Managerial overconfidence and firm profitability

ABSTRACT

This study examines how Chief Executive Officer (CEO) overconfidence affects profitability. Using United States data from 1992 to 2010, we find that firms with overconfident CEOs have a greater return on net operating assets (RNOA). To identify the sources of this higher performance, RNOA is partitioned into profit margin and asset turnover. This Dupont analysis reveals that higher RNOA of firms with overconfident CEOs comes from profit margin and is not associated with asset turnover. Our results also show that the earnings components of firms having overconfident CEOs better predict future earnings change. The results are robust to the different definitions of CEO overconfidence and profitability. Additional analyses show that CEO overconfidence is positively related with stock performances proxied by abnormal stock returns. Overall, our results suggest that CEO overconfidence is an important factor that contributes to higher and predictable performance of firms.

KEYWORDS:

• Overconfidence
• profitability
• DuPont components
• earnings predictability
• stock performance

Acknowledgments

Wooseok Choi acknowledges that this study is partially supported by Korea University Business School Research Grant.

Disclosure statement

No potential conflict of interest was reported by the authors.

AppendixDefinitions of variables

Table

Notes

1. According to the book, The varieties of economic rationality, ‘The role of the concept of economic rationality is essential to economics as a scientific discipline. It serves as the fundamental behavioural premise in every economic explanation, and in neoclassical economics it even guarantees the very existence of the discipline.’

2. Optimism and miscalibration are its major manifestations, but it is difficult to distinguish these two dimensions in reality, thus it is often used to represent both of them in archival papers (Skala 2008).

3. CEO overconfidence is considered a detectable attribute in the literature because it is measured through publicly available data, that is, Execucomp or articles in The New York Times, BusinessWeek and Financial Times, which stakeholders including investors can easily access (Goel and Thakor 2008; Hirshleifer, Low, and Teoh 2012).

4. RNOA is not affected by differences or changes in capital structure because it does not include financial assets in the denominator and deducts operating liabilities. Thus, it is appropriate to analyze profitability to business operations, and this is commonly used in the valuation literature (Fairfield and Yohn 2001; Nissim and Penman 2001; Penman and Zhang 2002).

5. As stated, overconfidence can have both positive and negative effects on firm profitability. Therefore, the results of this study do not suggest that overconfidence is always beneficial to firms. Rather, the results suggest that overconfidence, on average, contributes to firm profitability and the reliability of earnings in a positive way.

6. Although innovation and competitiveness through overconfidence can positively affect a firm’s performance, success in innovation generally involves many failures, which are directly related to overinvestment. Consequently, we are aware that CEO overconfidence can have both positive and negative effects on firm profitability.

7. Concerning the association between overconfidence and profitability, Patelli and Pedrini (2014) provide evidence that an optimistic tone in CEO letters reflects future performance. However, as the authors mentioned, the finding cannot be generalized to the overconfidence literature because the optimism of this paper does not directly indicate overconfidence and the result is only effective under tough macroeconomic conditions. The critical difference between our study and Patelli and Pedrini (2014) is that our study examines the association between overconfidence and firm profitability using the definition of overconfidence in the literature with a more general sample.

8. ROE is the return on book equity and RNOA is the return on net operating assets. FLEV and SPREAD indicate financial leverage and any associated returns (e.g. interest income/expense), respectively.

9. If overconfident managers do not achieve innovative success, their continuous overinvestment is likely to affect sales negatively. Thus, in this case their ATOs are expected to be more persistent than moderate managers.

10. In the robustness test, we also use two other proxies of CEO overconfidence. Our second measure is calculated by using the value of the CEOs’ in-the-money unexercised but exercisable options following Schrand and Zechman (2012). Third, we use media and press following prior research (Malmendier and Tate 2008; Hirshleifer, Low, and Teoh 2012; Hribar and Yang 2015). The results are similar for all three measures.

11. The threshold of 67% used in prior studies is derived from Hall and Murphy (2002). Their framework assumes that CEOs failing to exercise an executive option that is 67% in-the-money implies a constant relative risk-aversion parameter (CRRA) of three (ρ = 3). A risk-aversion parameter of three is at the low end of the reasonable range of estimates (Hall and Murphy 2002).

12. CEOs are typically risk-averse and undiversified and exposed to the risk of the stock options of their own firm. To decrease the risk of their stocks, CEOs minimize their holding periods of stock options by exercising vested options if the option is in the money. However, CEOs who are overconfident about their ability and firm performance may not exercise options until they are satisfied with the increased exercise prices (Hall and Murphy 2002; Malmendier and Tate 2005, 2008).

13. We thank the reviewer for his/her valuable suggestions which have led to significant improvement on the methodology.

14. Operating assets is calculated as total assets less cash and short-term investments. Operating liabilities is calculated as total assets less the long- and short-term portions of debt, the book value of total common and preferred equity, and minority interest.

15. Richardson et al. (2005) control for three components of total accruals. Specifically, they included change in working capital (ΔWC), net noncurrent operating assets (ΔNCO) and net financial assets (ΔFIN). We include these accruals variable to find our results robust. See Appendix for the detailed definitions of these variables.

16. Over-investment is the values of residuals from the following model in McNichols and Stubben (2008): INV_it = α₀ + α₁Q_it-1+ α₂Q_it-1*Qrt2_it-1 + α₃Q_it-1*Qr3_it-1+ α₄Q_it-1*Qrt4_it-1+ α₅CF_it+ α₆Growth_it-1+ α₇INV_it-1 + ε_it where INV is the capital expenditures; Qrt2, Qrt3, and Qrt4 are indicator variables that equal 1 if Q is in the second, third, and fourth quartiles of its industry-year distribution; CF is the cash flows; Growth equals the assets growth rate.

17. In an additional test, we examine H1 with different profit measures, including ROE, ROI and ROA in Table 7. The results show that overconfident CEOs report higher profits irrespective of the profit measure, confirming the finding with RNOA.

18. As a sensitivity test, we also conduct the test with all the control variables used in Table 3. The untabulated result confirms that the inclusion of additional control variables does not change the main results of Table 4.

19. It may seem that ATO results in Tables 3 and 4 are not consistent, but the effect of overconfidence on ATO level can be different from its influence on ATO persistence. As shown in the literature, CEO overconfidence has both positive and negative aspects regarding ATO (Ramiah et al. 2016; Zavertiaeva, López‐Iturriaga, and Kuminova 2018). This indicates that some overconfident CEOs can increase ATO by efficiently managing inventory, but other overconfident CEOs can engage in excessive investments, decreasing ATO. These contrasting effects can lead to the insignificant relationship in Table 3. However, ATO persistence can be affected by continuous efforts of overconfident managers to achieve their objectives. As Phua, Tham, and Wei (2018) and an article in Forbes (Sep, 2018) suggest, overconfident CEOs’ substantial influence on the stakeholders and their strong-self beliefs make unrealistic goals possible. These traits may contribute to the persistence of ATO by continuing the ATO change in one direction.

20. Execucomp provides estimated unexercised but exercisable options. In contrast, the data used in Malmendier and Tate (2005, 2008) are not publicly available.

21. Schrand and Zechman (2012) define overconfidence if the amount of delayed option is greater than the industry median. However, prior studies generally show that the mean value of the overconfident CEO is less than 50% (34% CEOs are classified as high optimism in Campbell et al. (2011) and 35.1% of CEOs are overconfident using Holder67 in Ahmed and Duellman (2013)). Considering these numbers of overconfident CEOs, we use the highest quintile rather than the median value. The results are similar whether we use the median or quartile classification.

22. Following prior literature, we use press articles from Wall Street Journal, New York Times, and BusinessWeek. OC_Press equals one when the articles describe certain CEOs as optimistic or confident.

23. See Appendix for variable definitions.

24. The untabulated results using one-year-ahead buy-and-hold stock returns, abnormal one-year-ahead buy-and-hold stock returns, and Tobin’s Q as the dependent variables are similar to the findings in Panel B of Table 8.

25. The patent data is downloaded from the NBER database (https://www.nber.org/patents/) following Hall, Jaffe, and Trajtenberg (2001) and Hirshleifer, Low, and Teoh (2012).

26. In Panel B, Inv equals one if CAPX investment, R&D investment and Ln(Patent) are greater than the highest quartile values, respectively.