Vishal Gupta is associate professor and Sandra Mortal is associate professor of Economics, Finance, & Legal Studies at the University of Alabama Culverhouse College of Commerce. This post is based on an article, forthcoming in Academy of Management Journal, by Professor Gupta; Professor Mortal; Bidisha Chakrabarty, Edward Jones Endowed Professor of Finance at St. Louis University; Xiaohu Guo, PhD Candidate in Finance at The University of Alabama; and Daniel B. Turban, Emma S. Hibbs/Harry Gunnison Brown Chair of Business and Economics at the University of Missouri.
Our research examines whether CFO gender affects the likelihood of irregularities in a firm’s financial statements. In the current climate of increased focus on women leaders, especially in corporate positions, an intriguing question of academic and popular interest is whether male and female managers are associated with different firm decisions and behaviors. We examine this in the context of financial statements, which are an important part of the firm’s communications with its stakeholders. Because information presented in financial statements informs the investment decisions of capital market participants, the accuracy of said statements is critical to the effective functioning of the economy. Unfortunately, irregularities in financial statements are commonplace, with some estimates suggesting that at least 5% of the annual revenues of US-based firms are lost to fraud.
Although a large multidisciplinary literature examines the causes and consequences of financial misreporting (Amiram, Bozanic, Cox, Dupont, Karpoff, & Sloan, 2018), it is based on companies “caught” engaging in financial misconduct. This is problematic because the large majority of firms that commit financial fraud are never caught. Dyck, Morse, & Zingales (2013) estimate that although approximately 15% of public firms engage in accounting fraud annually, less than 1% are caught. Consequently, much effort has been expended on developing tools to diagnose misreporting. Recently, Amiram, Bozanic, and Rouen (2015) offered the Financial Statement Deviation (FSD) score to capture the level of irregularities in financial statements by applying the 19th century mathematical formula known as the Benford’s Law to company financial filings. Benford’s Law describes the expected frequencies of leading digits in naturally occurring collections of numbers. The FSD Score measures the extent to which leading digit distributions in a firm’s financial statements deviate from Benford’s Law, such that higher scores (greater deviation from the expected distribution) signal financial irregularities. We use the Benford’s Law to examine firms’ financial statement irregularities conditional upon the gender of the CFO.
There are three distinct possibilities when it comes to predicting the effect of CFO gender on financial statement irregularities. One possibility is that women who reach the upper echelons of the corporate world tend to think and act like men in making firm decisions (Adams & Funk, 2012). As a result, corporate decisions will not be affected by the gender of the executive. A second possibility is that because of socialization and evolutionary differences, men and women tend to differ on various psycho-social qualities (Meyer-Levy & Loken, 2013), which result in higher aggressiveness, greater risk-taking, and lower ethicality among men than among women. These differences are then reflected in firm decisions, such as Huang and Kisgen’s (2013) research finding that firms with female CEOs and CFOs do less acquisitions than male CEOs and CFOs. Consequently, firms with male CFOs will have higher incidence of financial misreporting than firms with female CFOs. A third possibility is that women who make it to top management positions need to surmount more obstacles and barriers on their way up the organizational hierarchy (Hill, Upadhyay, & Beekun, 2015), so that female executives may be willing to take greater risks than their male counterparts. As evidence, consider Berger, Kick, & Schaeck’s (2014) intriguing finding that banks with management teams that have more female executives tend to have riskier portfolios (presumably to obtain higher returns). As a result, the likelihood of financial misreporting may be higher for firms with female CFOs compared to male CFOs.
Based on the literature about possible gender differences in ethics and risk-taking, we hypothesize that CFO gender will affect financial statement irregularities, such that firms with female CFOs will have lower likelihood of financial misstatements than firms with male CFOs. Additionally, as noted by managerial discretion theory, the influence senior executives have on corporate actions depends on the leeway—or discretion—available to them (Hambrick, 2007). Discretion depends, in large part, on powerful stakeholders (or constituents) who can constrain or magnify an executive’s effect on corporate outcomes. Thus, we also hypothesize that the effect of executive gender on firm behaviors will be stronger in situations of higher managerial discretion, such as when analyst coverage and institutional ownership is low. Taken together, therefore, our research examines whether and when CFO gender affects firm financial misreporting.
To test our hypotheses, we draw a sample of US-based large public firms by merging six different datasets: ExecuComp for CFO gender, Compustat for accounting data, CRSP for return volatility, Institutional Shareholder Services (ISS) for board data, IBES for analyst coverage, and Thomson Reuters for institutional holdings (13-F). Our sample contains 18,659 firm-year observations, from 1996 to 2016, covering 2,186 unique US-based firms with 1,550 firm-year female CFO observations and 339 unique female-CFO firms.
To compute FSD Score, we follow Amiram et al (2015). Specifically, we download the line items reported in the Balance Sheet, Income Statement, and Statement of Cash Flows in the annual report (Form 10-K), and generate the frequency distribution of leading digits for all line items, by recording the percentage of numbers that begin with a “1”, with a “2”, and so on. We then take the absolute value of the difference between the actual distribution and that predicted by Benford’s law. The FSD Score is the sum of these values across all nine digits.
We perform regression analyses to test our hypotheses. Our results show that firms with female CFOs have 2.6% lower FSD Score (our proxy for financial misreporting) compared to firms with male CFOs, which is an economically meaningful effect in comparable literature. Notably, this main effect is qualified by governance mechanisms such that CFO gender will be more closely associated with the likelihood of financial misreporting when there is lower institutional ownership and analyst coverage, both of which permit greater executive discretion. In all tests we control for an exhaustive set of possible confounding factors, including (but not limited to) firm size, firm age, CEO gender, board independence and board gender diversity. We conduct several additional analyses, including substituting financial restatements for FSD score, an alternative measure of financial misreporting, and find consistent results. Our results are also generally robust to employing alternative analytical approaches (including, random effects, fixed effects, and hybrid models, propensity-score matching and treatment-effects model), enhancing confidence in our findings.
As more women break through the proverbial “glass ceiling” to achieve top-level leadership positions, the issue of how executive gender may impact corporate decisions and actions has gained heightened salience. While much is known about the barriers preventing women from reaching senior management positions, little is understood about the ways in which corporate decisions of firms with female executives differ from those of their male counterparts once they are in such coveted positions. Our research shows that firms with women CFOs have significantly lower likelihood of financial misreporting than firms with male CFOs, and that these gender differences are stronger in situations of low monitoring from key stakeholders. We believe these results reveal an intriguing—and previously unrecognized—gender aspect to the efficacy of key governance mechanisms. The interface of executive gender and corporate governance, which is at the heart of our research, deserves greater attention in future scholarship. We hope our research, conducted in the context of large public firms in the US, inspires further inquiry into the influence of managerial attributes on additional corporate decisions beyond financial (mis)reporting, and in countries with sociocultural contexts different from the US.
The complete article is available for download here.