Towards understanding the direct and indirect effects of CEOs' transformational leadership on firm...

8/12/2019 Towards understanding the direct and indirect effects of CEOs' transformational leadership on firm innovation Don

1/13

Towards understanding the direct and indirect effects of CEOs'transformational leadership on rm innovation

Dongil (Don) Jung a , , Anne Wu b ,1, Chee W. Chow c,2a Department of Management, College of Business, San Diego State University, 5500 Campanile Dr., San Diego, CA 92182, USAb Department of Accounting, National Chengchi University, Wenshan, Taipei, Taiwanc School of Accountancy, San Diego State University, San Diego, CA 92182, USA

a r t i c l e i n f o a b s t r a c t

This study seeks to advance understanding of how transformational leadershipby topmanagers(CEOs) can affect theircompanies' innovativeness.We propose a model that includes bothdirecteffects and indirect effects moderated by aspects of organizational culture, structure, and theexternal environment. The predicted effects are tested with data collected through multiplesources on 50 Taiwanese electronics and telecommunications companies. The results supportthe expectation thata positive relationship exists between CEOtransformational leadershipandorganizational innovation. They also support most of the predicted moderating effects. Theimplications of these ndings for practice and research are delineated.

Published by Elsevier Inc.

Keywords:Transformational leadershipInnovationModerators

This study seeks to advance understanding of how transformational leadership by Chief Executive Of cers (CEOs) affectsinnovationat the organizational level. This topic is important because in today's global economy, rms areunder constantpressure toinnovate their products and services ( Andriopoulos & Lowe 2000; Perry-Smith, 2006; Puranam, Singh, & Zollo, 2006;Tierney, Farmer,& Graen,1999 ). Thus, for example, Leifer, O'Connor, & Rice (2001) have observed that the contemporary competitive landscape hasbeen and continues to be driven by technological revolution, globalization, hyper-competition, and extreme emphasis on price,quality, and customer satisfaction, requiring an increased recognition and focus on innovation as a strategic competence (p. 102).

Numerous studies have sought to identify factors that can stimulate organizational innovation, broadly de ned by Amabile &Conti (1999) as theimplementationor adoptionof new, useful ideas by people in organizations (p.630). Examplesof factors foundto have an effect include leadership ( Amabile, 1998 ; Mumford & Gustafson, 1998), intra-organizational networks and learningcapability ( Tsai, 2001 ), CEO pay (Balkin, Markman, & Gomez-Mejia, 2000 ), a creativity-conducive work environment ( Amabile,1998 ), job complexity and type of supervision ( Oldham & Cummings, 1996 ), and organizational culture and climate ( Mumford &Gustafson,1998 ). Among these myriad factors, managers' leadership behaviorhas been identi ed by many researchersas being oneof the most, if not the most, important ( Amabile, 1998; Jung, 2001; Mumford, Scott, Gaddis, & Strange, 2002 ).

Indeed, over the past two decades, many management scholars and practitioners have called for more adaptive leadership bytop business executives in responding to the rapid changes confronting today's organizations ( Bass, Avolio, Jung, & Berson, 2003 ).Corporate chieftains like Jack Welch of General Electric and Lou Gerstner of IBM have been cited as examples of how adaptiveleaders can successfully restructure and transform their organizations. The label transformational has been applied to a set of adaptive leadership behaviors held to be more effective than other leadership styles in enhancingorganizational innovation ( Lowe,Kroeck, & Sivasubramaniam, 1996; Gardner & Avolio, 1998; Howell & Avolio, 1993 ).

Despite being held to be a key driver of innovation at the organizational level, transformational leadership's effects have mostlybeen studied at the levels of individual employees or organizational subunits ( Mumford et al., 2002 ). The limitation of such a focus

The Leadership Quarterly 19 (2008) 582 594

Corresponding author. Tel.: +1 619 594 0208; fax: +1 619 594 3272.E-mail addresses: [email protected] (D.(D.) Jung), [email protected] (A. Wu), [email protected] (C.W. Chow).

1 Tel.: +886 2 2938 7128.2 Tel.: +1 619 594 5331.

1048-9843/$ see front matter. Published by Elsevier Inc.doi: 10.1016/j.leaqua.2008.07.007

Contents lists available at ScienceDirect

The Leadership Quarterly

j o u r n a l h ome p ag e : w ww.e l s ev i e r. co m / l o ca t e / l e aq ua
mailto:[email protected]:[email protected]:[email protected]://dx.doi.org/10.1016/j.leaqua.2008.07.007http://www.sciencedirect.com/science/journal/10489843http://www.sciencedirect.com/science/journal/10489843http://dx.doi.org/10.1016/j.leaqua.2008.07.007mailto:[email protected]:[email protected]:[email protected]


2/13

is that,unless thecreativebehaviorsand outputsof individuals andsubunitsarecoordinated toyieldorganizational-level outcomes,thecompany as a wholestill couldbe left without effectiveresponses tothe challenges ofa competitivemarketplace. A secondlimitation of extant researchis that it hasmainlyused subjective performancemeasures (e.g., self-reported andsupervisoryratings ofcreativityandinnovation), which are open to judgment and other biases. Although objective measures are not necessarily free of other forms of unmeasured systematic or random error, they are less subject to cognitive biases and can more directly capture the end-products of innovative efforts in an organization. Therefore, our empirical tests go beyond subjective assessments of organizational innovation toalso include objective measures. A further advance over prior research is that we identify andsimultaneously test a number of internaland external organizational attributes that may interact with leadership behaviors in affecting organizational innovativeness.

We focus on CEOs because they usually play key roles in determining organizational policies/processes as well as resourceallocations, yet relatively few studies have examinedhow they affect innovation at the organizational level. Speci cally focusing ontransformational leadership, Waldman & Yammarino (1999) have called for more research on how this type of leader behavioraffects organizational performance through various levels of mediators and moderators. Recently, several authors have undertakenwork of this nature. Zhu, Chew, & Spangler (2005) collected company-level data from 170 rms in Singapore. They foundperception of the CEO's transformational leadership to be positively associated with perceptions of organizational outcomes (e.g.,ability to attract essential employees ), and negatively associated with absenteeism. They also found that a company's practices inperformance appraisal, staf ng, training, and compensation systems fully mediated the relationship between CEO transforma-tional leadership and perceived organizational outcomes. Similarly, De Hoogh & her colleagues (2005) studied 73 small andmedium-sized companies, and found that perception of the CEOs' transformational leadership had a positive relationship withtheir direct reports' work attitude. While these studies did not focus on organizational innovativeness, they do provide support forexpecting that the CEO's leadership behavior can affect organizational level outcomes.

A prior study that did examine this relation is Jung, Chow & Wu (2003) . They posited that a positive direct relation would existbetween CEO transformational leadership and organizational innovation, and that this relation would be moderated by theorganization's extent of employee empowerment and climate of support for innovation. Survey data from 32 Taiwanese electronics/telecommunications rmssupported theexpected directeffect. They alsorevealed a statistically signi cant positive moderatingeffectfromclimate ofsupport for innovation.However, Junget al. (2003) were emphatic that their results needed validation. Speci cally, theynoted as limitations their relatively small sample size and their use of only one informant per rm for each measurement scale. Theyalsostressedtheneed to expand the scope of their organizational innovationmeasure and to augment thesetof moderating variables.

Our study is patterned after Jung et al. (2003) and advances over this prior study in several ways. First, our sample includes alarger number of rms, and we use multiple informants in each rm. Second, we use a more encompassing measure of organizational innovativeness. Finally, we test the moderating effects of a richer set of variables. In addition to empowerment andclimate for innovation which were included in Jung et al. (2003) , we propose moderating effects fromtwo organizational structuralattributes (centralization and formalization) and two attributes of the external environment (competition and uncertainty). In thefollowing subsections, we explicate the theoretical basis for the hypothesized direct and moderating effects. Following that, themethod section will describe how we use a multi-source and multi-dimensional approach to operationalize the variables.

1. Theoretical background and hypotheses

1.1. Transformational leadership and organizational innovation

Bass et al. (2003) have characterized transformational leadership as encompassing ve theoretically distinct components:charisma, idealizedin uence, inspirationalmotivation, intellectual stimulation, and individualizedconsideration. Theyexplainthatboth charisma andidealized in uence have to dowithserving as role models,with the former re ecting followers' attributions andthe latter capturing leaders' behaviors. Leaders with idealized in uence tend to place followers' needs over their own needs, sharerisks with followers, and exhibit adherence to a setof underlying principles andvalues. Inspirational motivation involves providingmeaning and challenge to followers' work, including inculcating visions of the future. Intellectual stimulation entails stimulatingfollowers to questionassumptions, reframeproblems,andto approacholdsituations in newways. New ideas andcreative solutions

are solicited from followers, and there is no public criticism of individuals' mistakes. Finally, leaders who provide individualizedconsideration areseenas ones whopayattentionto followers' individual needs forachievement andgrowth.They doso byactingascoach or mentor, and by creating learning opportunities as well as a supportive climate in which to grow.

Advocates of transformational leadership maintain that by means of their behavior, transformational leaders create personaland professional commitment from subordinates toward higher-level needs like self-esteem and self-actualization ( Bass, 1985;Gardner & Avolio,1998 ). This in turn increases the latter's intrinsic motivation, which has been identi ed as an important driver of employee creativity and rm innovation ( Oldham & Cummings, 1996; Amabile, 1998; Zhou, 2003 ). In addition, a transformationalleader's intellectual stimulation can facilitate unconventional and innovative thinking and working processes that lead to newknowledge and technology, which is fundamental to rm innovation ( Dougherty & Hardy, 1996 ). At the opposite end of thespectrum, transactional leadership focuses on maintaining the status quo and motivating people through contractual agreement(Bass,1985 ). This leadership style tends to emphasize extrinsic rewards, such as monetary incentives and promotion, as a means toincrease followers' motivation. Amabile, Conti, Coon, Lazenby, & Herron (1996) have speci cally identi ed transactional leadershipas being detrimental to creativity, and a recent meta-analysis by Bono & Judge (2004) has provided support for this proposition. It

found that as compared to transactional leadership, transformational leadership yielded higher employee job satisfaction andincreased their autonomous goal-directed motivation.

583D.(D.) Jung et al. / The Leadership Quarterly 19 (2008) 582 594


3/13

Along the same vein, other studies have found that leaders who display transformational behaviors are able to realign theirfollowers' values and norms, promote both personal and organizational changes, and help followers to exceed their initialperformanceexpectations (e.g., House& Shamir,1993; Jung & Avolio, 2000 .See Mumford et al., 2002 for a review).By and large, thesestudies have focused on leadership and outcomes at individual or organizational subunit levels using experimental settings and/orsubjective measures of creativity (e.g., subjective supervisor ratings). For example, Sosik, Avolio, & Kahai (1997) found that ascompared toother leadership types, transformational leadershipwasmoreeffectiveatencouraging followers to think outofthebox and to adopt generative and exploratory thinking processes that yielded more creative ideas and solutions. Keller (1992) discoveredthat transformational leadership displayed by project team leaders in a large R&D organization improved team performance. Shin &Zhou (2003) studied 290 employees and their supervisors from 46 Korean companies, and found a positive relation betweentransformational leadership and follower creativity. While Jung et al. (2003) did consider the link between CEO transformationalleadership and organizational innovativeness, the authors were emphatic that their exploratory study needed validation.

Another point to note about these prior studies is that few have gone beyond the relational aspects of transformationalbehaviors. Although leaders oftentimes promote creative thoughts and encourage innovation by interacting with followers (e.g.,challenging their long-held assumptions regarding their work approaches), CEOs may in uence innovation in yet other ways. Inparticular, Upper Echelons Theory posits that organizationaloutcomes bothstrategiesand effectiveness are re ections of thevalues and cognitivebases of powerful actors in the organization. (Hambrick & Mason,1984 , p. 193.) In other words, leaders in theupper echelons of an organization in uence organizational performance directly through their characteristics and behaviors andindirectly through the strategic choices they make. As the top executive and a powerful actor, the CEO is in a unique situation todetermine the organization's strategies in responding to the challenges of the environment, and CEOs with transformationalleadershipare likely to choose strategiesthat arechange- andgrowth-oriented in nature. Thus, wespecify the followinghypothesis:

Hypothesis 1. Organizational innovation is positively related to the CEO's extent of transformational leadership.

1.2. Moderating effects

Although we have made the case for CEOs' leadership behavior to directly in uence organizational innovation, it must also berecognized that such behavior occurs within the larger context of the organization. As such, the analysis would be incompletewithout alsoconsidering attributesof the organizationaland environmental contexts in which rm innovation occurs( Damanpour,1991; Scott & Bruce,1994;Mumford et al., 2002; Jung et al., 2003 ). Amabile & hercolleagues (1996) have labeled this the contextof creativity. In the following subsections, we developexpectationsof howspeci c attributesof organizational climate, structure, andthe external environment would moderate the relationship between CEO transformational leadership and rm innovation.

1.3. Attributes of organizational climate

1.3.1. Climate for innovationOrganizational climate is widely de ned as organizational members' shared perceptions of organizational policies, practices, and

procedures, both formal and informal indicative of the organization's goals and appropriate means to goal attainment. (Reichers &Schneider, 1990 , p. 22). Scott & Bruce (1994) have further explained that climate represents signals individuals receive concerningorganizational expectations for behavior and potential outcomes of behavior (p. 582). By serving as the gel that ties togetheremployees from different units and levels, an organization's climate can work for or against particular goals ( Daft, 2001 ). Sinceinnovation typically requires long-term investments and risk-taking ( Balkin et al., 2000 ), employees need to perceive that theorganization supports working on innovative yet more risky ideas and products rather than focusing on short-term pro t andimmediate nancial results ( Kanter, 1983 ). Thus, Mumford & Gustafson (1998) have argued that even when individuals havedeveloped the capacity for innovation, their willingness to undertake productive efforts maybe conditioned by beliefs concerning theconsequences of such actions in a given environment (p. 37). As such, when there is not a climate that supports or values innovativeinitiatives, a CEO's transformational leadership behaviors alone may only marginally stimulate employee efforts towards innovation.

Consistent with this view, a study of 78 managers by Howell & Avolio (1993) found that when there was a climate of support for

innovation within the organizational unit, there was a signi cant positive relationship between the intellectual stimulationprovided by the leader and the unit's innovative performance. But when support for innovation was absent, the relationship wasinsigni cant. Jung et al. (2003) have provided preliminary evidence consistent with climate for innovation moderating the linkbetween CEO transformational leadership and organizational innovation.

1.3.2. Empowerment Allowing subordinates autonomy and exibility in carrying out their duties can create a sense of ownership andcontrol over the

work to be performed ( Mumford et al., 2002 ), and Amabile et al. (1996) and Quinn (1985) have suggested that employees producemore creative work when they perceive more personal control over howto accomplish given tasks. Empirical support for this viewhas been provided by Zhou (2003) , who found that individuals generated the most creative ideas when they worked in a high taskautonomy environment. Jung & Sosik (2002) also have argued that people need to be empowered before they could developintrinsic motivation, which in turn promotes creative endeavors, while Damanpour (1991) has suggested that empoweringemployees facilitates innovation by increasing their awareness, commitment, and involvement. More speci cally related to

transformational leadership, Avolio, Zhu, Koh, & Bhatia (2004) have found that the effect of such leadership on employeecommitment depended on employees' sense of empowerment. Taken as a whole, the preceding discussion suggests that for the

584 D.(D.) Jung et al. / The Leadership Quarterly 19 (2008) 582 594


4/13


5/13

development process by selectively focusing on different competitive advantages such as marketing pro ciency, cross-functionalintegration, and technical synergy. Taken together, the preceding literature provides the basis for expecting a positive moderatingeffect of uncertainty on the relationship between CEO transformational leadership and organizational innovation.

1.5.2. CompetitionWe also expect competition to have a positive moderating effect. When an organization operates in a highly competitive

market, or when its members perceive that they are operating under competitive market conditions, they are more likely to acceptthe need for innovation ( Frambach & Schillewaert, 2002 ). In a highly competitive industry such as computers and consumerelectronics, for example, the value and uniqueness of knowledge-intensive resources can be swiftly lost to competitors (Balkinet al., 2000 , p.1118). As such, just like the case of environmentaluncertainty, competition can increase subordinates' responsivenessto stimuli for innovation.

However, competition also may reduce the slack resources (e.g., via reduced pro t margins) for supporting experimentationand innovation initiatives. Nohria & Gulati (1996, p.1246) de nedsuch resources as the pool of resources in an organization that isin excess of the minimum necessary to produce a given level of organizational output, and found an inverse U-shaped relationshipbetween slack resources and innovation. In total, then, while competition can increase attention to innovation, its net impact maybe dampened by a lower availability of resources for supporting long-term and risky initiatives.

Taken as a whole, our discussion above suggests our fourth hypothesis:

Hypothesis 4. Uncertainty and competition moderate the relationship between CEO transformational leadership andorganizational innovation such that the relationship is more positive among companies with higher levels of (a) environmentaluncertainty and (b) competition.

2. Method

2.1. Participants

We focused on a single industry so as to limit extraneous in uences due todifferent industry types. Our sample was comprisedof 50 Taiwanese companies from the electronics and telecommunications industry. This industry was selected because newproduct development and creative R&D efforts are critical for company survival due to the industry's rapid technological advancesand highly competitive markets ( Carey & Nahavandi, 1996; Balkin et al., 2000; Schilling & Hill, 1998 ).

To generate the sample, we used a database of published nancial statements (the Taiwan Economic Journal Taiwan DataBank) to identify all Taiwanese electronics and telecommunications companies that were traded either on the Taiwan StockExchange or over-the-counter. We randomly selected 50 of the 317 rms identi ed, and solicited their participation byapproaching a top manager (vice president and above) of each company with the aid of consulting and accounting rms. As

inducement, each rm was promised a report comparing it to the rest of the sample. As we encountered (initial) refusals,replacements were randomly selected. In total, we approached 62 companies, out of which 53 (85.5%) ultimately participated. Asexplained further below, three of these rms were later dropped due to inadequate data.

Multiple data sources were used to control commonresponse biases. Three sources were used to measure each company's levelof innovation: published nancial statements, a government report on patents granted, and nancial analysts specializing in theTaiwanese electronics and telecommunications industry. Information on leadership behavior and the contextual variables wasobtained from surveying managers with at least three years' work experience in their companies. Three top-level managersrepresenting various functional areas from each company were asked to complete Survey A, which measured their company'sdemographic characteristics (e.g., age, size) and the CEO's extent of transformational leadership. The respondents were selectedfrom those who worked closely with, and/or had frequent interaction with the CEO. This group also answered questions on theirorganizations' environmental uncertainty and degree of competition. A second group of three mid- to top-level managers per rmwas given Survey B that contained measures of empowerment and climate for innovation. A third group of managers (four mid- totop-level managers per rm) was given Survey C containing measures of centralization and formalization. In total, there were 141,

144, and 200 completed responses to Surveys A, B, and C.

2.2. Procedures

Each company provided a contact person (typically vice president or above). This person was visited personally by the secondauthor, whoexplained thegeneral natureof thestudywithout revealingthespeci c hypotheses. Then thecontact personwasprovidedwith copies of the three surveys andgiven explicit instructionson thenumber andtypeof managersneededto complete each version.Without exception, the contact persons expressed enthusiasm for the study and committed to following the instructions.

To encourage truthful responses in light of many questions' sensitive nature (e.g., the CEO's leadership behaviors, organizationalstructure andclimate), completecon dentialityand anonymitywas guaranteedforboth respondents andtheircompanies,and a postage-paid, self-addressed envelope was provided for direct return of completed surveys to us. Since responses were obtained from multiplemanagers in each company, all surveys provided to a company hada common identi cation code. The presence and purpose of this codewas explained in the survey instructions. Participants also were given the second author's contact information in case of questions.

Because the survey was administered in Chinese, we followed Brislin's (1986) recommendation of translation and back-translation to ensure conceptual equivalence between the original instruments (in English) and the Chinese versions. A number of



6/13

recent studies on leadership have adopted a similar approach (e.g., Shin & Zhou, 2003 ). All three surveys were rst translated intoChinese by a bilingual individual who was not told the objective of the study. Then, another bilingual person back-translated theseinto English without having access to the original instruments. Only a few minor changes to the Chinese surveys were triggered bycomparing the back-translated and original English versions.

2.3. Demographic statistics

Survey A contained a question on how frequently the respondent interacted with his/her rm's CEO. Thepossible answers wererarely, once in a while, sometimes, fairly often, and regularly. In three of the rms, at most one respondent had answeredwith a frequency higher than rarely or once in a while. These rms were dropped to preserve the integrity of the leadershipmeasure and accordingly, the usable responses to Surveys A, B, and C were reduced to 132, 135, and 188, respectively.

On average, the 50 rms remaining in the samplehadbeen in existence forabout 17.7 years (range: 5.2 to 37 years). The averagetenure of the CEOs was about 10 years and all had been with their company for at least 3.5 years, which was suf cient to span thethree-year period (2000 2002) from which we drew our innovation measures. The companies were relatively large. Only threehad fewer than 100 full-time employees. Twenty-three had between 101 and 1000 employees, and the remaining 24 companieshad more than 1000 full-time employees.

To assess the representativeness of our sample rms, we compared their 2002 sales and R&D as a percentage of sales to those of the 267 Taiwanese electronics and telecommunications rms not included in this study. These two measures were used becausethey are available from the publicly disclosed nancial statements (whereas attributes like leadership behavior and environmentaluncertainty arenot). There was no signi cant difference on the latter (mean values of 0.042 vs. 0.04, p=0.825), but the mean size of the sample rms was signi cantly larger than that for the non-sample rms (New Taiwanese Dollars: 22.792 billion vs.8.612 billion, p b 0.0001). Thus, our sample rms are more representative of the larger, and presumably more important/in uential,ones in the Taiwanese electronics and telecommunications industry.

The respondents' demographic pro les suggest that they were well quali ed to answer the survey questions. All indicated thatthey were either senior- or middle-levelmanagers. In terms of gender andage, the majority of theSurvey A respondentswere male(70%), and they were quite evenly split between 30 to 40 years old (41%) and 41 to 50 years old (45%). On average, they had beenworking for their company for 8.8 years (sd=6.22 years) and for the current CEO for 5.8 years (sd= 5.3 years). The vast majority(97%) had a college degree or above. Their average tenure in their current job was about 4.5 years and the majority (74%) indicatedthat they interacted with their CEO either fairly often or regularly. Similarly, most Survey B respondents were male (82%). Slightlyover half (57%) were between 41 and 50 years old, with another 39% being in the 30 to 40 age bracket. On average, they had beenworking for their company for 8.9 years (sd=5.8 years), with an average tenure at the current job of 4.3 years. The vast majority(98%) hada college degree or above. TheSurvey C respondents were slightly younger than theother two groups. About 66% of themwere between 30 and 40 years old, and their average lengths of organizational and positional tenure were 7.9 and 3.7 years,respectively. This group also hada more balanced gender ratio (65% male).Like theother two groups, they were well educated (97%had college or above levels of education).

2.4. Measures

The CEO's extent of transformational leadership was measured using Bass & Avolio's (1997) Multifactor LeadershipQuestionnaire (MLQ). The MLQ has been extensively used in prior research and is considered to be a well-validated measure of transformational leadership ( Awamleh & Gardner, 1999 ). Its construct validity has been demonstrated using Con rmatory FactorAnalysis (cf. Avolio, Bass, & Jung, 1999 ). Each participant was asked to rate four aspects of his/her CEO's behavior related to each of the ve transformational leadership dimensions (total=20 items). The 5-point response scale for each item was anchored by1= strongly disagree and 5= strongly agree. Sample items are My CEO talks optimistically about the future (for inspirationalmotivation) and My CEO gets me to look at a task from many different angles (for intellectual stimulation).

Employees' perception of empowerment was measured using a 12-item scale developed by Spreitzer (1995) . A sample item isI have signi cant autonomy in determining how I do my job.

All items were rated using a 7-point scale anchored by 1=

Verystrongly disagree and 7= Very strongly agree.

Existence of an innovation supporting organizational climatewas measured with a 22-item scale originally developed by Siegel& Kaemmerer (1978) , and later modi ed by Scott & Bruce (1994) . It contains two subscales (support for creativity and tolerance of differences). Sample items are This organization gives me free time to pursue creative ideas during the workday and Peoplearound here are expected to deal with problems in the same way. All items were rated on a 7-point scale anchored by 1= Verystrongly disagree and 7= Very strongly agree.

The centralization and formalization measures wereadapted fromsurveys developed by Khandwalla (1977) and Pugh, Hickson,Hinings, and Turner (1968) , and extensively used in prior research with adequate levels of reliability ( Chow, Shields, & Wu, 1999;Gordon & Narayanan, 1984 ). Minor modi cations were made to t the context of the current study. Speci cally, respondents wereasked to indicate the lowest managerial level to which authority has been delegated to make each of eight decisions (e.g., hiringand ring employees, preparation of operating procedures and project calendars). All questions were based on a 7-point scaleanchored by 1= Top management ; 4 = Middle management ; and 7= First line supervisors. The responses were reverse scored

to yield the centralization measure. The formalization question also focused on these eight decisions, and asked respondents toidentify the highest managerial level for which their companies had written manuals that specify how to make each decision.



7/13

Perceived competition and environmental uncertainty were measured by questions adapted from Khandwalla's (1977)instrument. Speci cally, we asked respondents to rate the intensity of competition in their industry on (a) obtaining inputs and (b)price. The 7-point response scale was anchored by 1= Of negligible intensity and 7= Extremely intense. For environmentaluncertainty, respondents were asked How stable/dynamic is the external environment facing your company in terms of theseaspects: (a) economic; (b) technological; and (c) political/regulatory? The 7-point response scale for each aspect was anchored by1= Very stable (changing slowly) and 7= Very dynamic (changing rapidly).

Having an accurate measure of organizational innovation is important because if organizations are misclassi ed on thisattribute, then inferences about different factors' effects also will be mistaken. In developing the innovation metric, we heededBalkin et al.'s (2000) recommendation that a composite measure be used to capture broad aspects of innovative activities. In total,we collected three measures to tap into different aspects of innovation. First was annual R&D expenditures as a percentage of salesrevenues over the three years prior to our survey (i.e., 2000 through 2002). Many authors (e.g., Hitt, Hoskisson, & Kim, 1997; Balkinet al., 2000 ) have argued that this ratio proxies for a rm's emphasis on innovation. The data were obtained from the sample rms'publicly disclosed nancial statements.

Second, sinceinputsmay not successfullyproduceoutputs,wealso includedthenumberofpatents that a company hadobtainedduringthe same three-year period. This measure has been commonly used by past research as an indicator of creative performance ( Oldham &Cummings,1996 ). The data came from the patent database of the Taiwanese Ministry of Economic Affairs' Intellectual Property Of ce.

Third, we contacted 20 nancial analysts in Taiwan actively following the electronics and telecommunications industry. Theseanalysts have a professional stake in understanding the fundamental strengths and weaknesses of each rm that they follow, andalso would have access to an extensive database for doing so ( Barron, Byard, Kile, & Riedl, 2002; Barth & Hutton, 2004 ). The reasonfor seeking their evaluations of each company was that, while the number of patents has face validity, it is limited to innovationsthat are amenable to patent application, and for which rms consider patents to offer meaningful property rights protection. Assuch, its scope excludes other fruits of innovation efforts, such as process and service improvements. Consider the case of DellComputer. Thecompany is well known for its innovative operational processes such as a made-to-order manufacturing systemandresponsive customer services, which have fueled its market dominance in the U.S. computer industry. Yet it seems unlikely thatinnovations of this nature would be manifest in the number of patents obtained.

Each analyst was asked two questions about each sample rm. The rst question was his/her familiarity with the company. Thethree possible answers were 1= Not familiar with this company at all ; 2= Somewhat familiar with this company ; and 3= Veryfamiliar withthis company. The second question asked fora rating of the company's innovativeness, de ned as theability to developnew products, services, processes or systems for ful lling customers' needs. The 7-point response scale was anchored by 1= Totallylacking in innovation ; 4= Average in innovation ; and 7= Highly innovative. Only analysts who answered at least a 2 to the rstquestionwere kept for each company. All sample companies had valid ratings from at least 10 different analysts. There was very highagreementamongtheir innovationratings, as indicatedby the reliability score (.96). Thus, weaveraged their ratings foreach company.

3. Results

We tested our hypotheses with the partial least squares (PLS) structural equations modeling technique ( Wold, 1985 ), which isincreasingly beingadopted by leadershipresearchers (e.g., Bass et al., 2003; Howell, Neufeld, & Avolio, 2005; Shamir, Zakay, Breinin,& Popper,1998 ). Thismethod does notmakeassumptionsabout (a)datadistributions toestimatemodel parameters, (b) observationindependence, or (c) variables metrics ( Barclay, Higgins, & Thompson, 1995 ). As compared to more traditional techniques likeregression, PLS also is less vulnerable to measurement errors in variables and the effects of outliers ( Wilcox,1998 ). These attributesmake PLSparticularlysuitable for testing multivariate main andindirect effects models. This is especially so for small samplesizes asin the present study ( Chin,1998; Sosik & Dworakivsky, 1998 ). Thereason is that regression requires that onehas more observationsthan variables for sample size suf ciency (e.g., at least 3 4 observations per variable to get reliable estimates of the errors). Incontrast, PLS converts the data into pseudo-variables which capture the variability in the data related to what one is trying topredict. Since there are fewer pseudo-variables than variables, fewer observations are needed ( Martens & Naes, 1989 ).3

PLS generates estimates of standardized regression coef cients (i.e., path coef cients) for the model paths, which can then beused to measure the relationships between latent variables (see Sambamurthy & Chin,1994 for more information on PLS). We useda jackkni ng procedure called blindfolding to evaluate the statistical signi cance of the path coef cients ( Sambamurthy & Chin,1994 ). This procedure omits a part of the data matrix for a particular variable and then estimates path coef cients associated withthat variable. The customary omission distance of 10 was used.

Since there were multiple raters of CEOleadership as well as themoderating variables, we rst tested within-company varianceusing James, Demaree & Wolfe's (1984) r wg procedure. We found at least 90% of the companies had a r wg value of .7 or higher for allof the scales. Interrater agreement based on intraclass correlations also showed acceptable ranges for all of the measured variables(transformational leadership= .88; empowerment= .78; climate for innovation= .84; centralization =74; formalization = .85;competition=.74; and uncertainty=.71). Based on these results, we aggregated our data to the company level and conducted allsubsequent data analyses at this level. 4

3 A more detailed discussion of the attributes and bene ts of PLS is available from the rst author.4 While most of our measures relate to the company as a whole (e.g., centralization, climate for innovation, competition), our empowerment scale focused on

the individual manager. As we had noted earlier, Klein et al. (2000) have shown that individual level measures can be aggregated to higher levels if there is a high

level of within-group agreement. Since our data satis ed this criterion, we believe that our aggregation approach is valid for testing empowerment as a rm-levelconstruct. A similar point applies to our measures of climate for innovation and CEO leadership.



8/13

Our analyses used the individual items as indicators for all scales except for transformational leadership, which used ve scalescores (i.e., charisma, idealized in uence, inspirational motivation, intellectual stimulation, and individualized consideration) asindicators (see Sosik & Dworakivsky,1998 for a similar approach). Given the large number of items used to measure the constructsof interest, we rst ran a preliminary PLS analysis with all of the survey items to test the scales' psychometric properties. Threecriteria were used to determine whether any item indicator should be retained. First, to ensure adequate reliability, the factorloadings of indicators associated with each construct had to be .60 or above ( Bagozzi & Youjae, 1988 ). Second, the composite scalereliability for each construct (an internal consistency estimate similar to alpha) had to exceed .70. Finally, the average varianceextracted by the latent constructs from their indicators had to exceed .50 ( Fornell & Larcker, 1981 ).

Applying these criteria led to the retention of 4 items for empowerment, 8 items for climate for innovation, 4 items forcentralization, 4 items for formalization, 2 items for competition, and 3 items for environmental uncertainty. Finally, the twooutput-based innovation measures (number of patents obtained and analysts' ratings) loaded onto a single construct consistently.Table 1 shows descriptive statistics and inter-correlations among the scales. 5

3.1. Results for the measurement component

Table 2 presents the factor loadings, composite scale reliabilities and average variance extracted for constructs that containedtwo or more items. Since PLS estimates its measurement and structural models simultaneously, we ran a full scale model thatcontains CEO transformational leadership as well as the six moderators as independent variables predicting rm innovation.Table 2 shows that all constructs satis ed all three retention criteria described earlier. Additional analyses were run to test theconstructs' convergent and discriminant validity by comparing the average variance extracted by each construct (diagonalelements) and correlations between constructs. The results showed that the items representing a construct share more variancewith each other than with other constructs in the model, thus indicating adequate convergent and discriminant validity ( Carmines& Zeller, 1979; Wold, 1985 ).6

3.2. Results for the structural component and hypotheses

In response to the ndings of past research (e.g., Miller et al., 1988 ), we included each company's size (number of full-timeemployees) as a control variable. Additionally, the CEO's organizational tenure was used to control for the possible effects of timesharpening or dulling the effects of his/her transformational leadership. The PLS analysis showed that the model as a wholeaccounted for about 54% of the variance in rm innovation. Company size was signi cantly and positively related to organizationalinnovation ( =.31, p b .001), while CEO tenure was signi cantly and positively related to subordinates' perception of transformational leadership ( =.21, p b .001).

5 The third innovation metric, annual R&D expenditures as a percentage of sales revenues, had a loading of .50 on its own, unique factor. The R&D measure hadcorrelations of .09 and .03, respectively, with the patent and analysts ratings measures, while the latter two had a statistically signi cant 0.50 correlation. Weconducted all of the analyses by using, alternately, the R&D measure and the aggregate of the latter two metrics (after converting each into a zero mean and unitvariance distribution). The same patterns of results were obtained, though the effects were generally smaller in absolute magnitude when the R&D measure wasused. Below, we only report the results for the aggregate measure for parsimony. Results based on the R&D measure are available from the rst author on request.The lack of a signi cant correlation between R&D and the other two innovation measures, together with the similar patterns of signi cant effects, suggest that the

variables in our model are applicable to both the input and output ends of organizational innovation.6 Tables for convergent and discriminant validity are not presented here to conserve space. They are available from the rst author upon request.

Table 1Descriptive statistics and inter-correlations among constructs ( n =50 companies)

Variable Mean SD Intercorrelations

1 2 3 4 5 6 7 8 9 10

1. CEO transformational l eadership 3 .74 .42 (.87)2. Empowerment 5.54 .49 .14 (.84)3. Climate of support for innovation 4.73 .69 .39 .24 (.90)4. Centralization 3.24 .80 .15 .08 .35 (.70)5. Formalization 1.98 .68 .03 .09 .02 .33 (.90)6. Uncertainty 5.17 .77 .05 .05 .03 .03 .16 (.71)7. Competition 6.06 .63 .37 .06 .09 .10 .14 .43 (.74)8. Organizational innovation a .01 .89 .09 .26 b .32 .20 .10 .31 .01 9. CEO tenure with his/her organization 9.95 6.68 .18 .12 .02 .01 .16 .02 .12 .16 10. Company size c 3.07 1.23 .08 .06 .08 .01 .08 .41 .10 .52 .08

Note: values on the diagonal (in parentheses) represent Cronbach alphas. Values off the diagonal are correlations between constructs.a The two output-based measures of organizational innovation (# of patents and analysts' ratings) were standardized into zero mean and unit variance

distributions before they were combined into an aggregate measure of company innovativeness.b p b .10; * p b .05; ** p b .01; *** p b .01.c Company size ranged from 1 (less than 100 employees), 2 (101 500), 3 (501 1000), 4 (1001 5000), and 5 (5001 1000) to 6 (over 10,000 employees).

http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


9/13

3.2.1. Direct effect The CEO's transformational leadership was positively and signi cantly related to organizational innovation ( =.15, t (49) =3.18,

p b .01). This nding provides support for Hypothesis 1.

3.2.2. Moderating effectsTo test the moderating effects, we took each of the six moderating variables in turn, and created high and low subgroups based

on the variable's mean score. Then we reran the model used to test the direct effect of CEO transformational leadership on rminnovation. That is, for each moderating variable, we estimated a path coef cient and standard error between transformationalleadershipand innovation for each sub-sample. Then, an unpaired t -test was used to test the moderating effect (See Chin, 2000 formore detailed information on how to test moderating effects with PLS, and Sosik, 2005 for a recent example of the identicalapproach to testing a moderating effect.) The path coef cients and t -statistics for each sub-group are shown in Table 3 .

Table 3 shows that all of the hypothesized moderating effects are supported except for empowerment. Speci cally, andconsistent with Jung et al. (2003) , the path coef cient from CEO transformational leadership to rm innovation was higher in the

high climate-for-innovation group than in the low climate-for-innovation group ( =.07 vs. =.09; t (48) =3.11, p b .01). Similarly,the path coef cient was higher in the low centralization than in the high centralization group ( =.44 vs. =.06; t (48)= 25.54, p b .001), and higher in the low formalization than in the high formalization group ( =.28 vs. = .11; t (48)= 6.99, p b .001). Thepath coef cient also was higher in the high uncertainty than in the low uncertainty group ( =.23vs. =.18) though in this case, thedifference was only marginally signi cant ( t (48)=1.78, p b .10). Finally, the path coef cient for the high competition group wassigni cantly higher than that of the low competition group ( =.24 vs. =.07; t (48)= 9.05, p b .001).

While empowerment also hada signi cant moderating role, thedirection of its effect was opposite to that hypothesized.Ratherthan being higher, the path coef cient was lower in the high empowerment group than in the low empowerment group ( = .09vs. =.11; t (48)=11.42, p b .001). The same directional difference also was found by Jung et al. (2003) though in their case, it wasnot statistically signi cant.

4. Discussion and implications

This study has tested the direct and moderated effects of CEO transformational leadership on

rm innovation. Our resultssupported a direct and positive effect of CEO transformational leadership on organizational innovation. The ndings further

Table 2Factor loadings, weights, composite scale reliability, and average variance extracted for assessing construct reliability

Construct Item Factor loading Weights of measures Composite scale reliability Average variance extracted

1. CEO transformational leadership Charisma .89 .35 .91 .82II .89 .27IM .85 .35IS .78 .12IC .65 .10

2. Empowerment EMP1 .64 .33 .88 .75EMP2 .82 .29EMP3 .89 .34EMP4 .83 .33

3. Climate of support for innovation CLIMATE1 .67 .07 .92 .78CLIMATE2 .75 .11CLIMATE3 .71 .17CLIMATE4 .82 .12CLIMATE5 .75 .16CLIMATE6 .80 .28CLIMATE7 .86 .21CLIMATE8 .81 .16

4. Centralization CENTRAL1 .75 .39 .80 .71CENTRAL2 .61 .32CENTRAL3 .72 .34CENTRAL4 .76 .36

5. Formalization FORMAL1 .90 .50 .85 .76FORMAL2 .75 .34FORMAL3 .76 .32FORMAL4 .68 .10

6. Competition COMPETE1 .90 .59 .88 .64COMPETE2 .86 .51

7. Uncertainty UNCERT1 .69 .24 .84 .73UNCERT2 .92 .71UNCERT3 .74 .34

8. Organizational innovation ZPATENT .83 .51 .81 .87ZANALYSTS .90 .65

Note: II = idealized in uence; IM = inspirational motivation; IS = intellectual stimulation; IC = individualized consideration; EMP = empowerment; CLIMATE =climate of support for innovation; CENTRAL = centralization; FORMAL = formalization; COMPETE = competition; UNCERT = uncertainty; ZPATENT = standardized(zero mean, unit variance) measure of number of patents; ZANALYSTS = standardized (zero mean, unit variance) measure of analysts' innovativeness ratings.



10/13

revealed moderating effects by two attributes of the environment (uncertainty and competition) and four rm-levelcharacteristics: climate of support for innovation, formalization, centralization, and empowerment, though in the last case, theeffect was opposite in direction to that hypothesized.

From a practice standpoint, our ndings not only underline the importance of managers tting their leadership behaviors to theorganizational context in which they work. They also support going beyond that to seek changes in attributes of the organizationalcontext. In the case of high-level managers like CEOs, they may have suf cient authority to in uence some of the variables thatmoderate the link between their leadership behavior and organizational innovation, thereby amplifying the effects of theirleadership behaviors ( House & Aditya, 1997; Miller et al., 1988; Pillai & Meindl, 1998 ). For example, a number of researchers havesuggested that top managers are the main architects of organizational climate by communicating what strategy to implement andhowthe goals of theorganizationarerelevantto theemployees'personalvaluesanddesires. In thecaseoforganizational innovation,when top managers reinforce its importance by recognizing and rewarding creativity while being tolerant of mistakes, employeesare more likely to pursue new ideas and product innovation ( Mumford et al., 2002; Sosik, Jung, Berson, Dionne, & Jaussi, 2004 ).

While environmental factors are less likely to be controllable by the managers of a company, they also can be used as leveragefor promoting change. In this regard, our nding of signi cant positive moderating effects from two environmental variables isworthy of note. Bass (1985) has argued that transformational leaders often emphasize crisis in order to bring about change, and wehad proposed that when employees perceive a high level of uncertainty and competition surrounding their organization, they aremore likely to be receptive of change and innovation initiatives. The recent turnaround of Samsung Electronics illustrates how thismay come about. When Yun Jong Young became the vice-chairman and CEO in 1997, the company had US$11 billion in debt and abrand name that was mainly associated with low-end consumer products. Within ve years, the company had turned aroundcompletely, earning US$5.9 billion on sales of US$33.8 billion in 2002. Many observers have attributed this transformation toYoung's leadership, with a vision to make Samsung a world-class company that could charge premium prices ( Business Week,2002 ). Toward this end, he continually emphasized the uncertainty in the high-tech industry as a way to motivate employees andto promote change. Our nding of positive moderating effects for environmental uncertainty and competition is consistent withtheir serving such a motivational role.

From a research perspective, our ndings underline the desirability of placing the empirical analysis of CEO leadership in itsorganizational and environmental context, rather than analyzing bivariate relationships in isolation. Among possible directions forfuture research, a topic worthy of scrutiny is our nding a negative moderating effect of empowerment on the relationshipbetween CEO transformational leadership and rm innovation. As Table 1 shows, even the zero-order correlation betweenempowerment and innovation is negative and statistically signi cant. Although our data precluded a search for underlying causes,a plausible explanation is the sourcing of our data from Taiwan, where cultural values are relatively high in power distance(Hofstede, 1997 ). According to Hofstede, people from a high power distance culture expect leaders to act strongly, and becomeuncomfortable when leaders try to delegate heavily. In a work setting, this translates into a preference for paternalistic leaders,with considerable dependence of subordinates on their superiors ( Adler, 2002 ). Thus, subordinates in a high power distanceculture may feel confused and frustrated when left alone to gure out what they need to do and how to accomplish their goals.

More generally, our

nding relating to empowerment may imply a need for transformational leaders to maintain a balancebetween letting people feel empowered, and providing structure and control by de ning goals and agenda ( Mumford et al., 2002 ).The optimal balance may well vary across cultures an area worthy of future research especially considering the increasingglobalization of economic activities.

Alternatively, thenegative relationship between empowerment and innovation could be due to theunique nature of a high-techindustry with rapidly changing technology. In the hyper- competitive consumer electronics industry from which our samplecompanies were drawn, even top-selling innovativeproducts can become obsolete relativelyquickly ( Bharadwaj& Konsynski,1997 ).The resultant need to continuously develop innovative products might require what Brown & Eisenhardt (1998) have referred to asstructuredchaos. Theyarguedthat there is a rede nedroleof leaders as architectsand cultural guardians, who need togobeyond thetraditional managerial responsibilities by carefully monitoring and controlling organizational recon guration processes. As such,there may be a threshold past which additional empowerment would hamper managers' ability to lead change. Sosik et al. (2004)alsostressed the importance ofmaintaining sucha balance in many high-tech orientedcompanies: Thechallenge forexecutives is toalso set the boundaries that help direct those creative individuals towards achieving innovation. In an R&D organization,

establishing guidelines and boundaries is important, and then giving people freedom to operate within those boundaries isimportant (p. 176). It would be worthwhile to explore if, and how, this balance differs between high-tech and other industries.

Table 3Moderating effects of cultural, structural, and environmental variables

Hypothesis and moderator Path coef cients from CEO transformational leadership to rm innovation ineach sub-group ( N =25 in each sub-group)

t -value of difference in path coef cients

H2a: climate for innovation Low climate for innovation = .07 High climate for innovation =.09 3.11 H2b: empowerment Low empowerment = .11 High empowerment = .09 11.42 H3a: central ization Low cent ralization = .44 High centralization =.06 25.54 H3b: formal ization Low formalization = .28 High formalization = .11 6.69

H4a: uncertainty Low uncertainty =.18 High uncertainty =.23 1.78 H4b: competition Low competition =.24 High competition =.07 9.05

* p b .10; ** p b .01; *** p b .001 (df=48).



11/13

Yet another possibility is the sourcing of our empowerment measures. Even if the mid- to senior-level managers who suppliedthe survey data perceived that they were being empowered by their CEO, they may not have in turn empowered their ownsubordinates. Since major breakthroughs or innovative products often come from front-line employees and research scientists/engineers, data from these other levels of theorganizationcan shed further light on the role of empowerment in theorganizationalinnovation process.

In addition to delving more deeply into speci c ndings of this study, there is room for expanding the scope of inquiry. First,although the current model has managed to explain a signi cant proportion (54%) of the variance in organizational innovation, ithas only included a subsetof the potentially relevant cultural, structural, and environmental variables. Future research is needed toexpand the set of variables and to examine how they independently and interactively in uence organizational innovation.Examples of other potentially relevant variables include rm strategy, performance measurement, and compensation policies, toname just a few. In particular, we had omitted the effects of individual-level factors. Yet often, organizational innovations andinnovative products are initiated by a small group of highly creative individuals ( Amabile et al., 1996; Mumford et al., 2002 ). Suchemployees' ability to experiment and take risks maydependon attributes of the organizational context, including the resource andtime constraints at work, and how performance is measured and rewarded. As such, there is room for additional research whichexamines how CEO leadership affects managers and employees at different levels in an organization, and how the latter interactamong themselves to enhance innovation. The framework proposed by Waldman & Yammarino (1999) could help to guide such anexamination of leadership effects across multiple levels of an organization.

Second, in analyzing cross-sectional data, our study can only reveal correlation but not causation. A longitudinal study thatrelates changes in leader behavior to changes in outcomes is more suited foruncovering causal relationships. Furthermore, we hadsuggested earlier that the effects of CEO leadership behaviors can be magni ed or dampened by selected attributes of theorganization, and a longitudinal study can illuminate the extent to which transformational leaders seek to recon gure theseattributes. Such a study would be further enriched by investigating the processes and pathways whereby effects arise, includingfurther disaggregating the variables under study. For example, are there differences in the nature of innovative efforts under morevs. less formalized work arrangements? In the case of empowerment where we had obtained an unexpected result, what are thevarious ways that employees respond to different levels of empowerment and in turn, how do these responses feed forward totheir innovation activities? Methods other than cross-sectional surveys, such as in-depth interviews and eld observation, aremore suited to answering such questions ( Yin,1989 ). And because innovation can occur not only in the products and services thatorganizations produce, but also in the processes through which people work, tapping into these aspects of the organization canincrease the completeness and richness of our understanding.

References

Adler, N. J. (2002). International dimensions of organizational behavior. Cincinnati, OH: South-Western.Amabile, T. M. (1998). How to kill creativity. Harvard Business Review , 76 (9), 7787.

Amabile, T. M., & Conti, R. (1999). Changes in the work environment for creativity during downsizing. Academy of Management Journal , 42 , 11541184.Amabile, T. M., Conti, R., Coon, H., Lazenby, J., & Herron, M. (1996). Assessing the work environment for creativity. Academy of Management Journal , 39, 11541184.Andriopoulos, C., & Lowe, A. (2000). Enhancing organizational creativity: the process of perpetual challenging. Management Science , 38 , 734743.Avolio, B. J., Bass, B. M., & Jung, D. (1999). Reexamining the components of transformational and transactional leadership using the Multifactor Leadership

Questionnaire. Journal of Organizational and Occupational Psychology , 72 , 441462.Avolio, B. J., Zhu, W., Koh, W., & Bhatia, P. (2004). Transformational leadership and organizational commitment: mediating role of psychological empowerment and

moderating role of structural distance. Journal of Organizational Behavior , 25 , 951968.Awamleh, R., & Gardner, W. L. (1999). Perceptions of leader charisma and effectiveness: the effects of vision content, delivery, and organizational performance.

The Leadership Quarterly , 10, 345373.Bagozzi, R. P., & Youjae, Y. (1988). On the evaluation of structural equation models. Academy of Marketing Science , 16 , 7494.Balkin, D. B., Markman, G. D., & Gomez-Mejia, L. R. (2000). Is CEO pay in high-technology rms related to innovation? Academy of Management Journal , 43 ,

1118 1129.Barclay, D., Higgins, C., & Thompson, R. (1995). The partial least squares (PLS) approach to casual modeling: personal computer adoption and use as an illustration.

Technology Studies , 2, 285309.Barron, O. E., Byard, D., Kile, C., & Riedl, E. J. (2002). High technology intangibles and analysts' forecasts. Journal of Accounting Research , 40 (2), 289 313.Barth, M. E., & Hutton, A. P. (2004). Analyst earnings forecast revisions and the pricing of accruals. Review of Accounting Studies , 9, 5996.Bass, B. M. (1985). Leadership and performance beyond expectations. New York: Free Press.

Bass, B. M., & Avolio, B. J. (1997). Full-range of leadership development: Manual for the Multifactor Leadership Questionnaire. Palo Alto, CA: Mind Garden.Bass, B. M., Avolio, B. J., Jung, D., & Berson, Y. (2003). Predicting unit performance by assessing transformational and transactional leadership. Journal of AppliedPsychology , 88 , 207218.

Bharadwaj, A., & Konsynski, B. R. (1997). Capturing the intangibles. InformationWeek , 649 , 7174.Bono, J. E., & Judge, T. A. (2004). Personality and transformational and transactional leadership: a meta-analysis. Journal of Applied Psychology , 89 (5), 901 910.Brislin, R. (1986). The wording and translation of research instruments. In W. J. Lonner & J.W. Berry (Eds.), Field methods in cross-cultural research (pp. 137 164).

Beverly Hills, CA: Sage.Brown, S. L., & Eisenhardt, K. M. (1998). Competing on the edge: Strategy as structured chaos. Boston, MA: Harvard Business Press.Burns, T., & Stalker, G. M. (1961). The management of innovation. London: Tavistock Publications.Business Week (2002, June 16). The Samsung Way (pp. 56 64).Carey, J. M., & Nahavandi, A. (1996). Using impression management to establish successful service-level agreements. In J. W. Beard (Ed.), Impression management

and information technology (pp. 51 64). Westport, CN: Quorum Books.Carmines, E., & Zeller, R. (1979). Reliability and validity assessment. Beverly Hills, CA: Sage.Chin, W. (1998). The partial least squares approach to structural equation modeling. In G. A. Marcoulides (Ed.), Modern methods for business research (pp. 295 336).

Mahwah, NJ: Erlbaum.Chin, W. W. (2000). Multi-group analysis with PLS. Retrieved March 3, 2003, from http://disc-nt.cba.uh.edu/chin/plsfaq/multigroup.htmChow, C., Shields, M., & Wu, A. (1999). The importance of national culture in the design of management controls for multi-national operations. Accounting,

Organizations and Society , 24, 441461.Daft, R. L. (2001). Organization theory and design. Cincinnati, OH: South-Western College Publishing.

http://disc-nt.cba.uh.edu/chin/plsfaq/multigroup.htmhttp://disc-nt.cba.uh.edu/chin/plsfaq/multigroup.htm


12/13


13/13

Sosik, J. J., Jung, D., Berson, Y., Dionne, S. D., & Jaussi, K. S. (2004). The dream weavers: Strategy-focused leadership in technology-driven organizations. Greenwich, CT:Information Age Publishing.

Spreitzer, G. (1995). Psychological empowerment in the workplace: dimension, measurement, and validation. Academy of Management Journal , 38 , 14421465.Tierney, P., Farmer, S. M., & Graen, G. B. (1999). An examination of leadership and employee creativity: the relevance of traits and relationships. Personnel

Psychology , 52 , 591620.Tsai, W. (2001). Knowledge transfer in intraorganizational networks: effects of network position and absorptive capacity on business unit innovation and

performance. Academy of Management Journal , 44 , 9961004.Waldman, D., & Yammarino, F. (1999). CEO charismatic leadership: levels-of-management and levels-of-analysis effects. Academy of Management Review , 24, 266285.Wilcox, R. (1998). How many discoveries have been lost by ignoring modern statistical methods? American Psychologist , 53 , 300314.Wold, H. (1985). Systems analysis by partial least squares. In P. Nijkamp, H. Leitner, & N. Wrigley (Eds.), Measuring the unmeasurable (pp. 221 252). Martinus

Nijhoff: Dordrecht.Yin, R. K. (1989). Case study research: Design and methods. Beverly Hills, CA: Sage.Zhou, J. (2003). Feedback valence, feedback style, task autonomy, and achievement orientation: Interactive effects on creative performance. Journal of Applied

Psychology , 83 , 261276.Zhu, W., Chew, I., & Spangler, W. (2005). CEO transformational leadership and organizational outcomes: the mediating role of human-capital-enhancing human

resource management. The Leadership Quarterly , 16 , 39 52.


Towards understanding the direct and indirect effects of CEOs' transformational leadership on firm...

Documents

Transcript of Towards understanding the direct and indirect effects of CEOs' transformational leadership on firm...