Executive Religiosity and Disclosure Tone Ambiguity of Annual Reports

Abstract

This paper examines the effect of C-suite executive religiosity on the disclosure tone ambiguity of corporate annual reports. The paper utilizes executive-level religiosity, disclosure tone, and financial data from a sample of 2515 publicly listed U.S. corporations. It applies fixed-effect regression analysis to show that the presence of religious executives within the C-suite team reduces the disclosure tone ambiguity of annual reports, as evidenced by a reduction in the number of negative and uncertain words within corporate annual reports. Subsample analyses show that religious CEOs and CFOs in the C-suite primarily drive the main findings, which is consistent with their heightened control over corporate annual report preparation processes post-SOX. The main finding holds across multiple robustness tests and suggests that the individual religiosity of C-suite executives can be an important determinant of a company’s disclosure tone-related choices. By utilizing the measure of executive-level religiosity, this study directly addresses recent calls for further research to examine additional personal and psychological factors beyond executive-level narcissism and political ideology that can influence top management personnel’s corporate disclosure tone-related choices. This study contributes to the literature by examining the influence of individual executive-level religiosity on the tonal sentiment of corporate communications, as represented by corporate annual reports.

1. Introduction

Within the social sciences, religiosity has always been a prominent topic of multidisciplinary academic interest. This interest primarily stems from the historical influence of religiosity on human rationale (Anderson, 1988). This influence generates a wide array of ethical, moral, political, psychological, and socio-economic implications for how modern-day social systems, and the individual components within those social systems, function (White et al., 2019). Hence, in the corporate world, religiosity can be an important factor affecting corporate decision-making, risk-taking, and choices. Prior research in various subdisciplines of the business literature has documented religiosity’s influence on overall corporate practices. In light of religiosity’s influence on human rationale and corporate practices, our study examines how the individual religiosity of C-suite executives affects corporate disclosure tone ambiguity.

The extant literature on disclosure tone has shown that the tone of corporate disclosures has important capital market implications for corporate stock performance, operating performance, and corporate risk-taking (Loughran & McDonald, 2011, 2016; Davis & Tama-Sweet, 2012; Bassyouny et al., 2022). This has created a substream of research that examines the various managerial characteristics affecting corporate disclosure tone beyond the traditional firm- and industry-level variables (Luo & Zhou, 2020; Bassyouny et al., 2022). Most of these studies focus on demographic, physical, and psychological managerial characteristics such as age, gender, narcissism, political ideology, etc., which affect corporate disclosure tone (Bochkay et al., 2019; Marquez-Illescas et al., 2019; Luo & Zhou, 2020; Arikan et al., 2023). Even though religiosity has been recognized as a credible psychological factor with a plethora of ethical, moral, and socioeconomic implications for corporate practices, research examining the impact of C-suite religiosity on any disclosure tone attribute is lacking. As a result, this research attempts to fill this gap. In the process, it addresses the findings of Bassyouny et al. (2022), who call for further research into additional personal and psychological factors of C-suite leaders, beyond the factors of narcissism and political ideology, which can influence their disclosure tone-related choices.

Our study’s main findings suggest that the individual religiosity of C-suite executives has important implications for a firm’s disclosure tone. Specifically, the presence of religious executives within the C-suite reduces the ambiguity levels of disclosure tone sentiment in corporate annual reports, as evidenced by a reduction in the number of negative and uncertain tonal sentiment-bearing words in corporate annual reports (10-K Reports). Additional subsample analysis shows that our main findings are primarily driven by the presence of religious CEOs and CFOs in the C-suite team. This is consistent with the central role of CEOs and CFOs in the post-SOX financial reporting and certification processes. The findings hold across multiple robustness tests involving entropy-balanced regressions. Overall, the results show that the religiosity of C-suite executives can have important implications for corporate decision-making and risk-taking in terms of how the tone of information appears in corporate annual reports.

This study contributes to the literature by expanding the extant corporate disclosure literature and by exploring a unique dimension within the broader financial accounting and corporate finance literature. Specifically, we show that the individual religiosity of C-suite executives can be an important determinant of the disclosure tone and tone ambiguity in corporate annual reports. This novel contribution expands the scope of the extant disclosure tone literature and creates a meaningful connection with the extant religiosity literature in business. Our study also addresses previous calls for further research to examine new additional psychological factors, such as executive-level religiosity, which can influence corporate disclosure tone choices beyond factors like executive-level narcissism and political ideology (Bassyouny et al., 2020; Luo & Zhou, 2020; Bassyouny et al., 2022; Arikan et al., 2023).

Additionally, our study contributes to the recent stream of literature on religiosity in business that studies the impact of individual religiosity within the C-suite on corporate decision-making and risk-taking. These studies narrow down a relatively broad location-based religiosity perspective to a more focused individual religiosity-based perspective (Baxamusa & Jalal, 2016; Nazrul et al., 2022). This approach helps address the direct effect of C-suite executives’ religiosity on corporate decision-making, as opposed to examining the religiosity-based effect indirectly with a local community-based lens (Baxamusa & Jalal, 2016; Chen et al., 2022; Nazrul et al., 2022). Additionally, the findings of prior location-based religiosity research tend to be primarily driven by the majority religious denomination within a local vicinity. This can create some noisy proxies, particularly in corporate organizations led by executives who belong to a minority religious denomination or executives who are not affiliated with any religious group. Hence, drawing interpretations that do not factor in the presence of individual-level religiosity within C-suite teams can create potential misclassification issues (Baxamusa & Jalal, 2016; Nazrul et al., 2022). Therefore, by examining the impact of direct and “cleaner” individual-based measures of religiosity on corporate decision-making and the disclosure tone in annual reports, this study contributes to emerging research within the broader religiosity-related literature in business. Also, by using the subset of BoardEx data containing publicly available information on executive-level religiosity and religious affiliations, our study expands the scope of use and interpretation of this unique data stream.

The remainder of the paper is organized as follows: Section 2 presents the background literature and hypothesis. Section 3 presents the materials and methods used. Section 4 discusses the research results. Section 5 presents our conclusions, study limitations, and future research.

2. Background Literature and Hypothesis

2.1. Religiosity Literature in Business

Most papers within the field of religiosity literature in business are typically clustered within the accounting and finance domain, and utilize location-based religiosity measures rather than individual ones (Nazrul et al., 2022). The overarching theme of prior research focuses on the existence of a positive association between religiosity and firm-level metrics of performance, value, and corporate quality practice (Nazrul et al., 2022). Firms operating in highly religious communities tend to have lower debt-financing costs, higher-quality accruals, less opportunistic earnings management, lower litigation risk, reduced audit fees, and less frequency of receiving going concern audit opinions (Grullon et al., 2009; Leventis et al., 2018; McGuire et al., 2012; Jiang et al., 2018; Omer et al., 2018). Firms operating in highly religious communities are less risk-centric and are likely to maintain long-term firm value while maximizing shareholder wealth (Hilary & Hui, 2009; Al Rahahleh et al., 2019). Similar positive findings have been reported in studies that examine the effect of C-suite executive religiosity on corporate decision-making, risk-taking, financial reporting quality, conservatism, disclosure readability, and overall corporate practice quality (Brenner, 2015; Baxamusa & Jalal, 2016; Cai et al., 2019; Iguchi et al., 2021; Chen et al., 2022; Nazrul et al., 2022; Zuo et al., 2022).

Cumulatively, the overwhelmingly positive findings of the location-religiosity-based and individual religiosity-based extant papers can be attributed to a broader reflection of the intrinsic dimension viewpoint of religiosity, which serves as our grand theory. The community-level social pressures that encourage firms to engage in value-enhancing corporate practices can be attributed to a form of herding behavior that the religiousness of their local communities creates (Cialdini & Goldstein, 2004). In other words, this represents the effect of the symbolization aspect of religious moral identity that the intrinsic dimension viewpoint of religiosity elucidates, encouraging the organizations within the community to be more socially responsible (Vitell et al., 2009). On the other hand, the value-enhancing and transparent corporate actions stemming from the individual religiosity of C-suite executives represent the internalization aspect of religious moral identity, as per the intrinsic dimension viewpoint of religiosity (Vitell et al., 2009). In other words, this highlights the sense of moral dedication and accountability that religiosity creates for humans at an individual level.

2.2. Disclosure Tone Research

The standard approach in the accounting and finance literature to measure the tonal sentiment of any document is to utilize word lists or dictionaries, where words representing a specific type of tonal sentiment are compared to the overall count of words within that document (Loughran & McDonald, 2011, 2014, 2016). Prior literature in the tonal analysis domain focuses on the tonal sentiment of mandatory and voluntary disclosures embedded in the annual reports or specific subsections of the reports, such as the MD&A section, the chairman’s statement or the president’s letter to shareholders, or other relevant sources of corporate information such as conference call scripts, press releases, news articles, and corporate announcements (Tetlock, 2007; Loughran & McDonald, 2011, 2020; Schleicher & Walker, 2010; Bassyouny et al., 2022; Alam et al., 2023).

The tone of corporate disclosures has been documented as having important, albeit mixed, capital market implications. A higher negative tonal sentiment in annual reports is generally associated with lower stock trading volumes, lower abnormal returns, higher stock volatility, higher analyst forecast dispersions, higher bankruptcy risk, and poor subsequent operating performance (Kothari et al., 2009; Loughran & McDonald, 2011, 2016; Davis & Tama-Sweet, 2012; Del Gaudio et al., 2020; Bassyouny et al., 2022). Conversely, a more optimistic tonal sentiment in annual reports is associated with higher stock trading volumes, higher stock returns, lower stock volatility, and lower analyst forecast dispersion (Kothari et al., 2009; Feldman et al., 2010; Loughran & McDonald, 2011, 2016; Davis & Tama-Sweet, 2012; Bassyouny et al., 2022). However, an abnormally positive tone and a higher amount of tonal uncertainty have been associated with greater corporate obfuscation and the use of impression management tactics (e.g., earnings management and positive framing), which increase the likelihood of litigation, fraud, and the issuance of going concern modified audit opinions (Rogers et al., 2011; Hossain et al., 2020; Bassyouny et al., 2022; S. Li et al., 2022; J. Li et al., 2024; Rezaee et al., 2024). Similarly, an abnormally negative tone has been used to deflate expectations, particularly in highly labor-unionized industries (Arslan-Ayaydin et al., 2021). Also, management could use such an abnormally negative tone to manipulate stock prices and engage in insider trading behavior on the purchasing side (Sandulescu, 2015; Choi, 2020; Xu & Qi, 2022; Cheng et al., 2023).

Several firm-level, industry-level, and managerial characteristic-based factors affect the disclosure tone of corporate annual reports. Firm-level factors include past performance, firm size, growth, volatility, CEO tenure length, and the degree of stock-based compensation, while industry-level factors include the affiliation of a firm with a known sin industry1, as well as the degree of labor unionization of the firm’s industry (Arslan-Ayaydin et al., 2016, 2021; Bochkay et al., 2019; Marquez-Illescas et al., 2019; Luo & Zhou, 2020; Bassyouny et al., 2022). From a managerial perspective, notable determinants of disclosure tone include managerial ability, age, gender, narcissism, and political ideology (Bochkay et al., 2019; Marquez-Illescas et al., 2019; Luo & Zhou, 2020; Yan et al., 2021; Arikan et al., 2023). In the specific psychological context, the religiosity of individual C-suite executives is mostly overlooked, yet it is a credible psychological trait that can influence the managerial decision-making process. Our study aims to address this gap, particularly given that a negative tone (barring abnormalities) and uncertain tone set forth by upper-level corporate leadership can have vastly unfavorable future corporate implications. Given the overwhelmingly positive implications of religiosity documented in the extant business religiosity literature, examining the prominent role played by individual C-suite religiosity in mitigating the level of negative and uncertain tone to manage future unfavorable corporate practice outcomes is the focus and motivation of this study.

2.3. Hypotheses

The extant theology and psychology literature often utilize the intrinsic dimension viewpoint of religiosity as a common way to explain how religious identities transform into life actions on an individual level (Donahue, 1985; Salsman et al., 2005; Vitell et al., 2009). This is attributable to the intrinsic dimension viewpoint of religiosity’s dual focus on internalization and the symbolization aspects of religion’s direct impact on moral identity, which is often overlooked by other religiosity-related theories (Vitell et al., 2009). The central argument of the intrinsic dimension viewpoint of religiosity underscores the role of religion in steering individuals to embrace a time-invariant form of morality that is not driven by extrinsic needs or opportunities (Middleton & Putney, 1962; Salsman et al., 2005; Hardy, 2006; Vitell et al., 2009). This time-invariant form of morality is intrinsically driven by an individual’s strict adherence to their religion’s core teachings, which are often viewed as sacred tenets (Middleton & Putney, 1962; Salsman et al., 2005; Hardy, 2006; Vitell et al., 2009). As the social impact of individual-level decisions and actions is a core focus of most religious teachings, the intrinsic dimension viewpoint argues that religious individuals are more socially conscious of the impact of their actions and decisions on their communities (Watson et al., 1984; Roccas, 2005; Salsman et al., 2005; Vitell et al., 2009; White et al., 2019). Given its robustness and prevalence as a theory, the intrinsic dimension viewpoint of religiosity has been mostly used in prior business research to examine the community-level or individual-level impact of religiosity on business decisions (Omer et al., 2018; Nazrul et al., 2022; Heubeck, 2024). Hence, we follow a similar approach by utilizing the intrinsic dimension viewpoint of religiosity as our grand theory.

According to the intrinsic dimension viewpoint, as community value maximization is at the forefront of most religious teachings (Johnson, 1959; Geyer & Baumeister, 2005; Salsman et al., 2005; Roccas, 2005; Vitell et al., 2009), the presence of individual religiosity in corporate leadership positions is likely to be associated with positive firm value outcomes and value-maximizing corporate practices. For instance, the prior literature reports a negative association between executive religiosity and lower financial reporting quality indicators (e.g., earnings management and fraud) in companies led by religious executives (Cai et al., 2019; Chen et al., 2022; Zuo et al., 2022). Also, a negative association can be found between executive religiosity and corporate risk aversion (Brenner, 2015; Baxamusa & Jalal, 2016). In contrast, positive associations have been reported between executive religiosity and annual report readability (Nazrul et al., 2022) and between executive religiosity and CSR activities (Iguchi et al., 2021).

Given the positive relationship between executive religiosity and firms’ value-maximizing corporate practice outcomes in the context of the tonal sentiment of corporate annual reports, we anticipate two broad outcomes. First, the presence of religious executives within the C-suite team is likely to reduce the number of negative tonal sentiment-bearing words in annual reports. This is because a high negative tonal sentiment in corporate documents has detrimental capital-market implications and adverse firm value outcomes (Kothari et al., 2009; Loughran & McDonald, 2011, 2016; Davis & Tama-Sweet, 2012; Del Gaudio et al., 2020; Bassyouny et al., 2022). Given the adverse long-term firm value implications, religious executives will likely put considerable effort into mitigating the negative tonal sentiment of the disclosures embedded within annual reports. This leads to our first hypothesis:

H1. 

Executive religiosity reduces the negative tonal sentiments of corporate annual reports.

The presence of religious executives within the C-suite is also likely to reduce the number of uncertain tonal sentiment-bearing words in annual reports. This is likely to be influenced by religious individuals’ social consciousness and their mindfulness of the impact of their actions on their communities, as emphasized by the intrinsic dimension viewpoint of religiosity. Driven by heightened levels of social consciousness and duty of care, religious individuals are more likely to undertake actions and make decisions, similar to the example of honorable merchants in ancient societies (Bott & Milkau, 2018). According to the intrinsic dimension viewpoint, this high duty of care that religious individuals are likely to display would, therefore, result in actions that are highly transparent, community-value-maximizing, and not purposefully ambiguous (Johnson, 1959; Geyer & Baumeister, 2005; Salsman et al., 2005; Roccas, 2005; Vitell et al., 2009; Bott & Milkau, 2018). A highly uncertain tonal sentiment in corporate documents has been associated with a higher likelihood of manipulative earnings management, earnings persistence issues, and fraud probes (F. Li, 2008; S. Li et al., 2022), along with a higher post-IPO volatility (Loughran & McDonald, 2013). Given that the collective impact of such tactics adversely affects a firm’s value, religious executives are likely to mitigate the uncertain tonal sentiment of the disclosures embedded in corporate annual reports. This leads to our second hypothesis:

H2. 

Executive religiosity reduces the uncertain tonal sentiments of corporate annual reports.

3. Materials and Method

3.1. Sample Selection

For this study, data were collected from publicly available sources, covering a sample period from 1999 to 2020. Individual religiosity data were collected from BoardEx. For the social organization memberships listed under individual BoardEx profiles, we only focused on the current or former religious organization memberships of C-Suite executives, as reported by Nazrul et al. (2022). Identifying religious organization memberships can be challenging because not all entities with religious-sounding names require their membership to have faith-based affiliations2. To address this issue, we used the GuideStar directory3, which provides a comprehensive list of religious organizations across prominent faiths, to manually collect keyword terms matching the religious social organizations listed in BoardEx. Then, using text-based machine learning and filtering algorithms, we refined the dataset to exclude religious-neutral organizations with religious-sounding names. After completing the data validation step, we ensured that our BoardEx data were cleaned to only select C-suite executives4. We then merged our cleaned BoardEx data with our 10-K disclosure tone and document characteristic control data, firm-level financial data, and location-based religiosity data. Disclosure tone and document characteristic-related controls were obtained from Loughran and McDonald’s website, which contains a summarized 10-K datasheet5. Firm-level financial data were obtained from Compustat, while location-based religiosity data were obtained from the Association of Religion Data Archives (ARDA)6.

Table 1. Sampling table, summary statistics, and Pearson correlations.

Panel A: Sampling Table
Details													Count
C-Suite Executives With Disclosed Religiosity													552
Total No. of C-Suite Executives													7409
CEOs With Disclosed Religiosity													494
Total No. of CEOs													6489
CFOs With Disclosed Religiosity													20
Total No. of CFOs													470
Total No. of Firm-Year Observations (n)													12,736
Total No. of Firms													2515
Panel B: Summary Statistics
	No. of Firm Year Observations (n)				Mean		SD		p25		Median		p75
Negative Tone	12,736				0.016		0.003		0.014		0.016		0.019
Uncertain Tone	12,736				0.012		0.002		0.010		0.012		0.014
Disclosed Religiosity	12,736				0.068		0.252		0		0		0
Gross File Size	12,736				14.978		1.463		13.906		14.722		16.382
No. of Exhibits in 10-K	12,736				10.705		5.071		7		11		14
Tangibility	12,736				0.269		0.237		0.085		0.187		0.392
Profitability	12,736				0.073		0.257		0.063		0.115		0.166
Size	12,736				6.904		2.092		5.422		6.968		8.289
Firm Age	12,736				24.346		17.086		10		20		36
Cash Flow Volatility	12,736				0.065		0.107		0.023		0.039		0.07
Book Leverage	12,736				0.252		0.282		0.065		0.219		0.367
Debt Issue	12,736				0.32		0.466		0		0		1
Capital Expenditure	12,736				0.048		0.058		0.016		0.031		0.058
SGA Expense	12,736				0.277		0.301		0.1		0.2		0.353
Location Religiosity	12,736				0.587		0.159		0.461		0.584		0.671
Panel C: Pearson Correlations
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)
Negative Tone (1)
Uncertain Tone (2)	0.727
Disclosed Religiosity (3)	−0.027	−0.042
Gross File Size (4)	0.363	0.555	−0.012
No. of Exhibits in 10-K (5)	0.514	0.493	0.003	0.648
Tangibility (6)	−0.039	−0.042	0.01	0.045	0.07
Profitability (7)	−0.063	−0.089	0.046	−0.014	0.011	0.132
Size (8)	0.262	0.301	0.038	0.328	0.341	0.17	0.37
Firm Age (9)	0.003	−0.032	0.047	0.234	0.21	0.066	0.179	0.419
Cash Flow Volatility (10)	0.002	−0.005	−0.04	−0.093	−0.097	−0.162	−0.598	−0.389	−0.189
Book Leverage (11)	0.162	0.154	0.036	0.16	0.14	0.194	−0.111	0.202	0.03	0.011
Debt Issue (12)	0.038	0.052	0.013	0.066	0.069	0.101	0.034	0.177	0.048	−0.06	0.226
Capital Expenditure (13)	−0.054	−0.018	−0.011	−0.035	−0.017	0.594	0.075	0.046	−0.049	−0.05	0.08	0.133
SGA Expense (14)	−0.097	−0.147	0.001	−0.13	−0.125	−0.239	−0.461	−0.416	−0.168	0.492	−0.016	−0.117	−0.075
Location Religiosity (15)	−0.049	−0.076	0.071	−0.04	−0.003	0.018	0.04	0.076	0.071	−0.03	0.054	0.037	−0.016	−0.007
(Source: Authors). Bolded values represent statistical significance at the 1% level.

’s Panel A shows our sampling table. In total, 12,736 firm-year observations, 2515 firms, and 7409 C-suite executives are covered in this study, as shown in Panel B. CEOs (CFOs) constitute 6409 (470) of the executives in our sample. Of the 7409 C-suite executives, 552 have identifiable/disclosed religious affiliations, of whom 494 are CEOs and 20 are CFOs.

Table 1’s Panel B shows that within our sample population of 10-K words, on average, 1.6 percent have a negative tonal sentiment, and 1.2 percent have an uncertain tonal sentiment, which is similar to the figures reported by Loughran and McDonald (2011). Panel B also shows that 6.8 percent of firm-year observations have disclosed religiosity at the C-suite level, which is close to the 6.6 percent figure reported by Nazrul et al. (2022).

The Pearson correlations from Table 1’s Panel C show that both negative and uncertain disclosure tones have a significant negative correlation with executive-level religiosity, supporting the subsequent findings within this study. While location religiosity has a significant positive correlation with C-suite religiosity, similar to that reported by Nazrul et al. (2022), the coefficient is only 7.10 percent, which is relatively modest. This modest correlation validates the argument that individual C-suite executives may act based on personal values stemming from the core themes of their religious teachings, which are not influenced by their community’s religiosity, as predicted by Hambrick and Mason (1984) and validated by Baxamusa and Jalal (2016) and Nazrul et al. (2022).

3.2. Methodology

We utilized the following baseline regression model:

Disclosure Tone = ß₁ Disclosed Religiosity + ß₂ Document Characteristic Controls + ß₃ Firm-Level Controls + ß₄ Location Religiosity + ε.

Disclosed Religiosity is an indicator variable that equals 1 if a firm is led by one or more religious C-suite executives in a certain year, as supported by their publicly identifiable religious organization affiliation data in BoardEx. The Document Characteristic Controls vector consists of the gross 10-K file size and a count for the number of exhibits within 10-Ks, which is used following the technique of Nazrul et al. (2022). The Firm-Level Controls vector includes standard measures for asset tangibility, profitability, firm size, firm age, cash flow volatility, book leverage, the new debt issuance dummy, capital expenditures, and SGA expenses, which were obtained based on other prior individual religiosity-based papers (Baxamusa & Jalal, 2016; Nazrul et al., 2022). Location Religiosity measures the overall number of religious individuals in a specific county relative to the total county population, based on the Association of Religion Data Archives (ARDA) data (Hilary & Hui, 2009). Detailed variable definitions are listed in Appendix A at the bottom. We added industry and year-fixed effects to all our regression specifications that follow the baseline regression specification.

4. Results

4.1. Overall Findings

Table 2. Impact of Executive Religiosity on Disclosure Tone.

Variables	Negative Tone	Uncertain Tone
Disclosed Religiosity	−0.021 ***	−0.017 ***
	(−3.011)	(−3.087)
Gross File Size	0.026 ***	0.017 ***
	(8.695)	(12.129)
No. of Exhibits in 10-K	0.014 ***	0.003 ***
	(29.129)	(9.461)
Tangibility	−0.019	−0.013 **
	(−1.315)	(−2.019)
Profitability	−0.041 ***	−0.040 ***
	(−3.632)	(−5.918)
Size	−0.003	0.009 ***
	(−0.984)	(13.070)
Firm Age	−0.010	−0.001 ***
	(−1.460)	(−13.207)
Cash Flow Volatility	0.079 ***	0.076 ***
	(2.699)	(4.390)
Book Leverage	0.049 ***	0.020 ***
	(5.607)	(4.663)
Debt Issue	−0.004 **	−0.002
	(−1.993)	(−1.108)
Capital Expenditure	−0.088 ***	0.044 *
	(−2.877)	(1.850)
SGA Expense	0.013	−0.005
	(1.291)	(−1.315)
Location Religiosity	−0.016	−0.004
	(−0.839)	(−0.697)
Constant	0.002	−0.134 ***
	(0.014)	(−6.460)
Industry FE	Y	Y
Year FE	Y	Y
Observations	12,736	12,736
Adj. R-squared	0.626	0.601

(Source: Authors). T-statistics are reported in parenthesis. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively.

examines the effect of executive-level religiosity on disclosure tone ambiguity, as measured by the proportional word-list counts of negative and uncertain tonal sentiment-bearing words relative to the total number of words appearing in a firm’s 10-K. Table 2 reports a highly significant and negative coefficient for Disclosed Religiosity for both the negative tone specification (coefficient = −0.021; t-statistic = −3.011) and the uncertain tone specification (coefficient = −0.017; t-statistic = −3.087). The coefficient from both specifications suggests that the presence of any C-suite executive with a disclosed religious identity reduces the corporate annual reports’ negative (uncertain) tonal sentiment by 2.1 percent (1.7 percent). The findings from both specifications support Hypotheses 1 and 2. This shows that the presence of religious executives within the C-suite team leads to an active firm-level effort to reduce the magnitude of the negative and uncertain tonal sentiments of corporate annual reports. This, in turn, helps protect and maximize future firm value as it mitigates the future detrimental capital market outcomes resulting from negative and uncertain tonal sentiments in corporate annual reports. The results hold true while controlling for Location Religiosity within both specifications, the coefficients of which are negative but insignificant. This shows that the firm’s headquarters location-based religiosity does not solely drive the impact of religion, which lends credence to the arguments presented by other individual religiosity-based papers in business (Brenner, 2015; Baxamusa & Jalal, 2016; Cai et al., 2019; Iguchi et al., 2021; Chen et al., 2022; Nazrul et al., 2022; Zuo et al., 2022). This also supports Hambrick and Mason’s (1984) upper echelons theory, which explains how the executives’ personal values and beliefs can predict their corporate actions and decisions. Our overall findings offer incremental evidence that the individual-level religiosity of executives affects corporate decision-making and mitigates the tonal sentiment ambiguity levels of corporate documents such as annual reports.

4.2. Subsample Analysis—Individual Executive Groups

In

Table 3. Impact of CEO Religiosity on Disclosure Tone.

Variables	Negative Tone	Uncertain Tone
Disclosed Religiosity	−0.024 ***	−0.018 ***
	(−3.022)	(−3.268)
Gross File Size	0.026 ***	0.017 ***
	(8.627)	(26.443)
No. of Exhibits in 10-K	0.014 ***	0.003 ***
	(29.215)	(15.553)
Tangibility	−0.018	−0.038 ***
	(−1.293)	(−10.978)
Profitability	−0.040 ***	−0.056 ***
	(−3.613)	(−12.544)
Size	−0.003	0.008 ***
	(−0.989)	(22.370)
Firm Age	−0.010	−0.001 ***
	(−1.482)	(−24.393)
Cash Flow Volatility	0.073 **	0.070 ***
	(2.407)	(6.897)
Book Leverage	0.050 ***	0.016 ***
	(5.618)	(6.387)
Debt Issue	−0.004 *	−0.005 ***
	(−1.949)	(−5.246)
Capital Expenditure	−0.089 ***	0.106 ***
	(−2.880)	(7.241)
SGA Expense	0.011	−0.027 ***
	(1.116)	(−9.517)
Location Religiosity	−0.019	−0.003
	(−1.004)	(−0.704)
Constant	0.005	−0.104 ***
	(0.030)	(−11.042)
Industry FE	Y	Y
Year FE	Y	Y
Observations	11,207	11,207
Adj. R-squared	0.624	0.543

(Source: Authors). T-statistics are reported in parenthesis. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively.

,

Table 4. Impact of CFO Religiosity on Disclosure Tone.

Variables	Negative Tone	Uncertain Tone
Disclosed Religiosity	−0.041 *	−0.016 **
	(−1.947)	(−2.247)
Gross File Size	0.034 ***	0.004
	(3.498)	(0.804)
No. of Exhibits in 10-K	0.012 ***	0.004 ***
	(4.833)	(6.731)
Tangibility	−0.038	−0.022
	(−1.005)	(−0.792)
Profitability	−0.066 *	−0.061 ***
	(−1.877)	(−3.691)
Size	0.012 **	0.015 ***
	(2.327)	(4.049)
Firm Age	−0.001	−0.001 **
	(−1.048)	(−2.670)
Cash Flow Volatility	0.069	0.058
	(0.975)	(1.282)
Book Leverage	0.053 *	−0.017
	(1.738)	(−0.848)
Debt Issue	−0.010	0.003
	(−0.962)	(0.476)
Capital Expenditure	0.091	0.100 **
	(0.875)	(2.677)
SGA Expense	0.002	−0.015
	(0.081)	(−1.523)
Location Religiosity	−0.004	−0.003
	(−0.076)	(−0.149)
Constant	−0.427 ***	0.002
	(−3.779)	(0.035)
Industry FE	Y	Y
Year FE	Y	Y
Observations	623	623
Adj. R-squared	0.527	0.466

(Source: Authors). T-statistics are reported in parenthesis. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively.

and

Table 5. Impact of Non-CEO & Non-CFO Religiosity on Disclosure Tone.

Variables	Negative Tone	Uncertain Tone
Disclosed Religiosity	−0.007	−0.012
	(−0.150)	(−1.568)
Gross File Size	0.051 ***	0.019 ***
	(6.560)	(5.224)
No. of Exhibits in 10-K	0.010 ***	0.002 ***
	(7.462)	(2.935)
Tangibility	−0.056	0.017
	(−1.555)	(1.016)
Profitability	−0.075 ***	−0.061 ***
	(−2.889)	(−5.168)
Size	0.008 **	0.011 ***
	(2.457)	(7.243)
Firm Age	−0.001 *	−0.001 ***
	(−1.824)	(−3.688)
Cash Flow Volatility	0.008	0.001
	(0.151)	(0.041)
Book Leverage	−0.048 *	−0.021 *
	(−1.959)	(−1.922)
Debt Issue	0.024 **	0.007
	(2.104)	(1.412)
Capital Expenditure	0.026	−0.067
	(0.256)	(−1.426)
SGA Expense	−0.029	−0.024 **
	(−1.387)	(−2.510)
Location Religiosity	−0.038	−0.011
	(−0.558)	(−0.702)
Constant	−0.574 ***	−0.146 ***
	(−5.427)	(−2.805)
Industry FE	Y	Y
Year FE	Y	Y
Observations	906	906
Adj. R-squared	0.624	0.494

(Source: Authors). T-statistics are reported in parenthesis. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively.

, we show three subsample analyses to examine which executive group within the management team drives the main results shown in Table 2. According to Section 906 of the Sarbanes–Oxley Act of 2002 (SOX), each corporate annual report containing periodic financial statements of public limited companies must accompany a written statement from their CEOs and CFOs certifying the content of the corporate annual report (Geiger & Taylor, 2003)7. Research has shown that CEOs have the most influence on corporate decision-making relative to other C-suite executives (Ke et al., 2019). However, in the context of financial reporting, CFOs also have an important oversight over the quality of earnings and financial data reported in corporate annual reports and the overall communication transparency of those numbers, particularly post-SOX (Bedard et al., 2014; Baker et al., 2019). Given the importance of CEOs and CFOs in the financial reporting and annual report certification processes mandated by the SEC (Geiger & Taylor, 2003), we gauged their individual religiosity’s impact on the disclosure tone of annual reports, as shown in Table 3 and Table 4. The cumulative impact of the individual religiosity of any non-CEO and non-CFO executive is depicted in Table 5.

Table 3 focuses explicitly on the CEO sample and reports a highly significant and negative coefficient for Disclosed Religiosity for both the negative tone specification (coefficient = −0.024; t-statistic = −3.022) and the uncertain tone specification (coefficient = −0.018; t-statistic = −3.268). The coefficients from both specifications suggest that the presence of any religious CEO with a disclosed religious identity reduces the negative (uncertain) tonal sentiment of corporate annual reports by 2.4 percent (1.8 percent). The findings from both specifications support Hypotheses 1 and 2. The findings also validate the extant findings of religious CEOs playing an influential central role in corporate decision-making and disclosure-related choices (Brenner, 2015; Baxamusa & Jalal, 2016; Cai et al., 2019; Iguchi et al., 2021; Chen et al., 2022; Nazrul et al., 2022).

Table 4 focuses on the CFO sample. In comparison to the CEO sample, although the statistical significance of the coefficients is slightly weaker, the coefficients for Disclosed Religiosity are statistically significant and negative for both the negative tone specification (coefficient = −0.041; t-statistic = −1.947) and the uncertain tone specification (coefficient = −0.016; t-statistic = −2.247). The coefficients from both specifications suggest that the presence of any religious CFO with a disclosed religious identity reduces the negative (uncertain) tonal sentiment of corporate annual reports by 4.1 percent (1.6 percent). The findings from both specifications support Hypotheses 1 and 2. The findings also validate the extant findings of CFOs playing an influential role in corporate disclosures and broader financial-reporting-related choices post-SOX (Geiger & Taylor, 2003; Bedard et al., 2014; Baker et al., 2019). They also show that CFOs’ input in the broader financial reporting process is heightened, particularly when they are religious.

Table 5 focuses on the non-CEO and non-CFO executive sample. Although the coefficient for Disclosed Religiosity remains negative for both the negative tone specification (coefficient = −0.007; t-statistic = −0.150) and the uncertain tone specification (coefficient = −0.012; t-statistic = −1.568), they are statistically insignificant, which contrasts with the overall sample and the CEO and CFO subsample of observations. This result reiterates the prominent roles of the CEO and CFOs in the financial reporting process (Geiger & Taylor, 2003). It revalidates the statistically significant results reported in Table 3 and Table 4 for the CEO and CFO subsample.

4.3. Robustness Tests

As shown in

Table 6. Impact of Executive Religiosity on Disclosure Tone (Entropy-Balanced Approach).

Variables	Negative Tone	Uncertain Tone
Disclosed Religiosity	−0.014 ***	−0.008 ***
	(−3.788)	(−4.710)
Gross File Size	0.004 **	0.018 ***
	(2.037)	(23.203)
No. of Exhibits in 10-K	0.012 ***	0.003 ***
	(15.002)	(10.162)
Tangibility	−0.007	−0.020 ***
	(−0.596)	(−4.155)
Profitability	−0.091 ***	−0.051 ***
	(−6.205)	(−7.104)
Size	0.013 ***	0.007 ***
	(8.821)	(10.032)
Firm Age	−0.001 ***	−0.001 ***
	(−9.442)	(−13.591)
Cash Flow Volatility	0.284 ***	0.096 ***
	(6.784)	(5.302)
Book Leverage	0.052 ***	0.018 ***
	(5.035)	(3.838)
Debt Issue	−0.012 ***	−0.003 *
	(−2.994)	(−1.918)
Capital Expenditure	−0.107 **	0.045 *
	(−2.080)	(1.861)
SGA Expense	−0.030 ***	−0.029 ***
	(−4.027)	(−8.479)
Location Religiosity	0.015	−0.001
	(1.315)	(−0.187)
Constant	0.005	−0.123 ***
	(0.159)	(−10.319)
Observations	12,736	12,736
R-squared	0.358	0.47

(Source: Authors). T-statistics are reported in parenthesis. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively.

,

Table 7. Impact of CEO Religiosity on Disclosure Tone (Entropy-Balanced Approach).

Variables	Negative Tone	Uncertain Tone
Disclosed Religiosity	−0.021 ***	−0.006 ***
	(−3.098)	(−3.440)
Gross File Size	0.037 ***	0.018 ***
	(8.339)	(22.455)
No. of Exhibits in 10-K	0.014 ***	0.003 ***
	(15.547)	(9.820)
Tangibility	−0.002	−0.025 ***
	(−0.044)	(−4.859)
Profitability	−0.060 **	−0.041 ***
	(−2.350)	(−4.978)
Size	−0.016 ***	0.006 ***
	(−2.736)	(7.967)
Firm Age	0.007	−0.001 ***
	(0.696)	(−13.932)
Cash Flow Volatility	0.124 *	0.109 ***
	(1.922)	(5.550)
Book Leverage	0.015	0.022 ***
	(0.770)	(4.399)
Debt Issue	−0.001	−0.005 **
	(−0.213)	(−2.523)
Capital Expenditure	−0.164 ***	0.055 **
	(−2.631)	(2.081)
SGA Expense	−0.047 **	−0.029 ***
	(−2.197)	(−8.273)
Location Religiosity	0.004	0.005
	(0.117)	(0.845)
Constant	−0.506	−0.125 ***
	(−1.633)	(−10.002)
Observations	11,207	11,207
R-squared	0.336	0.468

(Source: Authors). T-statistics are reported in parenthesis. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively.

and

Table 8. Impact of CFO Religiosity on Disclosure Tone (Entropy-Balanced Approach).

Variables	Negative Tone	Uncertain Tone
Disclosed Religiosity	−0.053 *	−0.013 **
	(−2.075)	(−2.093)
Gross File Size	0.020	0.005
	(1.626)	(0.852)
No. of Exhibits in 10-K	0.013 ***	0.004 ***
	(8.586)	(6.221)
Tangibility	0.043	−0.008
	(0.749)	(−0.320)
Profitability	−0.061 *	−0.052 ***
	(−1.830)	(−3.515)
Size	0.014 *	0.014 ***
	(1.779)	(3.860)
Firm Age	−0.002 *	−0.001 **
	(−1.897)	(−2.474)
Cash Flow Volatility	0.133	0.046
	(1.250)	(1.267)
Book Leverage	−0.015	−0.011
	(−0.506)	(−0.532)
Debt Issue	0.008	0.006
	(0.432)	(0.828)
Capital Expenditure	−0.087	0.034
	(−1.374)	(0.897)
SGA Expense	−0.004	−0.010
	(−0.125)	(−1.093)
Location Religiosity	−0.037	−0.048
	(−0.484)	(−1.398)
Constant	−0.214	0.031
	(−1.442)	(0.408)
Observations	623	623
R-squared	0.323	0.455

(Source: Authors). T-statistics are reported in parenthesis. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively.

, we used entropy balancing to validate our core findings from Table 2, Table 3 and Table 4. This robustness technique has been utilized in individual-religiosity-based accounting research, such as the work of Nazrul et al. (2022), and broader accounting and corporate finance-based research (Shroff et al., 2017; McMullin & Schonberger, 2020) to create a “balanced” comparison between the treatment and control groups. Entropy balancing achieves this by reweighting the observations in the control sample, such that the moments of the distributions of the matching variables for the reweighted control sample are indistinguishable from the moments of the distributions of these variables for the treatment sample (Hainmueller, 2012; Shroff et al., 2017). As the differences in covariates are adjusted before the regression analysis, the entropy-balanced regressions are, therefore, designed to attribute the observed differences in outcomes to the treatment or intervention rather than to the pre-existing differences among the groups (Hainmueller, 2012).

As shown in Table 6, we ran entropy-balanced regression on our overall sample, whereas we ran entropy-balanced regression specifications on our CEO and CFO samples, as shown in Table 7 and Table 8, respectively. Consistent with our earlier findings, the coefficient on Disclosed Religiosity is negative and significant for both the negative and uncertain tone specifications in all three tables. Hence, these findings revalidate the role of individual, executive-level religiosity on firm disclosure practices that mitigate the tonal ambiguity of corporate annual reports.

As shown in

Table 9. Impact of Executive Religiosity on Net-Negative Disclosure Tone.

	Overall Sample	CEO Sample	CFO Sample
Variables	Net Negative Tone	Net Negative Tone	Net Negative Tone
Disclosed Religiosity	−0.018 ***	−0.022 ***	−0.039 *
	(−2.973)	(−3.163)	(−1.919)
Gross File Size	0.021 ***	0.021 ***	0.027 ***
	(8.114)	(8.048)	(3.283)
No. of Exhibits in 10-K	0.012 ***	0.012 ***	0.010 ***
	(29.021)	(29.078)	(5.348)
Tangibility	−0.011	−0.011	−0.025
	(−0.897)	(−0.891)	(−0.634)
Profitability	−0.037 ***	−0.036 ***	−0.042
	(−3.796)	(−3.747)	(−1.228)
Size	−0.005 **	−0.005 **	0.008
	(−2.001)	(−1.984)	(1.521)
Firm Age	−0.008	−0.008	−0.000
	(−1.565)	(−1.586)	(−0.607)
Cash Flow Volatility	0.076 ***	0.072 ***	0.037
	(3.109)	(2.816)	(0.438)
Book Leverage	0.047 ***	0.048 ***	0.045
	(6.136)	(6.142)	(1.671)
Debt Issue	−0.004 **	−0.004 **	−0.012
	(−2.313)	(−2.296)	(−1.300)
Capital Expenditure	−0.085 ***	−0.085 ***	0.057
	(−3.181)	(−3.171)	(0.838)
SGA Expense	0.007	0.006	0.004
	(0.829)	(0.730)	(0.217)
Location Religiosity	−0.005	−0.008	−0.013
	(−0.343)	(−0.536)	(−0.172)
Constant	0.003	0.005	−0.362 ***
	(0.023)	(0.034)	(−3.763)
Industry FE	Y	Y	Y
Year FE	Y	Y	Y
Observations	12,736	11,207	623
R-squared	0.612	0.608	0.418

(Source: Authors). T-statistics are reported in parenthesis. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively.

, we conducted a further robustness test to examine how individual religiosity in the C-suite affects abnormal negative tone (known as net negative tone from hereon). As net negative tone has been associated with impression management tactics designed to deflate expectations or is known as a tool for engaging in insider trading behavior, we conducted this additional robustness test to examine whether the presence of a religious executive in the C-suite reduces the magnitude of abnormal negative tone in 10-Ks. We constructed our net negative tone measure using a similar approach to the extant subliterature in this domain (Sandulescu, 2015; Choi, 2020; Arslan-Ayaydin et al., 2021; Xu & Qi, 2022; Cheng et al., 2023). Then, we repeated our analysis on the overall sample, CEO subsample, and CFO subsample using net negative tone as our dependent variable. In line with our findings in Table 2, Table 3 and Table 4, the results show that the presence of religious executives within the C-suite significantly reduces the magnitude of net negative tone in corporate annual reports. This is consistent with the conjectures of the intrinsic dimension viewpoint of religiosity, which suggests that religious individuals are highly transparent and would not purposefully engage in the manipulative practices often associated with an abnormal or net negative tone.

5. Conclusions, Limitations, and Future Research

This study examines how individual religiosity within the C-suite team affects the disclosure tone ambiguity of corporate annual reports. It shows that the presence of individual religiosity has important implications for the disclosure tone ambiguity of corporate annual reports, as documented by a reduction in the level of negativity and uncertainty in those reports. Subsample analyses show that the religiosity of CEOs and CFOs tends to be the primary driver of the main effect documented in this study. By highlighting the impact of individual religiosity on disclosure-tone-related choices, this study extends the extant disclosure literature, particularly in the subdomain of psychological managerial traits affecting the tone of disclosures in corporate annual reports. Overall, the study adds to the growing psychological-factor-based managerial traits research in accounting and finance and shows that individual executive-level religiosity has important corporate decision-making implications, as evidenced by how such religiosity affects the tonal sentiment of corporate communication.

This study has a few caveats. Data are limited to BoardEx’s sample availability, which limits the data availability pre-1999. Additionally, Loughran and McDonald’s tonal sentiment dictionary primarily covers the English language and U.S. firms, which is also the focus of our research. Hence, if the study’s scope were to be expanded to a cross-country or cross-language setting, Loughran and McDonald’s tonal sentiment dictionary would require quite a few modifications and adaptations, which could be cumbersome, especially if this is performed individually for each different language. However, this limitation provides a few avenues for future research. Given the advent of machine learning in the field of computer science, the natural language processing (NLP) techniques commonly seen in the field of computer science research can be used to batch process and adapt Loughran and McDonald’s dictionary or other similar tonal sentiment dictionaries to multiple other languages. This can be a feasible way to expand the cross-cultural adaptability of such tonal sentiment dictionaries. Modifying tonal sentiment dictionaries via NLP techniques could create new cross-country avenues of research involving religiosity and disclosure tone sentiments in corporate reports. Using NLP techniques to expand the range of words being analyzed can also improve our understanding of the tonal sentiment of alternative, social-media-based corporate communication channels such as Twitter, Instagram, or TikTok, the popularity of which has grown drastically since the advent of Loughran and McDonald’s tonal sentiment dictionary.