misfit and milestones: structural elaboration and

r Academy of Management Journal2016, Vol. 59, No. 4, 1430–1450.http://dx.doi.org/10.5465/amj.2014.0526

MISFIT AND MILESTONES: STRUCTURAL ELABORATIONAND CAPABILITY REINFORCEMENT IN THE EVOLUTION

OF ENTREPRENEURIAL TOP MANAGEMENT TEAMS

AMANDA J. FERGUSONNorthern Illinois University

LISA E. COHENMcGill University

M. DIANE BURTONCornell University

CHRISTINE M. BECKMANUniversity of Maryland

We examine how “top management team (TMT) misfit,” or discrepancies between theTMT’s functional roles and the qualifications of themanagers who fill those roles, affectsthe evolution of TMT composition and structure in a longitudinal study of entrepreneurialventures. We distinguish two types of misfit—overqualification and underqualification—and study how each is associated with TMT changes. We further consider the moderatingeffect of firm development. Results reveal that underqualified TMTs hire newmanagers toreinforce existing capabilities whereas overqualified TMTs elaborate their role structures.However, achieving developmental milestones (i.e., obtaining venture capital fundingand staging an initial public offering) is a critical contingency to TMT change: absentthese milestones, firms neither hire new managers nor add roles, even when theyseemingly need to do so. These findings contribute to knowledge of how TMTs and newventures evolve by underscoring the importance of simultaneously attending to TMTcomposition and structure.

The top management teams (TMTs) of new ven-tures perform a constant balancing act between ful-filling immediate needs and seizing opportunities,and they must do so in a context of constrained re-sources. One factor they may work to balance is theTMT itself: specifically, matching their own skillswith the formal role structure in which they work.Although researchers frequently treat TMTexpertiseand roles as interchangeable, recent evidence sug-gests that they are conceptually and empiricallydistinct (e.g., Beckman & Burton, 2011; Bunderson &Sutcliffe, 2002; Crossland, Zyung, Hiller, & Hambrick,

2014). TMT expertise drives strategic decisions and isa signal to important stakeholders (Beckman, Burton,& O’Reilly, 2007; Boeker, 1997). Role structures canenable coordination and help firms manage dynamicenvironments (Bechky, 2006; Cohen, 2013; Sine,Mitsuhashi, & Kirsch, 2006). Both are worthy of in-vestigation independently and in conjunction.

When TMT composition and roles are not aligned,new ventures may have a need or an opportunity tochange. Misalignment between TMT compositionand role structure—which we label TMT misfit—signifies a gap between what the TMT intends todo and what it is capable of doing. A deficit of qual-ifications relative to the current role structure,or underqualification, reflects unmet skill require-ments, while an excess of qualifications relative tothe role structure, or overqualification, representsopportunity. In some instances, the TMT may needto augment its expertise because of weaknesses incomposition; in other situations, there may be an op-portunity to develop and elaborate the role structure

We thank Marc Gruber and the three anonymousAcademy of Management Journal reviewers, as well asPhanish Puranam, Elena Belogolovsky, John Hausknecht,Joe Broschak, and Bart Sharp for their invaluable com-ments and advice. We also thank participants in researchseminars at London Business School, Tulane (FreemanSchool of Business), University of Navarra (IESE), andBocconi for their feedback.

1430

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.

http://dx.doi.org/10.5465/amj.2014.0526

given the qualification on the team. Thus, we ask,“How and when do entrepreneurial TMTs adjusttheir people and roles in response to TMT misfit?”

Two readily apparent adjustments are: (1) hiringnew managers into the existing role structure to en-hance team capabilities, and (2) adding new roles torecognize and allow for fuller use of existing capa-bilities. We expect that TMT misfit will affect thelevel of both; however, these adjustments requiremanagerial attention and financial resources that areoften in short supply. Therefore, changes to TMTcomposition and structure likely vary with the newventure’s level of development. As new venturespass key developmental milestones such as obtain-ing venture capital financing or staging an initialpublic offering, they acquire financial-, human-, andsocial-capital resources (e.g., Gompers & Lerner, 2001;Hellmann & Puri, 2002; Wasserman, 2003) that fa-cilitate attracting newcomers and/or elaborating rolestructures. It may be that, without the resources ob-tained through these developmental milestones,firms are unable to obtain the expertise they need ortake advantage of opportunities. Thus, we ask a sec-ond question: “How do the patterns of response toTMTmisfit varywith the level of resources availablein the firm?”

We examine the effects of TMT misfit on TMTcomposition and structure in high-technology en-trepreneurial firms and consider these effects in lightof whether the firms have passed the developmentalmilestones of achieving venture capital financingand staging an initial public offering. Our findingsreveal that underqualified TMTs reinforce theircurrent expertise by hiring new managers withoutexpanding the role structure, whereas overqualifiedTMTs elaborate their structure without hiring newmanagers. We also find that firms are more likely torespond to the needs for new expertise revealed byunderqualified TMTs than to the opportunities forrole elaboration provided by overqualified TMTs.

Ventures that have not achieved developmentalmilestones, however, make few changes to the TMT,even if it is characterized by misfit. This presentsa professionalization paradox: even if a firm needsor wants to make changes that would signal itslevel of professionalization and so help it reach de-velopmental milestones that bring an influx of re-sources (Chen, Hambrick, & Pollock, 2008; Khaire,2010; Zott &Huy, 2007), it is unlikely to do sowhen itlacks the very resources it seeks to attract. In short,misfit TMTs need to balance their own composi-tion and structure, but are constrained in theirability to do so.

Our study contributes to entrepreneurship theoryby adding important nuances to arguments about theevolution of new venture TMTs. In particular, whileextant literature confounds people and roles whenconsidering the evolution of TMT characteristics,our approach not only acknowledges theoreticaldistinctions between team members and their roles,but also empirically models the match between thetwo to better explain the factors that drive change.Indeed, our study reveals that assuming alignmentbetween these characteristics is inaccurate and canmask some of the distinct processes that shape bothhiring and role additions in new venture TMTs.Further, while extant research considers the role offirm development in shaping TMTs, it does sowithout considering how this interacts with the in-ternal aspects of the TMT. Our study considerswhether and how managers and roles in new ven-tures may be misfitting, in conjunction with theventure’s current stage of development, to providea more complete understanding of TMT evolutionand its paradoxes. Beyond contributing to entrepre-neurship theory, however, highlighting TMT misfitand its effects also has implications for theories oforganizational role structures, upper echelons, andmisfit, and ultimately helps to answer the importantquestion of why TMTs “look the way they do”(Hambrick, 2007; Pettigrew, 1992).

THEORY AND HYPOTHESES

New Venture TMT Evolution

The characteristics of TMTs are rarely static, andnowhere might they change more than in new ven-tures. For example, in its earliest days, a newventure’sTMT might comprise a couple of college dropoutsor unemployed individuals (Dencker, Gruber, &Shah, 2009a, 2009b); however, over time, the firmmight bring in more seasoned professional man-agers (Hellmann & Puri, 2002). Such compositionalchanges have been tied to factors such as the team’sindustry experience or functional diversity (Boeker &Wiltbank, 2005), new venture growth (Boeker &Karichalil, 2002), and the achievement of develop-mental milestones such as obtaining venture capitalfinancing (Wasserman, 2003). Changes to TMTs in newfirms may also come in the form of altering rolestructures (Baker & Nelson, 2005). For example,Beckman and Burton (2008) found that new ventureTMTs that began with a limited set of functional po-sitions had difficulty developing more completerole structures later on. We argue that any changesto composition and structure follow from some

2016 1431Ferguson, Cohen, Burton, and Beckman

combination of need and opportunity for change,and that these needs and opportunities can becaptured by considering the fit between the TMT’sformal roles and the qualifications of the managerswho fill those roles.

TMT Misfit and TMT Change

Frequently, scholars assume that the functionalbackgrounds of top managers and the functionalroles of the TMT are brought into alignment at thetime of hiring. For example, Menz (2012) suggestedthat TMT roles are determined a priori by the CEOand then filled by suitable individual executives.However, in practice, the functional backgrounds ofthe top managers often diverge from the functionalrole structure of the TMT, particularly in new ven-tures. Some TMTs are comprised of top managerswho possess narrow functional expertise, despiteholding roles that indicate a broader range of re-sponsibility. Other TMTs may be made up of broadgeneralists, with prior functional experiences aboveand beyond the roles they currently hold (seeBunderson & Sutcliffe, 2002). Both scenarios suggestthat mismatches between TMT roles and people canand do occur, and we argue that these discrepancies(i.e., TMT misfit) may represent needs or opportu-nities for TMT change.

We define “TMTmisfit” as any situation in whichthe composite knowledge, skills, and abilities of topmanagers do not correspond to the set of roles thatthey occupy. As such, TMTmisfit is conceptualizedat the team level, assessing the set of functionalqualifications and role requirements of the TMT asa whole. This focus on misfit at the team level is adeparture from prior work in the person–environmentfit literature, which primarily focuses on matchesbetween individuals and their jobs or organiza-tions (Chatman, 1991; Edwards, 1991; Kristof-Brown,Zimmerman, & Johnson, 2005). However, a group-level conceptualization of TMT misfit is importanthere because it captures overall TMT capabilityrelative to its stated formal structure, and thus thepotential for change in people and roles.

TMT misfit may occur in two directions: that is,there may be an excess or a deficit of qualificationswith respect to existing TMT role structures. Again,wedepart from theperson–environment fit literaturehere, which largely focuses on underqualification(Erdogan, Bauer, Peiro, & Truxillo, 2011). However,considering both TMT overqualification and TMTunderqualification is important in that they repre-sent fundamentally different problems. With TMT

underqualification, stated goals cannot be met withthe qualifications and associated human capitalcurrently in the team, indicating an underlying needfor change. With TMT overqualification, the TMT iscurrently fulfilling its formalized role requirementsyet has extra qualifications. As such, the TMT couldpotentially achieve additional goals by fully capi-talizing on the qualifications of current managers(Erdogan et al., 2011), indicating an unrealized op-portunity for change.

To date, research on TMT change has studied twobroad phenomena: succession as an organizationalresponse to underqualified managers (Boeker &Karichalil, 2002;Hellmann&Puri, 2002;Wasserman,2003) and the hiring of experts into new roles (Baron,Burton, & Hannan, 1999; Zorn, 2004). Both ap-proaches conflate people with roles. Turnover studiestreat roles as static and examine hiring and exitswithin established roles. Studies of new functionalroles presume that these roles are filled by externalhires and fail to consider the possibility that theyarise as the result of reshuffling existing people. Ourgoal is to expand and add nuance to our under-standing of entrepreneurial TMTs by carefully ex-amining when and how TMTs evolve in responseto under- and overqualified management teams. Anecessary first step is to disentangle persons androles. Thus, we examine hiring net of adding newroles, a TMT change we term capability reinforce-ment, and the addition of new roles net of hiring,a TMT change we term structural elaboration.1

TMT underqualification: A need for change.TMT underqualification is a deficit of functionalqualification with respect to the role structure ofthe TMT, which results when there are too few man-agers with requisite capabilities currently in the firm(Cappelli, 2008; Groysberg & Lee, 2009). The mostlikely response tounderqualification in theTMTsofnew ventures is to hire newmanagers who have theneeded skills (seeHambrick & Crozier, 1985). Indeed,Cappelli (2008) argued that hiring outsiders to make

1 Although turnover studies sometimes combine TMThires and exits for an overall measure of compositionalchange (e.g., Boeker & Wiltbank, 2005), here we focuson hiring because it has a strong association with newventure growth (Eisenhardt & Schoonhoven, 1990; Gilbert,McDougall, & Audretsch, 2006) and is more within thecontrol of the TMT itself. However, we control for mana-gerial exit in our empirical models; this allows us to cap-ture hiring that increases the skill capabilities of the TMTbeyond the hiring that may occur to fill vacancies, forexample.

1432 AugustAcademy of Management Journal

up the difference between current and desired skillsmight be an effective strategy in situations with un-certain demand and short-term needs, which is con-sistent with those faced by new ventures.

In contrast, creating new structural roles maynot provide a viable or helpful response to under-qualification in a TMT. First, adding new roles to theTMT would add to rather than reduce capability re-quirements. Moreover, eliminating roles from thestructure, even ones that are not matched by the ca-pabilities of team members, is unlikely in light ofwhat we know about organizational inertia and mayhave broader negative implications. For example,suppose four engineers start a new venture, yet, indoing so, take on broad TMT roles (e.g., those in-cluding general management, business develop-ment, etc.). Altering the role structure to reflect theircurrent capabilitiesmay threaten their legitimacy bysignaling low TMT quality to potential investors.Functional breadth in TMT role structures speedsthe process of staging an initial public offering(Beckman & Burton, 2008) and heterogeneity in ex-pertise is valued by venture capitalists (Franke,Gruber, Harhoff, & Henkel, 2008), both of which ar-gue against altering role structures to reflect actual-versus-intended capability when a TMT is morenarrowly qualified than its stated roles. In such in-stances, we suspect that team members will takeadditional time to “grow into” their roles, perhapsthrough on-the-job experience or training (Bidwell &Briscoe, 2010; Hersch, 1995), while minimizingchanges to the current role structure. Indeed, theymay even divert resources from the activities asso-ciated with role additions so they can focus on de-veloping and making use of the somewhat limitedskills in the team. In short, TMT underqualificationshould result in capability reinforcement throughthe hiring of new managers but not structural elab-oration through the formation of new roles.

Hypothesis 1. TMT underqualification will posi-tively predict TMT hiring.

Hypothesis 2. TMT underqualification will nega-tively predict new TMT roles.

TMToverqualification:An opportunity for change.TMT overqualification is an excess of functionalqualifications with respect to the role structuresthat exist within the TMT. Such TMTs do not needto change their composition, as they are alreadyfulfilling the goals implied by their formal rolestructure, yet their excess qualifications presentopportunities that could be capitalized on with the

current team. As such, we suspect that any changesto these TMTs will come in the form of adding newroles to the structure to reflect this unutilized ex-pertise (Miner & Estler, 1985) rather than hiringnewcomers. Adding new roles enables the TMT tomore accurately signal the capability that existswithin the team, whichmay be important as the newventure TMT attempts to attract investors (see Frankeet al., 2008) or as it makes preparations for an ini-tial public offering (Chen et al., 2008). In addi-tion, new roles can confer status and recognition(Baron&Bielby, 1986; Baron&Pfeffer, 1994), whichmay induce overqualified managers to stay witha young firm.

At the same time,weexpect levels ofTMThiring tobe lower when levels of TMT overqualification arehigher. Consistent with this argument, Boeker andWiltbank (2005) found that there was less hiring andexit in new venture TMTs that were more function-ally diverse—a phenomenon highly correlated withhaving more skills in the team—even in the faceof expansion. Similarly, evidence has shown thatentrepreneurial ventures with greater functionalbreadth in their TMTs bring in fewer new employees(Dencker et al., 2009a). Therefore, TMT overquali-fication should result in structural elaboration throughthe addition of new TMT roles but not capability re-inforcement through hiring.

Hypothesis 3. TMT overqualification will nega-tively predict TMT hiring.

Hypothesis 4. TMT overqualification will posi-tively predict new TMT roles.

The Moderating Effects of Firm Development

Thus far, we have argued that the gaps that resultfrom TMT misfit reveal needs or opportunities thatcontribute to the evolution of TMTs in new ventures,without considering what resources are available tomake these changes. An important contingency fac-tor that may affect the relationship between TMTmisfit and TMT change, therefore, may be a newfirm’s stage of development, as indicated bywhetherit has achieved developmental milestones such asreceiving venture capital financing or staging aninitial public offering (e.g., Boeker & Karichalil,2002; Chen et al., 2008; Gompers & Lerner, 2001;Hellmann & Puri, 2002; Rubenson & Gupta, 1996;Wasserman, 2003, 2006). Each of these milestonesprovides resources that facilitate hiring and addingroles, suchas knowledgeabouthow to structure rolesand attract experienced top managers, the ability to


recruit thesemanagers, and the financialmeans to doso. For example, venture capitalists not only providecapital investments innew firmsbut also help recruitsenior managers through their professional contacts(Bygrave & Timmons, 1992). While these resourcesmay facilitate overall hiring and role additions, theymayalso intervene in the relationshipbetweenmisfitand TMT changes. Therefore, the effects of TMTmisfit on TMT hiring and role additions should beconsidered with respect to the firm’s level of devel-opment (see Figure 1).

We predicted that TMT underqualification wouldresult in capability reinforcement through increasedhiring but not structural elaboration through theaddition of new roles. Some executive hiring mayoccur even before developmental milestones bringincreased resources. For example, to help attractventure capital financing, teams may hire new-comers to ensure they have extensive industry ex-perience and diversity in educational backgrounds(Franke et al., 2008). New venture TMTs also go togreat lengths to ensure that they attract prestigiousexecutives and directors in the year prior to an initialpublic offering (Chen et al., 2008). However, becausehiring new managers often requires significant re-sources both in terms of financial costs and in termsof tapping into social networks to recruit qualifiedcandidates, we expect that the positive effect of TMTunderqualification onhiring should be even strongerwith higher levels of firm development.

We also predicted that TMT underqualificationwould negatively affect new TMT roles because

adding new roles does not solve the problems thatunderqualification presents. In fact, this responsewould likely add to rather than reduce capabilityrequirements if new roles were added, whereasnarrowing the role structure to reflect existing ca-pabilities may send negative signals to potentialstakeholders. Although achieving developmentalmilestones might provide knowledge and financialresources that could give new firms more flexibilityto address mismatches by changing their structures,it does little to resolve these concerns about capa-bilities and legitimacy. Indeed, legitimacy continuesto be a concern for both the TMT and its investors infirms that have passed such milestones (see Guler,2007; Zimmerman & Zeitz, 2002). Therefore, we ex-pect that the relationship between TMT under-qualification and the number of new roles created isunlikely to change even after reaching develop-mental milestones. In short, the achievement of de-velopmental milestones will moderate the effect ofTMT underqualification on hiring, but not that ofTMT underqualification on new TMT roles.

Hypothesis 5. The positive effect of TMT under-qualification on TMT hiring will depend on firmdevelopment, such that the effect will be strongerwith greater development.

Recall that, in situations of TMToverqualification,we expect to see an increase of new TMT roles anda decrease of new TMT hires. Overqualified TMTsmay want to add roles to capitalize on opportunitiescreated by overqualification, but doing so can be

FIGURE 1TMT Misfit and Firm Development as Predictors of New Venture TMT Evolution

H5 H6

H2

H3

H4

TMT

Underqualification

TMT

Overqualification

TMT Hiring

(i.e., capability

reinforcement)

New TMT Roles

(i.e., structural

elaboration)

Firm Development

-

-

+

+H1


a significant and costly undertaking (Stinchcombe,1965), requiring managerial time, expertise, and co-ordination (Sine et al., 2006). Overqualified TMTswill have some surplus capabilities that might beapplied to developing new roles; however, the re-sources gained by achieving developmental mile-stones can further facilitate role additions. Theinflux of cash and expertise may afford managersmore time and financial resources to make adjust-ments in the structure, and investors themselvesmayprovide expert coaching (see Baum & Silverman,2004; Hellmann, 2000) to help them more appro-priately signal their quality through improving anddeveloping the role structure. Thus, the positive ef-fect of TMToverqualification onnew roles should beeven stronger at higher levels of firm development.

We also argued that TMT overqualification woulddampen hiring since the qualifications of current TMTmembers already fulfill stated goals and anticipatedneeds could potentially be filled with these existingsurplus capabilities. Overqualification creates neithera need nor an opportunity to hire, and thus the effectof TMT overqualification on hiring should be inde-pendent of the increased resources associated withachieving developmental milestones. So, while theresources that come with achieving developmentalmilestones may themselves result in greater hiring(Davila, Foster, & Gupta, 2003), these additional re-sources would be unrelated to the effect of TMToverqualificationonhiring. In sum,weexpect that theachievement of developmental milestones will mod-erate the effect of TMToverqualification onnewTMTroles, but not that of TMToverqualification on hiring.

Hypothesis 6. The positive effect of TMT over-qualification on new TMT roles will depend onfirm development, such that the effect will bestronger with greater development.

METHODS

Data and Sample

We test these ideas by examining a sample of 167high-technology entrepreneurial firms in a singleregion that have been studied previously (Baron,Hannan, & Burton, 1999, 2001; Burton, Sørensen, &Beckman, 2002). Previous studies collected TMTdata on these firms and examined important ele-ments of entrepreneurship theory, such as teamfactors that lead to venture capital financing or initialpublic offering (Beckman & Burton, 2008; Beckmanet al., 2007) and factors that result in different stra-tegic choices (Beckman, 2006). The current study is

designed to examine the evolution of the people androle structures that comprise the TMT in these en-trepreneurial firms. Building from past studies thatdocument path dependence in TMT compositionand structure (Beckman & Burton, 2008) and positionimprinting of individual roles (Burton & Beckman,2007), the present study is an attempt to better un-derstand the mechanisms of imprinting and path de-pendence at the team level. We extend ideas from thefit literature and consider how types ofmisfit betweenall TMT roles and the people that occupy those rolesdrive subsequent changes to TMT composition andstructure.

The firms in the samplewere less than 10 years oldand had fewer than 10 employees when they wereinitially contacted. This database contains rich in-formation about the roles in the TMT and the man-agerswho occupy them, aswell as information aboutthe developmentalmilestones achieved by each firm(e.g., dates of venture capital funding and initialpublic offering). Information about the roles andmanagers included in the TMTs was gathered viainterviews, internal company documents, publicarchives, and extensive Web searches. As describedin Burton and Beckman (2007), the research teamtracked the career histories, start dates, initial jobtitles, job title changes, and departure dates of eachexecutive who ever held a TMT position of vicepresident (VP) or higher in one of the sampledfirms from firm founding to December 2000 oruntil acquisition or failure. The database containsinformation on 1,452 executives holding 1,918TMT roles.

Measures

Independent variables. Based on our conceptu-alization of TMT misfit, we needed a measure thatcould (a) assess misfit across the entire managementteam, rather than as an aggregation of individual-level misfit; (b) capture both overqualification andunderqualification instead of one combined mea-sure; and (c) be assessed on a set of objective quali-fications that could be observed in both roles andmanagers. TMT misfit was calculated by comparingthe number of times each functional area was listedin any job title for the entire top team with the totalnumber of managers with experience in those func-tional areas.

First, both current job titles and the career historiesof top managers at the hierarchical level of VP orabove were classified according to 11 differentfunctional areas: sales,marketing, customer support,


operations, finance, administration, human resources,strategic planning/business development, science/engineering, information systems, and general man-agement (Burton&Beckman, 2007). For example, oneof the TMTs in our sample included the followingjob titles (and functional classifications): president/chief executive officer (general management), VPof engineering (science/engineering), VP of interna-tional operations (operations), chief financial officer(finance),VPofmanufacturing (operations), andVPofmarketing (marketing). This team thus had one gen-eral management role, one science/engineering role,one finance role, one marketing role, and two opera-tions roles.

Next, the career histories of the topmanagers werealso classified according to these 11 functional areas.Executives were considered to have experience inthese areas if they had titles that included thesefunctions in any of their previous three jobs. Onemanager in our sample was recruited from a com-petitor to replace the VP of marketing. The new hirehad served for four years first as strategic marketingmanager and later as marketing director for a busi-ness unit. His prior position had been systems engi-neer at a different firm. We would code this personas having prior experience in both marketing andengineering.

Using these data,we then calculated the amount ofoverqualification or underqualification relative tothe current role structure at the team level. TMToverqualification is a count of the prior experiencesabove that required by the current roles of the teamacross all 11 functions. For example, if there are threetop managers who have operations experience butonly two current job titles that indicate an operationsfunctional role (e.g., the VP of international opera-tions and the VP of manufacturing in the sampleTMT described above), the TMT would have anoverqualification score of 1 for the operations func-tion. Overqualification was then summed across all11 functional areas to give an overallmeasure of howmuch more experience the group of top managershad relative to the current roles listed for the team(mean 5 2.45, SD 5 2.27, range 5 0 to 18).

TMT underqualification is a count of the roleslisted for the TMT that include functional areas forwhich the group of top managers lack prior experi-ence. For instance, a firm that has three roles withmarketing listed in its TMT titles but only two topmanagers with marketing experience would havean underqualification score of 1 for the marketingfunction. Underqualification was summed acrossthe 11 functions to indicate how much the group of

top managers lacked experience relative to the cur-rent roles listed for theTMT(mean50.47,SD50.92,range50 to 9).On average, TMTshadmanagerswithexperience in 1.84 functions (SD 5 0.90) while theroles in the team required capabilities in 1.24 func-tions (SD 5 0.48). When aggregated across all posi-tions, this produced a higher level of overqualificationthan underqualification.

After executives held their current roles for oneyear, a long enough time frame to learn from on-the-job experience (National Center for O*NETDevelopment, 2011), they were considered to haveexperience in that function (i.e., had they not other-wise had prior experience in their career history).Other specifications of the extent to which execu-tives learn on the job were examined as robustnesschecks. Finally, we lagged the TMT misfit variablesby 12 months, t 2 12, in order to allow the effects ofoverqualification andunderqualification on changesto TMT composition and structure to be realizedwithin the firms over the following year.

Both the TMT overqualification and TMT under-qualification measures assess the fit between thework deemed important by theTMT (e.g., as listed byTMT roles) and the ability of the team to carry outthat work (e.g., as assessed by the total amount ofprior functional experiences across all top managersin those roles). The assumption of this measure isthat the prior functional experience of one managermay be helpful to another manager who holds a dif-ferent role. However, it is possible that topmanagerswill not use their skills to fill in voids in other areaswithin the TMT. Therefore, we constructed an al-ternative measure of TMT misfit as a robustnesscheck. For this alternative measure, we consideredthe match between current job title and prior func-tional experience for each individualmanager beforeaggregating these scores to the team level, as in anadditive composition model (Chan, 1998). For ex-ample, in our sample TMT described above, theremay be one individual who has experience beyondthat required by his or her role (e.g., the chief oper-ating officer has experience in both general man-agement and marketing) and one individual wholacks experience required by his or her role (e.g., theVP of marketing has business development experi-ence but not marketing experience). Assuming allother individuals on the TMT have experiencecommensurate with their roles, this TMT wouldhave an overqualification score of 1 (i.e., the chiefoperating officer) and an underqualification score of1 (i.e., VP of marketing). Thus, unlike the team-levelmeasure in which the chief operating officer’s


marketing experience would “count” for the role ofVP of marketing, this measure considers the under-qualification of the VP of marketing without respectto the prior functional experience of any othermember of the team. These measures are slightlylarger in magnitude for both types of misfit, and as-sume that individuals’ experiences are not utilizedbeyond their specific roles. We report the resultsusing this measure in the robustness checks sectionfor comparison purposes.

Next, we assessed firm development by examiningwhether a firm had received venture capital fundingor staged an initial public offering. Both of theseevents are considered to be important developmen-tal milestones (e.g., Boeker & Karichalil, 2002; Chenet al., 2008; Hellmann & Puri, 2002; Rubenson &Gupta, 1996; Wasserman, 2003) and each representsthe receipt of additional resources that are broughtinto the new venture, in terms of financial capitaland expertise (Gompers & Lerner, 2001; Hellmann &Puri, 2002; Wasserman, 2003). Therefore, we dummycoded the firm’s achievement of each of these twomilestones at time t, and then summed across eventsto create an overall measure of level of firm develop-ment. This variable ranged from 0, in which a firmhad no venture capital funding and no initial publicoffering, to 2, in which a firm had received venturecapital funding and had staged an initial public of-fering. In our sample, 71% of the firms receivedventure capital funding and 53% went public.These numbers are high relative to start-ups in thegeneral population of start-ups, but our sampleconsists of high-technology firms in the 1990s, aperiod of incredible growth for entrepreneurial com-panies in the United States.

Dependent variables. Changes to the TMT in re-sponse to the needs or opportunities reflected inTMT misfit could come in the form of hiring newmanagers and/or adding new structural roles. Hiringnewmanagers expands theTMT in terms of numbersof managers and, more importantly for our argu-ments, in terms of capabilities. Since hiringmay alsooccur as new roles are created,weexamine thehiringof new managers into the TMT while holding theTMT structure constant. The second type of change,adding new roles, is an elaboration of structure. Newroles involve the addition of distinct tasks not pre-viously identified in the role structure. We are partic-ularly interested in the phenomenon of broadeningthe role structure within which current managersoperate. For example, a current manager may take ona new role of VP of business development, whichhad never before existed in the firm, and his or her

previous role (e.g., VPof sales andmarketing)might befilled by another manager, altogether eliminated, orsplit into roles for other managers. Since new rolescould also be filled by hiring new managers (e.g., anewcomer takes on a new role of VP of business de-velopment) (see Levesque, 2005; Miner, 1987), weexamine the addition of new roles to the TMTholdingTMT hiring constant.

First, we calculated TMT hiring as the sum of alltopmanagerswho appeared inmonth t butwhowerenot employed by the firm in month t 2 1. This de-pendent variable comprises the number of hires inthe TMT over a one-year period; that is, TMT hiringmeasured at time t is the sum of the new hires withinthe TMT over the subsequent 12 months (t 1 12).However, this variable is updated each month to re-flect hiring at different times in the prior year. Thisallowed us to capture hiring in a fine-grained way,such that the misfit variables are predicting 12months of hiring exactly one year after assessingTMT misfit. In contrast, yearly analysis could in-troduce greater variability in these time spans; forexample, examining misfit that occurred in Januaryof one year with hiring that occurred in Novemberof the following year, even though the time span isgreater than 12 months. Monthly updating thuscaptures the dynamic changes through hiring ob-served in these firms within the 12-month period forwhich they are counted and allows us to be moreprecise with respect to the time spans between in-dependent and dependent variables.

Next, we assessed new TMT roles as being addedwhen a job title appeared that had never beforeexisted in the firm (see Cohen &Broschak, 2013). Forexample, the addition of a VP of sales title that hadnever before existedwould be counted as a newTMTrole. However, if the firm later added titles such asexecutive VP of sales, VP of NorthAmerican sales, orVP of sales and marketing, these would not be con-sidered to be newTMT roles as they are variations ona previously existing role. In addition, splitting anexisting role (e.g., splitting a previously existing roleof VP of sales and marketing into VP of sales and VPof marketing) would not be considered a new role inour terminology as the need for the functional ex-pertise had already been formally articulated. Noneof the titles that existedduring the foundingmonthofthe firm were coded as new, since, prior to that date,the firm did not exist. Finally, we did not include“founder” as a job title in our coding scheme as thistitle does not imply a structural role or responsibilitywithin the TMT; rather, it indicates an individual’sstatus or past history with the firm. The first author


coded the job titles that existed within the firmsbased on the above coding scheme. A second coderapplied the coding scheme to half of the dataset inorder to verify that the coding scheme was appliedconsistently and reliably. Agreement between thetwo coders was acceptable (Cohen’s k 5 .87), so wewere confident in using this coding scheme to assessnew TMT roles. There were 1,918 job titles in thedataset, of which 40%were considered new roles, inthat no portion of the job title existed in the priorperiods. Like TMT hiring, new TMT roles werecounted over a one-year period for the analysis, butupdated monthly to reflect the fine-grained changesin TMT structure that occurred over that period.

Control variables. We controlled for industryto account for potential differences in opportunitystructures for firms within particular industries.For example, manufacturing has lower turnoverrates than high-technology firms in general (Burton &Beckman, 2007) and may also experience more sta-bility in roles. In our analyses, we found that dif-ferent industries matter for TMT hires and roleadditions, so we included five dummy variables toaccount for variations among firms in all six in-dustries. Firm age has been shown to relate to rolestructure (Baron, Burton, & Hannan, 1999), so wecontrolled for firm age as measured by the numberof months since founding. Firm age was lagged by12 months to match our misfit variables.

Size can affect potential for growth via new hiresand proliferation of new job titles (Baron & Bielby,1986; Strang & Baron, 1990), and team size may beparticularly relevant for entrepreneurial firms (Burton& Beckman, 2007; Eisenhardt & Schoonhoven, 1990).Therefore, we control for team size as the numberof executives in the TMT. Evidence suggests that itis important to control for organizational tenure(Cohen & Broschak, 2013), which we measured asboth the mean tenure among a firm’s TMT and thestandard deviation of tenure among top managers.Finally, because the level of functional diversity ina team may also influence whether positions orpeople are altered, we controlled for the functionaldiversity of the team by calculating Blau’s (1977)index for functional diversity for the executives’three most recent job titles. Team size, average tenureand tenure dispersion, and functional diversity werealso lagged by 12 months.

Next, we controlled for TMT exits, which couldaffect the amount of hiring and the addition of newroles in the current structure, by calculating the totalnumber of top managers who were employed by thefirm in month t 2 1 but not employed by the firm in

month t. Similar to the dependent variables de-scribed above, exits were considered over a one-yearperiod for the analysis butupdatedmonthly to reflectthe dynamic changes occurring in the compositionof the TMT over that period.

Finally, in order to capture distinct changes to theTMT in terms of hiring or roles, it is necessary tocontrol for the other potential dependent variablewhen estimating the effects of misfit on either of thechanges to the TMT. For example, we have arguedthat TMT overqualification should result in the elab-oration of structure through addition of new TMTroles for existing top managers. Any role changesshould be over and above those that occur whenhiring newcomers to fill new roles. Thus, whenestimating the effects of TMT misfit on new TMTroles, it is essential to control for TMT hiring in thesame period. Similarly, we control for new TMTroles when estimating the effects of TMT misfit onTMT hiring, which captures the reinforcement of ca-pabilities while keeping the role structure constant.

RESULTS

Descriptive statistics and correlations among studyvariables are presented in Table 1. Of note is thehigh amount of overqualification (mean 5 2.45,SD 5 2.27) relative to underqualification (mean 50.47, SD5 0.92) in these teams. Moreover, there isa small positive correlation between over- andunderqualification (r 5 .05), which indicates thatthese are independent types of misfit. Given rela-tively high correlations between some of our in-dependent variables and control variables (e.g.,TMT overqualification and functional diversity),we checked for the possibility of multicollinearity.First, we analyzed variance inflation factors, whichrevealed that no variables had scores higher than 10(the highest score was 4.17) (Hair, Anderson, Tatham,& Black, 1995; Neter, Wasserman, & Kutner, 1996).Next, we ran our models without each of the controlvariables that had high correlations with the indepen-dent variables (e.g., functional diversity, team size),and the exclusion of these controls yielded resultsconsistent with those presented below. Not surpris-ingly, TMT hiring and new TMT roles are highly cor-related, because hiring often occurs with the additionof new roles. Again, our focus here is on capabilityreinforcement through TMT hiring while controllingfor structural elaboration through adding new rolesand vice versa.

To estimate the effects of TMT misfit on com-position and structure, we performed monthly


panel-Poisson random effects regression analysespredicting counts of the number of new TMT hiresand newTMTroleswithestimates groupedby firm (seeTable 2). Poissonmodels are superior to ordinary leastsquares regression when estimating count data be-cause the distribution of the data are typically skewed(Allison, 2009), and are a standard approach foranalyzing panel count data in particular (Somaya,Williamson, & Lorinkova, 2008).

Hypothesis 1 and 2 suggested that TMT under-qualification would positively predict new TMThires and negatively predict new TMT roles. Models1 and 4 are our baseline models with only controlvariables. Models 2 and 5 add the main effects ofTMT underqualification on new TMT hires androles, revealing that TMT underqualification posi-tively predicts the number of hires and negativelypredicts adding roles, supporting these hypotheses.The practical significance of these findings is illumi-nated by the incident rate ratios from these models,which suggest that, for every unit increase in TMTunderqualification, there is a 7% increase in the in-cident rate of the number of new TMT hires anda 12%decrease in the incident rate of the number ofnew TMT roles. Considering these effects overtime, an average firm, with the mean level of TMTunderqualification, takes approximately 2.6 yearsto hire a newcomer and approximately 1.9 years toadd an additional role. In contrast, firms with highlevels of TMT underqualification (i.e., two stan-dard deviations above the mean) hire a newcomerabout 4 months sooner but wait about 6 monthslonger to add a new role.

Hypotheses 3 and 4 suggested that TMT over-qualificationwouldnegativelypredictTMThiringandpositively predict new TMT roles. Models 2 and 5show the main effects for TMT overqualificationon new TMT hires and roles, revealing that TMToverqualification negatively predicts the number ofhires and positively predicts adding roles, supportingboth hypotheses. Note that the coefficients for TMToverqualificationpredictinghires and roles are smallerthan those of TMT underqualification, a point we ex-plore further with our robustness checks. Practically,for every unit increase in TMT overqualification thereis a 3% decrease in the incident rate of the number ofnew TMT hires and a 3% increase in the incident rateof the number of new TMT roles. Considering theseeffects over time, an average firm, with the mean levelof TMT overqualification, takes approximately 2.6years to hire a newcomer and approximately 1.9 yearsto add an additional role. In contrast, firms with highlevels of TMT overqualification wait about 5 months

TABLE1

Descriptive

Statisticsan

dCorrelation

sam

ongStudyVariables

a

Variable

Mea

nSD

12

34

56

78

910

1112

1314

1516

1.TMThiring

0.67

1.12

2.New

TMTroles

0.46

0.82

0.70

3.Firm

dev

elop

men

t0.97

0.79

0.23

0.19

4.TMTunderqu

alificationb

0.47

0.92

0.16

0.08

0.06

5.TMTov

erqu

alificationb

2.45

2.27

0.16

0.08

0.53

0.05

6.Com

puterindustry

0.44

0.50

0.04

20.01

20.14

20.01

20.01

7.Man

ufacturingindustry

0.04

0.19

20.11

20.10

20.24

20.05

20.18

20.18

8.Med

ical

industry

0.14

0.34

0.02

0.07

0.20

0.03

0.01

20.35

20.08

9.Resea

rchindustry

0.03

0.17

20.10

20.10

20.22

20.06

20.11

20.16

20.04

20.07

10.S

emicon

ductor

industry

0.13

0.33

20.02

20.01

0.08

0.02

0.10

20.34

20.08

20.15

20.07

11.T

elec

ommunicationsindustry

0.23

0.42

0.04

0.04

0.14

0.01

0.06

20.48

20.11

20.22

20.09

20.21

12.F

irm

ageb

67.14

46.01

20.11

20.16

0.32

20.20

0.32

20.10

20.03

0.00

0.03

0.05

0.09

13.T

eam

size

b4.40

3.35

0.28

0.14

0.59

0.17

0.71

20.09

20.15

0.01

20.12

0.14

0.10

0.36

14.T

eam

tenure

b40

.70

29.18

20.24

20.23

0.01

20.31

0.00

20.09

0.11

20.09

0.22

0.03

0.01

0.71

20.02

15.T

eam

tenure

(SD)b

16.32

17.03

0.06

20.01

0.43

0.00

0.40

20.11

20.16

20.01

20.15

0.12

0.18

0.67

0.47

0.30

16.F

unctional

diversity

b0.63

0.24

0.13

0.07

0.46

20.03

0.57

20.08

20.26

0.01

20.07

0.14

0.13

0.30

0.54

0.05

0.44

17.T

MTex

its

0.54

1.05

0.36

0.10

0.35

0.05

0.43

0.03

20.10

20.02

20.08

0.01

0.05

0.25

0.55

0.02

0.31

0.29

aN

517

,659

observationsfrom

167firm

s.Correlation

s.

|.01|

aresign

ifican

tatp

,.05.

bLagge

dby

12mon

ths.


TABLE 2Results of Panel–Poisson Regression Analysis for TMT Hiring and New TMT Rolesa

TMT Hiring New TMT Roles

Variables Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

Computer industry 20.15 20.13 20.13 20.29 20.31† 20.31†

(0.18) (0.18) (0.18) (0.18) (0.17) (0.17)Manufacturing industry 21.91*** 21.81*** 21.84*** 21.87*** 21.73*** 21.80***

(0.40) (0.39) (0.40) (0.40) (0.39) (0.39)Medical industry 20.07 20.13 20.13 0.17 0.11 0.13

(0.23) (0.23) (0.23) (0.23) (0.22) (0.23)Research industry 22.48*** 22.37*** 22.41*** 24.39*** 24.20*** 24.23***

(0.54) (0.54) (0.54) (0.84) (0.84) (0.84)Semiconductor industry 20.08 20.10 20.08 20.05 20.03 20.02

(0.25) (0.25) (0.25) (0.25) (0.25) (0.25)Firm ageb 20.01*** 20.01*** 20.01*** 20.01*** 20.01*** 20.01***

(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)Team sizeb 0.02*** 0.02*** 0.02*** 20.00 20.01 20.01

(0.00) (0.01) (0.01) (0.01) (0.01) (0.01)Team tenureb 20.00* 20.00 20.00 0.01*** 0.00† 0.00*

(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)Team tenure (SD) b 0.02*** 0.02*** 0.01*** 20.00 0.00 0.00

(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)Blau’s indexb 0.15* 0.22*** 0.19** 0.08 20.11 20.03

(0.06) (0.06) (0.07) (0.07) (0.08) (0.08)TMT exits 0.19*** 0.20*** 0.21*** 20.12*** 20.14*** 20.14***

(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)TMT hiring — — — 0.51*** 0.51*** 0.51***

(0.01) (0.01) (0.01)New TMT Roles 0.58*** 0.58*** 0.58*** — — —

(0.01) (0.01) (0.01)Firm development 0.16*** 0.08* 0.10** 0.04

(0.03) (0.03) (0.03) (0.04)TMT underqualificationb 0.07*** 20.07*** 20.13*** 20.13***

(0.01) (0.02) (0.01) (0.01)TMT overqualificationb 20.03*** 20.02** 0.03*** 20.01

(0.01) (0.01) (0.01) (0.02)TMT underqualificationb 3 0.11*** —

Firm development (0.01)TMT overqualificationb 3 Firm — 0.03*

development (0.01)Constant 20.81*** 20.99*** 20.87*** 20.93*** 20.78*** 20.75***

(0.15) (0.15) (0.16) (0.15) (0.15) (0.15)Wald x2 6065.91*** 6115.83*** 6138.58*** 4464.54*** 4619.61*** 4627.85***Log likelihood 214088.77 214039.42 213999.09 211868.08 211804.53 211801.76Likelihood-ratio test 98.70*** 80.65*** 127.10*** 5.55*Observations 17,659 17,659 17,659 17,659 17,659 17,659Number of firms 167 167 167 167 167 167

a Standard errors in parentheses.b Lagged by 12 months.

***p , .001**p , .01*p , .05†p , .10


longer to hire a newcomer, but add a new role about3 months sooner.

Hypothesis 5 suggested that firm developmentmoderates the effect of TMT underqualification onTMT hiring. Model 3 shows that the coefficient onthis interaction term is positive, indicating that theslope of the effect of TMT underqualification on thenumber of new TMT hires becomes more positive asfirm development increases. Figure 2 shows the ef-fect of TMT underqualification on the expectednumber of new TMT hires for each number of de-velopmentalmilestones the firmhas achieved.Usinga procedure described by Dawson (2014) to evalu-ate simple slopes for interactions with nonlinearmodels, we discovered that, when firms have notachieved any developmental milestones, the effectof TMT underqualification on new TMT hires isnegative (b 5 2.07, p 5 .00). When the number ofmilestones equals one or two, the effect is positive(b 5 .04, p 5 .00; b 5 .15, p 5 .00). This supportsHypothesis 5.

Practically, the expected number of new TMThires increases with greater TMT underqualificationand the achievement of more milestones, and, whenfirms have not achieved anymilestones, they are notlikely to hire as a result of underqualification. Thislatter effect suggests that the support for Hypothesis1 is driven by those firms that have reached de-velopmentalmilestones. For example, atmean levelsof TMT underqualification, new ventures that havenot achieved any milestones are expected to hirea newcomer in approximately 3.1 years, whereasnew ventures that have achieved two developmentalmilestones hire about 9 months sooner. At highlevels of TMT underqualification (i.e., two standarddeviations above the mean), new ventures that havenot achieved any milestones are expected to hirea newcomer in approximately 3.6 years, whereasnewventures that have achieved twomilestoneshireabout 1.8 years sooner.

Hypothesis 6 suggested that firm developmentmoderates the relationship between TMT over-qualification and new TMT roles. Model 6 revealsthat the coefficient on the interaction term is positiveand significant, indicating that the slope of the effectof TMT overqualification on the number of newroles becomes more positive as firms pass more de-velopmental milestones. Figure 3 shows the effectsof TMT overqualification on the expected number ofnew TMT roles added for each number of de-velopmental milestones achieved. Simple slopeanalysis shows that the relationship between TMToverqualification and new TMT roles is negative

when the number of developmental milestonesequals zero (b 5 2.02, p 5 .00). However, this re-lationship is not significant when the firm hasachieved one developmental milestone (b5 .01, p5.28). Further, this relationship is positive when thefirm has achieved two developmental milestones(b 5 .04, p 5 .00). These results are consistent withHypothesis 6.

Practically, the expected number of new TMTroles increases with greater TMT overqualificationand the achievement of both venture capital financ-ing and initial public offering, and, when firmshave not achieved either of these milestones, theyare not likely to create new roles as a result of over-qualification. This latter effect suggests that thesupport for Hypothesis 4 is driven by those firms thathave reached developmental milestones. For exam-ple, at mean levels of TMT overqualification, newventures that have not achieved any milestones areexpected to add a new TMT role in approximately4.4 years, whereas new ventures that have achievedtwo developmental milestones add a new role about10 months sooner. At high levels of TMT over-qualification, new ventures that have not achievedany milestones are expected to add a new TMT rolein 4.8 years, whereas new ventures that have achievedtwo developmental milestones add a new role about1.8 years sooner.

Robustness Checks

We performed a number of additional analyses toverify the robustness of our results. First, to un-derstand how long misfit affects our outcomes, wetested alternative lag structures for the TMT misfitand control variables: an 18-month lag for TMTmisfit, to predict changes to TMT hiring and roles inthe following18months, anda24-month lag forTMTmisfit, predicting changes to TMT hiring and rolesover the following 24 months. Our results remainconsistent with these lags, with two exceptions.First, when we adopt the 18-month lag structure, theinteraction effect of TMT overqualification and firmdevelopment on new TMT roles in Model 6 is notsignificant (b 5 2.01, p 5 .19) Second, when weadopt either of these longer lag structures, the posi-tive main effect of TMT overqualification on newTMT roles inModel 5 is no longer significant. Again,however, this main effect is qualified by the signifi-cant interaction between TMT overqualification andfirm development on the number of new TMT rolesin the 24-month lag model, which is consistent withour results as currently presented.


To consider the possibility that some hiring androle additions may be created in anticipation of anupcomingmilestone, as “window dressing,”we alsoadopted a different event window with respect toour firm development variable. Specifically, wecreated a variable called “milestone anticipation”that equaled one for the 12-month period prior to thereceipt of venture capital financing or initial publicoffering (i.e., the 12-month period prior to the pointat which our firm development variable became oneinstead of zero). We found that this variable wasnegatively related to hiring (b 5 2.20, p 5 .01) andnew TMT roles (b 5 2.26, p 5 .00), indicating thatfirms were less likely to hire or add roles in the yearprior to achieving a developmentalmilestone than atother times. Moreover, including the milestone an-ticipation variable along with the firm developmentvariable in our current models did not change any ofthe signs or significance levels for the hypothesizedresults reported in Table 2. This suggests that, whilefirmsmay engage in other forms of window dressingin anticipation of a financing event, such as addingexecutives or directors with specific prestige (Chenet al., 2008), they do not substantially alter the roles

or composition of the TMT more generally. This isconsistentwith our finding that firms need resourcesto make changes to the TMT.

Second, we considered several alternative speci-fications of our TMTmisfit variables. Because misfitis not typically conceptualized at a collective teamlevel, we first considered an alternativemeasure thattreated misfit as position specific to account for theidea that the excess qualifications of one managermay not benefit those in another role. The pattern ofresults was identical with respect to the signs andsignificance levels of the coefficients for our pre-dictor variables. We also considered the possibilitythat on-the-job training for executives might takelonger than one year when updating the functionalexperience calculations for our misfit variables.Therefore, we adopted a three-year on-the-job train-ing rule of thumb (see Conger &Nadler, 2012), whichresulted in similar resultswith the exception that theinteraction between TMT overqualification and firmdevelopment on new TMT roles becomes insignifi-cant (b 5 .01, p 5 .49). Finally, we considered thepossibility that calculating functional experiencebasedon thepast threepositionsmayoveremphasize

FIGURE 2The Effects of TMT Underqualification and Firm Development on Predicted Number of New TMT Hires

0.8

0.6

0.4

0.2

00 1 2 3

TMT Underqualification

Pre

dic

ted

New

TM

T H

ires

Development = 0 Development = 1 Development = 2


distant past experience. As such, we calculatedfunctional experience using three alternatives: themost recent position only, the past three executive-level positions (vs. including experience gained innon-executivepositions), and themost recentexecutive-level position only. Across all of the models usingthe TMT misfit variables calculated under thesealternative assumptions, we found that the signsand significance levels of the coefficients for ourpredictor variables remain unchanged from thosepresented in Table 2.

Next, we ran Models 3 and 6 with the addition ofthe interaction terms that we did not hypothesizea priori. Specifically, we included the interactionbetween TMT overqualification and firm develop-ment when predicting new TMT hires and the in-teraction between TMT underqualification and firmdevelopment when predicting new TMT roles. Nei-ther of these interactionswas statistically significant,and their inclusion in the models did not changethe sign or significance levels of the hypothesizedeffects.

Finally, we examined two alternative modelspecifications. First, we estimated fixed effects

panel-Poisson models to further take into accounttime-invariant characteristics of firms. The resultsare consistent with the random effects modelspresented in Table 2, both in terms of the signs andsignificance levels of the coefficients of the pre-dictor variables. Second, we considered the possi-bility of excess zeros in the dependent variables.Because we have panel data, we could not use zero-inflated Poisson. Instead, we ran two-stage Poissonrandom effect hurdle models to predict the excesszeros and then the count of new TMT roles andhires (see Min & Agresti, 2005). First, we recodedthese dependent variables into dummy variablesand ran a panel-logit model to simulate runningthe first part of a zero-inflated Poisson inflatemodel. Next, we ran panel-Poisson models for thecounts of the dependent variables when they wereequal to one or above. After taking into account theexcess zeros in the panel-logit model, the panel-Poisson count models for dependent variables atone or abovewere largely consistentwith thepanel-Poisson count models that include zero countsreported in Table 2. For example, they showedthat TMT overqualification negatively predicts the

FIGURE 3The Effects of TMT Overqualification and Firm Development on Predicted Number of New TMT Roles

0.8

0.6

0.4

0.2

00 1 2 3 4 5

Development = 0 Development = 1

TMT Overqualification

Pre

dic

ted

New

TM

T R

oles

Development = 2

6 7 8 9


number of new TMT hires (at p 5 .06) and TMTunderqualification positively predicts the numberof new TMTs hires, although the latter main effectcontinues to be qualified by a positive interactionterm for TMT underqualification and firm devel-opment. In addition, TMT overqualification posi-tively predicts the number of newTMT roles (at p5.05) and TMT underqualification negatively pre-dicts the number of new TMT roles. However, theinteraction term between TMT overqualificationand firm development is no longer significant,which is inconsistent with Hypothesis 6.

In summary, these alternativemodels indicate thatour results are generally robust across a variety ofmodel specifications, variable operationalizations,and variable lag structures. The exception is the in-teractive effect of TMT overqualification and firmdevelopment on new TMT roles, which was notsupported in three robustness checks. Moreover, thecoefficients of TMT overqualification were smallerthan those of TMT underqualification. One expla-nation for these results is that the opportunity affor-ded by overqualification is less of a driver of changethan the need indicated by underqualification. Toexplore this empirically, we performed dominanceanalysis using thedominpackage inStata (Luchman,2013), which examines the relative importance oftwo predictor variables for every possible subset ofthe full model in which only one of the two pre-dictors is entered (Azen & Budescu, 2003). The gen-eral and conditional dominance statistics showedthat TMT underqualification reduced the Akaikeinformation criterion fit statistic by a greater amountthan did TMT overqualification, suggesting it is themore critical predictor of TMT changes.

DISCUSSION

We began this paper by noting the balancing actthat new venture TMTs perform between fulfillingimmediate needs and capitalizing on opportuni-ties in a context in which resources are in shortsupply, andwe examined this balancing act for thecomposition and structure of the TMT itself. Wetheorized that TMTmisfit, defined as discrepanciesbetween the qualifications of TMT managers andthe functional roles they fill, would influence hiringand the addition of new roles to the TMT, and thatthese relationships may be contingent upon the re-sources new firms gain by achieving developmentalmilestones.

We found that TMT misfit in the form of under-qualification resulted in increased hiring of newcomers

but not the addition of roles,whereas TMTmisfit in theform of overqualification resulted in the addition ofroles but not increased hiring of newcomers. However,firm development was an important determinant ofwhether TMT misfit resulted in such changes at all.Prior to achieving developmental milestones, firmsmade fewTMTchanges in response tomisfit.Onlyafterpassing developmental milestones did overqualifiedTMTs elaborate their structures by adding new rolesand underqualified TMTs reinforce their capabilitiesby hiring new people.

We also found that TMT underqualification wasa more robust predictor of TMT change than TMToverqualification. TMT underqualification occurswhen a team has fewer capabilities than those iden-tified in the formal role structure, and, as such, rep-resents an underlying need to add skills to achieveits implied goals. Teams characterized by over-qualification, in contrast, are currently fulfilling theimplied goals of their role structure and have excesscapabilities. These teams do not have an underlyingneed, but, rather, an opportunity to better recognizeand profit from the capabilities of their members. Assuch, the effects of TMT misfit appear to be asym-metrical: the need presented by underqualificationmay trump the opportunity presented by over-qualification when predicting TMT evolution.

These findings contribute to theories of entrepre-neurial firms in several ways. First, by modeling themisfit between TMT composition and structure innew ventures, we demonstrate that it is insufficientand possibly misleading to treat composition andstructure as interchangeable or to look at one in-dependently of the other.While a small set of studieshas examined the effects of one or the other inexplaining new venture TMT evolution, ours is thefirst to consider TMT misfit, the intersection of thetwo, in explaining how new venture TMTs evolve.

Second, in examining this interplay betweenmisfit and new venture development, we reveal animportant professionalization paradox in the evolu-tion of new venture TMTs: firms seem unable toreach the level of professionalization needed to at-tract resources without first having obtained thoseresources. Specifically, new venture TMTs often ex-perience mismatches between the composite experi-ences of their top managers and the roles thatthey occupy, and, as such, may need to changeroles or people to reach developmentalmilestones.However, their ability to make such changes maybe constrained until they can obtain the financialand social capital resources that the achievementof developmental milestones provides and which


facilitate further change. This professionalizationparadox may offer an additional explanation of whyit is rare for entrepreneurial firms to ever reach thesemilestones and more generally to succeed. Ourfinding on the lack ofwindowdressing in the form ofadding roles or members in advance of resourcemilestones further highlights this paradox. Unlikeresearch on other forms of window dressing (Chenet al., 2008), our sample includes firms that never getventure capital or go public, and our findings thusraise questions about whether window dressing isa strategy that firms follow or is an artifact of sam-pling successful firms. Once a firm obtains resources,this facilitates team changes (see also Boeker &Wiltbank, 2005), but not all firms are able to obtainresources. Our examination of how developmentalmilestones interact with TMT overqualification andunderqualification to influence TMT evolution thusharkens back to an earlier tradition of research thatemphasized organizational life stages (e.g., Greiner,1972), which established that processes can anddo operate differently at different stages in a firm’sdevelopment.

Further contributions to the entrepreneurship lit-erature include our findings that over- and under-qualification have asymmetrical effects on TMTevolution. Many entrepreneurial studies point to theimportance of thequality of the topmanagers for newventure success (e.g., Burton et al., 2002; Frankeet al., 2008;Hsu, 2007).Our study suggests, however,that overqualifiedTMTsmay fail to take advantage ofthe opportunities their excess capabilities afford. Forexample, because firms do not explicitly signal theirfull capabilities through the role structure, investorsmay underestimate the value of that new venture.Not being able to attract those resources, in turn,maymake it difficult to respond to that misfit.

Next, our study goes beyond previous entre-preneurship studies of growth in the TMT capabil-ities of entrepreneurial firms that count the overallsize of the team. Rather, there are different ways toexpand—by adding people or by adding roles—andthe twohavediffering implications for newventures.Thus, we differentiate whether adding team mem-bers is associated with adding breadth to the TMT inthe form of structural elaboration, orwhether it is theresult of adding capacity to areas of established ex-pertise by hiring additional senior managers. Ourresults suggest that simply adding more managers tothe count may have unintended negative conse-quences when those managers do not have the rightskills: adding excess skills may make future hiringless likely. Nor would it be accurate to treat the

addition of a new person into a new role as identicalto the addition of a new person into an existing role.Manager counts must be considered in the contextof TMT structural needs. Moreover, conceptuallydistinguishing structural elaboration and capabilityreinforcement as alternative approaches to TMTgrowth and professionalization points to avenues forfuture research and provides an overarching frame-work for understanding the literature that has ex-amined the emergence, diffusion, and performanceimplications of specific TMT roles such as the chiefoperating officer, chief financial officer, or VP ofhuman resources (Hambrick & Cannella, 2004;Welbourne & Cyr, 1999; Zorn, 2004).

Beyond informing literature on entrepreneurship,these results contribute more generally to researchon upper echelons. While upper echelons scholarsrecognize that there are costs tomismatches betweenthe experiences of top executives and structures inwhich they work (e.g., Burton & Beckman, 2007;Chen & Hambrick, 2012), to date there has been verylittle attention paid to the phenomenon of misfit atthe level of theTMT.This study examines a setting inwhich misfit is likely to occur in top teams (i.e., newventures), and shows that misfit acts both as a cata-lyst to change and a source of inertia in TMT struc-ture and composition. It ultimately answers callsto more fully explain the characteristics of TMTs(Finkelstein,Hambrick,&Cannella, 2009;Hambrick,2007) by examining these types of teams in theirearly stages.

Inaddition,wecontribute to theoriesof rolechange.Though there is substantial evidence of inertia invarious aspects of organizational and role struc-tures, eveninentrepreneurialorganizations(e.g.,Baronet al., 2001; Beckman & Burton, 2008; Burton &Beckman, 2007), it is also evident that these struc-tures must and do change over time. However, weknow relatively little about how and why those rolestructure changes happen (Cohen, 2013). Whenprevious research has explored change in roles, it ismost often as a response to individual motivations(e.g., Wrzesniewski & Dutton, 2001) or externalpressures (e.g., Zorn, 2004). However, we find thatstructural role elaboration also results from an in-ternal structural team factor, TMT overqualification,in which top managers have functional experiencesbeyond those required by the roles within the TMT.

Finally, we build on the person–job misfit liter-aturewith twomethodological shifts. Insteadof focusingon individual responses to misfit (e.g., dissatisfactionand turnover) (Kalleberg, 2008; Kristof-Brown et al.,2005),weexplorehowmisfit at a team level influences


changes to team composition and team structure. Assuch, this study contributes to a growing body of lit-erature on team-level conceptualizations of fit (e.g.,DeRue & Hollenbeck, 2007; Seong & Choi, 2014). Wealso investigate two distinct forms of misfit: over-qualification and underqualification. Person–jobmisfit literature has generally not distinguishedbetween these (Erdogan et al., 2011); however, do-ing so allows us to be more precise in explaininghow misfit alters role structures and people withinthe TMT.

Limitations and Future Directions

While this paper makes several unique contribu-tions, it has some limitations. First, it is difficult toknow from this studywhether the net effects ofmisfitare good or bad for the firm and its topmanagers. Forinstance, the addition of new roles can be seen asa form of individual reward (Baron & Bielby, 1986;Baron & Pfeffer, 1994) and as a way of taking ad-vantage of unforeseen organizational opportunities(Miner, 1987; Miner & Estler, 1985), but is also anadditional form of turbulence for already-turbulentnew ventures. Similarly, hiringmay bring additionalcapabilities and resources such as prestige into newfirms, but the costliness and potential imperma-nence of some hires may not translate into long-termgains in these areas (Chen et al., 2008). Future studiesshould thus incorporate measures of firm perfor-mance to better understand whether changes to theTMT as a result of misfit are ultimately beneficial ordetrimental as new ventures develop.

Next, though we observe patterns of change inTMTs due to misfit, it is unclear whether thesechanges to people and roles are intentional strategieson the part of top managers or more serendipitousprocesses. For example, elaborating the TMT struc-ture by adding new roles can be due to a deliberateeffort to recruit or retain specific individuals, or itcan evolve more organically around perceptionsof managerial expertise (Bunderson & Sutcliffe,2002; Miner, 1987). The results of this study revealthe overarching effects of misfit on changes to theTMT, but do not isolate the influence of the dif-ferent mechanisms that explain these effects (e.g.,need vs. opportunity vs. resources). Therefore, fu-ture studies might assess top managers’ percep-tions of TMT misfit to better understand whetherthey are recognizing and responding to misfit orwhether these changes are occurring in lieu ofa deliberate strategy to bring roles and people backinto alignment.

We differentiated between overqualification andunderqualification, but future work might exploredifferences even within these types of misfit. Forinstance, the effects of overqualificationonnewrolesmight differ between a TMT that has no marketingrole and people with marketing backgrounds anda TMT that has one marketing role and two peoplewith marketing experience. Would both be equallylikely to add new roles? We also considered TMTmisfit without respect to industry expectations ofa new venture’s composition and structure. Futureresearch could model fit using this external lens.

We found that TMT misfit only invites responseswhen new firms have achieved developmental mile-stones. Future research could examine factors thatstifle TMT advances even when firms have the re-sources associated with developmental milestones.For example, powerful founder CEOs (e.g., thosewithlarge ownership control) may resist TMT change (seeBoeker, 1992; Fischer & Pollock, 2004). In analysis notshown here, we examined the effects of three-wayinteractions between TMT misfit, firm development,and a dummy variable indicating that the CEO wasa founder on new TMT hires and roles. We found ev-idence consistent with the idea that a founder CEOmight constrain hiring in response to TMT under-qualification; however,moreprecisemeasuresofCEOpower should be used to provide definitive tests of thisidea. Further, the inability of firms to alter their com-position and structure without first achieving firmdevelopmentalmilestonesmaysuggest thereareotherfactors endogenous to the firm that drive both misfitand TMTchange.We see the impact ofmisfit on TMThiringandaddingnewroles, but it is impossible to ruleout that some other factor is driving both effects. Re-lated to this, our sample of firms has high rates ofobtaining venture capital and going public, higherthan one might expect with a sample of firms ina broader time period or geographic region. Futureresearch should consider firms inother regionswheremore firms fail to reach these milestones.

Although we find effects of TMT misfit on newhires and roles, the base rates of these events aresmall. For example, TMT underqualification increasesthe rate of new hires by 7%, but most firms in oursample (about 65%) do not hire anyone in a givenyear. However, the fact that minimally altering theindependent variable (e.g., a TMT is misfit by onemissing qualification) results in changes to the de-pendent variable (e.g., hiring), and that these effectscumulate over time, suggests that these results, ifnot numerically large, are theoretically important(Abelson, 1985; Prentice & Miller, 1992). Finally,


we examined a relatively small sample of firmsfrom a narrow range of industry sectors, in a par-ticular region, and at a moment in time of tremen-dous growth and economic success. Our resultsmay not generalize and future scholars must test theboundary conditions of our ideas.

Managerial Implications

This paper highlights the importance of misfit innew venture TMTs as a predictor of hiring and roleelaboration, particularly for firms that have achieveddevelopmental milestones. In these firms, TMT over-qualification may reduce the need to hire additionalmanagers to increase the capabilities of the TMT;however, these firms could add roles that betterincorporate top managers’ prior experiences tomore accurately signal the quality of the TMT.Thus, top managers should be flexible with regardto roles and the structure of work, particularlywhen there is overqualification among managersrelative to the current role structure. In addition,TMT underqualification is likely to result in morestability to role structures yet more change in thepeople who move into these roles. Specifically,newcomersmay need to be hired to support gaps inskills relative to these existing roles. While theseactions may represent strategic choices in teamcomposition, if the roles available within the teamare fixed (e.g., new roles are not added), this mayconstrain future rewards and opportunities. Inshort, this may be akin to a promotion paradox(Phillips, 2001), albeit at a different level of anal-ysis, where TMT underqualification provides short-term opportunities for new hires but perhaps lesslong-term development of TMT structures. In addi-tion, our results point to the importance of the de-velopmental stage: firms are only able to change theTMT, regardless of the level of misfit, after achievingsome developmental milestones. This points to theimportance of hiring people and creating roles witha good fit from the beginning, rather that hoping togrow into roles or add people over time.

CONCLUSION

In sum, this study reveals that TMT misfit actsas a catalyst for changes to roles and peoplewithinnew venture TMTs, but that its effects are de-pendent upon the type of misfit and the level offirm development. Once new ventures have achieveddevelopmental milestones, overqualified TMTselaborate their structural roleswhereas underqualified

TMTs hire new people. Prior to the achievement ofdevelopmental milestones, misfit TMTs are lesslikely to make these changes despite the need to doso. Ultimately, this study provides a more nu-anced understanding of the characteristics of en-trepreneurial TMTs by attending to phenomenathat exist at the intersection of composition andstructure.

REFERENCES

Abelson, R. P. 1985. A variance explanation paradox:When a little is a lot. Psychological Bulletin, 97:129–134.

Allison, P. D. 2009. Fixed effects regression models. LosAngeles, CA: Sage.

Azen, R., & Budescu, D. V. 2003. The dominance analysisapproach for comparing predictors in multiple re-gression. Psychological Methods, 8: 129–148.

Baker, T., & Nelson, R. E. 2005. Creating something fromnothing: Resource construction through entrepre-neurial bricolage.Administrative ScienceQuarterly,50: 329–366.

Baron, J. N., & Bielby, W. T. 1986. The proliferation ofjob titles in organizations. Administrative ScienceQuarterly, 31: 561–586.

Baron, J. N., Burton, M. D., & Hannan, M. T. 1999. Engi-neering bureaucracy: The genesis of formal policies,positions, and structures in high-technology firms. Jour-nal of Law Economics and Organization, 15: 1–41.

Baron, J. N., Hannan,M. T., & Burton,M. D. 1999. Buildingthe iron cage: Determinants of managerial intensity inthe early years of organizations. American Sociolog-ical Review, 64: 527–547.

Baron, J. N., Hannan, M. T., & Burton, M. D. 2001. Laborpains: Change in organizationalmodels and employeeturnover in young, high-tech firms.American Journalof Sociology, 106: 960–1012.

Baron, J. N., & Pfeffer, J. 1994. The social psychology oforganizations and inequality. Social PsychologyQuarterly, 57: 190–209.

Baum, J. A. C., & Silverman, B. S. 2004. Picking winners orbuilding them? Alliance, intellectual, and humancapital as selection criteria in venture financing andperformance of biotechnology startups. Journal ofBusiness Venturing, 19: 411–436.

Bechky, B. A. 2006. Gaffers, gofers, and grips: Role-basedcoordination in temporary organizations. Organiza-tion Science, 17: 3–21.

Beckman, C. M. 2006. The influence of founding teamcompany affiliations on firm behavior. Academy ofManagement Journal, 49: 741–758.


Beckman,C.M.,&Burton,M.D. 2008. Founding the future:Path dependence in the evolution of top managementteams from founding to IPO. Organization Science,19: 3–24.

Beckman, C. M., & Burton, M. D. 2011. Bringing organiza-tional demography back in: Time, change, and struc-ture in topmanagement teamresearch. InM.A.Carpenter(Ed.), Handbook of top management team research:49–70. Cheltenham, England: Edward Elgar.

Beckman, C. M., Burton, M. D., & O’Reilly, C. 2007. Earlyteams: The impact of team demography on VC fi-nancing and going public. Journal of Business Ven-turing, 22: 147–173.

Bidwell, M., & Briscoe, F. 2010. The dynamics of in-terorganizational careers. Organization Science, 21:1034–1053.

Blau, P. 1977. Inequality and heterogeneity: A primitivetheory of social structure. New York, NY: Free Press.

Boeker, W. 1992. Power and managerial dismissal: Scape-goating at the top. Administrative Science Quarterly,37: 400–421.

Boeker, W. 1997. Strategic change: The influence of man-agerial characteristics and organizational growth.Academy of Management Journal, 40: 152–170.

Boeker, W., & Karichalil, R. 2002. Entrepreneurial transi-tions: Factors influencing founder departure. Acad-emy of Management Journal, 45: 818–826.

Boeker, W., & Wiltbank, R. 2005. New venture evolutionand managerial capabilities. Organization Science,16: 123–133.

Bunderson, J. S., & Sutcliffe, K. M. 2002. Comparing al-ternative conceptualizations of functional diversity inmanagement teams: Process and performance effects.Academy of Management Journal, 45: 875–893.

Burton, M. D., & Beckman, C. M. 2007. Leaving a legacy:Position imprints and successor turnover in youngfirms. American Sociological Review, 72: 239–266.

Burton, M. D., Sørensen, J. B., & Beckman, C. M. 2002.Coming from good stock: Career histories and newventure formation. In M. Lounsbury & M. J. Ventresca(Eds.), Research in the sociology of organizations:Social structure and organizations revisited, vol. 19:229–262. Bingley, England: Emerald Group Publish-ing Limited.

Bygrave, W. D., & Timmons, J. A. 1992. Venture capital atthe crossroads. Boston, MA: Harvard Business Press.

Cappelli, P. 2008. Talent management for the twenty-firstcentury. Harvard Business Review, 86: 74–81.

Chan, D. 1998. Functional relations among constructs inthe samecontentdomain atdifferent levels of analysis:A typology of composition models. The Journal ofApplied Psychology, 83: 234–246.

Chatman, J. A. 1991. Matching people and organizations:Selection and socialization in public accounting firms.Administrative Science Quarterly, 36: 459–484.

Chen, G., & Hambrick, D. C. 2012. CEO replacement in turn-aroundsituations: Executive (mis)fit and itsperformanceimplications.Organization Science, 23: 225–243.

Chen, G., Hambrick, D. C., & Pollock, T. G. 2008. Puttin’ onthe Ritz: Pre-IPO enlistment of prestigious affiliates asdeadline-induced remediation. Academy of Man-agement Journal, 51: 954–975.

Cohen, L. E. 2013. Assembling jobs: A model of how tasksare bundled into and across jobs. Organization Sci-ence, 24: 432–454.

Cohen,L.E.,&Broschak, J. P. 2013.Whose jobsare these?Theimpact of the proportion of female managers on thenumberofnewmanagement jobs filledbywomenversusmen.Administrative Science Quarterly, 58: 509–541.

Conger, J. A., & Nadler, D. A. 2012. When CEOs step up tofail. MIT Sloan Management Review, 45: 50–56.

Crossland, C., Zyung, J., Hiller, N., & Hambrick, D. 2014.CEO career variety: Effects on firm-level strategic andsocial novelty.AcademyofManagement Journal, 57:652–674.

Davila, A., Foster, G., & Gupta, M. 2003. Venture capitalfinancing and the growth of startup firms. Journal ofBusiness Venturing, 18: 689–708.

Dawson, J. F. 2014. Moderation in management research:What, why, when, and how. Journal of Business andPsychology, 29: 1–19.

Dencker, J. C., Gruber, M., & Shah, S. K. 2009a. Individualand opportunity factors influencing job creation innew firms. Academy of Management Journal, 52:1125–1147.

Dencker, J. C., Gruber, M., & Shah, S. K. 2009b. Pre-entryknowledge, learning, and the survival of new firms.Organization Science, 20: 516–537.

DeRue, D. S., & Hollenbeck, J. R. 2007. The search for in-ternal and external fit in teams. In C. L. Ostroff &T. A. Judge (Eds.),Perspectives on organizational fit:259–285. Mahwah, NJ: Lawrence Erlbaum.

Edwards, J. R. 1991. Person–job fit: A conceptual integra-tion, literature review, andmethodological critique. InC. L. Cooper & I. T. Robertson (Eds.), Internationalreview of industrial and organizational psychology,vol. 6: 283–357. New York, NY: John Wiley & Sons.

Eisenhardt, K. M., & Schoonhoven, C. B. 1990. Organiza-tional growth: Linking founding team, strategy, envi-ronment, and growth among U.S. semiconductorventures, 1978–1988. Administrative Science Quar-terly, 35: 504–529.

Erdogan, B., Bauer, T. N., Peiro, J. M., & Truxillo, D. M.2011. Overqualified employees: Making the best of


a potentially bad situation for individuals and orga-nizations. Industrial and Organizational Psychology:Perspectives on Science and Practice, 4: 215–232.

Finkelstein, S., Hambrick, D. C., & Cannella, A. A., Jr. 2009.Strategic leadership: Theory and research on ex-ecutives, top management teams, and boards. Ox-ford, England: Oxford University Press.

Fischer, H. M., & Pollock, T. G. 2004. Effects of socialcapital and power on surviving transformationalchange: The case of initial public offerings. Academyof Management Journal, 47: 463–481.

Franke, N., Gruber, M., Harhoff, D., & Henkel, J. 2008.Venture capitalists’ evaluations of start-up teams:Trade-offs, knock-out criteria, and the impact of VCexperience. Entrepreneurship Theory and Practice,32: 459–483.

Gilbert, B. A., McDougall, P. P., & Audretsch, D. B. 2006.Newventure growth: A review and extension. Journalof Management, 32: 926–950.

Gompers, P., & Lerner, J. 2001. The venture capital revo-lution. The Journal of Economic Perspectives, 15:145–168.

Greiner, L. E. 1972. Evolution and revolution as organiza-tions grow. Harvard Business Review, 50: 37–46.

Groysberg, B., & Lee, L.-E. 2009. Hiring stars and theircolleagues: Exploration and exploitation in professionalservice firms.Organization Science, 20: 740–758.

Guler, I. 2007. Throwing good money after bad? Politicaland institutional influences on sequential decisionmaking in the venture capital industry. Administra-tive Science Quarterly, 52: 248–285.

Hair, J. F., Jr., Anderson, R. E., Tatham, R. L., & Black,W. C.1995.Multivariate data analysis (3rd ed.). NewYork,NY: Macmillan Publishing Company.

Hambrick, D. C. 2007. Upper echelons theory: An update.Academy of Management Review, 32: 334–343.

Hambrick, D., C., & Cannella, A. A., Jr. 2004. CEOs whohave COOs: Contingency analysis of an unexploredstructural form. Strategic Management Journal, 25:959–979.

Hambrick, D. C., & Crozier, L.M. 1985. Stumblers and starsin the management of rapid growth. Journal of Busi-ness Venturing, 1: 31–45.

Hellmann, T. 2000. Venture capitalists: The coaches ofSiliconValley. InW.Miller, C.-M. Lee,M.G.Hanock&H. Rowen (Eds.), The Silicon Valley edge: A habitatfor innovation and entrepreneurship: 276–294.Stanford, CA: Stanford University Press.

Hellmann, T., & Puri, M. 2002. Venture capital and theprofessionalization of start-up firms: Empirical evi-dence. The Journal of Finance, 57: 169–197.

Hersch, J. 1995. Optimal “mismatch” and promotions. Eco-nomic Inquiry, 33: 611–624.

Hsu, D. H. 2007. Experienced entrepreneurial founders,organizational capital, and venture capital funding.Research Policy, 36: 722–741.

Kalleberg, A. L. 2008. The mismatched worker: Whenpeople don’t fit their jobs. The Academy of Man-agement Perspectives, 22: 24–40.

Khaire, M. 2010. Young and no money? Never mind: Thematerial impact of social resources on new venturegrowth. Organization Science, 21: 168–185.

Kristof-Brown, A. L., Zimmerman, R. D., & Johnson, E. C.2005. Consequences of individuals’ fit at work: Ameta-analysis of person–job, person–organization,person–group, and person–supervisor fit. PersonnelPsychology, 58: 281–342.

Levesque, L. L. 2005. Opportunistic hiring and employeefit. Human Resource Management, 44: 301–317.

Luchman, J. N. 2013. DOMIN: Stata module to conductdominance analysis [Computer software]. ChestnutHill, MA: Boston College, Department of Economics.Retrieved January26, 2015, fromhttps://ideas.repec.org/c/boc/bocode/s457629.html.

Menz, M. 2012. Functional top management team mem-bers: A review, synthesis, and research agenda. Jour-nal of Management, 38: 45–80.

Min, Y., & Agresti, A. 2005. Random effect models for re-peated measures of zero-inflated count data. Statisti-cal Modelling, 5: 1–19.

Miner, A. S. 1987. Idiosyncratic jobs in formalized organiza-tions.Administrative Science Quarterly, 32: 327–351.

Miner, A. S., & Estler, S. E. 1985. Accrual mobility: Jobmobility in higher education through responsibil-ity accrual. The Journal of Higher Education, 56:121–143.

National Center forO*NETDevelopment. 2011.Summaryreport for 11-1021.00—General and operationsmanagers. RetrievedNovember 14, 2011, fromhttp://www.onetonline.org/link/summary/11-1021.00.

Neter, J., Wasserman, W., & Kutner, M. H. 1996. Appliedlinear regression models (3rd ed.). Chicago, IL: R. D.Irwin.

Pettigrew, A. 1992. On studying managerial elites. Stra-tegic Management Journal, 13: 163–182.

Phillips, D. J. 2001. The promotion paradox: Organiza-tional mortality and employee promotion chances inSilicon Valley law firms, 1946–1996. American Jour-nal of Sociology, 106: 1058–1098.

Prentice, D. A., &Miller, D. T. 1992.When small effects areimpressive. Psychological Bulletin, 112: 160–164.


https://ideas.repec.org/c/boc/bocode/s457629.html

https://ideas.repec.org/c/boc/bocode/s457629.html

http://www.onetonline.org/link/summary/11-1021.00

http://www.onetonline.org/link/summary/11-1021.00

Rubenson, G. C., & Gupta, A. K. 1996. The initial succes-sion: A contingency model of founder tenure. Entre-preneurship Theory and Practice, 21: 21–35.

Seong, J. Y., & Choi, J. N. 2014. Effects of group-level fit ongroup conflict and performance: The initiating role ofleader positive affect. Group & Organization Man-agement, 39: 190–212.

Sine,W.D.,Mitsuhashi,H., &Kirsch,D.A. 2006.RevisitingBurns and Stalker: Formal structure and new ventureperformance in emerging economic sectors.Academyof Management Journal, 49: 121–132.

Somaya, D., Williamson, I. O., & Lorinkova, N. 2008. Gonebut not lost: The different performance impacts of em-ployee mobility between cooperators versus competi-tors. Academy of Management Journal, 51: 936–953.

Stinchcombe, A. 1965. Social structure and organizations.In J. G. March (Ed.), The handbook of organizations:260–290. Chicago, IL: Rand McNally.

Strang, D., & Baron, J. N. 1990. Categorical imperatives:The structure of job titles in California state agencies.American Sociological Review, 55: 479–495.

Wasserman, N. 2003. Founder–CEO succession and theparadox of entrepreneurial success. OrganizationScience, 14: 149–172.

Wasserman, N. 2006. Stewards, agents, and the founderdiscount: Executive compensation in new ventures.Academy of Management Journal, 49: 960–976.

Welbourne, T. M., & Cyr, L. A. 1999. The human resourceexecutive effect in initial public offering firms.Academy of Management Journal, 42: 616–629.

Wrzesniewski, A., & Dutton, J. E. 2001. Crafting a job:Revisioning employees as active crafters of theirwork.Academy of Management Review, 26: 179–201.

Zimmerman, M. A., & Zeitz, G. J. 2002. Beyond survival:Achieving new venture growth by building legitimacy.Academy of Management Review, 27: 414–431.

Zorn, D.M. 2004. Here a chief, there a chief: The rise of theCFO in the American firm. American SociologicalReview, 69: 345–364.

Zott, C., & Huy, Q. N. 2007. How entrepreneurs use sym-bolic management to acquire resources. Administra-tive Science Quarterly, 52: 70–105.

Amanda J. Ferguson ([email protected]) is anassistant professor of management at Northern Illinois Uni-versity. She received her PhD in organizational behavior fromLondon Business School. Her research focuses on team com-position, intergroup relationships, and team effectiveness.

Lisa E. Cohen ([email protected]) is an associate pro-fessor of organizational behavior at McGill University’sDesautels Faculty ofManagement. She receivedher PhD inorganizational behavior from the University of California,Berkeley. Her research examines the allocation of tasksinto and across jobs and of jobs into organizational andopportunity structures.

M. Diane Burton ([email protected]) is an associateprofessor of human resource management at Cornell’s ILRSchool with a courtesy appointment in the Department ofSociology. She received her PhD fromStanfordUniversity.She is an organizational sociologist interested in innova-tion, entrepreneurship, and employment systems.

Christine M. Beckman ([email protected]) is a pro-fessor of management and organization at University ofMaryland’s Robert H. Smith School of Business. She re-ceived her PhD in organizational behavior from StanfordUniversity. Her research has focused on organizationallearning, interorganizational networks, and entrepreneur-ship, particularly on how collaborative relationships facili-tate organizational change.


mailto:[email protected]




Copyright of Academy of Management Journal is the property of Academy of Managementand its content may not be copied or emailed to multiple sites or posted to a listserv withoutthe copyright holder's express written permission. However, users may print, download, oremail articles for individual use.

misfit and milestones: structural elaboration and

Documents