Using 2,603 executive assessments, we study how CEO candidates differ from candidates for other top management positions, particularly CFOs. More than half of the variation in the 30 assessed characteristics is explained by 4 factors that we interpret as general ability [note: Factor 1 may or may not be intelligence], execution (vs. interpersonal), charisma (vs. analytical), and strategic (vs. managerial).
CEO candidates have more extreme factor scores that differ statistically-significantly from those of CFO candidates. Conditional on being considered, candidates with greater general ability and interpersonal skills are more likely to be hired. These and our previous results on CEO success suggest that boards overweight interpersonal skills in hiring CEOs.
…II. The 4 Factors: ghSMART grades each candidate on 30 specific characteristics, grouped into 5 general areas: Leadership, Personal, Intellectual, Motivational, and Interpersonal. Table A.1 describes the characteristics and provides ghSMART’s internal guidelines for behaviors that result in higher and lower grades. Characteristics are graded from D (lowest) to A+ (highest), depending on the extent to which the candidate’s personality reflects the particular characteristic. We convert letter grades to numerical ratings using the traditional scale. An “A” grade is coded as 4, “B” is coded as 3, and so on. The “+” and “−” modifiers add and subtract 0.3, so “A+” is coded as 4.3. Our results are robust to other coding schemes, such as giving a relatively larger score to the highest grades.
Table 4 reports average ratings for the characteristics. The ratings are similar across positions, with CFOs scoring slightly lower, on average. The ratings are highly correlated across characteristics, however, and it is not possible to include all of them as explanatory variables in a multivariate regression due to multicollinearity. Therefore, like KKS, we use factor analysis to identify the underlying variation in the characteristics.
Panels A & B of Table 5 report the results of the factor analysis.12 Panel A presents factor loadings for the first 4 factors.13 Panel B presents eigenvalues and the variation explained by the first 6 factors. A factor is considered valid if its eigenvalue exceeds one. The first 4 factors are valid and together explain 51.2% of the variation in the characteristics.
A. Factor Interpretations: The 4 factors have natural interpretations:
- Panel A of Table 5 shows that the first factor has positive loadings on all characteristics, and Panel B shows that it explains 32.2% of the variation in the characteristics. We interpret this factor as a candidate’s general ability. This structure reflects a general tendency of the characteristics to move together, which is common in personality studies, dating back to the “g factor” identified by 1904. The loadings on the first factor range from a 0.30 on Integrity to 0.74 on Proactive. Note that the different magnitudes do not indicate that Proactive is more important, in some sense, than Integrity. Formally, the relative magnitudes mean that for candidates with greater scores on the first factor, that is, candidates with greater general ability, this general ability manifests itself more strongly in their ratings on Proactive than on Integrity.
-
The second factor explains 9.9% of the total variation and has two distinct sets of factor loadings. Its positive loadings, in decreasing order, are for the characteristics Treats People with Respect, Open to Criticism, Listening Skills, and Teamwork. These characteristics appear to reflect candidates’ interpersonal skills. The second factor loads negatively on Aggressive, Fast, Proactive, and Holds People Accountable, characteristics that arguably reflect a candidate’s execution ability. Thus, the second factor distinguishes candidates with greater interpersonal skills from those with greater execution ability. Candidates with greater interpersonal skills have positive scores on this factor, and candidates with greater execution ability have negative scores. The prominence of this factor is consistent with 1993 and et al 2013, who analyze the tension between resolute and overconfident managers versus managers with empathy and interpersonal skills in their models of CEO types.
-
The third factor explains 5.2% of the variation. Panel A of Table 5 shows that it loads more negatively on Enthusiasm, Persuasion, Aggressive, Proactive, and Fast, which arguably describe more charismatic personalities, while it loads more positively on Analytical Skills, Attention to Detail, Organization, and Brainpower, which describe more analytical personalities. Thus, the third factor appears to distinguish candidates with more charismatic personalities (negative factor scores) from candidates with more analytical personalities (positive factor scores).
-
Finally, the fourth factor explains 3.9% of the variation. It loads more positively on Strategic Vision, Brainpower, Creative, and Analytical Skills, and more negatively on Holds People Accountable, Attention to Detail, Organization, Efficiency, and Removes Underperformers. We interpret this factor as assigning positive scores to candidates with a broader and more strategic focus, and negative scores to candidates with greater attention to detail and more managerial personalities.
…For all candidates, the patterns are largely similar to those in the other panels. For all firm sizes, hired candidates have more general ability and are less execution-oriented. The coefficients on general ability are statistically-significant for all company sizes, with the magnitude of the coefficient tending to increase with company size. Execution ability is statistically-significant for medium and large companies. For CEOs, the results are less clear-cut. Medium and Very Large companies hire candidates with statistically-significantly more general ability, while Small and Large companies hire candidates who are statistically-significantly less execution oriented.
…IV. Personalities and Subsequent Careers: …To evaluate the external validity of our results, we perform an out-of-sample analysis where we consider whether the assessments and factor structure predict candidates’ subsequent long-term careers. We consider whether the characteristics of non-CEO candidates, that is, candidates who were considered for a position other than a CEO position when initially assessed, predict whether these candidates subsequently become CEOs. Importantly, such subsequent promotions and job changes occur later, typically in other companies, and without the involvement of ghSMART. Hence, investigating whether the initially assessed characteristics predict future career progressions provides a test of the classifications of the personalities and hiring decisions of CEO, CFO, and COO candidates. As reported below, we find that the characteristics are indeed predictive of the candidates’ future careers, confirming the broader validity of the classifications.
For each candidate, we used LinkedIn, CapitalIQ, and other Internet searches to determine their subsequent career and whether the candidate later becomes a CEO, COO, or CFO. Table 8 presents the results. Panel A shows that 79% of CEO candidates eventually become CEOs, and 16% of non-CEO candidates become CEOs. Panels B & C report these percentages for CFO and COO candidates. Interestingly, Panel D of Table 8 shows that only 2% of candidates considered for CEO positions eventually become CFOs, and Panel E shows that only 10% of CFO candidates eventually become CEOs. There is sometimes a perception that CFOs are natural successors for CEOs, but this does not appear to be the case for our candidates, which is also consistent with our finding that CEO and CFO candidates have distinct personality traits.