«J Bus Econ (2015) 85:793–815 DOI 10.1007/s11573-014-0753-1 ORIGINAL PAPER New insights on CEO charisma attribution in companies of different sizes ...»
As a second contextual factor, we assume that a company’s ownership structure plays a crucial boundary role for the relationship between prior company performance and CEO charisma attribution. Evidence grounded in managerial capitalist theory demonstrates that ownership structure plays a critical role in CEO discretion and inﬂuence (Berle and Means 1932; Marris 1964; McEachern 1975;
Williamson 1964). For instance, differences in a company’s ownership structure
798 B. Michaelis et al.
have been found to inﬂuence diversiﬁcation (e.g., Hill and Snell 1988, 1989), accounting practices (e.g., Tosi et al. 1999), the monitoring of CEO compensation (e.g., Tosi and Gomez-Mejia 1989, 1994), CEO compensation (e.g., Fong et al.
2010; Gomez-Mejia et al. 1987; McEachern 1975; Williamson 1963), sources of annual CEO pay raises (e.g., Hambrick and Finkelstein 1995), emphasis on innovation (e.g., Hill and Snell 1988), corporate R&D spending (e.g., Baysinger et al. 1991), and ﬁrm performance (e.g., Hill and Snell 1988, 1989; Hu and Izumida 2008; Hunt 1986). Those differences may occur because companies with different ownership structures place different values on organizational culture (Howorth et al.
2010), management systems (Schachner et al. 2006), and performance outcomes (Ghobadian and O’Regan 2006).
Moreover, research in the SME context has demonstrated that ownership structure inﬂuences CEO status (Jayaraman et al. 2000), which is a major source of managerial discretion (Hambrick and Finkelstein 1987). CEOs in privately owned companies, for instance, have controlling block discretion to focus on long-term performance in keeping with their stronger interests in sustainable growth, compared with CEOs of publicly owned companies (Ling et al. 2008). Aligned with that reasoning, research demonstrates that publicly owned companies often focus on short-term performance based on shareholder and stock market demands for positive quarterly ﬁnancial ﬁgures. Family-owned companies, in contrast, often follow a more long-term strategy in terms of performance outcomes; when the family is the main shareholder, value turns to sustainable growth (Lumpkin et al.
2010). Furthermore, family-owned companies often have more long-term work relationships with their CEOs and focus more on integrity and sustainability (e.g., Le Breton-Miller and Miller 2006).
As such, we expect that prior company performance will be applied differently to CEO charisma attribution between the two ownership structures. Employees in publicly owned companies might take company performance as a main criterion for attributing CEO charisma, while employees in family-owned companies might use company performance as only one of many criteria because they focus more on sustainable growth and not on quarterly ﬁnancial ﬁgures. Thus, subordinates in publicly owned companies will over-attribute prior company performance to their CEOs when perceiving charismatic qualities. In sum, we propose a moderation
H3: The positive relationship between prior company performance and subsequent CEO charisma attributions will be stronger in publicly owned companies.
2.5 Hypotheses testing
To investigate our hypotheses, we applied a mixed-method strategy. Study 1 was designed as an experimental vignette study to examine the effects of prior company performance on subordinates’ attribution of CEO charisma. Study 2 was a ﬁeld study to validate the prior company performance—charisma attribution relationship
and to investigate the two hypothesized organizational-level boundary conditions:
company size and ownership structure.
3 Experimental design: study 1 Study 1 was a test of the proposed hypotheses and the causal direction of the assumed relationships.
3.1 Participants and procedures We recruited 92 public university business and psychology students; 78 were female (84.8 %), 14 were male (15.2 %), ranging from 19 to 49-years-old, with an average age of 23.03 (SD = 5.24); 81.5 % were undergraduates, 10.9 % were graduate students, 7.6 % were professionals. Data collection took place at the University of Heidelberg, Germany, between November 2008 and May 2009. The study administrator randomly approached students and offered them four course credit points in exchange for their participation (some courses at the University of Heidelberg require students to participate in experiments for course credits). Before handing students the paper–pencil version of the questionnaire, the administrator asked them to imagine that they were employees of a German software company.
3.2 Measures 3.2.1 Prior company performance Prior company performance was manipulated through two ﬁctional annual reports based on the quarterly report from SAP Corporation (a German software company) from the ﬁrst quarter of 2001. This report was randomly chosen and subsequently shortened, eliminating all references to the stock market, external factors, and the SAP Corporation. Furthermore, the report was reconstructed in terms of revenue and earnings before interest and taxes (EBIT). Two annual reports were generated: one for the success condition and one for the failure condition (see ‘‘Appendix’’). Each report concluded with a short notice about how the company succeeded compared with other companies. In the success treatment, the notice read: ‘‘Compared with similar companies in this branch, your company has achieved above average success.’’ This was designed to help the participants estimate the company’s success.
3.2.2 CEO charisma
We assessed CEO charisma (a = 0.92) by using the Conger–Kanungo scale of charismatic leadership (henceforth, the C–K-scale). We utilized a version that Rowold (2004) modiﬁed and translated into German. The scale consists of 27 items, constituting the ﬁve factors of the scale: sensitivity to member needs (six items), sensitivity to the environment (four items), personal risk (ﬁve items), 800 B. Michaelis et al.
unconventional behavior (ﬁve items), and strategic vision and articulation (seven items). We primarily considered the aggregated charisma value of all items of the C–K-scale. Sample items include ‘‘has vision; often brings up ideas about possibilities for the future’’ and ‘‘inﬂuences others by developing mutual liking and respect.’’ 3.2.3 Manipulation check We included a manipulation check after the annual reports to ensure effective treatment for prior company performance. We asked ‘‘How effective do you think your company is?’’ rated on a bipolar seven-point scale.
3.2.4 Control variables First, we included participants’ age (in absolute years) as a control variable because the relationship between the leader’s age and the age of subordinates can crucially determine the effectiveness of leadership behaviors (Kearney 2008). Second, subordinates’ characteristics are critical, particularly in charismatic leadership (Conger and Kanungo 1988a, b; Conger et al. 2000), so we recorded gender. Finally, we controlled for three educational levels using dummy variables: undergraduate student, graduate student, and professional.
3.3 Analysis and results
Table 1 summarizes the means, standard deviations, and intercorrelations among the dependent variables, the control variable, and the manipulation check measure.
The subscales of the C–K-scale were highly correlated, as expected.
For the independent variable of prior company performance, a t test of the manipulation check measure was conducted to determine whether the treatment was effective. The difference between company success and company failure was signiﬁcant (t = 20.31, p \ 0.001). Thus, it appears that the manipulation of prior company performance was successful. We conducted analyses of variance for testing the hypothesis. A two-way ANOVA on charisma ratings on the C–K-scale with participant age and gender as covariates revealed a signiﬁcant main effect of prior company performance: F (1, 83) = 5.73, p \ 0.05. As predicted, participants in the company success treatment considered their CEO to be more charismatic than did participants in the company failure treatment. Thus, Hypothesis 1 was fully supported (see Table 2).
4 Field study: study 2
n = 92 participants. Reliability values are in parentheses CKS Conger–Kanungo Scale of Charismatic Leadership * p \ 0.05 (two-sided) ** p \ 0.01 (two-sided) *** p \ 0.001 (two-sided)
n = 92 participants CKS Conger–Kanungo Scale of Charismatic Leadership * p \ 0.05 (one-sided) ** p \ 0.01 (one-sided) questioned whether our results would be transferable to real company settings.
Therefore, we undertook a ﬁeld study to validate the main ﬁnding—that prior company performance is related to CEO charisma attribution—and additionally explore the two hypothesized organizational-level boundary conditions: company size and ownership structure.
4.1 Participants and procedures
To test the prior company performance-CEO charisma attribution relationship as well as the two moderating hypotheses, we collected data, as part of a larger benchmarking study, from German-based companies that had fewer than 5,000 employees. The companies voluntarily applied to participate, and in recognition of their participation, they received a benchmarking report on their leadership and HR practices. An external professional agency managed the data collection procedure.
Initially, 111 companies agreed to participate, but 44 provided insufﬁcient data on the main study constructs, dropping the actual response rate to 62 % (n = 69). The remaining 69 companies had 23,319 employees overall (mean = 347; range 12–3265). Participating companies were from various industries such as production (22 %), service (54 %), grocery (4 %), wholesale (7 %), and ﬁnance (7 %). Thirtyseven of the ﬁrms were family-owned and 32 were non family-owned. We collected data from three sources to rule out risk of potential same-source bias (Podsakoff et al. 2003). First, we asked employees who reported directly to the CEO to rate their CEO’s charisma, among other items. The questionnaire took about 10 min to complete. On average, 4.41 persons per company were directly reporting to the CEO (ranging from 1 to 11 per company), were mostly male (65 %), and averaged 39-years-old. Second, members of the top management team (TMT) were surveyed to gauge information on the company performance for the previous year, among
New insights on CEO charisma attribution 803
other items. This questionnaire took about 10 min to complete. Ultimately, 198 TMT members provided data on their company performance (ranging from 1 to 15 TMT members per company; average = 2.9 members), were mostly male (85 %), and averaged 45-years-old. Finally, the companies’ top human resource representatives provided data on company size, ownership structure, and general information such as industry afﬁliation and other variables used as controls in our study. That questionnaire took about 15 min to complete.
4.2 Measures 4.2.1 Prior company performance We assessed prior company performance at time (t - 1) (a = 0.90; rwg = 0.85) by asking TMT members about the company’s performance in the previous year with three items gauging operational performance (employee productivity, efﬁciency of business procedures, and employee retention and ﬂuctuation) and three items gauging organizational performance (ﬁnancial performance, company growth, and return on investment) (Kunze et al. 2011). Following prior studies (Delaney and Huselid 1996; Wall et al. 2004), the perceived performance measure was benchmarked by asking the TMT members to assess their company’s performance in the prior year relative to the main competitors in their respective industries (1 = far below average; 7 = far above average). We averaged the item scores to form an overall performance score for each company.
4.2.2 CEO charisma
To measure the CEO charisma attribution (a = 0.89, rwg = 0.88), we used four subscales from the inventory developed by Podsakoff et al. (1990) with 18 items overall. Speciﬁcally, we considered the dimensions of providing a role model, articulating a vision, giving intellectual stimulation, and allowing individual consideration because they are aligned with the C–K-scale used in the experiment.
Applying that scale also allowed a cross-validation of the performance–charisma attribution relationship with a different operationalization. All items had the respondent’s direct supervisor as referent (sample item: ‘‘My direct supervisor is always seeking new opportunities for our company’’), which in our sample was always the CEO. We averaged all items to form an overall charisma attribution measure.
4.2.3 Company size The top HR representative in each company provided information about company size, measured by number of employees.
4.2.4 Company ownership structure As common practice in ownership research (i.e., Gomez-Mejia et al. 1987) we measured the ownership structure with a dummy variable. The top HR representative provided information about the business ownership, captured with a dummy variable (1 = \50 % family-owned company, 2 = [50 % family-owned company).
4.2.5 Control variables
As in the experimental studies, we controlled for the mean age of the respondents who answered the CEO charisma items, because other studies have suggested that age might inﬂuence the perception of leadership (Pastor et al. 2002). Additionally, we accounted for the average tenure and the percentage of women answering the leadership items, which might also affect charisma attribution (Conger et al. 2000).
Furthermore, to account for a potential inﬂuence of industry afﬁliation on the CEO charisma attribution, we controlled for ﬁve classes of industry (e.g., Dickson et al.
4.3 Analyses and results