re-examining strategic flexibility: a meta-analysis of its

British Journal of Management, Vol. 00, 1–21 (2020)DOI: 10.1111/1467-8551.12413

Re-examining Strategic Flexibility: AMeta-Analysis of its Antecedents,Consequences and Contingencies

Dennis Herhausen, Robert E. Morgan ,1,2 Danilo Brozovic 3

and Henk W. Volberda 4

Department of Marketing, KEDGE Business School, Domaine de Luminy, Marseille 13009, France,1Department of Marketing & Strategy, Cardiff Business School, Cardiff University, Colum Drive, Cardiff

CF10 3EU, UK, 2Copenhagen Business School, Solbjerg Plads 3, Frederiksberg, DK -2000, Denmark, 3Schoolof Business, University of Skövde, Högskolevägen 8, Skövde, SE -541 28, Sweden, and 4Department ofStrategic Management & Innovation, Amsterdam Business School, University of Amsterdam, Plantage

Muidergracht 12, Amsterdam 1018 TV, The NetherlandsCorresponding author email: [email protected]

Strategic !exibility (SF) is a concept that has evolved from strategy through other disci-plines, including management, marketing, innovation, entrepreneurship and operations.However, despite attempts to consolidate the domain of SF, there remain theoretical andempirical tensions underlying its antecedents, the consequences and contingencies. Basedon 106 independent samples reported in 98 different studies (n = 26,940 "rms), we pro-vide a meta-analytical examination of these tensions. We highlight and resolve severaldisagreements regarding the enablers, inhibitors and triggers of SF, and we reveal an ad-justed mean performance effect of 0.24. We further "nd that the measurement of SF,as well as some, but not all, dimensions of the environment, moderate the performanceeffect. Finally, an explorative analysis reveals that innovation outcomes and market out-comes mediate the positive relationship between SF and "nancial outcomes, in additionto a negative direct effect. These insights provide a comprehensive and coherent under-standing of the nomological network of SF and a stronger basis for further theorizing andconducting empirical research. Moreover, our "ndings help "rms to re"ne their strategyby implementing the right enablers that drive SF and to understand how and when theirinvestment in SF pays off.

Introduction

Over the last several decades, organizationaladaptation has become a focal topic in manage-ment research because of increasing irregularity,complexity, uncertainty and dynamism in mostmarkets and competitive environments (Posen andLevinthal, 2012; Stieglitz, Knudsen and Becker,2016). Many strategists and management schol-ars alike assert that strategic !exibility (SF) isfundamental to organizational adaptation. SF isde"ned as a "rm’s ability to be proactive or re-spond quickly to changing conditions, with a wide

variety of different internal and external options(Sanchez, 1995; Volberda, 1996).1 Consequently,researchers have a sustained and ongoing interestin the topic of SF (e.g. Brinckmann et al., 2019;Claussen, Essling and Peukert, 2018; Dai et al.,2018; see also Appendix 1 in the online SupportingInformation). Recent literature reviews identify

1We acknowledge that, recently, SF has sometimes beensubsumed with related concepts operating under differentaliases with similar de"nitions, ranging from synonyms– such as strategic adaptability, to antonyms – such asstrategic adherence and strategic rigidity (see Appendix1 in the online Supporting Information).

© 2020 TheAuthors. British Journal of Management published by JohnWiley & Sons Ltd on behalf of British Academyof Management. Published by John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 MainStreet, Malden, MA, 02148, USA.This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distri-bution and reproduction in any medium, provided the original work is properly cited.

2 Herhausen et al.

divergent theoretical perspectives and differentconceptual underpinnings to SF, which pinpointa series of unresolved tensions (Brozovic, 2018;Combe, 2012).

These tensions arise because the extant liter-ature on SF is heavily fragmented. Research onthis topic has developed concurrently in manyacademic "elds, such as strategy, management,marketing, innovation, entrepreneurship andoperations, with little theoretical and empiricalintegration (Combe, 2012). Consequently, severalpostulated enablers and inhibitors of SF havebeen subject to disagreement in the literature. Forexample, it has been argued that decentralization,de"ned as the degree to which decision-makingis delegated within the "rm, may either enable(Dai et al., 2018) or inhibit SF (Covin, Slevinand Schultz, 1997). Inhibitors of SF are a largelyunder-researched area, with several contradictoryresults reported. For example, some scholars claimyounger "rms to be more !exible (Nadkarni andHermann, 2010), while others posit that the de-gree of !exibility is not dependent on the age ofthe "rm (Brinckmann et al., 2019). Additionally,environmental variables are assumed to triggerSF (Brozovic, 2018), yet with limited empiricalinsights into whether this is truly the case.

Another major concern is the con!icting empir-ical evidence regarding the relationship betweenSF and performance. Evidently, proponents ofSF highlight that there is a price to pay for being!exible, one that might outweigh its bene"ts. Forexample, Das and Elango (1995) point out thatSF entails higher costs, increased stress amongemployees and a potential lack of focus. Con-sequently, a paradoxical challenge emerges formanagers seeking to build SF, since this actionmay generate losses that outweigh potential gains.This possibility is mirrored by inconsistent empir-ical "ndings on the SF−performance relationship.Previous studies report positive effects (Nadkarniand Herrmann, 2010), mixed effects (Grewal andTansuhaj, 2001), no effects (Brews and Hunt,1999) or even negative effects on performance(e.g. adherence to plans pays off for "rms; Covin,Slevin and Schultz, 1997).

These contradictory "ndings call for a thoroughexamination of the SF–performance relationship,in particular, because previous research has useddifferent SF and performance measures in differ-ent types of studies and across a large variety ofresearch contexts. For instance, contingency the-

ory asserts that SF would be bene"cial in dynamicenvironments, but not in stable environments – atheoretical postulate that is evident in empirical re-search (e.g. Claussen, Essling and Peukert, 2018).However, despite the multiplicity of different con-tingencies in previous studies, there is limited em-pirical evidence of the relative extent to whichvarious distinct facets of the environment deter-mine the performance effects of SF (Brozovic,2018). Moreover, triggers and contingencies occa-sionally overlap, depending on the theoretical po-sition (Combe, 2012), adding to their conceptualambiguity.

Overall, despite ongoing academic interest andvaluable contributions attempting to consolidatethe domain of SF, the literature clearly exhibitsunresolved theoretical tensions and disagreements,both across and within disciplines. We employmeta-analytic techniques for theory testing and ex-tension to address these important gaps in the lit-erature. Thus, the purpose of this paper is to bringmore clarity to the domain of SF by empiricallyresolving several existing tensions in the relevantliterature because meta-analysis is an indispens-able research tool for integrating and expandinga domain’s knowledge base (Hunter and Schmidt,2004). To this end, we integrate the fragmented lit-erature on SF using data obtained from 106 in-dependent samples reported in 98 different stud-ies (n = 26,940 "rms). We then use a combinationof theory, patterns in data (i.e. relationships fromexisting studies) and the interplay of theory anddata (i.e. consistent and inconsistent "ndings) todevelop a conceptual framework, which we thenempirically test in the meta-analysis by employ-ing a combination of bivariate correlation, meta-analytic structural equationmodelling (MASEM),meta-analytic regression analysis (MARA) and amoderator analysis.

Our work advances the academic understand-ing of SF in several ways. First, we bring seem-ingly distinct but naturally related streams of re-search on various antecedents of SF into sharperfocus through a quantitative meta-analytic syn-thesis. We resolve several tensions, consolidate the"eld and provide insights for further research. Sec-ond, we synthesize the existing empirical literatureon the consequences of SF to derive the mean per-formance effect of SF. This effect is positive andsupports theoretical perspectives that advocate for,rather than oppose, SF. Third, our analyses of con-tingencies provide guidance about methodological

© 2020 The Authors. British Journal of Management published by John Wiley & Sons Ltd on behalf of BritishAcademy of Management.

Re-examining Strategic Flexibility 3

considerations, "rm and sample characteristics, aswell as conditions of the environment that mitigateor enhance the SF–performance relationship. Fi-nally, our results suggest that the effect of SF on"nancial outcomes might be more nuanced thanpreviously theorized.

Our work also contributes to managers’ strate-gic decision processes and behaviours. In recentyears, strategists have been increasingly subjectto paradoxical challenges with environments thatpresent ever increasing discontinuities. For exam-ple, previously unremarkable macro environmentsnow represent some of the most formidable chal-lenges, along with digital transformation and thenew normal that summons strategists to ques-tion the status quo. Although it is organization-ally tempting to embed !exibility as a panacea,strategists can become paralysed by the prospectof not knowing how to approach this SF agenda.A strategy playbook is therefore required in or-der to move organizations out of the current so-called ‘ef"ciency paradigm’–whereby "rms seek toservice predictable demand behaviours, competingagainst direct and visible competitors with com-mon resources that can produce repeatable activ-ities, into the ‘exponential paradigm’ – whereby"rms move with agility and speed to both shapeand predict demand behaviours (Deloitte, 2016).Although this reality implores strategists to em-body SF, this can paradoxically often make theminert because of the uncertainty surrounding howto approach this.

Conceptual developmentDimensionality of strategic !exibility

The concept of SF has been addressed by schol-ars across the strategy, management, marketing,innovation, entrepreneurship and operationsdisciplines. This diversity, combined with a pro-liferation in both the theoretical approaches andthe empirical methods employed, has resulted ina ‘conceptual schizophrenia’ characterizing theSF literature (Bahrami and Evans, 2011: 37), asevident in the lack of cumulative theory develop-ment, multiple theoretical tensions and the needto pursue further attempts to consolidate the "eld.In reviewing the literature, we "nd that SF hasbeen conceptualized in various ways, as displayedin Table 1. By summarizing the common groundamong the de"nitions, one can attempt to ex-

trapolate SF’s dimensions and offer implicationsregarding the means by which it is performed. A"rm can exhibit SF by being reactive (respond-ing to change) and/or proactive (creating newopportunities) in terms of the variety of avail-able strategic options and/or the speed (timelyresponse) of pursuing a strategic option; it canalso react internally through resource deploymentand/or externally via competitive actions.Taken together, and similar to more recent con-

ceptual overviews (Brozovic, 2018; Combe, 2012),we thus consider SF a multidimensional con-struct.2 The dimensionality of SF has been empha-sized previously (see e.g. Bahrami andEvans, 2011;Brozovic, 2018; Evans, 1991), together with its an-tecedents (i.e. enablers, inhibitors and triggers), ashas its consequences and contingencies. We em-phasize the inherent theoretical tensions presentwithin every aspect when describing the theoreti-cal background of our study.

Antecedents of strategic !exibility

The antecedents of SF encompass enablers,inhibitors and triggers of SF, as identi"ed byBrozovic (2018). Previous studies stemming fromdifferent disciplines have identi"ed a plethoraof antecedents affecting SF and a number oftensions related to these antecedents. Based onexisting empirical studies, Table 2 summarizes thetensions related to various antecedents and prov-ides the main arguments for their expected in-!uence on SF.3 These studies have distinguishedbetween two categories of enablers: the strategicorientations of a "rm and its organizational de-sign. In terms of a "rm’s strategic orientation, theliterature frequently relates entrepreneurial orien-tation (Nadkarni and Narayanan, 2007), marketorientation (Johnson et al., 2003) and learningorientation (Santos-Vijande, Sanchez and Tres-palacios, 2012) to SF. Moreover, in terms oforganizational design, decentralization (Volberda,

2Please note that we are dependent on the operationaliza-tion of SF in the existing studies in our analysis. Giventhat a large majority of the studies operationalize SF asa single construct, we are unfortunately not able to dis-aggregate SF into its components and test those in ourmeta-analytic analyses.3We focus our discussion on antecedents that have fre-quently been empirically investigated. See Brozovic (2018)for a more exhaustive conceptual review of enablers, in-hibitors and triggers.


4 Herhausen et al.

Table1.

Multidimensionalconceptualizationofstrategic!exibility

Sanchez(199

5)Volberda(199

6)Hittetal.(19

98)

You

ng-Y

barraan

dWiersem

a(199

9)Grewal

andTa

nsuh

aj(200

1)

De!nition

Firm’sab

ility

torespon

dto

variou

sdeman

dsfrom

dyna

miccompetitive

environm

ents(p.1

38)

The

degree

towhich

anorga

nizatio

nha

savariety

ofman

agerialcap

abilitie

san

dthespeedat

which

they

canbe

activ

ated

(p.3

61)

The

capa

bilityof

the"rm

toproa

ctor

respon

dqu

ickly

tochan

ging

competitive

cond

ition

san

dthereby

developan

d/or

maintain

acompetitivead

vantage

(p.2

7)

The

ability

toad

aptto

environm

entalcha

nges,

tochan

gega

meplan

s,to

precipita

teintentiona

lchan

gesan

dto

continuo

uslyrespon

dto

unan

ticipated

chan

ges

(p.4

40)

Organ

izationa

labilityto

man

ageecon

omican

dpo

litical

risksby

prom

ptlyrespon

ding

ina

proa

ctiveor

reactiv

eman

nerto

marketthreats

andop

portun

ities

(p.7

2)Context

Prod

uctc

ompetition

Organ

izationa

lform

Com

petitiveland

scap

eStrategicallia

nces

Econo

miccrisis

Dimensions

Reactivity

Respo

ndingto

deman

dsfrom

theprod

uctm

arket

The

potentialfor

ady

namic

"tcanbe

reactiv

eRespo

ndingto

chan

ging

competitivecond

ition

sRespo

ndingto

chan

gesin

anallia

nce

Respo

ndingto

market

oppo

rtun

ities

Proactivity

–The

potentialfor

ady

namic

"tcanbe

proa

ctive

Proa

ctto

chan

ging

competitivecond

ition

sPrecipita

teintentiona

lchan

gesin

anallia

nce

Proa

ctto

market

oppo

rtun

ities

Variety

Largerrang

eof

alternative

prod

ucts

Variety

ofman

agerial

capa

bilities

Variety

inprod

ucto

fferings

Awidevarietyof

allia

nce

optio

nsDiverse

portfolio

ofstrategicop

tions

Speed

Reducingthetim

erequ

ired

tosw

itchbetw

een

prod

ucts

Speedat

which

man

agerial

capa

bilitiescanbe

activ

ated

Respo

ndingqu

icklyto

competitivecond

ition

sTim

elymod

i"catio

nan

dexitof

anallia

nce

Tim

eto

respon

dto

market

oppo

rtun

ities

Internally

Flexibleuseof

resourcesfor

prod

uctc

reation

Abilityto

adap

tcap

abilitie

sto

thedeman

dsof

the

environm

ent

Buildingdy

namiccore

competences

Internal

structural

adjustmentsin

anallia

nce

Develop

ing!exible

capa

bilities

Externally

Coo

rdinatingresourcesin

theprod

uctm

arket

Abilityto

in!u

ence

the

environm

ent

Exp

loiting

glob

almarkets

Ada

ptingto

environm

ental

chan

ges

Ada

ptingto

marketthreats

andop

portun

ities

Googlecitations

1,78

21,34

71,48

779

81,10

4

Note:Weconsider

themainconceptualizations

ofstrategic!exibilityba

sedon

theirim

pact

ontheliterature(i.e.measuredwith

Goo

glecitatio

nsin

Janu

ary20

19).



Table 2. Tensions related to the antecedents of strategic !exibility

Antecedent De"nition Arguments for enabling SF Arguments for inhibiting SF

Enablers of SFStrategicorientation

Entrepreneurialorientation

Emphasizing risky projectsand a proclivity topioneer

Entrepreneurial intent may becentral to improving strategicadaption and SF within a "rm(Nadkarni and Narayanan,2007)

Entrepreneurial orientation mayresult in an exploration trap,with an abundance ofalternatives leading to inertiabecause of limited processingcapacity (Levinthal and March,1993)

Marketorientation

Responding to customers’needs and competitors’actions

The generation, dissemination anduse of market information mayfacilitate and accelerate SF(Johnson et al., 2003)

Excessive market orientation mayconstrain the SF of a "rm(Grewal and Tansuhaj, 2001)

Learningorientation

Organization-wide activityto create and useknowledge

Learning may allow a moreeffective adjustment to changingconditions and thus promote SF(Santos-Vijande, Sanchez andTrespalacios, 2012)

Successful experiences are likely toreinforce themselves positivelyover time and may constrain SF(Nadkarni and Herrmann, 2010)

Organizationaldesign

Decentralization Degree to whichdecision-making isdelegated within the "rm

Sharing decision-making across anorganization may promote SF(Volberda, 1998)

Decision-sharing throughout anorganization may promoteadherence to prevailing strategies(Covin, Slevin and Schultz, 1997)

Slack resources Degree to whichuncommitted resourcesare available

Organizations withoutuncommitted liquid resourcesare less likely to adapt toenvironmental changes (Barkerand Barr, 2002)

Slack has a negative effective on"rms’ response to environmentalshifts and is therefore likely toreduce SF (Cheng and Kesner,1997)

Formal routines Degree to which formalrules governdecision-making

Standardized routines maygenerate more rapid responses toenvironmental changes and thuspromote SF (Shimizu and Hitt,2004)

Formal rules and routines aremeans for advancing ef"ciencyand may thus constrain SF(Schreyögg and Sydow, 2010)

Inhibitors of SFFirm size Total number of employees Larger "rms may have more

resources to devote to SF thansmaller "rms (Barker and Barr,2002)

Smaller "rms may be moredynamic than larger "rms, whichtend to have greater structuralinertia (Nadkarni andHerrmann, 2010)

Firm age Number of years since the"rm was founded

Older "rms may have moreresources to devote to SF thanyounger "rms (Barker and Barr,2002)

Younger "rms may shift theirstrategies more frequently thanolder "rms, which focus on thestatus quo (Nadkarni andHerrmann, 2010)

Past success A "rm’s prior success on themarket

Past success may be positivelyassociated with managers’environmental awareness andmay promote SF (Lant, Millikenand Batra, 1992)

Past success may reinforce thevalue of existing strategies anddiscourage SF (Nadkarni andHerrmann, 2010)

Triggers of SFEnvironmental

dynamismUnpredictability and

variation of change in theenvironment

Environmental changes mayincrease the likelihood ofstrategic reorientation and thustrigger SF (Lant, Milliken andBatra, 1992)

Organizations tend to maintain thestatus quo in high environmentaluncertainty, therebydisapproving of SF (Shimizuand Hitt, 2004)

(Continued)


6 Herhausen et al.

Table 2. (Continued)

Antecedent De"nition Arguments for enabling SF Arguments for inhibiting SF

Demanduncertainty

Degree of uncertainty aboutcustomer demand

Markets with more variation indemand may favour more!exible "rms (Claussen, Esslingand Peukert, 2018)

High uncertainty may createresistance to change anddiscourages SF (Skordoulis,2004)

Competitiveintensity

Degree of competition The strength and activity ofcompetitors may trigger SF(Grewal and Tansuhaj, 2001)

Competitive pressure may reducethe resources available and limitSF (Barker and Barr, 2002)

Notes:We report potential antecedents of SF that have been investigated in three or more empirical studies. Due to space restrictions,the table contains only selected references in favour of the arguments.

1998), slack resources (Barker and Barr, 2002)and formal routines (Shimizu and Hitt, 2004)have been related to SF. However, there has beendisagreement regarding these enablers of SF. Forexample, while formal routines have been positedto result in more rapid responses to environ-mental changes (Worren, Moore and Cardona,2002), they have also been identi"ed as a meansfor advancing ef"ciency rather than !exibility(Schreyögg and Sydow, 2010).

Compared to the enablers, the inhibitors of SFare largely under-researched within the literature(Brozovic, 2018). We identify "rm size, "rm ageand past success as potential inhibitors. However,in terms of "rm size, some scholars ascribe SFas a characteristic of large "rms because of thepotential availability of slack resources (Pauwelsand Matthyssens, 2004). In contrast, SF can beexhibited by smaller "rms speci"cally becausetheir resource adaptability is higher (Ebben andJohnson, 2005). Additionally, "rm age (Barkerand Barr, 2002) and past success (Lant, Millikenand Batra, 1992) have also been claimed to en-courage SF.

Finally, the literature asserts that changes in thebusiness environment may not only affect the SF–performance relationship, but also trigger SF. En-vironmental dynamism (Perez-Valls et al., 2016),demand uncertainty (Aaker and Mascarenhas,1984) and competitive intensity (Das and Elango,1995) may all act as external triggers of a "rm’s SF.Nevertheless, other scholars question whether thebusiness environment has these triggering effectson SF. For example, Shimizu and Hitt (2004) cau-tion that organizations tend to maintain the sta-tus quo in high environmental uncertainty, therebydisapproving of SF. Indeed, resistance to changeas a reaction to pursuing SF in a "rm is a practi-cally implicit challenge that must be overcome to

instil SF throughout the organizational hierarchy(Skordoulis, 2004).

To summarize, the extant literature offers aplethora of potential enablers, inhibitors and trig-gers of SF, but there are disagreements and con-troversies regarding how much (if at all) some ofthese factors actually affect SF.Another signi"canttension encompassing SF relates to the differentialeffects of SF, each of which are aligned with dif-ferent theoretical paradigms.

Consequences of strategic !exibility

The theoretical tensions surrounding the conse-quences of SF are summarized in Table 3. Whenbalancing the competing theoretical positions andconsidering the typical stance of the managerialliterature, most researchers argue that SF is ben-e"cial for "rms (Brozovic, 2018). When theorizingon its positive performance effect, some scholarsespouse SF as a dynamic capability because it em-phasizes the !exible use of resources and the re-con"guration of processes (Johnson et al., 2003).Others consider SF to be a means of increasingcontrol by maintaining a portfolio of strategic op-tions (Dalziel, 2009). Adopting an organizationaldesign perspective, other scholars have emphasizedthe bene"ts that arise from the degree to which anorganization maintains a repertoire of managerialcapabilities and the speed at which they can be ac-tivated (Wang, Senaratne and Ra"q, 2015).

However, in contrast to these viewpoints, otherresearchers argue that SF does not improvedesired outcomes or may even reduce a "rm’s per-formance. For example, the !exibility–ef"ciencytrade-off, which describes the distinction be-tween ef"cient and !exible strategies (Ebben andJohnson, 2005), illustrates a long-standingdilemma in organization theory. Thompson (1967)



Table 3. Tensions related to the consequences of strategic !exibility

Theoretical perspectives advocating SF Theoretical perspectives opposing SF

Dynamic capabilities! SF is promoted by "rms’ latent abilities to renew, augmentand adapt their routines over time (Teece, Pisano andShuen, 1997).

! SF, in exploiting and controlling resources, explains whysome "rms move more quickly into new niches and arethus more successful (Eisenhardt and Martin, 2000).

Resource-based view! Investments in SF require a "rm to possess surplusresources, thereby reducing the "rm’s ef"ciency and itscorresponding performance (Amit and Schoemaker,1993).

! Valuable resources are developed over time throughexperience, and experience is gained through adherence tostrategic actions (Lant, Milliken and Batra, 1992).

Real options reasoning! Having a portfolio of strategic options predisposes a "rmto bene"t from upside opportunities and mitigates itsdownside risk, thus improving performance (McGrath,1997).

! Firms can manage their performance trajectory by meansof real options, thereby engaging in SF, enabling them tosequence, stage and potentially reverse commitment to astrategy (Trigeorgis, 1996).

Evolutionary theory! The switching costs or costs of changing trajectories andacquiring knowledge unrelated to the asset base can behigh (Henderson and Clark, 1990).

! A capability is not valuable unless a reliable andconsistent ‘practice’ has evolved over time (Schreyögg andSydow, 2010).

Organizational design! SF depends on the "rm’s coordination capabilities inapplying resources to alternative courses of action(Volberda, 1996).

! The loose coupling of routines and functions limits thecost of change, improves responsiveness to environmentalshifts and increases performance (Sanchez, 1995).

Population ecology! Reliable and repeated patterns of action are a necessaryprecondition for organizational survival and higherperformance (Hannan and Freeman, 1977).

! Firms reduce their chances of survival when they engagein transformation and SF (Hannan and Freeman, 1977).

Industrial organization theory! The concept of new competition suggests that "rmspursue positional advantage within an industry bystrategic !exibility and adaptation (Best, 1990).

! SF enables "rms to reposition their product–marketstrategies to sustain a competitive advantage (Harrigan,1985).

Strategic planning school! Strategic planning improves "rm performance throughincreased continuity and ef"ciency (Brews and Hunt,1999).

! SF may distract managers from making the long-termcommitments needed to implement long-term strategiessuccessfully (Ghemawat and Ricart i Costa, 1993).

Notes:Due to space restrictions, the table contains only selected statements that represent the views of both advocates and opponents ofSF. These statements do not summarize the underlying theoretical streams, nor do they attempt to represent all aspects of the differenttheoretical streams.

calls this the ‘paradox of administration’ (p. 15),whereby managers need to reconcile, for example,the organization of routine tasks with tasks thatare non-routine. Ef"ciencies in delivering strategyrequire the "rm to maximize its output fromexisting resources, while !exibility requires it torecon"gure resources (Ghemawat and Ricart iCosta, 1993). Pursuing ef"ciency will therefore in-stil replicable routines that fundamentally oppose!exibility (Weigelt and Sarkar, 2009).

The suggestion here is not necessarily that ef-"ciency is superior to !exibility, but that mix-ing both strategies can lead to underperformance(Ebben and Johnson, 2005). Firms may make atrade-off between !exibility and ef"ciency, but "ndthat they achieve only mediocre performance asa result (Eisenhardt, Furr and Bingham, 2010).

Moreover, "rms may vacillate between strategicoptions, thereby wasting resources and risking fail-ure (Shimizu and Hitt, 2004).To conclude, the SF–performance relationship

has been described as complex, and recent stud-ies called for more research illuminating thesetensions (Claussen, Essling and Peukert, 2018).Moreover, despite themultiplicity of different con-tingencies in previous studies, there is limited em-pirical evidence of the relative extent to which var-ious distinct facets of the environment determinethe performance effects of SF (Brozovic, 2018).

Contingencies of strategic !exibility

We have previously emphasized the tensionssurrounding environmental dynamism, demand


8 Herhausen et al.

Figure 1. Meta-analytic research framework

uncertainty and competitive intensity as triggersof SF. However, the extant literature has also re-garded these three variables as contingencies in theSF–performance relationship, adding to the am-biguity surrounding these variables. Treating themas moderators of the strength of the relationshipbetween SF and performance (Brozovic, 2018;Combe, 2012), the consolidating literature on SFhas recommended even more research about them,despite some valuable insights (e.g. Grewal andTansuhaj, 2001; Nadkarni and Narayanan, 2007).

Thus, SF is likely to become more importantwhen an organization is faced with a high rateof change in the environment (Nadkarni andNarayanan, 2007). Similarly, high levels of en-vironmental dynamism are suggested to favour!exibility (Johnson et al., 2003). As for demand un-certainty, Harrigan (1985) "nds that SF is a meansto manage demand volatilities. More speci"cally,demand uncertainty creates dif"culty in devisingstrategic plans; SF should be more useful in theseuncertain markets (Claussen, Essling and Peukert,2018; Lee and Makhija, 2009). In other words, de-mand uncertainty moderates the relationship be-tween SF and performance (Grewal and Tansuhaj,2001). The same understanding is advanced forcompetitive intensity, whereby it is recommendedto develop SF in order for "rms to adapt tohighly competitive environments (Hitt et al., 1998;Nadkarni and Narayanan, 2007). Indeed, therelationship between SF and "rm performanceis extensively moderated by competitive intensity(Guo and Cao, 2014). In addition, previous re-search has further indicated that "rm size and "rmage may determine the performance effects of SF,with smaller "rms and younger "rms bene"ttingmore from SF (Segaro, Larimo and Jones, 2014).Consequently, we consider these contingencies inour analyses.

Our meta-analytical research framework is pre-sented in Figure 1. While the framework itself isgrounded in theory, the articulation of SF’s ex-pected antecedents, consequences and contingen-cies is driven by a combination of the portrayedtheoretical tensions and patterns in existing empir-ical "ndings. Thus, the goal of this meta-analysis isto empirically determine the nomological networkthat encircles SF.

Data collection

We followed the suggestions and guidelines pro-vided by Aguinis et al. (2011) and Bergh et al.(2016) for our meta-analysis. Our objective was toidentify all articles that report empirical data onthe outcomes and antecedents of SF. The key im-petus for SF research was Aaker and Mascaren-has’s (1984) article. Consequently, we adopt this asour census date and consider empirical research re-ported during the period from 1984 to April 2019.To identify relevant studies, we used "ve comple-mentary search strategies. First, we explored elec-tronic databases such as ABI/INFORM Global,EconLit, Google Scholar, JSTOR and SSRN, us-ing the following search terms: strategic !exibility,strategic adaptability, strategic rigidity and strate-gic adherence. Second, we manually searched 19scholarly journals, including the British Journal ofManagement, Academy of Management Journal,Journal of Management and Strategic Manage-ment Journal. Third, we explored the reference listsof all identi"ed articles and traced all sources thatcited them. Fourth, in an effort to identify relevantunpublished studies, we searched ProQuest DigitalDissertations and conference proceedings for theannual meetings of, among others, theAcademy ofManagement and Strategic Management Society.



Finally, requests were posted on the listservs of theAcademy of Management and Strategic Manage-ment Society to elicit unpublished research.

Studies were selected for inclusion on the basisof three criteria. First, we included only empiri-cal studies that reported an outcome statistic thatallowed for the computation of a correlation co-ef"cient (e.g. Hatum and Pettigrew, 2006 was ex-cluded). Second, a study had to report on relation-ships involving one or more operationalizations ofSF (or a related construct with a similar de"ni-tion), as well as relevant antecedents and/or out-comes. Third, only studies that measured SF at the"rm level were included. Accordingly, a numberof studies could not be included, because they fo-cused on !exibility at the individual or team level(Stewart and Barrick, 2000), or because the resultswere reported only in multivariate models.4

Using studies identi"ed up to the end of 2018,these efforts yielded a "nal sample of n = 26,940"rms examined in 106 independent samples from98 studies, consisting of 92 published studies and 6unpublished studies. Appendices 2 and 3 in the on-line Supporting Information provide details on allstudies, and Appendix 4 displays funnel plots forthe performance effect; the plots suggest that nei-ther publication bias nor other potential sourcesof plot asymmetry affect the obtained results.

To reduce coding error, we prepared a protocolthat speci"ed the information to be extracted fromeach study, and two coders extracted the data fromeach study. The level of agreement between the twocoders was high (94%), and discussion between thecoders helped clarify all disagreements. The codersfurther extracted data on effect sizes, measure-ment characteristics and measure reliability statis-tics, study sample sizes and other study character-istics. Following best-practice suggestions, we useda four-step approach to analyse the data: (1) bi-variate associations, (2) MASEM of antecedents,(3) MARA of consequences and (4) moderationanalysis of contingencies.

4We contacted the authors of nine studies with requests toprovide correlation tables of their studies. These effortsyielded four additional studies that were included in themeta-analysis.

Analyses and resultsBivariate relationships

First, we estimate the bivariate relationshipsamong all key constructs. As suggested by Aguiniset al. (2011), we use Hunter and Schmidt’s (2004)random effects approach at the effect-size level,which allows for the correction of statisticalartefacts and thus provides a relatively accurateestimate of the average strength and variance ofa relationship in the population of interest. Weuse artefact-corrected effect sizes, transformedinto Fisher’s z coef"cients and weighted by theestimated inverse of their variance. This correc-tion gives more weight to more precise estimatesbefore converting them back to correlation coef-"cients. For each bivariate relationship, we reporta 95% con"dence interval of the sample-weightedreliability-adjusted averaged correlation. We alsocalculate the ‘fail-safe N’, which indicates thenumber of non-signi"cant and unavailable studiesthat would be needed tomake the cumulative effectsize non-signi"cant to further evaluate a potentialpublication bias. We test the homogeneity of thedistribution of effect size using the Q-statistic.Heterogeneity for all relationships was signi"cant.However, the standard deviations are all below thecommon cut-off value for an acceptable standarddeviation in a population (Cortina, 2003). Thus,analyses of the main effects are warranted. Foreach focal relationship, we apply the trim-and-"llmethod to assess publication bias and impute themissing values. None of the imputed relation-ships changed in direction or signi"cance, and allbivariate relationships are reported in Table 3.

Antecedents of strategic !exibility

The bivariate relationships between antecedentsand SF in Table 3 are largely in line with our theo-retical expectations. However, while primary stud-ies investigating the antecedents of SF focus onlyon subsets of potential enablers, inhibitors andtriggers of SF, the aggregated study effects canbe used to assess simultaneously the effects ofvariables that prior research have not consideredjointly (Bergh et al., 2016). Therefore, we useMASEM, a two-stage procedure, to estimate theunique contribution of each of the antecedentsof SF (Cheung and Chan, 2005). During the "rststage, we conduct several meta-analyses, relating


10 Herhausen et al.

Table 4. Bivariate relationships

ka n ps pw p pʹ SDp 95% CIp var% Q

Antecedents of SFEnablers of SFStrategic orientationEntrepreneurial orientation 11 2,424 0.32 0.44 0.51 – 0.24 0.45/0.57 5.47 207.00Market orientation 27 7,952 0.27 0.32 0.40 – 0.12 0.36/0.44 15.10 233.34Learning orientation 10 1,859 0.34 0.34 0.40 – 0.17 0.33/0.48 12.16 81.03Organizational designDecentralization 10 1,726 0.22 0.23 0.28 – 0.19 0.18/0.38 12.35 113.63Slack resources 11 2,004 0.19 0.19 0.24 – 0.21 0.14/0.34 10.49 146.38Formal routines 11 3,768 0.15 0.10 0.12 0.10 (2) 0.17 0.02/0.22 8.65 160.63Inhibitors of SFFirm size 39 9,994 0.00 0.01 0.01 – 0.12 −0.16/0.19 20.70 209.97Firm age 31 6,914 −0.01 −0.02 −0.03 – 0.09 −0.18/0.12 34.53 103.63Past success 15 1,842 −0.13 −0.04 −0.04 – 0.16 −0.26/0.17 23.39 66.38Triggers of SFEnvironmental dynamism 25 7,966 0.11 0.18 0.22 – 0.28 0.17/0.27 3.72 902.49Demand uncertainty 14 4,489 0.06 0.01 0.00 – 0.27 −0.33/0.33 4.18 440.73Competitive intensity 16 3,729 0.22 0.25 0.30 – 0.21 0.24/0.37 7.65 268.20Consequences of SFPerformance effect 181 47,604 0.25 0.21 0.24 0.24 (5) 0.19 0.23/0.27 8.11 2,495.87Financial outcomes 72 23,954 0.12 0.12 0.13 – 0.17 0.10/0.17 9.59 968.60Innovation outcomes 50 10,916 0.33 0.32 0.38 – 0.17 0.34/0.41 9.85 438.01Market outcomes 42 8,302 0.33 0.32 0.37 0.34 (2) 0.14 0.33/0.41 15.65 240.77Other outcomes 17 4,432 0.30 0.28 0.33 – 0.18 0.27/0.38 8.47 210.71

Notes: k = number of effect sizes; n = total sample size; ps = simple mean in population; pw = sampling-error-corrected mean inpopulation; p = bias-corrected mean in population; pʹ = trim-and-"ll-corrected mean in population and number of missing studies(if applicable); SDp = estimated standard deviation of corrected correlations in population; 95% CIp = 95% con"dence interval forp; var% = percentage of observed variance accounted for by statistical artefacts; Q = Q-test for homogeneity in the true correlationacross studies.

all constructs of interest to each other, using datafrom the primary studies in our database. We thuscalculate a complete correlation matrix, includingthe effect sizes of all relationships (see Appendix 5in the online Supporting Information). In the sec-ond stage, we apply structural equation modellingon the correlation matrix.

Given the variability in sample sizes associ-ated with each correlation coef"cient in the meta-analytic correlation matrix, we base this analysison the harmonic mean of the sample sizes com-prising each entry in the meta-analytic correlationmatrix (Viswesvaran and Ones, 1995). The "t in-dexes for the MASEM model in Table 5 suggesta good "t for the model (χ2/d.f. = 23.66/26; RM-SEA = 0.00; AGFI = 0.99; SRMR = 0.02). All ofthe relationships hold when the statistically non-signi"cant antecedents are removed.Moreover, us-ing trim-and-"ll-corrected effect sizes, the param-eter estimates do not change.

Regarding enablers of SF, we "nd that, consis-tent with Schweiger et al. (2019), entrepreneurial

orientation (β = 0.33, p < 0.01), market orienta-tion (β = 0.22, p < 0.01) and learning orientation(β = 0.30, p < 0.01) all are positively associatedwith SF, indicating the importance of the strategicorientation of a "rm. Turning to aspects of orga-nizational design, we "nd that decentralization (β= 0.54, p < 0.01) and formal routines (β = 0.28,p < 0.01) are positively associated with SF, whileslack resources (β = −0.22, p < 0.01) is negativelyassociated with SF. This is an unexpected "nding,considering that previous studies have emphasizedthat "rms need uncommitted, liquid resources toenable SF (Barker and Barr, 2002).

When considering the inhibitors of SF, we "ndthat past success (β = −0.21, p < 0.01) is indeednegatively associated with SF, and there is no sig-ni"cant relationship between "rm size and SF (β= −0.01, n.s.) and even a positive relationship be-tween "rm age and SF (β = 0.06, p < 0.01). These"ndings highlight that not all commonly acknowl-edged inhibitors of SF can be traced in existingempirical studies. Finally, the "ndings related to



Table 5. MASEM results: multivariate antecedents of strategic !exibility

β SE 95% CI t-Ratio p-Value

Enablers of SFStrategic orientationEntrepreneurial orientation 0.33 0.02 0.28, 0.37 13.66 >0.001Market orientation 0.22 0.02 0.18, 0.27 9.46 >0.001Learning orientation 0.30 0.03 0.23, 0.36 10.01 >0.001Organizational designDecentralization 0.54 0.04 0.46, 0.61 14.72 >0.001Slack resources −0.22 −0.02 −0.27, −0.18 −9.10 >0.001Formal routines 0.28 0.04 0.19, 0.37 6.66 >0.001Inhibitors of SFFirm size −0.01 0.02 −0.06, 0.03 −0.57 0.566Firm age 0.06 0.02 0.02, 0.09 2.82 0.005Past success −0.21 0.02 −0.25, −0.16 −10.10 >0.001Triggers of SFEnvironmental dynamism −0.24 0.03 −0.31, −0.18 −7.79 >0.001Demand uncertainty −0.06 0.02 −0.11, −0.03 −3.21 0.001Competitive intensity 0.38 0.03 0.31, 0.45 12.19 >0.001Harmonic mean 1,317χ2/d.f. 23.66/26RMSEA 0.00AGFI 0.99SRMR 0.01R2 0.42

Notes: Results are based on a meta-analytic correlation matrix. The standardized path coef"cients equal effect sizes. Error variancesfor each construct indicator were "xed at zero. SE = standard error; CI = bias-corrected con"dence interval.

the triggers of SF reveal that competitive inten-sity is positively associated with SF (β = 0.38,p < 0.01). Contrary to our expectations derivedfrom the literature, environmental dynamism (β =−0.24, p < 0.01) and demand uncertainty (β =−0.06, p < 0.01) are negatively associated with SF.Discovering that both environmental dynamismand demand uncertainty do not trigger SF high-lights an important gap between theory and em-pirical "ndings in existing studies.

Consequences of strategic !exibility

The bivariate relationships in Table 4 supportthe claim that SF is positively related to per-formance consequences. The correlations betweenSF and "rm performance range from −0.59to 0.87, and the correlation frequency is nor-mally distributed (Zskewness = −0.30, Zkurtosis =0.92, pKolmogorov–Smirnov < 0.05). Most of the SF–performance correlations in the database are pos-itive (85%), and the adjusted mean correlation ofthe focal performance relationship is 0.24 (95% CI= 0.23–0.27).

We next investigate the speci"c conditions un-der which the performance effect of SF might belarger or smaller. We consider measurement ofconstructs, as well as methodological and samplecharacteristics, as potential in!uencers of the SF–performance relationship (e.g. Karna, Richter andRiesenkampff, 2016). Regarding the measurementof SF, we code all operationalizations according tothe six dimensions displayed in Table 1. In line withother meta-analyses (e.g. Rubera andKirca, 2012),we group the various operationalizations of "rmperformance into innovation outcomes, marketoutcomes and "nancial outcomes (see Appendix 6in the online Supporting Information). We furthercode the year of data collection, whether the studyis published, the ISI impact factor and secondarydata versus survey data collection, as well as thegeographic region of the sample and whether thesample consists of large "rms or SMEs.We then use MARA (Van den Noortgate et al.,

2013), a special type of weighted least squaresregression analysis designed to assess the relation-ship between effect size and potential moderatorvariables by modelling heterogeneity in the effectsize distribution. The homogeneity tests for the


12 Herhausen et al.

Table 6. MARA results: performance effect of strategic !exibility

Sample γ SE t-Ratio p-Value

Intercept 0.27 0.02 15.02 <0.001Measurement of SFReactivity 88% of studies −0.02 0.02 −1.00 0.320Proactivity 50% of studies 0.05 0.02 2.68 0.009Variety 89% of studies −0.01 0.02 −0.38 0.708Speed 40% of studies 0.01 0.02 0.32 0.754Internal 81% of studies 0.05 0.02 2.50 0.015External 71% of studies 0.04 0.02 2.81 0.006Measurement of performanceLagged measurement 14% of studies −0.06 0.03 −1.82 0.073Innovation outcomes 28% of studies 0.05 0.02 3.38 0.001Market outcomes 23% of studies 0.07 0.02 3.93 <0.001Other outcomes 9% of studies 0.02 0.02 1.08 0.285Methodological controlsYear of data collection Mean: 2005.94 0.06 0.02 3.02 0.003Published study 88% of studies 0.03 0.02 2.06 0.043ISI impact factor Mean: 2.36 −0.02 0.02 −1.50 0.139Secondary data (vs. survey data) 17% of studies −0.02 0.03 −0.81 0.418Sample controlsEuropean sample 27% of studies 0.00 0.02 0.03 0.977Asian sample 19% of studies −0.04 0.03 −1.73 0.088Mixed sample 8% of studies 0.01 0.01 0.88 0.380Large "rms (vs. SME) 47% of studies 0.03 0.02 1.15 0.254Between-study R2 0.39Within-study R2 0.59Number of studies (Level 3) 88Number of effects (Level 2) 181

Notes: We use US samples (35% of studies) and "nancial outcomes (40% of studies) as base categories. Standardized results with robuststandard errors are reported, and all slope coef"cients were treated as "xed. SE = standard error.

Table 7. Moderator analysis of the environment and "rm characteristics

Correlation with performanceeffect

Partial correlation withperformance effect

k rsimple rcorrected rsimple rcorrected

EnvironmentEnvironmental dynamism 45 0.31* 0.34* 0.43** 0.46**Demand uncertainty 18 0.02 0.10 0.39 0.41Competitive intensity 32 0.23 0.30 0.12 0.20Firm characteristicsFirm size 24 0.25 0.28 −0.09 −0.09Firm age 43 −0.08 −0.09 0.29 0.25

Notes: *p < 0.05 (two-tailed). Please note that the signi"cance tests take the sample size k into account. Partial correlations control forall signi"cant effects from the MARA analysis (year of data collection, published study, measurement in terms of proactivity, internaland external attribution, as well as innovation and market outcomes).**p < 0.01.

correlations involving the SF–performance rela-tionship indicate that moderator variables mayexplain heterogeneity in the effect sizes (χ2

181= 2,495.87, p < 0.01). We model the reliability-corrected correlations in the moderation analysis

as a linear function of the determinants andperform a three-level analysis to account forwithin-study error correlation between effect sizesand dependence of effect sizes within studies (Vanden Noortgate et al., 2013). Therefore, we specify



the following model:

Yijk = γ000 + γ0l0

n∑

l=1Xl,jk + r00k + u0jk + eijk,

(1)where Yijk represents the reliability-corrected cor-relations of effect j in study k, γ 000 is a constant,γ 0l0 represents parameter estimates of the deter-minants, Xl,jk represents variable matrices of thedeterminants (moderators), r00k is the study-levelresidual error term, u0jk is the effect-level resid-ual error term and eijk is the measurement-levelresidual error term. The results of the hierarchicallinear regression appear in Table 6.

The positive SF–performance relationship isstronger when SF is measured in terms of proac-tively creating new opportunities (γ = 0.05, p <0.01), internal resource deployment (γ = 0.05, p <0.05) and external competitive actions (γ = 0.04,p < 0.01), compared to when it is not. MeasuringSF in terms of reactivity (responding to change),variety (increasing options) and speed (timely re-sponse) does not affect the SF–performance rela-tionship. We further "nd that the performance ef-fect of SF is stronger when performance is mea-sured in terms of innovation outcomes (γ = 0.05,p < 0.01) or market outcomes (γ = 0.07, p < 0.01)rather than "nancial outcomes. Finally, the SF–performance relationship is stronger both in newerstudies (γ = 0.06, p < 0.01) and in published stud-ies (γ = 0.03, p < 0.05).

Contingencies of strategic !exibility

The lack of available correlations involving in-teraction terms precludes moderation analysesof the theoretically relevant contingencies withMARA. Therefore, consistent with other meta-analyses (e.g. Blut et al., 2015), we report the par-tial correlation between the moderator and the ef-fect size while controlling for all signi"cant effectsfrom the MARA analysis. Table 7 shows that onlyenvironmental dynamism signi"cantly and posi-tively moderates the SF–performance relationship(rcorrected = 0.46, p < 0.01). While we "nd direc-tional support for the moderating effects of de-mand uncertainty and competitive intensity, theseeffects are not statistically signi"cant (the signi"-cance tests in all analyses take the sample size kinto account). Moreover, neither "rm size nor "rmage moderate the SF–performance relationship.

Post-hoc analysis: from strategic !exibility to"nancial outcomes

Both the bivariate results in Table 4 and theMARA results in Table 6 reveal that SF has ahigher performance effect when "rm performanceis measured as innovation or market outcomesrather than as "nancial outcomes. In line withother meta-analyses (e.g. Rubera andKirca, 2012),we thus explore the potential chain of effects, fromstrategic !exibility to "nancial outcomes, via inno-vation and market outcomes. On the one hand, SFenables "rms to anticipate market demands andto develop innovative products and services (Neilland Rose, 2007), thereby increasing performancein terms of innovation outcomes (Kortmann et al.,2014). On the other hand, SF enables "rms torespond and adapt to changing market demands(Johnson et al., 2003), thereby increasing perfor-mance in terms of market outcomes (Segaro, La-rimo and Jones, 2014). Finally, improvements inboth innovation andmarket outcomes should sub-sequently increase "nancial outcomes (Rubera andKirca, 2012).Considering the direct effect of SF on "nancial

outcomes, SFmay enable "rms to possess a varietyof different plans, decisions and strategies unre-lated to innovation or the "rm’s market position(e.g. Evans 1991). This variety, however, introducescosts and potential bene"ts needed to exceed thenecessary investments, as highlighted by the oppo-nents of SF in Table 3. For example, when assetsare committed for the purpose of SF, the "rmforgoes its short-term earning potential, introduc-ing the additional burden of opportunity costs(Shimizu andHitt, 2004). Thus, one may expect nodirect effect from SF on "nancial outcomes; rather,one may expect that the innovation and marketoutcomes will mediate the "nancial effects of SF.We use MASEM to gain insights into the chain

of effects between SF and "nancial performance(Bergh et al., 2016). As Figure 2a summarizes, ananalysis of the expected mediation model did notresult in an adequate model "t. Given this, andconsistent with the modi"cation indices, we revisethe model, as shown in Figure 2c. This model "tsthe data well (χ2/d.f. = 19.44/44; RMSEA = 0.03;AGFI = 0.99; SRMR = 0.02) and demonstratesthat SF has a direct positive effect on innovationoutcomes (β = 0.35, p < 0.01), a small direct pos-itive effect on market outcomes (β = 0.17, p <0.01) and a small direct negative effect on "nancial


14 Herhausen et al.

(a) Expected Model

χ2/d.f. = 1,544.49/5; RMSEA = 0.24; AGFI = 0.69; SRMR = 0.11

(b) Competing Model

χ2/d.f. = 3,737.21/7; RMSEA = 0.31; AGFI = 0.46; SRMR = 0.16

(c) Best Fitting Model

χ2/d.f. = 19.44/4; RMSEA = 0.03; AGFI = 0.99; SRMR = 0.02

Figure 2. MASEM results: from strategic !exibility to "nancial outcomesNotes: We control for "rm size and "rm age in all models. Dominant effects are in bold. MASEM results display standardized pathcoef"cients, which equal effect sizes. We use the harmonic mean (n = 5,575) for estimation purposes. For clarity of presentation, these"gures do not include correlations or endogenous error terms.



outcomes (β = −0.06, p < 0.01). We use a boot-strapping procedure with 1,000 samples to test forindirect effects. The results show a positive effectof SF on market outcomes through innovationoutcomes (β = 0.17, p < 0.01; 95% CI = 0.16–0.19) and a positive effect of SF on "nancial out-comes through innovation outcomes and marketoutcomes (β = 0.21, p < 0.01; 95% CI = 0.19–0.23). The overall effect of SF on "nancial out-comes, which combines both the direct and indi-rect effects, is positive (β = 0.15, p < 0.01; 95% CI= 0.13–0.18). These results suggest that SF is in-deed an important part of a "rm’s competitive ar-moury, but its effect on "nancial outcomes mightbe more nuanced than previously theorized.

Discussion and conclusions

Given the importance of SF in the contemporaryeconomic reality and its more recent manifesta-tions in the form of agility, it is important to takestock of the research in this area and provide guid-ance for its development in the future. We addressthis gap and employ a meta-analytic approach todevelop and test amodel of the antecedents, conse-quences and contingencies of SF rooted in theoryand existing studies. In doing this, we contributeto the further consolidation of the "eld of SF byconnecting the seemingly distinct but naturally re-lated research streams of strategy, management,marketing, innovation, entrepreneurship and op-erations, and we resolve several disagreements andtheoretical de"cits identi"ed in the literature.Morespeci"cally, our results offer guidance for schol-ars and managers alike by providing an overviewof SF’s multidimensionality (Table 1), highlight-ing theoretical tension in the literature (Tables 2and 3), summarizing empirical generalizations re-garding key relationships (Table 4), testing an over-all model of the antecedents of SF (Table 5), ex-amining the consequences of a more nuanced SF(Table 6 and Figure 2) and identifying contingen-cies for the performance effect (Table 7). While ourresults resolve some theoretical tensions, we alsohighlight gaps between theoretical developmentsand related empirical "ndings in existing studies.

Research implications

Table 8 summarizes our core "ndings, theoreti-cal and managerial implications and future re-

search directions. Several potential antecedents ofSF have proved to be controversial in the litera-ture (see Table 2). We address these tensions andoffer insights into the key enablers, inhibitors andtriggers of SF. Our multivariate results reveal thatboth the strategic orientation and the organiza-tional design of a "rm enable SF. Interestingly,decentralization has a higher impact on SF thanany other enabler (χ2

diff = 8.75, p < 0.01). Thus,shared decision-making across an organization ap-pears to be crucial to promote SF. We further "ndthat slack resources negatively affect SF.While this"nding contradicts the common claim that orga-nizations without uncommitted, liquid resourcesare less likely to adapt to environmental changes(Barker and Barr, 2002), it is in line with Sanchez(1995), who caution that resources should be !exi-ble rather than slack. Following his logic, resourcesurplus implies they are not utilized optimally andef"ciently, while resource !exibility implies that re-sources can swiftly be re-utilized to underpin newstrategic options.The analyses also reveal that past success in-

hibits SF, while "rm size and "rm age do notinhibit SF. These insights extend the knowledgeon the inhibitors of SF, previously an under-researched area, and provide resolution to tensionson whether "rm size and "rm age inhibit SF. Fi-nally, competitive intensity triggers SF, while envi-ronmental dynamism and demand uncertainty ex-hibit a negative relationship with SF. Discoveringthat neither environmental dynamism nor demanduncertainty trigger SF challenges the existing liter-ature and raises some potential controversies, espe-cially in suggesting that SF is achieved more easilyin highly competitive environments. However, tak-ing into consideration the multidimensionality ofthe environment and the results of the contingencyanalyses, the whole picture is complex. Evidently,more research on the triggers and contingenciesof SF, especially on environmental dynamism andcompetitive intensity, is needed.Our "ndings also address the con!icting empiri-

cal evidence regarding the relationship between SFand performance (see Table 3). We "nd a generalmoderately positive effect size of 0.24 for the SF–performance relationship, thereby resolving theongoing debate over whether SF has performance-enhancing bene"ts for "rms or whether it can evenharm "rms. Our "ndings support the argumentssubmitted by proponents of SF, namely that,in today’s competitive landscape, "rms’ success


16 Herhausen et al.Table8.

Overviewofcore"ndings,implicationsandfutureresearchdirections

Issue

Core"n

ding

sTheoretical

implications

Man

agerialimplications

Futureresearch

directions

Antecedents

!Boththestrategicorientation

andtheorga

nizatio

nald

esign

ofa"rm

enableSF

.Con

trary

toexpectations,slack

resourcesdo

note

nableSF

.!Pa

stsuccessinhibitsSF

,while"rm

size

and"rm

age

dono

tinh

ibitSF

.!Com

petitiveintensity

triggers

SF,w

hileenvironm

ental

dyna

mism

anddeman

dun

certaintyhave

anegativ

erelatio

nshipwith

SF.

!Resolving

severaltension

sin

theliteraturerelatedto

SF(i.e.theissues

ofstrategic

plan

ning

/form

alroutines

and

slackresources).

!Extending

know

ledg

eon

the

inhibitors

ofSF

and

resolvingseveraltension

sin

theliterature(i.e."rm

size

and"rm

age).

!Discovering

that

neith

erenvironm

entald

ynam

ism

nor

deman

dun

certaintytrigger

SFchalleng

estheliterature.

!Strategicorientations,

decentralizationan

dform

alroutines

areim

portan

tenablers

ofSF

.Slack

resourcesarenegativ

ely

relatedto

SF.

!Awarenessthat

pastsuccess

constrains

SF,w

hile"rm

size

and"rm

agedo

not,ha

sop

posing

effectson

SF.

!Po

sitiv

easpect

ofcompetitive

intensity

inachievingSF

(but

note

nviron

mentald

ynam

ism

anddeman

dun

certainty).

!Exa

miningtheim

portan

ceof

additio

nala

ntecedents(e.g.

techno

logicalo

rientatio

nor

digitalcap

abilitie

s).

!Con

sidering

meso-levela

ndmicro-le

vela

ntecedentsof

SF.

!Further

exam

inationof

environm

entald

ynam

ism

and

competitiveintensity

aspo

tentialinh

ibito

rsof

SF.

!Investigatingtherelatio

nship

betw

eenad

ditio

nal"

rmcharacteristicsan

dSF

(e.g.

strategicconsensus).

Consequences

!Overall,

SFha

sapo

sitiv

eeffect

on"rm

performan

ce.

!The

performan

ceeffect

islarger

whenSF

ismeasured

interm

sof

proa

ctively

creatin

gnewop

portun

ities,

internal

resource

deploy

ment

andexternal

competitive

actio

ns.

!The

inno

vatio

n-an

dmarket-relatedbene"tsof

SFou

tweigh

thecostsassociated

with

SF,leading

toapo

sitiv

eSF

effect

on"n

ancial

performan

ce.

!Resolving

along

-stand

ing

disagreemento

nthena

ture

oftheSF

–perform

ance

relatio

nship.

!Con

gruencewith

previous

literatureclaimingthat

SFas

aconstructiscomplex,w

ithmultip

ledimension

s.!Highlightingthemod

elthat

best"tswith

existin

gem

piricalresearchin

explaining

nuan

cesin

the

SF–p

erform

ance

relatio

nship.

!SF

isim

portan

toverall,

yetit

shou

ldno

tberega

rded

asa

universalsolutionagainsta

llkind

sof

competitivethreats.

!Im

portan

ceof

being

proa

ctivean

dconsideringSF

both

ininternal

resource

deploy

menta

ndexternal

competitiveactio

ns.

!Morefocuson

inno

vatio

nou

tcom

esan

dmarket

outcom

eswhenevalua

ting

thebene"tsof

SF.

!Investigatingthecostsan

dbene"tsof

differentS

Fdimension

s.!Con

sidering

meso-levela

ndmicro-le

velcon

sequ

encesof

SF,suchas

orga

nizatio

nal

clim

atean

dtheeffect

ofSF

onem

ploy

ees.

!Differentia

tingbetw

een

short-an

dlong

-term

consequences

(e.g.

inno

vativ

enessvs.

orga

nizatio

nalsurvival).

Contingencies

!Env

iron

mentald

ynam

ism

increasesthe

SF–p

erform

ance

relatio

nship.

Nosign

i"cant

effectsfor

deman

dun

certaintyan

dcompetitiveintensity.

!Su

ggestin

gamorenu

anced

view

onthena

ture

ofenvironm

entala

nd"rm

contingenciesaffectingthe

consequences

ofSF

than

previouslyconsidered.

!Bene"tsof

SFarelarger

inhigh

lydy

namic

environm

ents.D

eman

dun

certaintyan

dcompetitive

intensity

dono

taffectS

Fbene"ts.

!Ta

king

amoredifferentia

ted

view

onenvironm

ental

contingencies.

!Investigatingcontingencies

notc

apturedhere

(e.g.

techno

logicalturbu

lence).



depends on their ability to address changing con-ditions and to constantly adapt their strategiesaccordingly.

From a contingency point of view, our resultsalso indicate that neither the aspects of the en-vironment nor the "rm characteristics affect theSF–performance relationship in the way we mightexpect based on previous studies conducted inparticular settings (e.g. Grewal and Tansuhaj,2001). While environmental dynamism positivelymoderates the SF–performance relationship, nosigni"cant effects were found for either demanduncertainty or competitive intensity. These "nd-ings suggest a more nuanced view on the natureof environmental contingencies affecting theconsequences of SF than previously considered.

Moreover, we demonstrate that con!icting "nd-ings in the literature can be at least partly at-tributed to the conceptualizations of SF employedin earlier studies. In order to appreciate the widercontribution of SF, it must be acknowledged thatit is a multidimensional concept, something thatmany earlier studies failed to recognize suf"ciently.We "nd that the performance effect is larger whenSF is measured in terms of proactively creatingnew opportunities, internal resource deploymentand external competitive actions. It is surprisingto "nd that neither variety nor speed (timely re-sponse) affect the SF–performance relationship.Our analyses suggest that for many "rms, the costof developing and maintaining strategic optionsmay outweigh the bene"ts.

We also identi"ed the avenues through which SFin!uences "rm performance. We "nd that SF hasa negative direct effect on "nancial performance,therefore providing prima facie support for the op-ponents of SFwhen the ‘bottom-line’performanceof "rms is considered. However, when the indirectpositive effect of SF through innovation outcomesand market outcomes is considered, SF’s overallimpact on "nancial outcomes is positive, withan effect size of 0.15. This overall positive effectsuggests that, all else being equal, the innovation-and market-related outcomes of SF (positiveindirect effect) outweigh the costs associated withSF (negative direct effect), leading to a positiveSF effect on "nancial performance. However, itshould be stated that the small-to-moderate effectsize of the SF–performance relationship does notjustify the overly enthusiastic appraisal of SF inthe normative literature. Thus, while our "ndingsattest to the overall importance of SF, we caution

against regarding SF as a panacea for all kinds ofcompetitive threats.Although we were able to study only "rm-level

effects, there is a dearth of literature devoted tothe micro foundations and micro-level variablesin the study of SF. Individual-level phenomenawith regard to C-suite executives (such as lead-ership variables) should be examined in order toidentify the individual- and also group/team-leveldynamics that enable SF to !ourish (such asmotivations and engagement). Moreover, futurestudies in SF should adopt multi-level approachesin which micro-, meso- and macro-level phenom-ena can be examined simultaneously, along withtheir interactions and interdependencies. Equally,studies should consider the extent to which SFcan be differentially geared within organizations.By this we mean that SF can be treated in anambidextrous manner whereby parts of the or-ganization are exposed to the demands of SFwhereas other parts are not. From this form ofintra-organizational research, modest forms of!exibility can be considered exploitative whereassubstantive forms can be described as exploratory.This is apparent in the ambidexterity literaturebut not explicitly from a SF perspective.

Managerial implications

In 1964 the average tenure of "rms on the S&P500 index was 33 years. By 2016 this reduced to24 years and is projected to be a mere 12 years by2027 (Anthony et al., 2018). This equates to halfof these large US$6bn or more "rms disappear-ing by 2027. Various reasons account for this at-trition, not least of which is "rms’ lack of strategic!exibility in aligning their offerings to the market.Our results provide important insights for man-agers faced with questions regarding SF to remaincompetitively vital, in order to strategically !our-ish over time.Table 8 provides a series of actionable manage-

rial practices that are available for strategists to en-gage in seeking to realize SF. Our results from themultivariate analysis of antecedents of SF demon-strate the importance of entrepreneurial orienta-tion, market orientation, learning orientation, de-centralization and formal routines in enabling a"rm to become !exible. Thus, engaging in these"rm-wide orientations demands the ability to en-gage in ambidextrous and polydextrous behavioursthat allow these different plates to be consistently


18 Herhausen et al.

and simultaneously spun. However, managers alsoneed to be aware that slack resources can haveunintended consequences with regard to SF andthat past success further inhibits SF, while "rmsize and "rm age do not have opposing effects. Fi-nally, the positive effect of competitive intensityon SF (but not of environmental dynamism anddemand uncertainty) highlights that competitivepressure will motivate "rms to become more !exi-ble. These different "ndings indicate a repertoire ofnuanced managerial interventions and not a ‘one-size-"ts-all’ approach. Clearly, managers need tobe aware of the complacency effect that is oftencharacterized by exemplary past performance. Theappetite "rms have to engage in success needs to re-main once they become successful. Ensuring thatstructural, leadership and cultural transformationremains aligned with the evolving marketplace iscritical, and maintaining a focus on the compet-itive pressures will keep this intensity to performhigh.

Moreover, our meta-analytical "ndings demon-strate that SF has an overall positive impact on"rm performance. However, not all "rms canextract the highest gains from SF. As such, the"ndings on the contingencies provide useful man-agerial insights pertaining to which "rms shouldinvest more or less in SF. While SF appears to beespecially valuable for "rms that face high levelsof environmental dynamism, "rms facing lowerlevels of dynamism should be more cautious aboutwhether their investment will pay off. However,demand uncertainty and competitive intensity didnot affect the SF–performance relationship andare thus not helpful to managers in terms of deter-mining what level of SF may be appropriate. Wealso found that "rms should invest in SF regard-less of their size and age, as these characteristicsdid not affect the SF–performance relationship.

Our "ndings further reveal that managersshould be cognisant of the total impact that SFcan have on "rm performance, rather than focus-ing on its impact on "nancial outcomes alone.We examine a more comprehensive model ofthe mechanisms that mediate the SF–"nancialoutcomes relationship than those tested in theextant literature. We thus suggest that measures ofthe mediating variables – innovation and marketoutcomes – may be useful for managers in termsof tracking the impact of SF on overall perfor-mance. Only if managers undertake such trackingcan the full potential of SF be captured. A myopic

focus on "nancial outcomes may underestimatethe importance of SF.

Limitations and directions for future research

Despite its contributions, this study has some im-portant limitations that should be consideredwheninterpreting our "ndings. First, any meta-analysisis constrained by the nature and scope of the origi-nal studies. Thus, our study is de facto limited to ex-amining the effects of variables that were availablein existing studies.Moreover, because the primary-level studies were not experimental in nature, thereported results do not provide direct and unequiv-ocal evidence regarding causality. Indeed, the vastmajority of the studies have used cross-sectionalmethods to examine SF. However, most researchquestions concerning SF – even if not framed assuch – address issues of change. Consequently, werecommend that future studies use longitudinal orexperimental designs in order to overcome thesemethodological weaknesses and capture the causaleffects of SF. Another important concern is thatthe relationships reported in the original studiesmay be positively biased because of an oversam-pling of successful "rms; "rms that failed to be!exiblemaywell have exited themarket andwould,therefore, not have been captured by the studies inour sample.

In terms of measurement, many studies haveemployed a unidimensional measure of SF, thusignoring the multidimensional composition of theconstruct. Their empirical "ndings may have un-derestimated the extent of SF because they did notconsider that SF can be a combination of variousfactors: the ability to be both reactive and proac-tive and to have internal and external options,which enable a "rm to take a variety of actions andto do so rapidly. Thus, multidimensional measuresof SF should be employed in order to better cap-ture the relationships under scrutiny. Moreover,most studies investigating the SF–performance re-lationship under different environmental condi-tions have employed a unidimensional measure ofthe environment. A shortcoming of the argumentthat more !exibility is needed in more dynamicmarkets is that this fails to consider the multidi-mensionality of the environment and the con!ict-ing empirical "ndings. Therefore, further researchshould detail the effects of SF under varying envi-ronmental conditions.



Despite signi"cant progress in understandingSF and its nomological network, several knowl-edge gaps remain, and future research directionsarise from these gaps. Longitudinal studies thatdifferentiate clearly between short-term and long-term outcomes would also be valuable – this wouldyield a greater understanding of SF’s effects on"rm performance. In addition, while the currentfocus in the strategy literature is on the externalconsequences of SF, it is important to note thatSF may affect several internal organizational fac-tors that, in turn, may affect innovation, marketand "nancial outcomes. Further research shouldtherefore investigate themediating role of the costsof experimentation, inef"ciency, stress among em-ployees and a potential lack of focus in the SF–performance relationship. Due to data limitations,we could not examine individual- or team-level as-pects of SF. Thus, we only focused on the "rm levelwhile neglecting the micro and meso level of SFsuch as leadership and motivation, commitmentand focus among employees. Future studies shouldtake a more multi-level approach, in which boththe micro and meso level of SF are involved.

In summary, this meta-analysis develops anintegrative framework of the antecedents, con-sequences and contingencies of SF by resolvingseveral tensions and con!icting "ndings in the lit-erature. By highlighting which enablers, inhibitorsand triggers have the strongest impact on buildingSF, and by clarifying when and how SF in!uencesthe performance of "rms, we provide managerialguidance to bene"t from SF and conceptual clarityfor future research on SF.

Acknowledgements

We are grateful for comments provided byDon Bergh, Mike Hitt, Dave Ketchen, MichaelLeiblein, Neil Morgan, Sucheta Nadkarni, Hansvan Oosterhout and Charles Snow on previousdrafts of this paper. We also appreciate the co-operation of authors from whom we requestedadditional data.

References

Aaker, D. A. and B. Mascarenhas (1984). ‘The need for strategic!exibility’, Journal of Business Strategy, 5, pp. 74–82.

Aguinis, H., D. R. Dalton, F. A. Bosco, C. A. Pierce and C. M.Dalton (2011). ‘Meta-analytic choices and judgment calls: im-

plications for theory building and testing, obtained effect sizesand scholarly impact’, Journal of Management, 37, pp. 5–38.

Amit, R. and P. J. Schoemaker (1993). ‘Strategic assets and orga-nizational rent’, Strategic Management Journal, 14, pp. 33–46.

Anthony, S. D., S. P. Viguerie, E. I. Schwartz and J. vanLandeghem (2018). Corporate Longevity Forecast: CreativeDestruction is Accelerating. Boston, MA: Innosight.

Bahrami, H. and S. Evans (2011). ‘Super-!exibility for real-timeadaptation’, California Management Review, 53, pp. 21–39.

Barker, V. L. and P. S. Barr (2002). ‘Linking top manager attri-butions to strategic reorientation in declining "rms attemptingturnarounds’, Journal of Business Research, 55, pp. 963–979.

Bergh,D.D.,H.Aguinis, C.Heavey,D. J.Ketchen, B.K. Boyd, P.Su, C. L. L. Lau andH. Joo (2016). ‘Usingmeta-analytic struc-tural equation modelling to advance strategic management re-search: guidelines and an empirical illustration via the strategicleadership–performance relationship’, Strategic ManagementJournal, 37, pp. 477–497.

Best, M. H. (1990). The New Competition: Institutions of In-dustrial Restructuring. Cambridge, MA: Harvard UniversityPress.

Blut,M., N. Chowdhry, V.Mittal and C. Brock (2015). ‘E-servicequality: a meta-analytic review’, Journal of Retailing, 91,pp. 679–700.

Brews, P. J. and M. R. Hunt (1999). ‘Learning to plan and plan-ning to learn: resolving the planning school/learning school de-bate’, Strategic Management Journal, 20, pp. 889–913.

Brinckmann, J., J. Villanueva, D. Grichnik and L. Singh (2019).‘Sources of strategic !exibility in new ventures: an analysisof the role of resource leveraging practices’, Strategic En-trepreneurship Journal, 13, pp. 154–178.

Brozovic, D. (2018). ‘Strategic !exibility: a review of the lit-erature’, International Journal of Management Reviews, 20,pp. 3–31.

Cheng, J. L. and I. F. Kesner (1997). ‘Organizational slack andresponse to environmental shifts: the impact of resource allo-cation patterns’, Journal of Management, 23, pp. 1–18.

Cheung, M. W. L. and W. Chan (2005). ‘Meta-analytic struc-tural equation modeling: a two-stage approach’, PsychologicalMethods, 10, pp. 40–64.

Claussen, J., C. Essling and C. Peukert (2018). ‘Demand vari-ation, strategic !exibility and market entry: evidence fromthe US airline industry’, Strategic Management Journal, 39,pp. 2877–2898.

Combe, I. A. (2012). ‘Marketing and !exibility: debates past,present and future’, European Journal of Marketing, 46,pp. 1257–1267.

Cortina, J. M. (2003). ‘Apples and oranges (and pears, oh my!):the search for moderators in meta-analysis’, OrganizationalResearch Methods, 6, pp. 415–439.

Covin, J. G., D. P. Slevin and R. L. Schultz (1997). ‘Top man-agement decision sharing and adherence to plans’, Journal ofBusiness Research, 40, pp. 21−36.

Dai, Y., J. C. Goodale, G. Byun and F. Ding (2018). ‘Strategic!exibility in new high-technology ventures’, Journal of Man-agement Studies, 55, pp. 265–294.

Dalziel, M. (2009). ‘Forgoing the !exibility of real options: whenand why "rms commit to investment decisions’,British Journalof Management, 20, pp. 401–412.

Das, T. K. and B. Elango (1995). ‘Managing strategic !exibility:key to effective performance’, Journal of GeneralManagement,20, pp. 60–75.


20 Herhausen et al.

Deloitte (2016). Unlocking the Flexible Organisation: Organisa-tional Design for an Uncertain Future. London: Deloitte De-velopment LLC.

Ebben, J. J. and A. C. Johnson (2005). ‘Ef"ciency, !exibility, orboth? Evidence linking strategy to performance in small "rms’,Strategic Management Journal, 26, pp. 1249–1259.

Eisenhardt, K. M. and J. A. Martin (2000). ‘Dynamic capa-bilities: what are they?’, Strategic Management Journal, 21,pp. 1105–1121.

Eisenhardt, K. M, N. R. Furr and C. B. Bingham (2010). ‘Mi-crofoundations of performance: balancing ef"ciency and !ex-ibility in dynamic environments’, Organization Science, 21,pp. 1263–1273.

Evans, J. S. (1991). ‘Strategic !exibility for high technology ma-noeuvres: a conceptual framework’, Journal of ManagementStudies, 28, pp. 69–89.

Ghemawat, P. and J. E. Ricart i Costa (1993). ‘The organizationaltension between static and dynamic ef"ciency’, StrategicMan-agement Journal, 14, pp. 59–59.

Grewal, R. and P. Tansuhaj (2001). ‘Building organizational ca-pabilities for managing economic crises: the role of marketorientation and strategic !exibility’, Journal of Marketing, 65,pp. 67–80.

Guo, H. and Z. Cao (2014). ‘Strategic !exibility and SME per-formance in an emerging economy: a contingency perspective’,Journal of Organizational Change Management, 27, pp. 273–298.

Hannan, M. T. and J. Freeman (1977). ‘The population ecologyof organizations’, American Journal of Sociology, 82, pp. 929–964.

Harrigan, K. R. (1985). ‘Exit barriers and vertical integration’,Academy of Management Journal, 28, pp. 686–697.

Hatum, A. andA.M. Pettigrew (2006). ‘Determinants of organi-zational !exibility: a study in emerging economy’,British Jour-nal of Management, 17, pp. 115–137.

Henderson, R. M. and K. B. Clark (1990). ‘Architectural inno-vation: the recon"guration of existing’, Administrative ScienceQuarterly, 35, pp. 9–30.

Hitt, M. A., B.W. Keats and S.M. DeMarrie (1998). ‘Navigatingin the new competitive landscape: building strategic !exibilityand competitive advantage in the 21st century’, Academy ofManagement Executive, 12, pp. 22–42.

Hunter, J. E. and F. L. Schmidt (2004). Methods of Meta-analysis. Correcting Error and Bias in Research Findings, 2ndedn. London: Sage.

Johnson, J. L., R. P-W. Lee, A. Saini and B. Grohmann (2003).‘Market-focused strategic !exibility: conceptual advances andan integrative model’, Journal of the Academy of MarketingScience, 31, pp. 74–89.

Karna, A., A. Richter and E. Riesenkampff (2016). ‘Revisitingthe role of the environment in the capabilities–"nancial perfor-mance relationship: a meta-analysis’, Strategic ManagementJournal, 37, pp. 1154–1173.

Kortmann, S., C. Gelhard, C. Zimmermann and F. T. Piller(2014). ‘Linking strategic !exibility and operational ef"ciency:the mediating role of ambidextrous operational capabilities’,Journal of Operations Management, 32, pp. 475–490.

Lant, T.K., F.Milliken andB. Batra (1992). ‘The role of manage-rial learning and interpretation in strategic persistence and re-orientation: an empirical exploration’, Strategic ManagementJournal, 13, pp. 585–608.

Lee, S.-H. and M. Makhija (2009). ‘Flexibility in internation-alization: is it valuable during an economic crisis?’, StrategicManagement Journal, 30, pp. 537–555.

Levinthal, D. A. and J. G.March (1993). Themyopia of learning,Strategic Management Journal, 14, pp. 95–112.

McGrath, R. G. (1997). ‘A real options logic for initiating tech-nology positioning investments’,Academy of Management Re-view, 22, pp. 974–996.

Nadkarni, S. and P. Herrmann (2010). ‘CEO personality, strate-gic !exibility and "rm performance: the case of the Indianbusiness process outsourcing industry’, Academy of Manage-ment Journal, 53, pp. 1050–1073.

Nadkarni, S. and V. K. Narayanan (2007). ‘Strategic schemas,strategic !exibility and "rm performance: the moderating roleof industry clockspeed’, Strategic Management Journal, 28,pp. 243–270.

Neill, S. and G. M. Rose (2007). ‘Achieving adaptive endsthrough equivocality: a study of organizational antecedentsand consequences’, Journal of Business Research, 60, pp. 305–313.

Pauwels, P. and P. Matthyssens (2004). ‘Strategic !exibility inexport expansion: growing through withdrawal’, InternationalMarketing Review, 21, pp. 496–510.

Perez-Valls, M., J. Cespedes-Lorente and J. Moreno-Garcia(2016). ‘Green practices and organizational design as sourcesof strategic !exibility and performance’, Business Strategy andthe Environment, 25, pp. 529–544.

Posen, H. E. and D. A. Levinthal (2012). ‘Chasing a moving tar-get: exploitation and exploration in dynamic environments’,Management Science, 58, pp. 587–601.

Rubera, G. and A. H. Kirca (2012). ‘Firm innovativeness and itsperformance outcomes: a meta-analytic review and theoreticalintegration’, Journal of Marketing, 76, pp. 130–147.

Sanchez, R. (1995). ‘Strategic !exibility in product competition’,Strategic Management Journal, 16, pp. 135–159.

Santos-Vijande, M. L., J. Á. L. Sanchez and J. A. Trespalacios(2012). ‘How organizational learning affects a "rm’s !exibil-ity, competitive strategy and performance’, Journal of BusinessResearch, 65, pp. 1079–1089.

Schreyögg, G. and J. Sydow (2010). ‘Organizing for !uidity?Dilemmas of new organizational forms’,Organization Science,21, pp. 1251–1262.

Schweiger, S. A., T. R. Stettler, A. Baldauf and C. Zamu-dio (2019). ‘The complementarity of strategic orientations: ameta-analytic synthesis and theory extension’, Strategic Man-agement Journal, 40, pp. 1822–1851.

Segaro, E. L., J. Larimo and M. V. Jones (2014). ‘Internation-alisation of family small and medium sized enterprises: therole of stewardship orientation, family commitment cultureand top management team’, International Business Review, 23,pp. 381–395.

Shimizu, K. and M. A. Hitt (2004). ‘Strategic !exibility: organi-zational preparedness to reverse ineffective strategic decisions’,Academy of Management Review, 18, pp. 44–59.

Skordoulis, R. T. (2004). ‘Strategic !exibility and change: anaid to strategic thinking or another managerial abstraction’,Strategic Change, 13, pp. 253–258.

Stewart, G. L. and M. R. Barrick (2000). ‘Team structure andperformance: assessing the mediating role of intrateam pro-cess and the moderating role of task type’, Academy of Man-agement Journal, 43, pp. 135–148.



Stieglitz, N., T. Knudsen and M. C. Becker (2016). ‘Adaptationand inertia in dynamic environments’, Strategic ManagementJournal, 37, pp. 1854–1864.

Teece, D. J., G. Pisano and A. Shuen (1997). ‘Dynamic capabil-ities and strategic management’, Strategic Management Jour-nal, 18, pp. 509–533.

Thompson, J. D. (1967). Organizations in Action: Social Sci-ence Bases of Administrative Theory. New York: McGraw-Hill.

Trigeorgis, L. (1996). Real Options: Managerial Flexibilityand Strategy in Resource Allocation. Cambridge, MA:MIT Press.

Van den Noortgate, W., J. A. López-López, F. Marín-Martínezand J. Sánchez-Meca (2013). ‘Three-level meta-analysis of de-pendent effect sizes’,Behaviour ResearchMethods, 45, pp. 576–594.

Viswesvaran, C. and D. S. Ones (1995). ‘Theory testing:combining psychometric meta-analysis and structuralequations modelling’, Personnel Psychology, 48, pp. 865–885.

Volberda, H.W. (1996). ‘Toward the !exible form: how to remainvital in hypercompetitive environments’,Organization Science,7, pp. 359–374.

Volberda, H. W. (1998). Building the Flexible Firm: How to Re-main Competitive. Oxford: Oxford University Press.

Wang, C. L., C. Senaratne and M. Ra"q (2015). ‘Success traps,dynamic capabilities and "rm performance’, British Journal ofManagement, 26, pp. 26–44.

Weigelt, C. and M. B. Sarkar (2009). ‘Learning from supply-sideagents: the impact of technology solution providers’ experi-ential diversity on clients’ innovation adoption’, Academy ofManagement Journal, 52, pp. 37–60.

Worren, N., K. Moore and P. Cardona (2002). ‘Modularity,strategic !exibility and "rm performance: a study of thehome appliance industry’, Strategic Management Journal, 23,pp. 1123–1140.

Young-Ybarra, C. and M. Wiersema (1999). ‘Strategic !exibilityin information technology alliances: the in!uence of transac-tion cost economics and social exchange theory’,OrganizationScience, 10, pp. 439–459.

Dennis Herhausen is Associate Professor of Marketing at KEDGE Business School. His research fo-cuses on marketing strategy implementation, customer relationships, digital marketing and channelmanagement. His work has been published in the British Journal of Management, Journal of Market-ing, Journal of Retailing, Journal of Service Research, Journal of Business Research and Academy ofManagement Best Paper Proceedings, among others.

Robert E. Morgan holds the Sir Julian Hodge Chair and is a Professor of Marketing and Strategyat Cardiff University. He is a Strategy Research Foundation Scholar and hold visiting professorshipsincluding VUUniversiteit Amsterdam and Copenhagen Business School. His research focuses on "rmcapabilities, new product development and technology management from a strategy perspective. Hiswork has appeared in the British Journal of Management, Strategic Management Journal, Journal ofManagement Studies, Journal of theAcademy of Marketing Science, Journal of ServiceResearch,Strate-gic Entrepreneurship Journal and Journal of Product Innovation Management, among others.

Danilo Brozovic is Senior Lecturer in Business Administration at the University of Skövde, Sweden.His research interests include strategic !exibility, service marketing and the intersection between thesetopics. His research has been published in the International Journal of Management Reviews, Journalof Service Theory and Practice, Journal of Services Marketing andMarketing Theory, among others.

Henk W. Volberda is Professor of Strategic Management and Innovation at Amsterdam BusinessSchool, University of Amsterdam. His research has been published extensively in several internationalpeer-reviewed journals, includingAcademy of Management Journal,Global Strategy Journal, Journal ofInternational Business, Journal of Management, Journal of Management Studies,Management Science,Organization Science, Strategic Entrepreneurship Journal and Strategic Management Journal.

Supporting Information

Additional supporting information may be found online in the Supporting Information section at the endof the article.

SupportingMaterial


re-examining strategic flexibility: a meta-analysis of its

Documents