Reporting Concerns About Earnings Quality: An Examination of Corporate Managers

There are several noteworthy examples from practice that highlight the various issues that CFOs face when discovering and reporting fraud (e.g., Segarra 2014; McCann 2017).

In regard to the reporting environment, Italy has enacted whistleblower protection legislation with features that are consistent with that of other countries. For example, similar to requirements mandated by the U.S. Sarbanes–Oxley Act of 2002, Italian publicly traded companies that want to receive a positive assessment of their internal control systems are required to adopt whistleblowing processes that allow employees to report irregularities or violations anonymously. Similarities also exist between Italy and other countries in regard to earnings management. Burgstahler et al. (2006) and Van Tendeloo and Vanstraelen (2008) provide insights into earnings management in European firms, and Leoni and Florio (2015) provide a comparison of the U.S. and Italian earnings management literatures.

https://​www.​coso.​org/​Documents/​COSO-Fraud-Risk-Management-Guide-Executive-Summary.​pdf.

Extant research suggests that managers are more apt to report concerns internally over externally (e.g., Robertson et al. 2011; Brink et al. 2013). Consistent with this notion, demographic data from our participants (presented in Table 1 and discussed in the Participants section) illustrate a strong preference for reporting internally vs. externally. Given this preference, we designed our experiment such that participants chose to report internally first and then considered reporting externally only after “inside the company nothing was done in response to your concern.” We believe this design choice reflects the decision-making process managers would employ in practice.

For the initial sample of 951 companies, 132 were public (13.88%; 132/951), while 819 were private (86.12%; 819/951). The final sample of 204 companies consisted of 37 public companies (18.14%; 37/204) and 167 private companies (81.86%; 167/204).

According to a 2010 EY survey of 669 CFOs from Europe, the Middle East, India, and Africa, only 27% had obtained the MBA degree (http://​www.​ey.​com/​Publication/​vwLUAssets/​Estudio_​DNA_​CFOs_​2010/​$FILE/​DNA_​CFOs_​2.​pdf).

Burgstahler et al. (2006) document that, within the European Union, private companies are much more prevalent than public companies and private companies exhibit higher levels of earnings management. In Italy, approximately 360 companies are publicly listed (http://www.borsaitaliana.it/homepage/homepage.htm), 7500 companies are owned by the State (http://www.panorama.it/economia/aziende/aizende-pubbliche-quanto-costano-stato), and 5.3 million companies are privately owned (http://www.digital4.biz/pmi/approfondimenti/quasi-53-milioni-le-imprese-in-italia_4367215623.htm). We acknowledge that private company managers do not face the pressures associated with widely dispersed investors, which may affect their willingness to report red flags (e.g., the market reaction to restatements is not as forefront in their minds as it would be for managers of publicly traded companies). To investigate differences between private and public company participants, we compare the responses provided by private and public company managers in our most extreme condition (PRESSURE high and RED FLAGS present or Condition 4 as described in Figure 1 and Table 3). Mean responses for CONCERN, INTERNALLY, and EXTERNALLY are not significantly different (p’s > .05) between private and public company managers.

Given that our study extends the work of Dichev et al. (2013), we compare our demographic data in Table 1 to that obtained from the survey participants in Dichev et al. (2013) who worked for private U.S. companies (see Table 1 in Dichev et al. (2013), where 54.93% of participants worked for a private company). Like our sample, the private company CFOs in Dichev et al. (2013) are most likely managing a manufacturing company with between $100-$499 million in sales. As one would expect, given that Italy is a smaller market than the U.S., our participants report a higher proportion of foreign sales. Our participants are also slightly younger, but have greater experience in their position. Last, the percentage of our participants with a public accounting background (41.32%) is very similar to the 41.26% observed by Dichev et al. (2013).

The experimental instrument was provided online and in Italian, the native language of the participants. The instrument was first developed in English. To develop the Italian version of the instrument, we followed the translation-back procedures outlined by Brislin (1986). Specifically, in the first stage, the experimental instrument was translated from English to Italian by one of the authors who is fluent in both languages. Then, another independent academic translated the Italian version back to English (back-translated English version). The original and back-translated English versions were then compared, and all discrepancies resolved by the translators. In a second stage, to assure that the material would be realistic and understood by respondents, the instrument was carefully pre-tested. First, the instrument was reviewed by Italian academic scholars to assess the clarity of the instrument. Afterwards, a pilot study was also conducted with a group of accounting managers from three Italian companies (with their input being incorporated into the instrument). Finally, the final instrument was reviewed once more by a panel of three Italian academic scholars.

Participants were post-experimentally asked to recall: 1) the difference between the company’s net income and cash flow from operations; and 2) the difference between the company’s sales growth and growth in NFMs (measured via scales, where 1 = “Very small” and 7 = “Very large”). Non-tabulated results indicate that those in the RED FLAGS present conditions rated both differences to be significantly larger than those in the RED FLAGS not present conditions (both p’s < 0.01). In addition, both of these measures have a significant impact on CONCERN, but the difference between the company’s net income and cash flow from operations appears to have a slightly stronger influence (t = 3.92, p = 0.001 vs. t = 2.23, p = 0.027 for the NFM red flag). All tests reported in the text and the tables are two-tailed.

Our manipulations of PRESSURE and RED FLAGS would be more likely associated with net income being overstated than understated (e.g., pressure related to the return on assets, net income substantially higher than cash flow from operations). Indeed, only three participants in the PRESSURE high/RED FLAGS present condition (Condition 4 in Table 3) indicated concerns that earnings were understated (our tests of hypotheses are robust to excluding these three participants from our analyses). However, to avoid demand effects, we provided our participants with the option to respond that net income was either understated, very accurate, or overstated.

We measured discussing concerns with the corporate controller with following: Based on your preliminary review, you stated that the 20XX net income for Tecno may be overstated/understated. To what extent would you discuss this concern with your corporate controller in charge of consolidating the divisions’ financial statements? The response scale was the same as INTERNALLY. We measured discussing concerns with the corporate controller prior to discussing concerns with the CEO (INTERNALLY). However, given our one-time access to participants, we were unable to discern the sequential process and iterations our participants would have followed given their experimental condition (e.g., having several meetings with the corporate controller before meeting with the CEO). Examining the sequential processes and iterations involved in reporting concerns over earnings quality internally and externally represents a fruitful avenue for future research.

Interestingly, even though shorter tenure and a CPA background are associated with a higher likelihood of reporting externally, we do not observe significant correlations between these two factors and responses to our red flag manipulation checks (see Footnote 10). Thus, managers with shorter tenure and CPA backgrounds were no more likely to identify the red flags, but were more apt to report their concerns externally.