document5

TRANSFORMATIONAL LEADERSHIP’S ROLE IN PROMOTINGCORPORATE ENTREPRENEURSHIP:

EXAMINING THE CEO-TMT INTERFACE

YAN LINGGeorge Mason University

ZEKI SIMSEKUniversity of Connecticut

MICHAEL H. LUBATKINUniversity of Connecticut

andEM Lyon

JOHN F. VEIGAUniversity of Connecticut

Research about transformational CEOs’ impact on firm-level outcomes, particularlycorporate entrepreneurship, has been equivocal, partially because the underlyingmechanisms remain largely unexplored. Given that the individuals most closely in-fluenced by a firm’s CEO are its top management team (TMT) members, we focus on theCEO-TMT interface as a salient intervening mechanism. We posit that transforma-tional CEOs influence TMTs’ behavioral integration, risk propensity, decentralizationof responsibilities, and long-term compensation and that these TMT characteristicsimpact corporate entrepreneurship. Data from 152 firms supported most of our hy-pothesized links, underscoring how the CEO-TMT interface helps explain transforma-tional CEOs’ role in promoting corporate entrepreneurship.

Transformational leaders are drawn by the needto transform individuals, teams, and firms by goingbeyond the status quo and, in so doing, affect theirfirms’ ability to innovate and adapt. Widely exam-ined as a multifaceted meta-construct, transforma-tional leadership consists of exhibition of four in-terdependent and mutually reinforcing attributes.These are (1) charisma: creating and presenting anattractive vision of the future; (2) inspirational mo-tivation: energizing followers to go beyond self-interest; (3) intellectual stimulation: stimulatingfollowers to challenge assumptions and view prob-lems from new perspectives; and (4) individualizedconsideration: focusing on follower developmentby providing support, encouragement, and coach-ing (Bass, 1985).

Described as the “givers” and “definers” of adap-tive organizational culture (Waldman & Yam-marino, 1999), CEOs who are transformational

leaders are believed to induce organization mem-bers to constantly anticipate and adapt to environ-mental change (Jung, Chow, & Wu, 2003; Waldman,Javidan, & Varella, 2004). For example, Kotter sug-gested that a firm’s entrepreneurial proclivity isenhanced to the extent that a transformational vi-sion seeps into the very fiber of the firm to become“the way we do things around here” (1995: 652).Transformational CEOs are also believed to en-hance this proclivity by being enthusiastic aboutinnovation (Howell & Higgins, 1990) and by show-ing how volatility in the firm’s competitive envi-ronment can be turned into a vision of opportunity(Avolio, Zhu, Kho, & Puja, 2004).

It would seem to follow then that transformation-ally led firms are more likely to engage in corporateentrepreneurship, defined by Zahra (1996) as thesum of a firm’s product innovation, business ven-turing, and strategic renewal activities. Yet re-searchers have been equivocal about the generalimportance of CEOs’ roles in affecting such firm-level outcomes. Some have suggested that CEO im-pact may be, at best, minimal (e.g., House & Aditya,1997; Meindl, 1998) or concluded that “the top

We thank Nandini Rajagopalan and three anonymousreviewers for their constructive comments and sugges-tions, and Peter Gianiodis and Gideon Markman for theirgenerous feedback on a draft.

� Academy of Management Journal2008, Vol. 51, No. 3, 557–576.

557

Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holder’s expresswritten permission. Users may print, download or email articles for individual use only.

team, rather than the top person, has the greatesteffects on organizational functioning” (O’Reilly,Snyder, & Boothe, 1993: 10), whereas others haveviewed CEO impact as being more directed at or-ganizational aspects of a firm that may in their turnaffect firm outcomes (Peterson, Smith, Martorana,& Owens, 2003; Thomas, 1988). According to thislatter view, such equivocality reflects the manydirect and indirect influences of CEOs on theirfirms, especially given the number of possible in-tervening mechanisms that influence firm-leveloutcomes and that CEOs are likely to affect.

With the exception of Peterson and colleagues(2003), however, researchers have paid scant atten-tion to intervening mechanisms. Indeed, the upperechelons theory’s argument (Hambrick & Mason,1984) may have conceptually constrained suchpursuits by not precisely distinguishing betweenthe impact of a firm’s CEO and that of the firm’s topmanagement team (TMT) (cf. O’Reilly et al., 1993).Therefore, we propose and test a model that ex-plores the mechanisms underlying transforma-tional CEOs’ main effects. Specifically, in keepingwith Peterson and colleagues’ (2003) research,which applied a five-factor model to examine theeffects of CEO personality on a selected set of TMTconstructs in nine firms, our model focuses on theCEO-TMT interface. We complement their study byusing this interface to explain the effects of CEOtransformational leadership on the pursuit of cor-porate entrepreneurship. Although it is not centralto our theoretical model, to more didactically de-velop the latter we borrow Hambrick’s (1994)framework identifying four general elements (pro-cess, structure, composition, and incentives) thatunderlie TMT effects on firm-level outcomes. Fol-lowing this framework, and informed by insightsfrom the upper echelons and corporate entrepre-neurship literatures, we identify and then hypoth-esize four salient TMT characteristics that are bothinfluenced by transformational CEOs and influencecorporate entrepreneurship. We report tests of ourmodel using data from a multisource survey ofCEOs and members of their TMTs in 152 firms.

THEORETICAL BACKGROUND

To date, upper echelons research has suggestedthat TMTs influence various firm-level behaviorsand outcomes (see Carpenter, Geletkanycz, andSanders [2004] for a review). TMT characteristicshave been linked by such research to the order andtiming of new product moves (Srivastava & Lee,2005), product innovation (Bantel & Jackson, 1989),and strategic change (Wiersema & Bantel, 1992).However, Jackson (1992), among others (e.g., Kli-

moski & Koles, 2001; Peterson et al., 2003), ob-served that upper echelons research typically treatsCEOs as members of their TMTs, averaging theircharacteristics to assess such things as overall TMTdemographic composition. As a result, such re-search has paid little attention to the CEO-TMTinterface, or the common boundary between thesetwo sets of actors, to explain the effects of CEOs onfirm-level outcomes (Peterson et al., 2003).

This oversight is curious in that it is well under-stood that CEOs are uniquely responsible for select-ing, evaluating, rewarding, motivating, and coach-ing members of their TMTs. In fact, there are atleast two reasons why, from a leadership theoryperspective, the CEO-TMT interface should be par-ticularly salient. First, CEOs are expected to have agreater impact on those with whom they work di-rectly, unlike group leaders at lower levels, whoseinfluence is likely to be constrained and mitigatedby periodic interventions from senior managers(Zaccaro & Klimoski, 2002). Second, and equallyimportant, the often distorted and unrealistic “greatman” views of CEOs are less likely at this interface,because here the CEO’s leadership style is assessedand interpreted by close direct reports, rather thanby those who are more socially distant in a firm(Katz & Kahn, 1978; Shamir, 1995). Indeed, Ham-brick argued that TMT characteristics do not occurin isolation, but rather are significantly traceable toCEO leadership style, suggesting that “the topgroup leader has a disproportionate, sometimesnearly dominating influence, on the group’s vari-ous characteristics and outputs” (1994: 180).

In keeping with Hambrick’s (1994) argument, ourthesis is that a CEO’s transformational proclivityimpacts his or her firm’s engagement in corporateentrepreneurship by shaping characteristics of theTMT. For example, we reason that transformationalCEOs can influence corporate entrepreneurship byencouraging their TMTs to be more responsive tonew opportunities and associated risks and morecommitted to initiating and supporting entrepre-neurial initiatives. Such a CEO can also have aninfluence because, the more responsive and com-mitted the TMT is to a firm’s entrepreneurial vi-sion, the more apt team members are to marshaltheir subordinates’ efforts toward that vision.

HYPOTHESES

Hambrick (1994) proposed a generalized frame-work to capture the essential elements that under-lie the effects of a TMT on firm-level outcomes, andhe suggested that considering these elements inconcert provides a more comprehensive andgrounded starting point than many of the previous

558 JuneAcademy of Management Journal

TMT studies focusing on only part of them. Theseelements include a TMT’s process, or the nature ofinteraction among TMT members; structure, or theorganization within which TMT members enacttheir roles; composition, or the collective charac-teristics of TMT members; and incentives, or thecompensation that TMT members receive. Ham-brick described each element in very broad termsso as to accommodate specific characteristicswithin each. Drawing from the entrepreneurial andupper echelons literatures, we identify four spe-cific TMT characteristics that, we reason, are di-rectly influenced by transformational CEOs and, aswell, are salient to a firm’s pursuit of corporateentrepreneurship. As detailed below, we posit be-havioral integration as a process characteristic, de-centralization of responsibilities as a structuralcharacteristic, risk-taking propensity as a composi-tional characteristic, and long-term compensationas an incentive-based characteristic.

Behavioral Integration

Behavioral integration refers to the degree towhich a TMT engages in mutual and collectiveinteraction. Developed by Hambrick (1994, 1995,1998, 2005), this team process construct parsimo-niously recasts a TMT’s social and task processesinto a meta-construct consisting of one social di-mension (the level of the team’s collaborative be-havior) and two task dimensions (the team’s quan-tity and quality of information exchanged and itsemphasis on joint decision making). Hambrick(1994) reasoned that these mutually reinforcingprocesses, when taken in concert, better capture aTMT’s level of wholeness and unity of effort thando process constructs such as cohesion, social in-tegration, and communication quality. For exam-ple, Hambrick argued that although a purely cohe-sive team may experience “groupthink” (Janis,1983), a behaviorally integrated team should be lessapt to encounter this problem because task pro-cesses, such as extensive sharing of information,should offset this downside of too much cohesive-ness. Thus, the meta-dimensional nature of behav-ioral integration prevents it from attributing moreimport to a single process dimension than iswarranted.

We reason that the attributes of transformationalCEOs can facilitate the achievement of both thesocial and task processes that underlie a TMT’sbehavioral integration. Regarding social processes,studies have shown that by articulating and trans-mitting a strong sense of vision and mission, aswell as inspiring followers to high levels of collec-tive satisfaction, transformational leaders raise fol-

lowers’ social identification, which motivates themto base their self-concepts and self-esteem partly ontheir belonging to the group (Trice & Beyer, 1993).Such social identification has been found to arousepride, respect, trust, and loyalty among followers(Shamir, House, & Arthur, 1993), which shouldmotivate them to cooperate more fully with eachother, engage in more intensive exchange, and con-tribute more to their team (Shamir et al., 1993).Supporting this argument, empirical research hassuggested that teams with leaders who are less neu-rotic or possess more positive views of theworld—a personality trait predictive of transforma-tional leadership (Bono & Judge, 2004)—experi-ence greater social integration (Peterson et al.,2003). Regarding task processes, transformationalleaders are expected to challenge followers to viewproblems from a diversity of perspectives (Bass,1985; Sosik, 1997). Diverse perspectives help ateam avoid cognitive homogeneity and simplifica-tion, both of which tend to reduce the quality ofinformation exchanged and obviate the need forjoint decision making. Moreover, transformationalleaders seek followers’ participation by highlight-ing the importance of cooperation in performingcollective tasks (Jung, Chow, & Wu, 2003). Thus,

Hypothesis 1. CEO transformational leader-ship is positively associated with top manage-ment team behavioral integration.

Furthermore, when a TMT is behaviorally inte-grated, it is more likely to contribute to a firm’spursuit of corporate entrepreneurship. In part, thiscan be explained by Hambrick’s (1994) observationthat when TMTs lack behavioral integration, eventhough the individual members may possess all theinformation, insights, and energies needed to dotheir own jobs, they are unable, or disinclined, toengage in internal information exchange, collabo-ration, and mutual adjustment. Thus, “althoughideas are formed in the minds of individuals, suchinteractions typically play a critical role in devel-oping ideas,” so that absent integration, TMTs areless likely to develop a “community of interaction”(Nonaka, 1994: 15). Nonaka further argued thatwithout a functioning “community” at the top, anorganization cannot effectively develop and exploitnew knowledge, which suggests that corporate en-trepreneurship will suffer. Complementing this ar-gument, Hambrick (1998) suggested that firms withless behaviorally integrated TMTs are more likelyto be impaired by tactical impediments (e.g., un-realized economies of scope, poor coordinationbetween departments, and the like), as well asbureaucratic impediments (e.g., difficulties inreaching consensus on strategic changes).

2008 559Ling, Simsek, Lubatkin, and Veiga

Conversely, by sharing decision making, infor-mation, and effort, behaviorally integrated TMTsare better able to shorten the time needed to de-velop a collective understanding of environmentalchanges, such as customers’ changing needs, andformulate corresponding responses (Hambrick,1998; Smith, Smith, Olian, Sims, O’Bannon, &Scully, 1994). Further, a TMT’s unity of effort andinvolvement in decision making can increase mem-bers’ commitment to new entrepreneurial initia-tives and stimulate their enthusiasm to “sell” theseinitiatives to subordinates, thereby facilitating thetransformation of entrepreneurial ideas into action-able outcomes for the firm (Damanpour, 1991;Stevenson & Jarillo, 1990). In keeping with thisargument, Hambrick (1998) found that firms withbehaviorally integrated TMTs had less difficultyorchestrating transformations to enhance productinnovation, international sales, and cooperativeproduction, and Lubatkin, Simsek, Ling, and Veiga(2006) found that a TMT’s behavioral integrationsignificantly contributed to its firm’s ability to pur-sue new strategic initiatives. Thus,

Hypothesis 2. Top management team behav-ioral integration is positively associated withcorporate entrepreneurship.

Decentralization of Responsibilities

Similar to Hage and Aiken’s (1967) notion ofauthority hierarchy, decentralization of responsi-bilities refers to the degree of concentration of de-cision making in regard to task and operationalactivities within a TMT. In brief, the more respon-sibility over day-to-day tasks and tactical decisionsa firm’s CEO delegates to the members of its TMT,the more responsibilities are decentralized. Thus,this structural characteristic is intended to captureeach member’s flexibility and latitude in makingtactical and operational decisions, as opposed tothe TMT process construct of joint decision makingthat is reflected in behavioral integration and asso-ciated with team-level strategic decisions (Baum &Wally, 2003).

We reason that when a TMT is led by a transfor-mational CEO, responsibilities for specific keytasks needed to achieve important strategic deci-sions are more likely to be distributed among TMTmembers because such leaders seek to encourageself-management and self-development among fol-lowers (Barling, Moutinho, & Kelloway, 2000). Inparticular, by providing followers with greater lat-itude and a greater sense of responsibility, transfor-mational leaders more fully empower and intellec-tually stimulate team members (Dvir, Eden, Avolio,

& Shamir, 2002). And, by being more attentive tofollowers’ individual needs, these leaders enhanceindividual team members’ willingness to take re-sponsibility. Put differently, Avolio and Gibbson(1988) have gone so far as to argue that a major goalof transformational leaders is to develop followers’self-management and self-development skills by al-lowing them to make and implement actions with-out direct supervision or intervention.

Hence, transformational CEOs tend to be morelikely to structure their TMTs in such a way as toencourage decentralization of responsibilitiesamong TMT members as a direct manifestation oftheir leadership style (Zhu, Chew, & Spangler,2005). Research has provided some additional evi-dence to support this tendency by suggesting that“highly agreeable leaders,” as transformationalCEOs have been labeled (Judge & Bono, 2000; Ru-bin, Munz, & Bommer, 2005), tend to promote de-centralization of power within their teams (Peter-son et al., 2003). Indeed, research has shown thattransformational leaders not only empower teammembers by giving them high autonomy (Smith,Montagno, & Kuzmenko, 2004), but also increasetheir perceptions of the amount of authorizedpower they have (Ozaralli, 2003). Therefore, weexpect:

Hypothesis 3. CEO transformational leader-ship is positively associated with top manage-ment team decentralization of responsibilities.

Relative to their CEO, senior managers are closerto the product-markets, customers, and operationaldetails that serve as the stimuli for exploiting inno-vative ideas. In a more centralized structure, how-ever, TMT members may have little motivation torecognize new market opportunities or tacticalproblems, because they lack the authority to act onthem without the CEOs’ approval (Rickards, 1985).Further, the time and effort associated with obtain-ing CEO approval may also impair the firm’s re-sponsiveness. Not surprisingly, therefore, Caruana,Morris, and Vella (1998) found that centrali-zation limited firms’ entrepreneurial behavior,and Damanpour (1991) found a negative relation-ship between centralization and organizationalinnovation.

In contrast, in more responsibility-decentralizedTMTs, decision-making authority is delegated tothose who have greater experience and expertise todiagnose problems and implement solutions intheir domain. Amabile (1995) argued that such del-egation is fundamental to a firm’s ability to inno-vate, because it helps the firm better reap eachmanager’s knowledge. In addition, because delega-tion gives individuals more control over how they


accomplish tasks, it increases their intrinsic moti-vation and facilitates their more creative endeavors(Basadur, 2004). In keeping with this line of rea-soning, Atuahene-Gima (2003) reported that decen-tralization was positively related to new productdevelopment in fast-changing environments. Thus,

Hypothesis 4. Top management team decen-tralization of responsibilities is positively asso-ciated with corporate entrepreneurship.

Risk Propensity

A TMT’s shared preference for risky growth op-portunities is likely to be enhanced by transforma-tional CEOs, because such leaders possess both avision and a measured degree of optimism aboutchange, and they tend to communicate inspiration-al messages that both challenge TMT members tothink “outside of the box” and to instill in them theconfidence that obstacles can be overcome (Bass,1985). These messages should not only mitigateTMT member reluctance to pursue growth oppor-tunities with reasonable risks (Amabile, Schatzel,Moneta, & Kramer, 2004), but also engender inthem a heightened sense of team efficacy, as well asindividual efficacy (Jung & Sosik, 2002), whichshould further enhance the TMT’s willingness totake risks (Zhao, Siebert, & Hills, 2005). Put differ-ently, transformational leaders, as passionate, butnot reckless, advocates for change are better able toconvince followers to frame decisions in an un-skeptical manner and to buy into a risky vision,based at least as much on trust in the leader as incritical analysis (Flynn & Staw, 2004). Further,these leaders tend to emphasize teaching andcoaching that necessitates followers acceptingsome risk, along with the potential for failure, as apart of the developmental process (Nystrom, 1993)Thus, when a transformational CEO stimulates aTMT’s members’ efforts to be innovative and cre-ative by questioning the team’s assumptions, re-framing problems, and encouraging considerationof new ways to approach existing situations, theleader helps to alleviate some of their concernsassociated with such undertakings (Amabile et al.,2004), thereby enhancing the TMT’s propensity totake risk. For all these reasons, we hypothesize:

Hypothesis 5. CEO transformational leader-ship is positively associated with top manage-ment team risk propensity.

Entrepreneurial activities like innovation, ven-turing, and strategic renewal entail considerablerisk, because time, effort, and resources must beinvested before the distribution of their returns is

known, but they also entail the potential for con-siderable return. Consequently, on the basis of theassumption that TMT members’ collective valuesguide the team’s strategic choices (Finkelstein &Hambrick, 1996), it would follow that the greater aTMT’s shared preference for risky growth opportu-nities, the more apt the firm is to engage in corpo-rate entrepreneurship. Simply put, a TMT with agreater propensity for risk taking is more likely tofocus on the potential benefits of risk-entailing en-trepreneurial activities, and more risk-averse TMTsare more likely to focus on the potential losses(Sitkin & Weingart, 1995). Moreover, this greateracceptance of risk might also cascade down tothose whom TMT members manage, thus furthersupporting and enhancing the firm’s entrepreneur-ial proclivity. Supporting this prediction, Knight,Durham, and Locke (2001) found a positive rela-tionship between managerial risk taking and inno-vative task performance, and Gilley, Walters, andOlson (2002) concluded that TMT risk taking had astrong, positive influence on firm innovativeness.Thus, we hypothesize:

Hypothesis 6. Top management team risk pro-pensity is positively associated with corporateentrepreneurship.

Long-Term Compensation

With the exception of Goodwin, Wofford, andWhittington (2001), who suggested that rewards forperformance may be tied to transformational lead-ership, discussion of the use of contingent rewardshas generally been absent from research on trans-formational leadership. This may be because theserewards are viewed as being central to the constructof transactional leadership (Bass, 1990). Neverthe-less, this absence is curious, since effective leaders,regardless of their leadership style, are believed tocarefully manage performance-based incentives.

We reason that what distinguishes transactionalleaders from transformational leaders is theirchoice of performance time horizon. Specifically,transactional leaders, because they seek to monitorand control rather than inspire followers, are morelikely to base rewards on immediate, short-termperformance, and transformational leaders aremore likely to base rewards on longer-term perfor-mance. Indeed, research has shown that when man-agers and employees are rewarded only for theiraccomplishments during a short, fixed time period(e.g., via monthly or semiannual performance re-views), they are primarily motivated to expend justthe effort required by the job and make little effortto realize their full potential (Gomez-Mejia, 1992;


Henderson & Fredrickson, 1996). Clearly, such mo-tivational outcomes, especially among members ofa TMT, are incongruent with a transformationalleadership style and expectations. Said differently,because transformational leaders are guided by theprinciple of encouraging followers to exceed ex-pectations and generate the highest levels of per-formance for the collective (Dvir et al., 2002),they are more apt to base rewards on long-termexpectations.

At the same time, compared to short-term com-pensation, long-term compensation demands thatfollowers place great trust in their leader, becausethey are aware of the time lag between initiatingaction and receiving rewards. Long-term compen-sation also calls for the existence of a visioningmechanism that helps followers see how theiractions can subsequently influence future firmperformance (Cogliser & Brigham, 2004). Conse-quently, given that the ability to provide vision andbuild trusting relationships is more characteristicof transformational than transactional leaders(Bass, 1985), we reason that transformational CEOsare more likely to facilitate the implementation of acompensation system that makes TMT rewardscontingent upon a firm’s long-term performance.Thus,

Hypothesis 7. CEO transformational leader-ship is positively associated with the extent towhich a top management team’s compensationis based upon the long-term performance ofits firm.

Drawing from agency theory, researchers havesuggested a link between the performance objec-tives upon which a manager’s compensation is con-tingent and the manager’s willingness to pursuegrowth-oriented, risk-taking initiatives, like inno-vation, venturing, and strategic renewal (Rajago-palan, 1997). Specifically, it has been recognizedthat short-term compensation clearly links manage-rial rewards to a firm’s short-term performance andgenerates no further risk or commitment on the partof the manager, since the value of a short-termreward is not affected by how well the firm does inthe future (Rajagopalan, 1997). As such, managerswhose compensation is contingent upon the firm’sshort-term performance are less likely to be moti-vated to pursue entrepreneurial initiatives, whichinvolve long lead times between investment andeventual pay-off. Indeed, some researchers (Jones &Butler, 1992; Zahra & Hayton, 2002) have proposedthat managers whose compensation is tied to short-term performance objectives tend to view entre-preneurial initiatives as threatening to their per-formance appraisal and employment, because

short-term performance matrixes such as net in-come are typically adversely affected by suchinitiatives (Rappaport, 1978).

In contrast, however, managers whose compen-sation emphasizes their firms’ long-term perfor-mance are believed to have greater incentive topursue such initiatives, because their personal gaindepends strongly upon the company’s future wel-fare. In other words, when TMT members are eval-uated and compensated on the basis of a longertime horizon, they should embrace a long-term ori-entation, perceive greater ownership of and com-mitment to entrepreneurial initiatives, and conse-quently be more predisposed to pursuing suchundertakings (Black & Scholes, 1973). Consistentlywith this view, Waegelein (1988) found a negativeassociation between short-term bonus plans andR&D expenditures, and Rappaport (1978) found apositive association between compensation plansthat were tied directly to a firm’s long-term perfor-mance and R&D expenditures. Similarly, Holthausen,Larcker, and Sloan (1995) found that the proportionof total compensation tied to a firm’s long-termperformance was positively related to future inno-vation. From this, we expect:

Hypothesis 8. The extent to which a top manage-ment team’s compensation is based upon thelong-term performance of its firm is positivelyassociated with corporate entrepreneurship.

METHODS

Data Collection Techniques and Sample

We chose small-to-medium-sized firms (SMEs)as a sampling frame, because firms of this size (i.e.,employing no more than 500 individuals) pre-sented a more direct litmus test for our hypotheses.Compared to large firms, SMEs have fewer inter-vening levels of management that can dilute theinfluence of their CEOs and TMTs on firm-leveloutcomes. Moreover, SMEs are less constrainedthan large public firms by extraneous influences,like those coming from a powerful board of outsidedirectors, capital markets, and the strategic andadministrative challenges of competing with mul-tiple divisions.

We identified 795 SMEs in New England as ourinitial population, using Dun & Bradstreet’s (D&B’s)Million Dollar Database in 2004. Then, to induceCEOs from this population to participate, we sent aletter endorsed by the director of the regional SmallBusiness Development Center to each CEO. Theletter explained the research project, encouragedparticipation, and indicated that we would followup by telephone. We then began contacting CEOs to


request participation and, if they agreed, to sched-ule meetings. One hundred ninety-three CEOsagreed. During our meetings, the nature of thestudy was further explained, and each CEO re-ceived a survey to complete and return via a post-age-paid envelope. Then, following Smith and col-leagues (1994), we asked these CEOs to identify allof their TMT members and to send them a memoencouraging them to participate along with a sur-vey and postage-paid return envelope. After ex-cluding incomplete surveys, surveys from the firmsthat had less than a 50 percent intrateam responserate, and surveys from CEOs with less than threeyears tenure, we had usable responses from 152firms’ CEOs and 416 of their TMT members, repre-senting 20 percent of the firms in the original sam-pling frame.

The CEOs in our final sample had held theirpositions for 14 years, on average, and had spent anaverage 19 years with their firms. Their firms em-ployed an average of 62 people, had been in busi-ness for 24 years, and reported median sales of $5.1million. All but 14 of the 152 sample firms wereprivately held. Classified in terms of the first twodigits of the North American Industry Classifica-tion System (NAICS), the firms primarily repre-sented three industries: manufacturing (56%),scientific and technical services (18%), and con-struction (12%); the remaining firms were spreadout over several different industries. A paired com-parison test indicated no significant differences infirm age, size, or industry between firms whoseCEOs agreed to participate in our study and thosethat did not.

Finally, the sample firms’ TMTs averaged 4.5members and, on average, 81 percent of the mem-bers of each TMT responded to our survey, a ratethat compares favorably with that of most upperechelons studies (cf. Geletkanycz, 1998; Smith etal., 2005). Moreover, the data showed a high levelof stability in the composition of the sampledTMTs, with 91.4 percent of the surveyed membershaving served on their respective teams for at leastthree years.

Measures

Transformational leadership. This was mea-sured with the Multifactor Leadership Question-naire (MLQ Form 5X-Short) developed by Bass andAvolio (1995) and shown by Avolio, Bass, and Jung(1999) to possess convergent and discriminant va-lidity. Specifically, we asked the members of eachTMT, except the CEO, to evaluate how frequentlytheir CEO engaged in four components of transfor-mational leadership—charisma, inspirational mo-

tivation, intellectual stimulation, and individual-ized consideration—on a scale ranging from 1 (“notat all”) to 5 (“frequently, if not always”).

A one-way analysis of variance (ANOVA) usingfirm affiliation as the independent variable wasperformed on each item as a means to determine ifthere was greater variability in the ratings betweenorganizations than within organizations. A signifi-cant (p � .001) F for each item legitimized ouraggregating the individual team member scores. Wealso used James, Demaree, and Wolf’s (1993) inter-rater reliability coefficient (rwg) to examine the in-tragroup reliability of responses. A value greaterthan or equal to .70 indicates good agreementwithin a group. The average intragroup reliabilityof this scale was .82, further legitimizing the aggre-gation of individual team member scores. The reli-ability for the overall transformational leadershipscore was .90. Appendix A gives the results of aconfirmatory factor analysis for this and all theother constructs in our study and also provides thecomplete scale for each construct.

Behavioral integration. We used the nine-itemscale developed and validated by Simsek, Veiga,Lubatkin, and Dino (2005). Those researchersfound this measure, which they designed to cap-ture a TMT’s level of collaborative behavior, infor-mation exchange, and joint decision making, tohave content, construct, and convergent validity.All TMT members, including the CEOs, assessedthese team processes using a scale ranging from 1(“strongly disagree”) to 5 (“strongly agree”). AnANOVA suggested that individual scores could beaggregated to the team level (p � .001). The averagerwg for the scale was .80. The overall measure ofTMT behavioral integration had a Cronbach’s alphaof .90, which compares favorably to the alpha of .85reported by Simsek and colleagues (2005).

The other main variables, which are describednext, were all rated on the same five-point scaleused for behavioral integration.

Decentralization of responsibility. This seven-item scale was adapted by Sutcliffe (1994) from ascale originally developed by Glick, Huber, Miller,Doty, and Sutcliffe (1990). These items capturedthe extent to which team members perceive theyhave responsibility and authority regarding sevenaspects of day-to-day operations. All TMT mem-bers, including the CEOs, completed this scale.Again, ANOVA (p � .001) and rwg (average rwg �.87) results legitimized the aggregation of individ-ual team member scores. The reliability of the mea-sure was .92.

Risk propensity. Following Gilley et al. (2002),we measured TMT shared preference for riskygrowth opportunities using six items. Again, all


TMT members, including the CEOs, were asked tocomplete this scale, and ANOVA results supportedaggregating individual scores to the team level (p �.001). The average intragroup reliability (rwg) forthe TMT risk propensity scale was .86. However,the third item, “Our team prefers to carefully ana-lyze a situation before moving,” which was reverse-scored, had a low squared multiple correlation (.27)with other items and, thus, was dropped from fur-ther analyses. The reliability of the five-item mea-sure was .78.

Long-term compensation. We used Balkin andGomez-Mejia’s (1990) three-item measure. Thethree items asked TMT members, includingCEOs, to evaluate the executive compensation intheir firms. ANOVA (p � .001) and rwg (averagerwg � .84) results again supported aggregatingindividual scores to the team level. The overallmeasure reliability was adequate (� � .85).

Corporate entrepreneurship. We used Zahra’s(1996) 16-item scale, which broadly measures afirm’s entrepreneurial activities on three dimen-sions: innovation (creating and introducing prod-ucts, production processes, and organizationalmethods), venturing (expanding operations in ex-isting or new markets), and strategic renewal(changing the scope of business and/or its compet-itive approaches). Team members, including theCEOs, were asked to rate their firms’ actual, ratherthan preferred, entrepreneurial activities. Each di-mension consisted of 5 items, except for strategicrenewal, which had 6 items.

An ANOVA again confirmed (p � .001) that theindividual scores could be aggregated to the grouplevel. The average intragroup reliability for thisscale was .83. An examination of aggregated datashowed that for one item of strategic renewal (“Thefirm has divested several unprofitable businessunits”), 33 percent of the cases lacked data, indi-cating that this item was not applicable to many ofthe participant firms. Therefore, we dropped thisvariable, reducing the number of items from 16 to15. The overall measure had a reliability of .88.

Zahra (1996) demonstrated the validity of thisscale by showing that it significantly correlatedwith Miller’s (1983) corporate entrepreneurship in-dex and an objective indicator consisting of a firm’sR&D spending, number of new products, and salesgrowth. Like Zahra, we obtained an objective mea-sure of sales growth from D&B’s Million DollarDatabase (the only performance measure listed inthis database) for 128 of our sample firms for 2005.We correlated this objective measure with our mea-sure of corporate entrepreneurship and found, likeZahra, a positive and significant association (r �.27, p � .01).

Covariates. To reduce the variance caused byother factors that are extraneous to the researchquestion, we included firm size and firm age in thestudy. We also controlled for TMT size, the averageof TMT members’ team tenure, TMT diversity inteam tenure, education level, and functional back-ground, as well as CEO tenure, since all have beenexamined in previous TMT studies (Simsek et al.,2005; Smith et al., 1994). Following Allison (1978),we used the coefficient of variation to measuretenure diversity by dividing each team’s standarddeviation by the team’s mean. We asked respon-dents to report the highest educational degree theyhad attained (i.e., high school, baccalaureate, mas-ter’s, Ph.D., or J.D.). As did Smith and colleagues(1994), we transformed the answers on highest de-gree into years of formal education and then com-puted each group’s coefficient of variation to esti-mate education-level diversity. We also askedrespondents to indicate which of the following cat-egories reflected their functional specialty: finance,accounting, information system, personnel, generalmanagement, marketing, operations, research anddevelopment, or general counsel/secretary. Wethen used Blau’s (1977) index, which assesses thenumber of significant categories in a distributionand how individuals are dispersed over such cate-gories, to compute functional background diver-sity. Firm size, firm age, TMT size, the average ofteam tenure, and CEO tenure were square-root-transformed to achieve normality.

We also controlled for a firm’s unabsorbed slackand its past performance (i.e., before the focal year),as has previous research on corporate entrepre-neurship (Zahra, 1996). Unabsorbed slack was mea-sured by all TMT members using a four-item mea-sure (Simsek, Veiga, & Lubatkin, 2007), of whichtwo items were from Chattopadhyay, Glick, andHuber (2001), and two were from Nohria and Gulati(1996). These items captured the extent to which afirm: (1) has had plentiful resources to produce itsproducts and/or service, (2) has had abundant re-sources for training and rewarding employees toactively think about changes or new business prob-lems, (3) has made a great deal of resources avail-able for experimental projects, and (4) has hadmore resources than promising ideas for using all ofits resources. The average intragroup and overallmeasure reliabilities were adequate (rwg � .84; � �.76). Since objective data on the performance ofsmaller firms are generally not available, we usedtheir CEOs’ evaluation of past firm performance.We asked the CEOs to compare their firms’ pastperformance with that of their major competitorson profitability and growth, using an eight-itemscale from Covin, Prescott, and Slevin (1990). Cron-


bach’s alpha for the measure was .95. Finally, wecontrolled for type of industry and environmentaluncertainty to rule out the influence of a firm’sexternal environment on corporate entrepreneur-ship. As noted above, the firms in our sample werecategorized into four industries: manufacturing,scientific and technical services, construction, andother; we then dummy-coded variables for the firstthree industries. Environmental uncertainty wasevaluated by all TMT members using a four-itemscale reported by Waldman, Ramirez, House, andPuranam (2001); however, our reliability test re-sulted in two items being dropped so that an ac-ceptable reliability (� � .75) could be achieved.The two remaining items gauged the extent towhich a firm’s major industry was (1) very riskyand such that “one false step can mean the firm’sundoing” and (2) very “stressful, exacting, hostile,and hard to keep afloat [in].” The intragroup reli-ability was .79.

ANALYSES AND RESULTS

We used maximum-likelihood structural equa-tion modeling (SEM) to test our model’s hypothesesbecause SEM allows estimation of multiple associ-ations, simultaneously incorporates observed andlatent constructs in these associations, and ac-counts for the biasing effects of random measure-ment error in the latent constructs (Shook, Ketchen,Hult, & Kacmar, 2004). We adopted the two-stepapproach to SEM outlined in Anderson and Gerb-ing (1988), as numerous researchers, such as Hoyleand Panter (1995), have recommended. The firstphase of this approach involves using a confirma-tory factor analysis (CFA) model to fit to the ob-served data. The second phase involves comparinga sequence of nested structural models to gain in-formation concerning the structural model that bestaccounts for the covariances observed between theexogenous and endogenous constructs. Below, wereport results from both phases and results for eachindividual hypothesis.

Phase 1: Confirmatory Factor AnalysisModel Results

For the initial CFA, each latent variable in SEMneeds to be explicitly assigned a metric or a mea-surement range (Kline, 1998). We did so by settinga path for each latent variable to 1.0. Factor load-ings were also set equal to 1.0 for nonlatentvariables.

Multiple indexes were used to assess the fit ofeach model. The criteria examined included chi-square (�2) and the comparative fit index (CFI),

incremental fit index (IFI), Tucker-Lewis index(TLI), and root-mean-square error of approximation(RMSEA). When a hypothesized model fully cap-tures the data from a sample population, the CFI,IFI, and TLI are expected to have values of 1.0, andthe RMSEA, a value of 0.0. Although standards forsuch indexes are difficult to establish, a value of .90or higher for the CFI, IFI, and TLI and a value of .08or lower for the RMSEA are typically suggested asindicating adequate fit (Hu & Bentler, 1999). Thevalues on the fit indexes indicated our measure-ment model had adequate fit (�2[422, n � 152] �679.50, p � .001, CFI � .91, IFI � .91, TLI � .91,and RMSEA � .05).

As Hair, Anderson, Tatham, and Black noted,once an overall CFA model has been accepted,“Each of the constructs can be evaluated separatelyby: (1) examining the indicator loadings for statis-tical significance and (2) assessing the construct’sreliability and variance extracted” (1998: 652). Re-sults for our CFA indicate that the relationshipbetween each indicator and its respective variablewas statistically significant (p � .001), verifying theposited relationships among indicators and con-structs, and thus, convergent validity.

For evidence of discriminant validity, weexamined bivariate interitem correlations. Table 1,which summarizes means, standard deviations,and correlations among all study variables, pro-vides some initial evidence of discriminant valid-ity. No interfactor correlation is above the recom-mended level of .65 (Tabachnick & Fidell, 1996:86). At the same time, we verified that for eachlatent variable the average variance extracted by itsmeasure was larger than its shared variance withany other latent variable. From this result, we con-cluded that problems created by a lack of discrimi-nant validity were not likely to bias our data.

Phase II: Sequence of Nested StructuralModels Results

The second phase of Anderson and Gerbing’s(1988) approach involves making contrasts be-tween sequences of nested structural models to ob-tain information concerning a better-fitting struc-tural model that better accounts for the observedcovariances among the latent constructs. To gain acomplete understanding of our individual hypoth-eses, we tested four nested structural models. Inevaluating these models, we followed the sugges-tions of Joreskog (1993) and Bollen (1989) to assess(1) model fit using various fit indexes and (2) thesignificance of the completely standardized pathestimates, as a test of the model’s hypotheses.

The first structural model examined is the


TA

BL

E1

Mea

ns,

Sta

nd

ard

Dev

iati

ons,

and

Cor

rela

tion

sam

ong

Var

iabl

esa

Var

iabl

esM

ean

s.d

.1

23

45

67

89

1011

1213

1415

1617

1819

1.C

EO

tran

sfor

mat

ion

alle

ader

ship

3.55

.62

2.B

ehav

iora

lin

tegr

atio

n3.

51.4

4.5

2***

3.D

ecen

tral

izat

ion

ofre

spon

sibi

liti

es3.

14.5

8.4

6***

.59*

**

4.R

isk

pro

pen

sity

2.96

.49

.22*

*.2

9***

.29*

**5.

Lon

g-te

rmco

mp

ensa

tion

3.22

.73

.50*

**.5

0***

.46*

**.2

4**

6.C

orp

orat

een

trep

ren

eurs

hip

2.70

.56

.35*

**.2

7***

.35*

**.3

7***

.31*

**

7.F

irm

size

61.6

950

.98

�.0

1�

.10

.11

�.0

1.0

5.0

58.

Fir

mag

e24

.25

19.2

9.0

4.0

1�

.03

�.1

3.0

8�

.12

�.0

99.

TM

Tsi

ze4.

481.

63.0

5�

.07

�.1

1�

.02

�.1

1.1

2.1

3�

.07

10.

TM

Tte

nu

re9.

515.

19.0

3.1

7*.1

1.0

7.2

7**

.01

.01

.23*

*�

.17*

11.

Div

ersi

tyof

team

ten

ure

.61

.31

�.0

3�

.04

�.0

8�

.02

�.0

1.0

2.0

4.1

2.1

8*.0

1

12.

Div

ersi

tyof

edu

cati

onle

vel

.11

.08

�.0

4.0

1�

.09

�.0

4.0

4�

.11

�.0

3�

.05

.12

�.1

1�

.05

13.

Div

ersi

tyof

fun

ctio

nal

back

grou

nd

.48

.23

.01

�.0

4�

.13

�.1

2�

.17*

�.0

6.1

4.1

0.3

5***

�.1

3.1

4.0

3

14.

CE

Ote

nu

re14

.10

8.76

.02

.14

.03

�.0

1.1

9*.0

4�

.18*

.14

�.0

8.6

2***

.33*

**�

.10

�.1

015

.U

nab

sorb

edsl

ack

2.53

.62

.32*

**.2

7**

.13

�.0

4.4

2***

.16*

�.1

3†.1

2�

.04

.27*

*.0

8.1

4.0

5.2

3**

16.

Pas

tfi

rmp

erfo

rman

ce3.

32.7

4�

.12

�.1

3�

.02

.11

�.0

4�

.01

�.0

6.0

6.0

6.0

1.1

3.0

3�

.01

.20*

*�

.01

17.

Ser

vice

ind

ust

ryb

.18

.38

.06

.04

.12

.06

�.1

2.1

7*�

.10

�.1

9*.0

3�

.15†

�.0

4�

.09

�.1

1�

.09

�.2

1**

.11

18.

Con

stru

ctio

nin

du

stry

b.1

1.3

2.0

6.0

6�

.05

�.1

0.0

9�

.15†

.06

�.0

6�

.11

.09

�.0

2.2

0**

�.0

1.0

4.1

5†�

.05

�.1

7*

19.

Man

ufa

ctu

rin

gin

du

stry

b.5

6.5

0�

.11

�.0

6�

.11

�.0

1.0

4.0

1.1

5*.0

3�

.03

�.0

3.0

1�

.09

.08

�.0

8.0

5�

.13

�.5

2***

�.4

2***

20.

En

viro

nm

enta

lu

nce

rtai

nty

2.96

.70

�.1

5†.0

4.0

8�

.09

�.0

2�

.06

.23*

*�

.04

�.0

9.1

4†�

.14

.04

�.0

9.0

4�

.12

�.0

5�

.03

.16*

.02

an

�15

2.b

Cod

edas

ad

um

my

vari

able

.†

p�

.10

*p

�.0

5**

p�

.01

***

p�

.001

covariates model (model 1 in Table 2), which spec-ifies the influences of all team-, firm- and environ-ment-level covariates on the model’s five endoge-nous constructs. These covariates accounted for 27percent of the variance in corporate entrepreneur-ship, 19 percent of the variance in behavioral inte-gration, 15 percent of the variance in decentraliza-tion of responsibilities, 6 percent of the variance inrisk propensity, and 30 percent of the variance inlong-term compensation. Model 2 specifies the ef-fects of transformational leadership on the fourTMT variables without linking transformationalleadership directly to corporate entrepreneurship.Supporting Hypotheses 1, 3, 5, and 7, transforma-tional leadership was positively associated withand explained additional variance in behavioralintegration (.61, p � .001; �R2 � .29), decentraliza-tion of responsibilities (.58, p � .001; �R2 � .26),risk propensity (.27, p � .01; �R2 � .07), and long-term compensation (.57, p � .001; �R2 � .25). Thenext model (model 3) added the effects of four TMTdimensions on corporate entrepreneurship. Sup-porting Hypotheses 4, 6, and 8, respectively, decen-

tralization of responsibilities (.26, p � .05), riskpropensity (.24, p � .05), and long-term compensa-tion (.34, p � .01) were all associated with corpo-rate entrepreneurship, but contrary to Hypothesis2, behavioral integration was not. The four TMTcharacteristics explained additional significantvariance in corporate entrepreneurship (�R2 � .10)

In model 4, the path from transformational lead-ership to corporate entrepreneurship was added.As shown, model 4 (�2[499, n � 152] � 853.40, p �.001, CFI � .92, IFI � .92, TLI � .91, and RMSEA �.05) was slightly superior to the more constrainedmodel 3 (�2[500, n � 152) � 858.57, p � .001, CFI �.91, IFI � .91, TLI � .90, and RMSEA � .06). Usinga two-tailed test, we found the difference in chi-squares of 5.17 (�df � 1) between the models wassignificant (p � .05). This finding suggests that, inaddition to having the hypothesized effects throughTMT characteristics, transformational CEOs alsohave a direct influence on corporate entrepreneur-ship (.21, p � .05; �R2 � .05). Figure 1 presents thepath coefficients in this best-fitting model—thepartially mediated model 4.

TABLE 2Summary of Fit Indexes for Contrasts Based on the Hypothesized Modela

Model df �2 CFI IFI TLI RMSEA Comparison ��2 �dfR2 in Corporate

Entrepreneurship

Covariates onlyModel 1 508 1,018.99*** .83 .84 .75 .09 .27

Covariates plusModel 2: transformational leadership

to behavioral integration,decentralization of responsibility,risk propensity, and long-termcompensation

504 880.93*** .87 .88 .81 .08 Model 2 vs. 1 138.06*** 4

Model 3 (full mediation):transformational leadership tobehavioral integration,decentralization, risk, andcompensation; behavioralintegration, decentralization, risk,and compensation to corporateentrepreneurship

500 858.57*** .91 .91 .90 .06 Model 3 vs. 2 22.36*** 4 .37

Model 4 (partial mediation):transformational leadership tocorporate entrepreneurship(controlled); transformationalleadership to behavioralintegration, decentralization, risk,and compensation; behavioralintegration, decentralization, risk,and compensation to corporateentrepreneurship

499 853.40*** .92 .92 .91 .05 Model 4 vs. 3 5.17* 1 .42

a n � 152. Variable names are abbreviated in models 3 and 4. Covariates include firm size; firm age; TMT size; average TMT tenure; TMTdiversity in team tenure, education level, and functional background; CEO tenure; unabsorbed slack; past firm performance; industry type;and environmental uncertainty.

* p � .05*** p � .001


Post Hoc Analyses

To further verify our findings and gain additionalinsight, we conducted a series of post hoc analyses.First, we tested for the possible influence of commonmethod bias on our findings. Appendix B describesthese analyses. The results suggested that our find-ings were not significantly influenced by such bias.Second, we reestimated the model depicted in Figure1 by excluding CEO scores to determine whether ornot our findings were sensitive to the inclusion/omis-sion of CEO scores in our measure of the four TMTcharacteristics. Again, we arrived at the same patternof significant results (results are available on request).

Our third post hoc analysis was stimulated bythe suggestion in some previous research thattransformational leadership may have a darkside, particularly if charisma is high. For exam-ple, Kark, Shamir, and Chen (2003) argued that aleader’s high charisma may cause followers tobecome dependent on the leader for guidance,and Agle, Nagarajan, Sonnenfeld, and Srinivasan(2006) argued that charisma can be associatedwith a dysfunctional form of narcissism. Conse-quently, to explore this contrarian view, wetested for a possible inverse U-shaped relation-

ship between CEO transformational leadership,the four TMT characteristics, and corporate en-trepreneurship using multiple regression andfound no curvilinear effects (results are availableon request). In addition we tested for a possibledirect relationship between charisma and ourmodel’s other core constructs, but again found nosignificant relationships. This analysis furtherconfirms Bass’s (1985) contention that a leadercan display charisma without being transforma-tional. Moreover, a purely charismatic leader,one who leads followers by invoking blind obe-dience or habituated subordination, is fundamen-tally different from the transformational leaderwho encourages followers to address challenges,think creatively, and develop themselves (Avolioet al., 2004). In sum, because charisma is only acomponent of the multifaceted construct, we con-clude that any dysfunctional consequences ofcharisma are likely mitigated by the other at-tributes of transformational leadership.

Finally, given that behavioral integration is a ratherbroad measure of team process, one speculative ex-planation for the unexpected null Hypothesis 2 find-ing is that this construct is not capturing the sort of

FIGURE 1Final Modela

a Standardized parameter estimates. n � 152. This is a simplified version of the actual model. It does not show indicators, error terms,exogenous factor variances, nor control variables. Solid arrows represent hypothesized paths; dotted arrow represents a path that wasnot hypothesized.

* p � .05** p � .01

*** p � .001


information exchange, collaboration, and joint deci-sion making that is directly relevant to corporate en-trepreneurship. To gain additional insights, we alsoexplored the possibility of an indirect association inthe form of a moderating effect. We considered thispossibility in view of a remark made by Hambrick(2005), who, reflecting on recent advances made inupper echelons research, speculated as follows:“TMT characteristics will predict organizational out-comes only in proportion to the degree that TMTbehavioral integration exists. That is, behavioral inte-gration is a key moderator of the basic upper echelonsrelationships” (2005: 121–122). Absent theoreticaland empirical support, however, we did not for-mally hypothesize such relationships, but wetested for them.

The results are intriguing. Although behavioralintegration did not moderate the relationship be-tween risk-taking propensity and corporate en-trepreneurship or that between long-term com-pensation and corporate entrepreneurship, itpositively moderated the relationship betweendecentralization of responsibilities and corporateentrepreneurship (p � .05). A plot of the interac-tion term (available from the authors) revealed amore positive decentralization-corporate entre-preneurship relationship for firms with higherlevels of behavioral integration (above the samplemean) than for firms with lower levels. At firstglance, it appears that TMT dynamics conduciveto corporate entrepreneurship require both inte-gration and decentralization elements. Behav-ioral integration ensures the quality and com-pleteness of decision making and increases teammembers’ commitment to final decisions, and de-centralization allows team members sufficientauthority to implement those decisions. Thus,when these two TMT characteristics are in sync,they appear to exert a significant influence oncorporate entrepreneurship.

DISCUSSION

Although extant research has been equivocal aboutthe importance of CEOs in affecting firm-level out-comes, it nevertheless seemed intuitive to us that atransformational CEO plays an important role, partic-ularly when it comes to promoting corporate entre-preneurship. Building on upper echelons research,which has documented the influence of TMTs onfirm-level outcomes but has largely ignored the inter-face between CEO and TMT, our conceptual modelpositioned a firm’s TMT as a pivotal interveningmechanism between a transformational CEO and cor-porate entrepreneurship. Our study’s central question

was not if CEOs play a role (because we assumed theywould), but rather, how they do so.

Our findings from a multisource survey of CEOsand members of their TMTs in 152 firms suggest amultifaceted answer. Specifically, we find that trans-formational CEOs play a significant role in directlyshaping four salient TMT characteristics, includingbehavioral integration, decentralization of re-sponsibilities, risk-taking propensity, and long-term compensation. In keeping with Peterson andcolleagues’ (2003) previous research, therefore, ourfindings demonstrate the distinct role played bytop executives at the CEO-TMT interface. More-over, our findings suggest that the recognition ofthis interface by upper echelons theorists couldprove especially fruitful in further specifying theirmodels. Along this line of thinking, it would beinteresting to explore other salient CEO attributesthat might shape TMT characteristics, such as thelevel of executive hubris that is embedded in theirdeeply held core self evaluations (Hiller & Ham-brick, 2004).

Our findings also suggest that with the exception ofbehavioral integration, three TMT characteristics, de-centralization of responsibilities, risk-taking propen-sity, and long-term compensation, were significantlylinked to corporate entrepreneurship. Given that oneof our study’s most important findings is that TMTsare an important intervening mechanism throughwhich the influence of transformational CEOs on cor-porate entrepreneurship is pronounced, future stud-ies might build upon this idea. For example, otherTMT characteristics, such as a team’s entrepreneurialorientation (Lumpkin & Dess, 1996), may need to beexamined to further explain transformational CEOs’impacts on corporate entrepreneurship.

In addition, since our study was only the second toisolate the CEO-TMT interface, and given the paucityof research and theory, we decided to focus on mod-eling main effects. Future research might address fin-er-grained and alternative conceptualizations of therelationships between the variables specified in ourmodel. For example, although we did not find evi-dence to support a direct link between behavioralintegration and corporate entrepreneurship, in ourpost hoc analysis we did find that behavioral integra-tion positively moderated the influence of decentral-ization of responsibility on corporate entrepreneur-ship. Although this finding is exploratory, it doesraise the broader issue for future research as to themore precise nature of the relationships betweenTMT characteristics and firm-level outcomes.

Our findings also showed a direct link betweentransformational leadership and corporate entre-preneurship, even after a CEO’s shaping influ-ence on his or her TMT is accounted for. We did


not hypothesize this link, but we neverthelessfind the support for it to be intriguing. It suggeststhat, in addition to the CEO-TMT interface, othersalient intervening mechanisms may be operat-ing. One particularly interesting avenue might beto examine a multilevel interface model as anattempt to elaborate more fully on how a trans-formational CEO’s shaping influence on the TMTcascades downward in a firm. Research in thisvein could address such intriguing questions asthese: To what extent does a transformationalCEO also influence the transformational behav-iors of the other TMT members? To what extentdo transformational TMT members encourageand facilitate, among those reporting to them,further entrepreneurial initiatives?

Another interesting avenue might be to examinetransformational CEO impact on organizational cul-ture, which is expected to nurture creative efforts andfacilitate diffusion of learning within an entire organ-ization, as an influence mechanism apart from TMTelements. Alternatively, future researchers might alsoexamine transformational CEO influence on recruit-ment and selection of employees. For example, po-tential employees who share the essential values andbeliefs embodied in transformational leadership maybe more likely to be attracted to firms whose culturalvalues are consistent with theirs. In turn, such em-ployees may also be more predisposed to contributeto as well as support their firm’s entrepreneurial ini-tiatives. Clearly, future research like this, and studiesusing archival measures of corporate entrepreneur-ship, might well uncover a number of transforma-tional CEOs’ “trace effects” that, like light through aprism, further amplify the entrepreneurial proclivityof firms.

In addition to its theoretical contribution, our studyhas important practical implications. As Dess andPicken (2000) emphasized, the business environmentin the 21st century requires organizations to be con-tinuously innovative by harnessing the collectiveknowledge, skills, and creative efforts of firm mem-bers. As our findings have shown, a transformationalCEO is one important driver of such outcomes. Fur-thermore, given that several aspects of leadership be-havior can be learned or adjusted (Kirkbride, 2006),our findings suggest that organizations can improvetheir corporate entrepreneurship by helping CEOs todevelop and display transformational leadership be-haviors through training and mentoring.

We believe our findings are robust, in that we tooka number of recommended steps to mitigate concernsabout informant bias, nonresponse bias, commonmethod variance, and measurement error. None ofthese steps suggested a bias. We are mindful, ofcourse, that as in most examinations of firm-level

effects, facets of our research design likely limit theextent to which we can place full confidence in theresults. However, given that our tests revealed nobias, we view these limits as acceptable. We were alsoable to rule out the possibility of charisma playing adysfunctional role by increasing follower depen-dency, as some have suggested. Perhaps—as Ham-brick (1994)—reasoned, senior executives in a TMTare generally more experienced and confident as wellas charged with greater responsibilities than arelower-level subordinates and, therefore, TMT mem-bers are less susceptible to such influence, espe-cially when their boss is encouraging and support-ive and stimulates them to challenge assumptionsand go beyond their own self-interest.

Finally, although our theoretical model impliescausality, it should not be inferred because ourstudy was correlational and did not involve themanipulation of variables. As such, the present re-search design cannot rule out the possibility ofreverse/reciprocal causality; for example, greatercorporate entrepreneurship might encourage a CEOto act more transformationally in a kind of virtuouscycle. Extant theory and research have not madethis case—yet assuming one could make it, only awell-designed longitudinal study could test for it.However, even this design would be subject to con-cern about exogenous influences that may play arole over time. It also remains an open question asto whether or not our findings can be generalized tolarge firms. As noted earlier, small-to-medium-sized firms differ from large ones in that they havefewer intervening levels of management and areless constrained by extraneous influences. Thus,although our findings offer a reasonable point ofdeparture for examining large public firms, futureresearch is needed to examine the veracity of ourfindings in that context. On the other hand, giventhat small firms represent a vital component ofmost nations’ economies and are, by far, the mostubiquitous form of business organization in theUnited States, generating about 70 percent of alljobs in the country (Small Business Association,2003), we believe our findings make a timely andrelevant contribution.

In sum, this study puts CEOs back into upperechelons research by elaborating on the often ig-nored CEO-TMT interface and demonstrates thattransformational CEOs are capable of shaping TMTcharacteristics that are salient drivers of corporateentrepreneurship. Of course, whether or not trans-formational CEOs are the sine qua non of theirfirms’ entrepreneurial proclivity remains an open,yet intriguing, question.


REFERENCES

Agle, B. R., Nagarajan, N. J., Sonnenfeld, J. A., & Sriniva-san, D. 2006. Does CEO charisma matter? An empir-ical analysis of the relationships among organization-al performance, environmental uncertainty, and topmanagement team perceptions of CEO charisma.Academy of Management Journal, 49: 161–174.

Allison, P. D. 1978. Measures of inequality. AmericanSociological Review, 43: 865–880.

Amabile, T. 1995. Within you, without you: The socialpsychology of creativity, and beyond. In B. M. Staw(Ed.), Psychological dimensions of organizational be-havior: 423–443. Englewood Cliffs, NJ: Prentice-Hall.

Amabile, T. M., Schatzel, E. A., Moneta, G. B., & Kramer,S. J. 2004. Leader behaviors and the work environmentfor creativity: Perceived leader support. LeadershipQuarterly, 15: 5–32.

Anderson, J. C., & Gerbing, D. W. 1988. Structural equa-tion modeling in practice: A review and recom-mended two-step approach. Psychological Bulletin,103: 411–423.

Atuahene-Gima, K. 2003. The effects of centrifugal andcentripetal forces on product development speedand quality: How does problem solving matter?Academy of Management Journal, 46: 359–373.

Avolio, B. J., Bass, B. M., & Jung, D. 1999. Reexaminingthe components of transformational and transac-tional leadership using the Multifactor LeadershipQuestionnaire. Journal of Organization and Occu-pational Psychology, 72: 441–462.

Avolio, B. J., & Gibbson, T. C. 1988. Developing transfor-mational leaders: A life span approach. In J. A.Kanungo & R. N. Kanungo (Eds.), Charismatic lead-ership: The elusive factor in organizational effec-tiveness: 276–308. San Francisco: Jossey-Bass.

Avolio, B. J., Zhu, W., Koh, W., & Puja, B. 2004. Transfor-mational leadership and organizational commitment:Mediating role of psychological empowerment andmoderating role of structural distance. Journal of Or-ganizational Behavior, 25: 951–968.

Balkin, D. B., & Gomez-Mejia, L. 1990. Matching com-pensation and organizational strategies. StrategicManagement Journal, 11: 153–169.

Bantel, K., & Jackson, S. E. 1989. Top management andinnovations in banking: Does the demography of thetop team make a difference? Strategic ManagementJournal, 10: 107–124.

Barling, J., Moutinho, S., & Kelloway, E. K. 2000. Trans-formational leadership and group performance:The mediating role of affective commitment.School of Business, Queen’s University, Kingston,Ontario.

Basadur, M. 2004. Leading others to think innovativelytogether: Creative leadership. Leadership Quar-terly, 15: 103–121.

Bass, B. M. 1985. Leadership and performance beyondexpectations. New York: Free Press.

Bass, B. M. 1990. Bass & Stogdill’s handbook of lead-ership (3rd ed.). New York: Free Press.

Bass, B. M., & Avolio, B. J. 1995. Multifactor LeadershipQuestionnaire (form 5x-short). Menlo Park, CA:Mind Garden.

Baum, J. R., & Wally, S. 2003. Strategic decision speedand firm performance. Strategic Management Jour-nal, 24: 1107–1129.

Black, F., & Scholes, M. 1973. Pricing of options andcorporate liabilities. Journal of Political Economy,81: 637–654.

Blau, P. M. 1977. Inequality and heterogeneity. NewYork: Free Press.

Bollen, K. A. 1989. Structural equations with latentvariables. New York: Wiley.

Bono, J. E., & Judge, T. A. 2004. Personality and transfor-mational and transactional leadership: A meta-anal-ysis. Journal of Applied Psychology, 89: 901–910.

Carpenter, M. A., Geletkanycz, M. A., & Sanders, G. 2004.Upper echelons research revisited: Antecedents, el-ements, and consequences of top management teamcomposition. Journal of Management, 30: 749–778.

Caruana, A., Morris, M. H., & Vella, J. 1998. The effect ofcentralization and formalization on entrepreneur-ship in export firms. Journal of Small BusinessManagement, 36: 16–29.

Chattopadhyay, P., Glick, W., & Huber, G. P. 2001. Orga-nizational actions in response to threats and oppor-tunities. Academy of Management Review, 44:937–956.

Cogliser, C. C., & Brigham, K. H. 2004. The intersection ofleadership and entrepreneurship: Mutual lessons tobe learned. Leadership Quarterly, 15: 771–799.

Covin, J. G., Prescott, J. E., & Slevin, D. P. 1990. Theeffects of technological sophistication on strategicprofiles, structure and firm performance. Journal ofManagement Studies, 27: 485–510.

Damanpour, F. 1991. Organizational innovation: A meta-analysis of effects of determinants and moderators.Academy of Management Journal, 34: 555–590.

Dess, G. G., & Picken, J. C. 2000. Changing roles: Leader-ship in the 21st century. Organizational Dynamics,29(4): 18–33.

Dvir, T., Eden, D., Avolio, B. J., & Shamir, B. 2002. Impactof transformational leadership on follower develop-ment and performance: A field experiment. Academyof Management Journal, 45: 735–744.

Finkelstein, S., & Hambrick, D. C. 1996. Strategic lead-ership: Top executives and their effects on organ-izations. St. Paul: West.

Flynn, F. J., & Staw, B. M. 2004. Lend me your wallets:The effect of charismatic leadership on external sup-


port for an organization. Strategic ManagementJournal, 25: 309–330.

Geletkanycz, M. A. 1998. The salience of “culture’s con-sequences”: The effect of cultural values on top ex-ecutive commitment to the status quo. StrategicManagement Journal, 18: 615–634.

Gilley, M., Walters, B., & Olson, B. 2002. Top manage-ment team risk taking propensities and firm perfor-mance: Direct and moderating effects. Journal ofBusiness Strategies, 19: 95–114.

Glick, W. H., Huber, G. P., Miller, C. C., Doty, D. H., &Sutcliffe, K. M. 1990. Studying changes in organiza-tional design and effectiveness: Retrospective eventhistories and periodic assessments. OrganizationScience, 1: 293–312.

Gomez-Mejia, L. 1992. Structure and process of diversifica-tion, compensation strategy, and firm performance.Strategic Management Journal, 13: 381–397.

Goodwin, V. L., Wofford, J. C., & Whittington, J. L. 2001.A theoretical and empirical extension to the trans-formational leadership construct. Journal of Orga-nizational Behavior, 22: 759–774.

Hage, J., & Aiken, M. 1967. The relationship of centraliza-tion to other structural properties. AdministrativeScience Quarterly, 12: 72–92.

Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C.1998. Multivariate data analysis. Upper SaddleRiver, NJ: Prentice-Hall.

Hambrick, D. C. 1994. Top management groups: A con-ceptual integration and reconsideration of the teamlabel. In B. M. Staw & L. L. Cummings (Eds.), Re-search in organizational behavior, vol. 16: 171–214. Greenwich, CT: JAI Press.

Hambrick, D. C. 1995. Fragmentation and the other prob-lems CEOs have with their top management teams.California Management Review, 37(3): 110–128.

Hambrick, D. C. 1998. Corporate coherence and the topmanagement team. In D. C. Hambrick, D. A. Nadler,& M. L. Tushman (Eds.), Navigating change: HowCEOs, top teams, and boards steer transformation:123–140. Boston: Harvard Business School Press.

Hambrick, D. C. 2005. Upper echelon theory: Origins,twists and turns, and lessons learned. In K. G. Smith& M. A. Hitt (Eds.), The Oxford handbook of man-agement theory: The process of theory develop-ment: 109–127. Oxford, U.K.: Oxford UniversityPress.

Hambrick, D. C., & Mason, P. A. 1984. Upper echelon:The organization as a reflection of its top managers.Academy of Management Review, 9: 193–206.

Henderson, A. D., & Fredrickson, J. W. 1996. Information-processing demands as a determinant of CEO com-pensation. Academy of Management Journal, 3:575–606.

Hiller, N. J., & Hambrick, D. C. 2004. Conceptualizingexecutive hubris: The role of (hyper-)core self-eval-

uations in strategic decision-making. Strategic Man-agement Journal, 26: 297–319.

Holthausen, R. W., Larcker, D. F., & Sloan, R. G. 1995.Business unit innovation and the structure of exec-utive compensation. Journal of Accounting andEconomics, 19: 279–313.

House, R. J., & Aditya, R. M. 1997. The social scientificstudy of leadership: Quo vadis? Journal of Manage-ment, 23: 409–473.

Howell, J. M., & Higgins, C. A. 1990. Champions oftechnological innovation. Administrative ScienceQuarterly, 35: 317–341.

Hoyle, R. H., & Panter, A. T. 1995. Writing about struc-tural equation models. In R. H. Hoyle (Ed.), Struc-tural equation modeling: Concepts, issues andapplications: 158–176. Thousand Oaks, CA: Sage.

Hu, L., & Bentler, P. M. 1999. Cutoff criteria for fit in-dexes in covariance structure analysis: Conventionalcriteria versus new alternatives. Structural Equa-tion Modeling, 6: 1–55.

Jackson, S. E. 1992. Consequences of group compositionfor the interpersonal dynamics of strategic issue pro-cessing. In P. Shrivastava, A. Huff, & J. Dutton (Eds.),Advances in strategic management, vol. 8: 345–382. Greenwich, CT: JAI Press.

James, L. R., Demaree, R. G., & Wolf, G. 1993. rwg: Anassessment of within-group inter-rater agreement.Journal of Applied Psychology, 78: 306–339.

Janis, I. L. 1983. Groupthink: Psychological studies of pol-icy decisions and fiascos. Boston: Houghton Mifflin.

Jones, G., & Butler, J. 1992. Managing internal corporateentrepreneurship: An agency theory perspective.Journal of Management, 18: 733–749.

Joreskog, K. G. 1993. Testing structural equation models.In K. A. Bollen & J. S. Long (Eds.), Testing structuralequation models: 294–316. New York: Wiley.

Judge, T. A., & Bono, J. E. 2000. Five-factor model ofpersonality and transformational leadership. Jour-nal of Applied Psychology, 85: 751–765.

Jung, D., Chow, C., & Wu, A. 2003. The role of transfor-mational leadership in enhancing organizational in-novation: Hypotheses and some preliminary find-ings. Leadership Quarterly, 14: 525–544.

Jung, D., & Sosik, J. J. 2002. Transformational leadershipin work groups: The role of empowerment, cohesive-ness, and collective-efficacy on perceived group per-formance. Small Group Research, 33: 313–336.

Kark, R., Shamir, B., & Chen, G. 2003. The two faces of trans-formational leadership: Empowerment and dependency.Journal of Applied Psychology, 88: 246–255.

Katz, D., & Kahn, R. L. 1978. The social psychology oforganizations. New York: Wiley.

Kirkbride, P. 2006. Developing transformational leaders:The full range leadership model in action. Industrialand Commercial Training, 38: 23–32.


Klimoski, R. J., & Koles, K. L. K. 2001. The chief execu-tive officer and top management team interface. InS. J. Zaccaro & R. J. Klimoski (Eds.), The nature oforganizational leadership: Understanding the per-formance: 219–269. San Francisco: Jossey-Bass.

Kline, R. B. 1998. Principles and practice of structuralequation modeling. New York: Guilford Press.

Knight, D., Durham, C. C., & Locke, E. A. 2001. The rela-tionship of team goals, incentives, and efficacy to stra-tegic risk, tactical implementation, and performance.Academy of Management Journal, 44: 326–338.

Kotter, J. P. 1995. Leading change: Why transformationefforts fail. Harvard Business Review, 73(2): 59–67.

Lubatkin, M. H., Simsek, Z., Ling, Y., & Veiga, J. F. 2006.Ambidexterity and performance in small- to medi-um-sized firms: The pivotal role of top managementteam behavioral integration. Journal of Manage-ment, 32: 646–672.

Lumpkin, G. T., & Dess, G. G. 1996. Clarifying the entre-preneurial orientation construction and linking it toperformance. Academy of Management Review,21: 135–172.

Meindl, J. R. 1998. The romance of leadership as a fol-lower-centric theory: A social constructionist ap-proach. In F. Dansereau & F. J. Yammarino (Eds.),Leadership—The multiple-level approaches: Con-temporary and alternative: 427–458. Stamford, CT:JAI Press.

Miller, D. 1983. Entrepreneurship correlates in threetypes of firms. Management Science, 29: 770–791.

Nohria, N., & Gulati, R. 1996. Is slack good or bad forinnovation? Academy of Management Journal, 39:1245–1264.

Nonaka, L. 1994. A dynamic theory of organizational knowl-edge creation. Organization Science, 5: 714–737.

Nystrom, P. C. 1993. Organizational cultures, strategies,and commitment in health care organizations. HealthCare Management Review, 18(1): 43–49.

O’Reilly, C. A., Snyder, R. C., & Boothe, J. N. 1993.Executive team demography and organizationalchange. In G. P. Huber & W. H. Glick (Eds.), Orga-nizational change and redesign: Ideas and insightsfor improving performance: 147–175. New York:Oxford University Press.

Ozaralli, N. 2003. Effects of transformational leadershipon empowerment and team effectiveness. Leader-ship and Organization Development Journal, 24:335–344.

Peterson, R. S, Smith, D. B., Martorana, P. V., & Owens,P. D. 2003. The impact of chief executive officer per-sonality on top management team dynamics: Onemechanism by which leadership affects organizationalperformance. Journal of Applied Psychology, 88: 795–808.

Podsakoff, P., MacKenzie, S., Lee, J., & Podsakoff, N. 2003.Common method biases in behavioral research: A crit-

ical review of the literature and recommended reme-dies. Journal of Applied Psychology, 88: 879–903.

Rajagopalan, N. 1997. Strategic orientations, incentiveplan adoptions, and firm performance: Evidencefrom electric utility firms. Strategic ManagementJournal, 18: 761–785.

Rappaport, A. 1978. Executive incentives versus corporategrowth. Harvard Business Review, 56(4): 81–88.

Rickards, T. 1985. Stimulating innovation—A systemsapproach. London: Pinter.

Rubin, R. S., Munz, D. C., & Bommer, W. H. 2005. Leadingfrom within: The effects of emotion recognition andpersonality on transformational leadership behavior.Academy of Management Journal, 48: 845–858.

Shamir, B. 1995. Social distance and charisma: Theoret-ical notes and an exploratory study. LeadershipQuarterly, 6: 19–47.

Shamir, B., House, R. J., & Arthur, M. 1993. The motiva-tional effects of charismatic leadership: A self-conceptbased theory. Organization Science, 4: 577–594.

Shook, C. L., Ketchen, D. J., Hult, G. T. M., & Kacmar, K. M.2004. An assessment of the use of structural equationmodeling in strategic management research. StrategicManagement Journal, 25: 397–404.

Simsek, Z., Veiga, J. F., & Lubatkin, M. 2007. The impactof the competitive environment on corporate entre-preneurship: Toward understanding discretionaryslack’s pivotal role. Journal of Management Stud-ies, 44: 1398–1424.

Simsek, Z., Veiga, J. F., Lubatkin, M., & Dino, R. N. 2005.Modeling the multilevel determinants of top man-agement team behavioral integration. Academy ofManagement Journal, 48: 69–84.

Sitkin, S., & Weingart, L. R. 1995. Determinants of riskydecision-making behavior: A test of the mediatingrole of risk perceptions and propensity. Academy ofManagement Journal, 38: 1573–1592.

Small Business Association. 2003. Administrative eco-nomic plan (release # 03–02). Washington, DC: SBA.

Smith, B., Montagno, R. V., & Kuzmenko, T. N. 2004.Transformational and servant leadership: Contentand contextual comparisons. Journal of Leadershipand Organizational Studies, 10: 80–91.

Smith, K. G., Smith, K. A., Olian, J. D., Sims, H. P.,O’Bannon, D. P., & Scully, J. A. 1994. Top manage-ment team demography and process: The role ofsocial integration and communication. Administra-tive Science Quarterly, 39: 412–438.

Sosik, J. J. 1997. Effects of transformational leadership andanonymity on idea generation in computer-mediatedgroups. Group and Organization Management, 22:460–487.

Srivastava, A., & Lee, H. 2005. Predicting order and tim-ing of new product moves: The role of top manage-


ment in corporate entrepreneurship. Journal ofBusiness Venturing, 20: 459–481.

Stevenson, H. H., & Jarillo, J. C. 1990. A paradigm ofentrepreneurship: Entrepreneurship management.Strategic Management Journal 11(summer specialissue): 17–27.

Sutcliffe, K. M. 1994. What executives notice: Accurateperceptions in top management teams. Academy ofManagement Journal, 37: 1360–1378.

Tabachnick, B. G., & Fidell, L. S. 1996. Using multivar-iate statistics. New York: Harper Collins CollegePublishers.

Thomas, A. B. 1988. Does leadership make a difference toorganizational performance? Administrative Sci-ence Quarterly, 33: 388–400.

Trice, H. M., & Beyer, J. M. 1993. The cultures of workorganizations. Englewood Cliffs, NJ: Prentice-Hall.

Waegelein, J. F. 1988. The association between the adoptionof short-term bonus plans and corporate expenditures.Journal of Accounting and Public Policy, 7: 43–63.

Waldman, D. A., Javidan, M., & Varella, P. 2004. Charis-matic leadership at the strategic level: A new appli-cation of upper echelon theory. Leadership Quar-terly, 15: 355–380.

Waldman, D. A., Ramirez, G. G., House, R. J., & Puranam,P. 2001. Does leadership matter? CEO leadershipattributes and profitability under conditions of per-ceived environmental uncertainty. Academy ofManagement Journal, 44: 134–143.

Waldman, D. A., & Yammarino, F. J. 1999. CEO charis-matic leadership: Levels-of-management and levels-of-analysis effects. Academy of Management Re-view, 24: 266–285.

Wiersema, M., & Bantel, K. 1992. Top management teamdemography and corporate strategic change. Acad-emy of Management Journal, 35: 91–121.

Zaccaro, S. J., & Klimoski, R. 2002. The interface of lead-ership and team processes. Group and OrganizationManagement, 27: 4–13.

Zahra, S. A. 1996. Governance, ownership, and corporateentrepreneurship: The moderating impact of indus-try technological opportunities. Academy of Man-agement Journal, 39: 1713–1735.

Zahra, S. A., & Hayton, J. C. 2002. Organizational cul-ture and entrepreneurship in family firms: Aresource-based analysis. Paper presented at the Re-search Workshop in Family Business, The WhartonSchool, University of Pennsylvania, Philadelphia.

Zhao, H. S., Seibert, S. E., & Hills, G. E. 2005. The medi-ating role of self-efficacy in the development of en-trepreneurial intentions. Journal of Applied Psy-chology, 90: 1265–1272.

Zhu, W., Chew, I. K. H., & Spangler, W. D. 2005. CEOtransformational leadership and organizational out-comes: The mediating role of human-capital-enhanc-ing human resource management. Leadership Quar-terly, 16: 39–52.

APPENDIX AResults of Confirmatory Factor Analysis for Measures

Constructs and Itemsa �

Factor

1 2 3 4 5 6

Transformational leadership: The extent to which a CEO . . . .90Talks about . . . values . . . .71Instills pride . . . .86Specifies the importance . . . .84Goes beyond self-interest . . . .75Build . . . my respect . . . .87Considers the moral . . . .72Displays power and confidence . . . .60Emphasizes the collective mission . . . .83Talks optimistically . . . .60Talks enthusiastically . . . .79Articulates a compelling vision . . . .77Expresses confidence . . . .81Reexamines critical assumptions . . . .83Seeks differing perspectives . . . .70Suggests different angles . . . .74Suggests new ways . . . .73Spends time teaching . . . .72Treats me as an individual . . . .66Considers my needs . . . .64Helps my strengths . . . .76

a These are word indicators only, not the complete MLQ items. The items were reproduced by special permission of the Publisher, MINDGARDEN, Inc. www.mindgarden.com, from the “Multifactor Leadership Questionnaire for Research” by Bernard M. Bass and Bruce J.Avolio. Copyright 1995 by Bernard M. Bass and Bruce J. Avolio. All rights reserved. Further reproduction is prohibited without thePublisher’s written consent.


APPENDIX A(Continued)

Constructs and Itemsa �

Factor

1 2 3 4 5 6

Behavioral integration: The extent to which TMT members . . . .90Let each other know when their actions affect another team member’s work .67Have a clear understanding of the job problems and needs of other team members .71Discuss their expectations of each other .71Volunteer to help some team members, who are busy, to manage their workload .87Are flexible about switching responsibilities to make things easier for each other .88Are willing to help each other complete jobs and meet deadlines .87Are effective in developing high-quality ideas .62Are effective in generating high-quality solutions .68Are effective in making decisions that require high levels of creativity and

innovativeness.53

Decentralization of responsibilities: The extent to which the CEO decentralizes thedecision making regarding . . .

.92

Entry into new market segments .77Changing policies that affect a portion of the firm .85Hiring midlevel management personnel .83Making capital expenditures greater than 1% of our firm’s annual budget .81Altering responsibilities of first-line managers .75Changing the way our firm serves the customers/clients .73Making changes in the way our firm produces its products/services .73

Risk propensity: The extent to which the TMT . . . .78Has a strong preference for high-risk projects .59Views bold acts as useful and common practice .62Favors the tried and trueb �.69Has a tendency to follow competitors instead of introducing new products

ourselves firstb�.61

Prefers to let other firms in our industry assume the risk of product or processinnovations before adopting them in our firmb

�.60

Long-term compensation: The extent to which executive compensation . . . .85Focuses top managers’ attention on the long-term (two or more years) goals of the

firm.67

Rewards top managers for short-term accomplishments during a fixed time period(e.g., semiannual or annual firm performance reviews)b

�.71

Recognizes that long-term firm results are more important than short-term firmresults

.74

Corporate entrepreneurship: The extent to which the firm . . . .88Has spent heavily (well above the industry average) on product development .62Has introduced a large number of new products to the market .72Has acquired significantly more patents than its major competitors .66Has pioneered the development of breakthrough innovations in its industry .69Has spent on new product development initiatives .58Has entered new markets .71Has established or sponsored new ventures .73Has found new niches in current markets .71Has financed start-up business activities .59Has created new semi and autonomous units .58Has changed its competitive approach (strategy) for each business unit .62Has recognized operations, units, and divisions to ensure increased coordination

and communication among business units.62

Has redefined the industries in which it competes .63Has introduced innovative human resource programs .71Has been first in the industry to introduce new business concepts and practices .62

a These are word indicators only, not the complete MLQ items. The items were reproduced by special permission of the Publisher, MINDGARDEN, Inc. www.mindgarden.com, from the “Multifactor Leadership Questionnaire for Research” by Bernard M. Bass and Bruce J.Avolio. Copyright 1995 by Bernard M. Bass and Bruce J. Avolio. All rights reserved. Further reproduction is prohibited without thePublisher’s written consent.

b These items are reverse-scored.


APPENDIX B

Addressing Common Method Bias

Given our reliance on self-reports, we took severalsteps to mitigate and detect potential common methodbias. To mitigate the biasing effects, in the relationshipbetween CEO transformational leadership and TMT char-acteristics, we excluded CEOs from the assessment of theformer while including them for the latter. In otherwords, the respondents for the transformational leader-ship and TMT variables were partially different. Second,we collected the data for each core variable from multiplerespondents rather than for a single respondent per firm.This procedure could reduce the method biases causedby respondents’ individual affect or mood (Podsakoff,MacKenzie, Lee, & Podsakoff, 2003). Third, we varied thescale anchors and format in the questionnaire. This way,the method biases caused by commonalities across mea-sures were minimized (Podsakoff et al., 2003).

To further detect the possible influence of commonmethod variance, we conducted several additional testsof our model. In the first test, we reassessed the modelpresented in Figure 1 by only utilizing responses fromthe CEOs on corporate entrepreneurship (�2 [499, n �152] � 868.34, p � .001, CFI � .90, IFI � .91, TLI � .90,and RMSEA � .06), and in the second test we only usedresponses from TMT members other than the CEO (�2

[499, n � 152] � 859.12, p � .001, CFI � .92, IFI � .91,TLI � .91, and RMSEA � .05). Both tests generatedresults that were consistent in direction and significancewith what we found when taking responses from entireTMTs inclusive of the CEOs. In the third test, followingPodsakoff and colleagues’ (2003) recommendation, were-estimated the model in Figure 1 with all the indicatorvariables loading on a general common method factor.Results indicate that although the general commonmethod factor did improve model fit, none of the indi-vidual path coefficients corresponding to the relation-ships between the indicators and the general methodfactor were significant. The coefficient estimates for hy-pothesized paths were similar to those obtained earlier.These findings are consistent with research on commonmethod bias indicating that although method bias may be

present, it does not always significantly affect results orconclusions (Podsakoff et al., 2003). Results of theseanalyses are available from the authors.

Yan Ling ([email protected]) is an assistant professor ofmanagement at George Mason University. She receivedher Ph.D. from the University of Connecticut. Her re-search interests revolve around managerial issues insmall- to medium-sized companies, including top man-agement teams, leadership, strategic entrepreneurship,family ownership, and corporate governance.

Zeki Simsek ([email protected]) is an as-sistant professor of management at the Management De-partment at the University of Connecticut. He receivedhis Ph.D. from University of Connecticut. His currentresearch interests are in the area of strategic managementand strategic entrepreneurship, with emphasis on topmanagement teams, corporate governance and valuation,and entrepreneurial initiatives.

Michael H. Lubatkin ([email protected])is the Thomas, John, and Bette Wolff Family Chaired Pro-fessor of Strategic Entrepreneurship at the University ofConnecticut and a professor of Management at EM Lyon. ADBA from the University of Tennessee, Dr. Lubatkin spe-cializes in corporate diversification issues as they pertain toproblems of core competence transfer and risk managementbetween divisions and across national borders, and in cor-porate governance issues as they pertain to family andentrepreneurial firms.

John F. Veiga ([email protected]) earnedhis DBA from Kent State University; he holds the North-east Utilities Chair in Business Ethics and is a Board ofTrustees Distinguished Professor and the head of theManagement Department at the University of Connecti-cut. His current research interests focus on top manage-ment teams, teleworking, strategic entrepreneurship,managerial ethics and people with disabilities.


document5

Documents

firms ceo

transformational ceos

ceo impact

transformational ceos

firms ability

transformational vision

effects of ceo personality

firms product innovation