Seeing is believing?

Recent studies in psychology and neuroscience document the effect of facial trustworthiness perceptions on observers' subsequent decision-making process. Specifically, neuroscience research finds that the amygdala in the human brain plays an important role in interpreting facial trustworthiness. We make trust-related decisions by triggering autonomic responses to emotional stimuli. Several business studies, in particular in the areas of economics and finance, have applied the findings in neuroscience to various settings, and provided evidence that a trustee's facial trustworthiness may affect a trustor's subsequent decision-making process in peer-to- peer lending, legal defense, government contracting, and various trust games.

Facial trustworthiness in an audit setting

How important is the issue of facial trustworthiness in an audit setting? We are particularly motivated by the requirement of the Public Company Accounting Oversight Board (PCAOB) Auditing Standards (AS) that auditors should communicate with management to assess the risks of material misstatement (see AS 2401 for an overview). This communication involves face-to-face interactions between auditors and client firms' executives. One can therefore expect that during initial and subsequent interactions, auditors may develop perceptions of client executives' facial trustworthiness and incorporate such perceptions into their subsequent decision-making processes.

The purpose of our study is to investigate whether and, if so, how auditors incorporate facial trustworthiness of Chief Executive Officers (CEOs) and Chief Financial Officers (CFOs) into the pricing of their audit services, when making audit engagement contracts with their client firms.

The literature of psychology provides evidence suggesting that people tend to extend less trust to an individual with lower facial trustworthiness. Building on this evidence, one can therefore expect that auditors may be less trusting of the business information provided by client firms' CEOs and CFOs with lower facial trustworthiness, and accordingly perceive a higher audit risk for such clients, necessitating a higher fee. To provide large-sample, systematic evidence on this prediction, we propose and test the hypothesis that auditors charge a higher fee to client firms with untrustworthy-looking CEOs and CFOs to compensate for the higher perceived audit risk.

We further expect that at the general level, gaining more relevant information about an individual's past behavior may weaken the impact of perceived facial untrustworthiness when making strategic decisions related to that individual. Auditors accumulate more experience and knowledge about client firms' business facts as audit tenure increases. One can therefore predict that auditors with longer tenure may tend to rely more on their understanding of client firms' underlying risk factors and thus rely less on their initial perception of CEOs' and CFOs' facial trustworthiness. To test this prediction, we propose our second hypothesis that length of auditor tenure weakens the impact of CEOs' and CFOs' facial trustworthiness on audit fee.

How auditors assess inherent risk

In theory, audit fee is determined by audit efforts and litigation risks. Empirical evidence has shown that auditors incorporate management integrity into their judgment and decision making when assessing inherent risks and control risks associated with audit engagements. Auditing standards emphasise the importance of assessing management's values and integrity throughout the audit. AS 2110 requires auditors to make judgments about the integrity of their clients' management team, including CEOs and CFOs, based on face-to-face communication and interactions. The International Standard on Auditing 260 also requires the auditor team to enhance auditor client communications by establishing an effective two-way communication throughout the audit engagement process. AS 2401 also asserts the importance of understanding CEOs and CFOs, stating that "management has a unique ability to perpetrate fraud because it frequently is in a position to directly or indirectly manipulate accounting records and present fraudulent financial information."

Public Company Accounting Oversight Board requirements

In line with the fraud triangle, the Public Company Accounting Oversight Board (PCAOB) claims that the occurrence of fraudulent behavior is related to the personality and ethical values of top executives; when they encounter unethical situations, managers' attitudes may affect their judgment or decision making, thus influencing the likelihood of fraudulent behaviour. Consequently, the PCAOB requires auditors to identify top managers' personal beliefs or ethical standards as a risk factor for fraudulent financial reporting. During face-to-face inquiries and communication about auditing issues, auditors may cognitively develop their perceptions about top managers' ethical values and level of integrity based on factors that they observe, such as the facial trustworthiness of management. Nevertheless, AS 1015 requires auditors to "neither assume that management is dishonest or honest" when exercising professional skepticism.

Exercising professional skepticism

Auditors should assume no bias in management's representations based on forecasts rather than actual results, and exercise their professional skepticism during audit engagement. However, prior studies suggest that auditors' perceived management integrity has an effect on judgment. Specifically, auditors who view management as being of high integrity exhibit a lack of professional skepticism, while auditors with lower trust in client's management subsequently make higher assessments of fraud risk. Thus, trustworthy-looking management may affect auditor's professional skepticism. Auditors need to rely on the information provided by management to determine clients' audit risks, audit hours, and their bid for the audit engagement. In practice, auditors interact significantly with their clients' management teams when making audit fee decisions. Management, especially CFOs, plays an important role in selecting auditors and in determining the information to be distributed to audit committees and auditors. Management tends to spend a significant amount of time interviewing and evaluating potential auditors. Thus, auditors may rely heavily on their communications with client firms' management when judging audit risks and estimating audit hours to make audit fee decisions.

The role of external validation

Auditors tend to seek external validation of data when they perceive management integrity to be low. Auditors respond to the assessed risks of material misstatement by adjusting the nature, timing, and extent of auditing procedures, for example, as reflected in audit staffing allocations (AS 2301). One key factor determining audit fee may be an audit firm's staffing efficiency, since staff is a significant and scarce resource and audit pricing is mainly based on the hourly rates of auditors assigned to an audit engagement. When auditors perceive higher risks of material misstatement, they need to allocate more staff hours to identify issues and corroborate audit evidence. They also need to perform substantive testing and increase sample size or perform analytical procedures at a more detailed level (AS 2301). Therefore, the facial trustworthiness of CEOs and CFOs could affect auditors' pricing decisions. To summarise, auditors may associate untrustworthylooking executives with low management integrity and high audit risk, which leads to an audit fee premium.

What makes a face trustworthy?

Using the latest machine-learning-based facial-feature-point detection techniques developed in neuroscience, psychology, and computer science, we construct a proprietary facial-trustworthiness database for CEOs and CFOs of US listed companies. First, we develop a computer algorithm to collect CEO and CFO pictures from Google Images, based on the full name of each executive and his/her affiliated firm during the sample period of 2001–2014. Second, following the neuroscience and psychology literature, we use a face detector to identify the facial features in the CEO and CFO pictures, thereby calculating a rich set of facial-trustworthiness measures, comprising: (i) angle of inner eyebrow ridge; (ii) face roundness; (iii) chin width; and (iv) nose-tolip distance. We then construct a composite facial-trustworthiness index for each executive. Our analysis focuses mainly on this composite index, though it also considers individual facial features separately. This approach is based on the prior finding that people tend to interpret an individual's face as an integrated whole. We perform two validity checks to confirm that our computer-measured facial trustworthiness index is a valid and reliable measure of individuals' assessments of facial trustworthiness.

Auditors charge higher audit fees based on looks

We collect audit fee data from Audit Analytics, financial statement data from Compustat, and corporate governance data from RiskMetrics. Our final sample comprises 4,411 firm-year observations for 1,179 CEOs and 1,360 CFOs from 845 firms in the sample period of 2001–2014. The results of our regression analyses show that auditors charge a higher audit fee to firms whose CFOs have lower facial trustworthiness, indicating that auditors do factor client CFOs' facial trustworthiness into their pricing decisions. We find, however, that CEOs' facial trustworthiness is not significantly related to auditors' pricing decisions. This is consistent with the view that CFOs play a more important role than CEOs in communicating with auditors and in shaping financial reporting strategies.

We find that the effect of CFOs facial trustworthiness on audit fee is weakened as auditor tenure becomes longer; this suggests that, during multiple rounds of audit engagements, auditors tend to develop a comprehensive understanding of their clients' internal control environment and business operation, and thus rely less on their initial perceptions of management facial trustworthiness when evaluating audit risks.

The implications of personalised trust

Our study contributes to the accounting literature and the auditing profession in several ways. First, to the best of our knowledge, this is the first study that adopts a large sample to provide systematic evidence on auditors' use of facial trustworthiness. Our study, therefore, extends the auditing literature documenting the impacts of executives' individual characteristics on audit fee. Second, we provide further support to the literature on CFOs' crucial role in financial reporting by documenting that auditors consider the facial trustworthiness of CFOs, but not CEOs, in pricing their audit services. Third, the results of our study extend the literature documenting the important role of perceptions based on executives' facial appearance in firm valuation and executive compensation by presenting empirical evidence that CFOs' facial trustworthiness affects auditors' assessment of audit risk and, thus, audit fee. Finally, but importantly, our study develops a novel methodology to measure facial trustworthiness using recent image processing techniques developed by prior neuroscience, computer science, and psychology studies. Given the scarcity of prior accounting and auditing research on the role of facial trustworthiness as executives' personal attribute, we recommend further research on this innovative individual-level trustworthiness measurement and the implications of personalised trust in various business accounting and auditing settings.

Conclusion

By using machine learning-based facial-feature-point detection techniques, which are well-developed in the field of computer science, we construct a novel database of executives' facial trustworthiness for US listed companies and investigate whether and how auditors incorporate the facial trustworthiness of client firm executives into their audit pricing decisions. Our results suggest that auditors tend to charge an audit fee premium to firms with untrustworthy-looking CFOs, although the impact of CFOs' facial trustworthiness on audit fee is weakened as auditor tenure increases. To our knowledge, our study is the first to provide large-sample, systematic evidence on the correlation of CFOs' facial trustworthiness with audit fees.