits of whistleblowing. The availability of an anonymous reporting channel (Kaplan and Schultz 2007), the strength of evidence (Brink, Eller, and Gan 2015b), other employees' awareness of the act (Brink et al. 2015b), and materiality (Brink et al. 2015a) may further effect how a whistleblower views the costs of reporting. Literature suggests characteristics of the organization such as organizational climate and structure and appropriateness of whistleblowing matter (Gao and Brink 2017; Near and Miceli 1995). As a result of data availability, funding, and public offices dedicated to the protection of the public (such as the SEC), most literature tends to concentrate on larger for-profit firms when considering whistleblowing.
Because whistleblowing intentions-related theories rely heavily on individual and organizational characteristics, it is important to understand the operational and power structures of organizations. Further, these characteristics tend to differ by the size of the organization and may differ by the stakeholders affected. These different characteristics lead to an exploration of stakeholder theory. Differences in fraud reporting may result from reporting controls being based on dissimilar types of stakeholders or dissimilar organizational sizes. Further, legislated and natural differences exist between for- and nonprofit organizations. Thus, the path to whistleblowing intent may differ, in part, due to context.
While whistleblowing-related theories focus on intent, stakeholder salience theory focuses on identifying key stakeholders. Mitchell, Agle, and Wood (1997) suggest that stakeholder identification be viewed through three lenses: power, legitimacy, and urgency. Power can be wielded externally (by shareholders or donors), while legitimacy (granted by pre-existing structure, rules, or those at the top of an organization) and urgency (created through a special or emergency situation) tend to be created from within or from unique circumstances. Core stakeholders possess characteristics of all three aspects (power, legitimacy, and urgency) and are critical in helping an organization to succeed. When comparing for-profit versus nonprofit organization type, there are key differences among core stakeholders. Identifying core constituencies provides insight into potential pressures or motivations relating to intentions to whistleblow. However, identifying key stakeholders can pose challenges in a nonprofit environment (Knox and Gruar 2007). Given the closeness between donors, those receiving services, and those providing services, those within the organization need to assume multiple roles to be effectively engaged in their work (Bissola and Imperatori 2016). The combination of an increasingly varied responsibility structure with different and sometimes competing stakeholders can lead to less efficiency when problems arise (Krashinsky 1997).
Hull and Lio (2006) suggest several firm goals as the main drivers of these differences, but they key in on organizational structure. Nonprofits are generally structured with a greater variety of legal choices and lower levels of politically mandated financial reporting responsibilities. These structural differences can foster a creative environment, ultimately leading to greater managerial care across the firm on behalf of employees, who often become important stakeholders themselves (Jensen 2001). Due to their passion for the cause, nonprofit employees typically have greater personal investment in the mission of the nonprofit. These employees often have higher levels of nonmonetary orientation and receive less overall compensation (Ridder and McCandless 2010).
Thus, stakeholder salience theory suggests organizations function more efficiently when the firm addresses the needs of multiple constituencies simultaneously (Andriof, Waddock, Husted, and Rahman 2017). As a result, there are natural differences in how for- and nonprofit organizations operate (Knox and Gruar 2007). Also, because individual power and dominance can change based on the size of the organization (Harrington 2006), the actions of individuals can be directly and indirectly affected by the size of the firm. Likewise, whistleblowing theory investigates the secondary effects of size (Reuben and Stephenson 2013) and organization type (Yallapragada, Roe, and Toma 2010), suggesting those in larger organizations are generally more likely to whistleblow and are less tolerant of dishonest behavior.
Further, stakeholder differences lead to nonprofits being monitored by donors (Parsons 2007), often representing a more personal influence in the organization (compared to for-profit shareholders). Also, nonprofits are often small (Casey 2016) in comparison to for-profits, which often leads to outsized influence of the owner/manager and complex relationships that challenge principal agent theories applied to the sector (Coule 2015). Thus, the combination of organization type and size leads to different levels of information asymmetry. As a result, the informational advantage and responses to this advantage can differ dramatically, even when two different (type, size) organizations are faced with a seemingly similar problem.
Despite the depth of research surrounding the agency issues associated with organization types, there is a dearth of knowledge regarding how organization type and firm size directly affect whistleblowing. Further, literature regarding how the combination of organization type and organization size influence fraud reporting is limited. Given the natural differences due to their stakeholders and magnitudes, we hypothesize that there are relationships between the effect of organizational type and magnitudes on fraud reporting intention.
Size is often viewed as a control or second-order effect when studying whistleblowing (Miceli and Near 1985; Greenlee, Fischer, Gordon, and Keating 2007; Lee and Fargher 2013), perhaps because studies investigating size suggest conflicting results in the literature. Miceli and Near (1992) suggest that large firms may be less dependent on one individual for a given job, leading to the belief one individual may not have much influence in changing wrongdoing. In contrast, Keenan (2000) suggests that smaller firms exhibit a strong sense of involvement and are more personally affected by wrongdoing, leading to factors that may explain an increase in the quantity of individual whistleblowers. This increase may be due to elevated levels of moral reasoning at smaller firms (Weber 1990), resulting in greater levels of ethical conduct by owners and employees, especially within family run organizations (Vallejo-Martos and Puentes-Poyatos 2014). However, in a study spanning 30 years (Longenecker, Moore, Petty, Palich, and McKinney 2006), no significant differences in ethical decisions were noted between small and large firms. Conversely, Lee and Fargher (2013) find that larger firms tend to have stronger reporting control structures, resulting in higher incidence of whistleblowing. Similarly, within accounting, the controls of larger firms are theoretically tighter than those of smaller counterparts. It is possible inconsistency across geographic regions requires large, international firms to implement firm-wide changes discovered in reaction to differences in situations and policies in place at the local offices (C. Albu, N. Albu, and Alexander 2014). We see an example of this difference in larger accounting firms. Research suggests larger accounting firms are found to have control polices and structures for whistleblowing more often than are smaller accounting firms, likely due to international firms trying to meet global firm requirements (Brennan and Kelly 2007). However, it should be noted that these studies took place within the context of samples comprised of for-profit employees.
As this issue remains unresolved, we rely on stakeholder salience theory to form a hypothesis. Namely, the number, scale, and scope of core stakeholders are likely increased within a larger organization than within a smaller one. Although one individual may wield more influence at any given smaller organization, the effect is highly variable across small organizations. In contrast, the core stakeholder demands within larger organizations tend to be more stable overall. As a result, it seems this dilemma in literature can be resolved in part through the application of stakeholder salience theory, which suggests that as organizations are larger and spread out, individual power is often lessened, and control structures become more fully developed to manage the organization. Thus, part one of this application would suggest the first hypothesis (where magnitude is measured separately as size based on number of employees and geographic reach of the organization):
As organization magnitude increases, employees will be more likely to whistleblow than they would as organization magnitude decreases.
It seems likely that the decisions of for- and nonprofit employees will differ, due to their different stakeholders, given a similar situation. Literature suggests that for-profit core stakeholders tend to involve shareholders, debt holders, management, and Wall Street expectations (Purnell and Freeman 2012). Given that many of these core stakeholders are external to the organization, it seems likely that employees would take on an additional custodial role as part of their day-to-day duties. Nonprofits tend to have core stakeholders being individual recipients, groups such as smaller, less-organized community soup kitchens or outreach, or more fully formed groups supported and in collaboration with local and state governments (Knox and Gruar 2007). Donors to the nonprofits and core stakeholders are often distinguished as external stakeholders (Van Puyvelde, Caers, Du Bois, and Jegers 2012). In addition, research suggests that the closeness of the bond between the potential whistleblower, those being reported, and the severity of the act influence whistleblowing intention (King 1997). Yet, King's (1997) research also finds that participants follow the chain of command in whistleblowing and calls for research on industry effects and firm type. Rothschild (2013) researches retaliation for whistleblowing by firm type, finding no significant difference. Still, there is an overall lack of investigation of the actual intention to whistleblow within the nonprofit community.
While nonprofit organizations do have external core stakeholders in the form of donors, from an employee perspective, the links between performance and donor activity are often not as explicit as the ties between performance and investor activity within for-profit organizations. Additionally, the enforcement culture codified within the for-profit community and the relative lack of studies about whistleblowing decision making within the nonprofit community lend evidence toward for-profit organizations being more likely to experience whistleblowing, despite the effects of retaliation. This evidence, combined with whistleblowing theory and stakeholder salience, suggests the second hypothesis:
Employees of for-profit organizations will be more likely to whistleblow than employees of nonprofit organizations.
Organization Magnitudes and Organization Type
Given the conflicting evidence about fraud reporting based on organizational size alone, it seems prudent to think about the effect that magnitude might have if combined with another characteristic of the organization. Despite the lack of information regarding nonprofit whistleblowing, inherent differences in the types of for- and nonprofit organizations seem likely to result in different decision making by their employees. For example, smaller nonprofit firms often have fewer filing requirements (McCarthy, Shelmon, and Mattie 2012) and less conservative financial reporting demands from external stakeholders (Krishnan et al. 2006). Conversely, larger (Vermeer et al. 2014) and international (Radebaugh, Gray, and Black 2006) for-profit organizations often experience high levels of regulation and increased financial reporting quality resulting from stakeholder demands. As a result, the combination of stakeholder salience theory and whistleblowing theory suggests that a large for-profit organization will be the most likely to have whistleblowers. The reverse is also likely true: small, nonprofit organizations, given the lack of resources and the desire to avoid this type of publicity, will be the least likely to have whistleblowers. However, as nonprofits become international in geographic reach, it is likely they will increasingly be structured similarly to a for-profit in terms of whistleblowing policy. This idea can be formally stated as the third hypothesis:
The effect of organization magnitude on employee intention to whistleblow is stronger within nonprofit organizations than within for-profit organizations.
III. RESEARCH METHOD
We utilize a survey in which (n = 153) fulltime workers answer questions about the organization for which they work and their likelihood of whistleblowing. We believe this is important for two reasons. First, there is a surprising lack of survey- and experiment-based whistleblowing literature measured from populations comprised of the people performing jobs related to those examined in this study. Second, we believe work experience is necessary to evoke the rich, real-world experience our participants have in order to more accurately measure real responses. The variables of interest are organization type (for-profit versus nonprofit) and magnitude (measured as both the size based on number of employees and geographic reach of the organization). We present both asset misappropriation and financial statement fraud scenarios. We randomize the order of presentation to control for information bias introduced through one method alone. We alter each scenario to further differentiate each and reduce potential information carryover bias. We ask the same questions of all participants because both types of fraud occur in all organizations. Few studies incorporate both types of fraud in their design. Fewer still look at differences in decisions that might result given each type of fraud in for-profit and nonprofit organizations. We pre-tested data on members of both communities and collected responses and feedback.
Because actual whistleblowing is challenging to observe, whistleblowing intent is often used as a proxy, or a prelude, to the decision process (Taylor and Curtis 2010). Once intent is established as a proxy, some researchers conclude that students are suitable proxies for capturing these data (Sonnier, Lassar, and Greene 2016). However, when significant numbers of such students lack meaningful experience to inform their decisions, they are not capable of measuring costs such as the likelihood of senior management's actions and potential personal losses. Thus, student whistleblowing data are often twice removed from the source of the study. In studies that do use students as proxies, those students who have some experience tend to be most studied (Brennan and Kelly 2007; Brink et al. 2015a). In studies that use students without experience, the author tends to explicitly acknowledge the results might not hold for those with experience in the workplace (Elias 2008). Given these findings, it is unsurprising that there is a lack of literature surrounding a combination of nonprofit and for-profit employee whistleblowing decision making.
As a result, we chose to use individuals with industry experience instead of students. A large survey company recruited participants on our behalf.2 To participate in the survey, participants were required to be at least 18 years of age, have worked for their current employer for at least one year, be employed fulltime, and be currently employed by either a for-profit or nonprofit organization. Selected demographic information can be found in Table 1. Participants are 71 percent female and are 46 years old (mean age). Further, 73 percent of participants hold a bachelor's or master's degree. On average, they have worked fulltime for 20 years and have worked for their current employer for ten years.3
Materials and Procedures
A large survey company sent an email to individuals who, based on information in their database, potentially qualified for participation in this study. Participants accessed the online survey through a link. Participants answered qualification questions about their employment. Those who passed proceeded to respond to the rest of the questionnaire. Participants were asked more questions about their employment, were asked to indicate their likelihood of whistleblowing in both a financial statement fraud scenario and an asset misappropriation scenario, answered attention check questions about the type of fraud, and then finished with demographic questions.
Detailed variable definitions are included in Appendix A.
Organization type is a measured variable with two levels: for-profit or nonprofit. Type is determined by participant responses to a question about the type of organization in which they are currently employed in a fulltime capacity.
Organization magnitudes is also a measured variable. We measure magnitude in two different ways that are independent of each other. First, we look at the size based on number of employees. In order to increase the comparability of our work within literature, we use the European Commission (EC 2017) definition4 to delineate large (250 or more employees) organizations from the rest of the sample. Further, this same definition denotes small and micro categories as having less than 50 employees. Accordingly, for purposes of analysis, we perform measurement splits at the 250 employee and the 50 employee levels. Thus, large (non-small) can be measured as 250 or more employees and as 50 or more employees to determine if differences exist. Second, we examine the geographic reach of the organization as being either local, state, regional, national, or international. Geographic reach splits are determined by participant responses to questions about the reach of the organization by which they are employed.
The primary dependent variable is intention to report to the whistleblowing website. Specifically, participants respond to the question “How likely are you to report the discovery to the whistleblowing website?” on a scale of 1 to 7, anchored with “extremely unlikely” and “extremely likely.” We utilize a website as a whistleblowing medium due to its increasing prevalence in the marketplace (Lowry, Moody, Galletta, and Vance 2013). Further, recent data internationally and from the ACFE indicate that tip reporting through a website is both preferred to and more common than by telephone (Henricksson 2017; ACFE 2018). As whistleblowing to the SEC is not an option for nonprofits, it was not listed as a whistleblowing option for nonprofit participants.
Tests of Hypotheses
Multiple statistical tests are conducted to analyze our hypotheses and inform our results. We start with an analysis of variance (ANOVA)5 where we analyze organization type (for-profit or nonprofit) and magnitude as measured by size based on number of employees (small or large) (Table 2). We then conduct a linear regression where we analyze organization type (for-profit or nonprofit) and magnitude as measured by geographic reach (local, state, regional, national, and international). We recode the categorical geographic reach variable to dummy variables and use the local category as the reference category for the analysis (Table 3). All participants received both financial statement and asset misappropriation scenarios. Supplemental analysis indicates no significant differences in whistleblowing intentions between the financial statement fraud and asset misappropriation scenarios. Thus, the following results are presented for responses to the financial statement fraud scenario. No participant responses were dropped as a result of this analysis. All results are presented with two-tailed p-values.
The first hypothesis states that as organization magnitude increases, employees will be more likely to whistleblow. Magnitude is first measured by size being small or large, based on the number of employees being less than 250 compared with 250 or more employees. Panel A of Table 2 indicates that size significantly predicts intention to whistleblow (F = 12.072; p < 0.001). Panel C of Table 2 shows that individuals at large firms are significantly more likely to whistleblow (M = 5.22, SD = 1.948) compared to those at small firms (M = 4.08, SD = 1.994). Magnitude is then measured by the organization's geographic reach. State, regional, national, and international are separately included in the regression where local is the reference category. As seen in Panel A of Table 3, state (t = 1.964; p = 0.026), regional (t = 1.697; p = 0.046), national (t = 2.836; p = 0.003), and international (t = 4.029; p < 0.001) all significantly predict intention to whistleblow. The data and results generally support the first hypothesis.
The second hypothesis states that for-profit employees will be more likely to whistleblow than will nonprofit employees. While the ANOVA in Table 2 indicates that organization type alone is not a significant predictor of intention to whistleblow, when controlling for local organizations, the regression in Panel A of Table 3 indicates that organization type does significantly predict intention to whistleblow (t = 2.528; p = 0.006). The data and results are consistent with expectations. The data indicate that for-profit organizations are 0.65 points (on a seven-point Likert-type scale; untabulated) higher than nonprofit organizations in whistleblowing intention when controlling for local organizations.
The third hypothesis states that magnitude and organization type will interact, and the effect will be strongest for nonprofit organizations. When considering the effect of size, the ANOVA in Table 2 indicates that this interaction is significant (F = 3.856; p = 0.026). Further, the regression in in Panel A of Table 3 shows that the interaction using geographic reach is significant for international (t = −4.622; p < 0.000). Panel B of Table 3 displays the regression results when the data are split by organization type. When interpreting the regression results, it appears that the international reach respondents drive the interaction effect for both for-profit as well as nonprofits.
The data suggest those in international nonprofit organizations have significantly lower intentions to report compared to those in local nonprofit organizations, whereas those in international for-profit organizations have significantly higher intentions to report than those in local for-profit organizations. Panel D of Table 4 shows the means and standard deviations of the for-profit and nonprofit results in each geographic reach category. The international reach nonprofit participant mean reporting was unexpectedly low. As such, international reporting participants were explored in greater detail in the supplemental analysis section. Overall, the number of employees appears to have more explanatory power in determining intention to whistleblow in both for-profit and nonprofit organizations.
Because fraud type might be driving these differences, we further investigate the difference between responses to the financial statement fraud and asset misappropriation scenarios. Further, because size initially seems less explanatory for nonprofits than for for-profits, we explore different size groupings. Table 5 indicates that the influence of size differs for those in for-profit organizations compared to those in nonprofit organizations when measuring the intention to whistleblow. In particular, those in smaller (n < 250) for-profit organizations are significantly (p < 0.001) less likely to whistleblow than are those in large (n ≥ 250) for-profit organizations. Conversely, those in nonprofit organizations show no significant (p > 0.10) differences in their intention to whistleblow, regardless of fraud type when size is defined as having 250 or less versus more than 250 employees.
However, due to their nature, nonprofit organizations generally tend to be smaller. Thus, we further examined the influence of size as measured at the 50 or less versus more than 50 employee level. Once again, in for-profit enterprises size is significant (p < 0.01). Those in larger organizations are significantly more likely to whistleblow than those in smaller organizations. Yet marginal significance (p < 0.10) also exists to suggest size may matter to nonprofit organizations as well. Those in nonprofits with 50 or more employees were an entire category (greater than one-point on the seven-point Likert-type scale) of increased likelihood compared to those with less than 50 employees.
While the overall results of a positive association of geographic reach and whistleblowing hold for those in nonprofits, international nonprofits again reduce the effect of the overall result for those in nonprofits with n ≥ 50 employees. Table 4, Panel D displays the results of our analysis. Delving deeper into the data, the finding that those in international nonprofit organizations where n ≥ 50 are significantly less likely to report than those in local nonprofit organizations appears at odds with the overall finding of whistleblowing intentions increasing for those in n ≥ 50 nonprofit organizations (p = 0.001). Given the majority (n = 11 out of n = 13 total) of our international nonprofit participants are in n ≥ 50 size organizations, we examine this subset to find reasons why. International participants by sector (Table 4, Panel E) were examined to understand the responses. Panel E of Table 4 shows that the unexpectedly low reporting mean for the international nonprofit participants is driven by participants in the Education (n = 6; M = 2.00) and Other (n = 1; M = 1.00) sectors. The variation from expectation in the international nonprofit mean reporting seems to be driven by these sectors.6
Finally, the influence of fraud type on fraud reporting as measured across a variety of sizes and industries, except for cases limited to international organizations, was largely statistically insignificant. Thus, these results provide support to previous research suggesting fraud type may be less important than the act of fraud itself when it comes to the intention to whistleblow.
Overall, the results of the supplemental analysis support the earlier findings. Both organization type and organization magnitude significantly predict the intention to whistleblow. However, it is important to note that nonprofits may need to key in on EC delineations for “small” firm size (splitting the data at the 50 employee mark) when studying small versus large decision making. Finally, many of the interactions remain consistent, and they are exemplified by the results presented in Panels A and B of Table 5.
All demographic variables were tested as potential covariates. No covariates were found to be statistically significant in the models and, as such, were not included in further analysis.
V. LIMITATIONS AND CONCLUSION
This study has five primary limitations. First, this investigation uses a survey presenting a hypothetical scenario. Despite using real fulltime employees who are able to rely on their lived experiences to inform intent and behavior, the study does not capture actual decisions as they occur. While this study uses fulltime employees, as opposed to students or hypothetical employees, it is possible that other unmeasured characteristics of our sample influence our findings. Second, our population is 85 percent female in the nonprofit sector we sampled (compared to 75 percent representation in the overall U.S. nonprofit sector) and is skewed toward experience, education, United States citizens, and local and international organizations. Despite including and testing these variables as controls, it is possible these skews make the study less generalizable. However, this skew is aligned with the design of this the study, which is intended to discover differences in the United States for nonprofit whistleblowing. Third, while this study has sufficient power on which to perform an analysis of the data, future archival research confirming these findings would be useful to cross-validate the results. The data indicate counter-intuitive results from the small, international nonprofit sector. This may be due to differences between the Education sector and Others. However, further research is necessary to determine if these differences hold in a larger sample. Fourth, participants are given both fraud scenarios. Despite there being large differences in the materiality and severity of both scenarios, our participants reacted similarly. However, the presentation order was randomized, and no statistical significance was found between the ordering and presentation. Analysis by industry is also consistent. It is possible the mere presentation of details created unconscious bias in ways that could not be captured. A final limitation is related to the ACFE data that were used to identify the magnitude of whistleblowing and fraud in various organizations. The ACFE conducts a biannual survey of its members and, as such, captures fraud information that may not be reflective of the population at large. This study does attempt to control for these limitations through both rigorous design and data analysis.
This study is the first to look at the differences between fulltime employee whistleblowing behavior differences between small and large for- and nonprofit organizations. In alignment with the ACFE (2018) report, the data suggest that employees of for-profit organizations are more likely to whistleblow than are employees of nonprofit organizations. Further, as magnitudes (measured both as size measured by number of employees and geographic reach of the organization) increase, reporting generally increases as well. These findings are consistent with our hypotheses and the line of research suggesting that larger, for-profit international firms tend to be more regulated, organized, and financed to support the whistleblowing process. However, differences occur in international nonprofit organizations (being less likely to report) and within nonprofits when size is measured by number of employees splitting at 50.
As a result, we extend literature in four important ways. First, we contribute to whistleblowing literature regarding size and type. Previous research is conflicted in respect to size and understudied concerning type. This study helps resolve these issues in important ways. The data suggest that type alone may be insufficient to reach conclusions regarding whistleblowing differences; however, when type is combined with size, data from the results help to resolve these differences. Further, we contribute to the overall domain of nonprofit whistleblowing, a domain that is lacking.
Second, we extend stakeholder salience theory to explicitly consider differences in organizational type given similar situations, establishing additional considerations for information asymmetry to be applied within stakeholder salience theory. Third, we expand research on nonprofits relating to geographic reach. Fourth, we discover key similarities (and differences) in how fraud magnitudes influence whistleblowing based on organization type.
Irvin (2005) suggests that removal of state registrations and regulations relating to nonprofit organizations will not likely affect fraud, since the nonprofits in the sample behaved similarly regardless of regulation. Yet our data suggest differences in fraud reporting intentions within the nonprofit group. Further, we observe differences when comparing with for-profit counterparts, suggesting regulation may at least have a role to play in fraud reporting. We further find that, without magnitudes, the story appears incomplete. Because those at small firms or firms with less geographic reach may have less regulatory scrutiny, less anonymity, and/or less standardization, any of these factors could lead to an increased perception of personal risk and cost in whistleblowing. Thus, studies that examine characteristics associated with firm magnitudes resulting in differing whistleblowing behaviors would be useful to expand knowledge of how other theories intersect. For example, a study looking at magnitudes in firms where the board used either employee count or organizational geographic reach to determine how to audit and/or design different whistleblowing pathways would be useful.
The issues relating to the intersection of organizational type and magnitudes can result in information imbalances in the real world that favor owners and donors when the organization is small or has little geographic reach but favors boards and shareholders when the organization is large or public. As a result, future research delineating and measuring the effects of magnitudes and type would be useful to suggest changes nonprofit and smaller firms might adopt. It is possible that those in nonprofit firms are invested in the mission of the organization and are willing to overlook problems in order to create what they see as a greater good. A study exploring public service motivation would be useful in expanding this idea. Further, a study looking at differences in for- and nonprofit fraud and whistleblowing would be informative about whether internal controls have any influence on nonprofit whistleblowing. Whistleblowing literature in accounting has several large gaps that remain unaddressed in the nonprofit community.
As the analysis deepens into geographic reach, the results become even more intriguing. For example, those in our study appear more likely to whistleblow at nonprofit organizations that are at the local, regional, or national level. Additional data should be collected to see if these effects are significant. While some measure of endogeneity might be shared between small firms and local firms, there is a relative consistency of the results across both dimensions of magnitude. It is possible that small, local for-profit organizations have no whistleblowing policies, or, if they do, such policies might be viewed as ineffective if the design and controls are all under the direction of one person. Nevertheless, this is an area without much research or regulation—small, local for-profit organizations may have one of the largest opportunities to commit fraud in our society. Future research that examines geographic scope may find interesting differences in the actions of employees who are contemplating whistleblowing. Further, research investigating the lack or inclusion of whistleblowing and regulatory requirements by geographic reach on the intention to misreport might be useful to help literature begin to address this missing dimension. Further, practice would be helped by knowing the extent to which changes in the types of whistleblowing required might influence employee behavior.
Bradley (2015) suggests the presence of engaged employees will result in lower levels of fraud. Given key findings from this study suggesting a tendency to whistleblow in international for-profit organizations, but a tendency not to whistleblow in international nonprofit organizations, we wonder if employee engagement is somehow being lost as reach increases for nonprofits, specifically in the Education sector. Perhaps employees can see direct results for themselves between what they do and the impact they make locally. Yet, this distinction may not be as pronounced if the work they do impacts people and places they may never see. If so, engagement may be lower internationally in nonprofits, resulting in decreased whistleblowing intentions compared to those in international for-profit organizations. Future research should consider exploring this idea.
It is also possible that due to the scale and geographic reach of international nonprofit firms, employees feel overwhelmed and are striving to take care of local affairs. International nonprofits in the Education sector may need to focus on ensuring consistency in whistleblowing policies and in stressing the importance through written policies and procedures at the local level. Whistleblowing literature would be well served by studies seeking to understand the factors and overarching theories that might result in international education organization employees failing to act.
Supplemental analysis suggests whistleblowing intention differs significantly for those in nonprofit organizations based on the number of employees in the firm. The data are consistent in suggesting employees in larger (both n ≥ 50 compared to n < 50 and n ≥ 250 compared to n < 250) for-profits being more likely to whistleblow. Yet, for nonprofits, the results are more nuanced. There is a significantly higher intention for those in larger nonprofits to whistleblow, but only when comparing size based on the EC definition of “small” (n ≥ 50 compared to n < 50). This suggests the majority of nonprofits in the United States, whose size is less than 50 employees and whose leader may have strong influence, may be helped by the creation of a nationalized, anonymous website available to small nonprofits.
Finally, supplemental analysis suggests a lack of significant difference in the effects of fraud type. Thus, our data offer support to the idea that researchers may be able to study either type of fraud effectively across organizations. This is unexpected, given stakeholder salience theory would suggest differences in response because the magnitude of the loss differs dramatically between the two fraud scenarios. Further, literature implies the differences due to the number of employees and geographic reach should be more dramatic across scenarios.
The data suggest differences in the behavior of employees at each organization type. In particular, those at smaller firms may consider adopting some of the standardized practices employed at the larger organizations in their area. If processes are not truly external and anonymous, the processes could be changed to be such. Our participants generally preferred whistleblowing through an external, anonymous website. However, pilot study data suggested people prefer a variety of whistleblowing outlets. Thus, while having an anonymous website would likely be the most useful for organizations that are lacking one, ensuring that additional channels are available would increase the odds that an employee reports malfeasance when discovered. Given stakeholder differences in nonprofits versus for-profits, the data suggest it may be most useful to adapt policies from international for-profit peers redesigned with international nonprofit stakeholders and employees in mind.
APPENDIX A Variable Definitions
Dependent Variable = “How likely are you to report the discovery to the whistleblowing website?” Participants indicated their response on a Likert-type scale where 1 represented extremely unlikely and 7 represented extremely likely.
Organization Type = whether the participant is currently employed fulltime by a for-profit (coded as 0) or nonprofit (coded as 1) organization.
Size = whether the participant is currently employed fulltime by an organization with fewer than 10 employees, 10 to 49 employees, 50 to 249 employees, or 250 or more employees. Table 2, Panel C splits size on 250 employees and Table 2, Panel D splits size on 50 employees.
Geographic Reach = whether the participant is currently employed fulltime by an organization with a local, state, regional, national, or international presence. Local is the reference group, and thus is an excluded variable.
The ACFE provides a biannual survey to its members that asks them to report on fraud situations that they have encountered. As nonprofits are less likely to employ a CFE due to the financial costs associated with an investigation, as well as a desire to avoid publicity that may impact future donor behavior, the statistics provided by the ACFE regarding fraud at nonprofit organizations are likely understated.
The experiment was approved by the Institutional Review Board (IRB) for Human Participants at the university where online administration of the study occurred.
A survey company was paid to collect respondents. From this, 567 surveys were sent and started, and 234 surveys were completed by participants passing the qualifications. As the conducted study was a survey and not an experiment, we did not have manipulations and thus no manipulation checks. However, participants were asked to answer questions related to the information they read to gauge attention. To check that participants were attentive to the information provided about the financial statement fraud scenario, participants were asked, “What type of unethical act was occurring in the previous scenario?” The options were theft of assets, corruption, and financial statement fraud. Forty-four of the 234 participants did not provide the correct answer of financial statement fraud. To check that participants were attentive to the asset misappropriation scenario, participants were also asked, “What type of unethical act was occurring in the previous scenario?” The options were false invoices, corruption, and financial statement fraud. Fifty-three of the 234 participants did not provide the correct answer of false invoices. In total, 81 participants failed one of the checks (65 failed one, 16 failed both) and were removed from the results. This result is consistent with prior research: DeZoort, Houston, and Hermanson (2003) report a 36 percent check failure rate, and Bishop, DeZoort, and Hermanson (2017) report a 44 percent failure rate.
For further information regarding the details and process of the background of the EC small and large categories, please see Schmiemann (2008).
To assess the influence of the abnormally low responses obtained in the international nonprofit category, Tables 2 and 3 were analyzed excluding the seven abnormal responses. The results displayed in Panels A and B of Table 2 and Panel A of Table 3 did not change significantly for size in Table 2 or reach in Table 3.